Documenti di Didattica
Documenti di Professioni
Documenti di Cultura
Edited by
Jacob I. Kleiman
Integrity Testing Laboratory Inc. Markham,
Toronto, Canada
A C.I.P. Catalogue record for this book is available from the Library of Congress.
Published by Springer,
P.O. Box 17, 3300 AA Dordrecht, The Netherlands.
www.springer.com
Introduction xi
Acknowledgements xiii
Organization xv
Testing of Spacecraft Materials for Long Flights in Low Earth Orbit 167
L. S. Novikov, V. N. Chernik, S. F. Naumov, S. P. Sokolova,
T. I. Gerasimova, A. O. Kurilyonok, and T. N. Smirnova
xi
xii INTRODUCTION
Jacob Kleiman
Chairman/Organizing Committee/ICMSE-7
Integrity Testing Laboratory Inc.
20 January, 2005
ACKNOWLEDGEMENTS
Jacob Kleiman
Integrity Testing Laboratory Inc.
Conference Chairman
xiii
ORGANIZATION
Organizing Committee
Program Committee
xv
xvi ORGANIZATION
Session Chairs
Opening Session
Moderator: Jacob Kleiman, ITL Inc./UTIAS, Canada
1. Introduction
Since solar cells are key elements to provide spacecraft with electric energy in
orbit, the degradation in their properties could directly influence the working
condition and lifetime of spacecraft. The electric properties of solar cells would
be degenerated under the space radiations [1]. It is important to improve the
radiation resistance of the solar cells [25]. In order to evaluate the performance
of the solar cells in the geostationary Earth orbit, it is necessary to characterize
the effects of charged particles with the energy less than 200 keV, which exist in
large amount in the Earths radiation belts. Also, when spacecraft passes into the
Earth shadow, the solar cells are subjected to the effect of low temperatures. The
1
J.I. Kleiman (ed.), Protection of Materials and Structures from Space Environment, 18.
C 2006 Springer. Printed in the Netherlands.
2 ZHENGYU HU ET AL.
aim of this study was to examine the change in electric properties of the silicon
solar cells under the radiations of protons and electrons with <200 keV at 77 K,
as well as the radiation damage mechanism.
2. Experimental
The sample of the backfield silicon solar cell is schematically shown in figure 1,
which is 250 m thick and has the base layer resistivity of 1012 cm. The sam-
ples with the size of 20 20 mm2 were not covered with glass. The phosphorus
doped Si emitter junction depth from the surface is 0.2 m. The parameters of
samples before irradiation are shown in table 1.
The ground-based simulation equipment used in this study can simulate the
radiations of solar electromagnetic rays, electrons and protons in orbit, indepen-
dently and simultaneously. The irradiation energy of protons and electrons was
chosen as 60180 keV, the flux 1.2 1012 cm 2 s1 and the fluence from 2 1013
to 2 1016 cm2 . The homogeneity of the irradiation flux is more than 95% in
the area of 100 100 mm2 by scanning the proton and electron beam. During the
irradiations, the chamber was kept at vacuum 104 Pa and temperature of 77 K.
The electric properties of the samples were characterized by I V curves, which
were measured before and after the irradiations. In order to analyze the radiation-
induced defects, DLTS analysis (deep level transient spectroscopy) was carried
out in the temperature range of 77450 K.
VI VII
V IV I. Ohmic contact region on back surface;
II. p+ doped region;
III. p base region;
III IV. pn junction region;
V. n+ top region;
VI. Attenuating reflection coating;
II VII. Gate electrode
I
Figure 2. The I V curves for the samples irradiated with (a) 60 and (b) 170 keV protons
Figure 2(a) and 2(b) show the I V curves for the samples irradiated with 60 and
170 keV protons, respectively. The changes in the normalized short-circuit current
Isc /I0 and the normalized open-circuit voltage Voc /Vo are given as a function
of fluence of protons with various energies, as shown in figures 3(a) and 3(b).
With increasing both fluence and energy, the short-circuit current Isc decreases
gradually. The irradiation of protons also leads to an obvious decrease of the
open-circuit voltage Voc , but the degradation extent depends on the fluence and
energy of the protons. After the irradiation by protons with 150170 keV, the Voc
is noticeably degraded at lower fluences, and then remains almost unchangeable
1.0 1.0
77 k 77 k
0.9 60 keV 0.9 60 keV
150 keV 150 keV
Normalized V oc
Normalized I sc
TABLE 2. Variation of the characteristics for the samples before and after 60 keV proton
irradiation (AM0)
Jsc
Characteristics Voc (V) Isc (mA) (mA.cm2 ) FF EFF (%) Rs ()
with the increase of fluence or even shows a recovery after a fluence higher than
5 1015 cm2 . Table 2 shows that after the irradiation by protons with 60 keV
to a fluence of 1 1016 cm2 , all the characteristics including the Voc , Isc , Jsc ,
FF, and EFF decrease, while the internal resistance in series increases remarkably
(almost 10 times higher). This implies that the concentration of carriers inside the
samples is obviously reduced due to the proton irradiation.
Figure 4 shows the I V curves for the samples before and after irradiation by 60
and 180 keV electrons. It is obvious that there is very little change in the I V
curves. However, the situation is quite different, if the I V curves are examined in
situ during the exposure. With increasing the electron fluence, the normalized Voc
gradually decreases (see figure 5(a)), while the normalized Isc increases until the
fluence less than 1 1016 cm2 (see figure 5(b)). Figures 6(a) and 6(b) show the
changes in the Voc and Isc with time in vacuum after the electron irradiation, respec-
tively. It is demonstrated that with increasing the time, the Voc rises and the Isc drops
again. Both the Voc and Isc tend to return gradually to the original level for the unir-
radiated samples. This phenomenon implies that a recovery must be considered
in evaluating the electron radiation effect for the backfield Si solar cells in orbit.
140
120
100
80
I (mA)
before irradiation
60 keV
60
11016 e/cm2
40 180 keV
11016 e/cm2
20
0
0 100 200 300 400 500 600
V (mV)
Figure 4. Variation of I V curves for the samples before and after electron irradiation
RADIATION EFFECTS OF PROTONS AND ELECTRONS 5
1.0 1.5
170 keV
0.9 1.4
Normalized Isc
Normalized Voc
0.8 1.3
77 K
0.7 1.2 170 keV
1.1
0.6
1.0
0.5 15 15 16 16 15 15 16 16
0.0 4.010 8.010 1.210 1.610 0.0 4.010 8.010 1.210 1.610
Electron fluence (e/cm ) 2
Electron fluence (e/cm2)
(a) (b)
Figure 5. The normalized Voc (a) and normalized Isc (b) measured in situ as a function of fluence
for the samples irradiated with 170 keV electrons
Figure 7 shows the normalized I V curves for the samples under the proton,
the electron and the combined irradiations with 170 keV. Notice that the damage
effect of protons is much larger than that of the electrons (see curves 1, 2, and
4). The effect of the combined radiation of protons and electrons does not show
simple additivity. The degradation extent due to the proton radiation is larger than
that of the combined radiation under the same proton fluence and lower electron
fluence (see curves 3 and 4). In contrast, the combined radiation results in more
severe damage under the same proton fluence and higher electron fluence (see
curves 4 and 5).
1.4
1.0 170 keV
1.3 1.51016 e/cm2
0.9 after irradiation
Normalized Voc
Normalized Isc
1.2
0.8 77 K
170 keV
1.51016 e/cm2 1.1
0.7
after irradiation
0.6 1.0
0.5 0.9
0 20 40 60 80 100 120 140 160 0 20 40 60 80 100 120
Time (min) Time (min)
(a) (b)
Figure 6. The normalized Voc (a) and normalized Isc (b) measured in situ as function of time in
vacuum after electron irradiation with 170 keV
6 ZHENGYU HU ET AL.
1
1.0 170 keV
I'
and 31015 p/cm2
3 4- 51015 p/cm2
0.4 4 5- 51015 e/cm2
5 and 51015 p/cm2
0.2
0.0
0.0 0.2 0.4 0.6 0.8 1.0
V'
Figure 7. The normalized I V curves for the samples under the proton, the electron, and the
combined irradiations with 170 keV
Figure 8 shows the DLTS spectra for the samples irradiated with 150 keV protons
for various fluences. With increasing fluence, the DLTS signals are reduced and
the peak width increases. The effect of proton energy on the DLTS spectrum is
shown in figure 9. The DLTS spectrum shows a more symmetrical distribution
for the 60 keV proton irradiation, while exhibits an obvious asymmetry after the
irradiation with 150 keV protons. In addition, the DLTS signals in the temperature
range of 125230 K are increased noticeably. The change in the DLTS peak
height can be related to the variation of the concentration of the defects at deep
energy levels due to proton irradiation. According to the DLTS analysis results, the
1
Intensity
1- 150 keV 2
21016p+/cm2
2- 150 keV
21013p+/cm2
Intensity 1- 60 keV
21016 p+/cm2
2- 150 keV
21016 p+/cm2
primary defects induced by the protons are believed to belong to the H1 type with
the +0.45 eV energy level. The H1 type defects are generally referred to as the
combination of radiation-induced vacancies with boron atoms, namely the [V+B]
defects [6]. Under the <200 keV proton irradiation, a large amount of vacancies
could be formed in the base region close to the pn junction of the backfield Si
cell samples, and the vacancies would interact with the nearby boron atoms in the
silicon, forming the [V+B] type defects. The interaction of the radiation-induced
vacancies with boron atoms can provide the thermodynamic driving force for the
segregation of boron atoms in the base region close to pn junction. As a result, a
resistance layer with high concentration of the [V+B] defects near the pn junction
could be formed due to the proton radiation. The formation of such a resistance
layer can contribute to the decrease in electric properties for the backfield Si solar
cells irradiated with <200 keV protons.
4. Conclusions
Solar cells are key elements of the system providing the electric energy in space-
craft. It is of theoretical and practical significance to thoroughly study the damage
effects of charged particles on solar cells under space environment. The samples
for the backfield Si solar cells were irradiated by protons and electrons with energy
of <200 keV. It is found that the proton irradiation leads to remarkable decreases
in the short-circuit current and open-circuit voltage. No obvious changes appear
in the electric properties after irradiation by electrons. The radiation effects of the
electrons are recoverable after the irradiation. The effect of the combined radiation
of protons and electrons does not show simple additivity. The damage extent of
8 ZHENGYU HU ET AL.
proton radiation is larger than the combined radiation under the same proton and
the lower electron fluence. In contrast, the combined radiation could result in more
severe damage if the electron fluence is higher. The damage of the backfield Si
solar cells is primarily caused by the protons, and could be related to the formation
of the [V+B] type defects at +0.45 eV level and the resistance layer in the base
region close to the pn junction.
Acknowledgment
The financial support of the National Basis Research Foundation of China under
Grant # G19990650 is greatly appreciated.
References
1. Hisamatsu, T., Kawasaki, O., and Tsukamoto, K. (1998) Radiation Physics and Chemistry 53,
25.
2. Yeh, Y. C. M., Chu, C. L., Krogen, J., and Ho, F. F. (1996) In Proceedings of the 25th IEEE
Photovoltaic Specialists Conference, IEEE, New York, p. 187.
3. La Roche, G. J., Schultze, W., Rizo, I., and Bogus, K. (1991) Assessment of Third Generation
Solar Cells, NASA N92-13223, pp. 487494.
4. Barabanenkov, M. Y., Gyulai, J., Leonov, A. V., and Mordkovich, V. N. (2001) Nuclear Instrument
and Methods in Physics Research B 174(3), 304310.
5. Hu, Z., He, S., and Yang, D. (2004) Nuclear Instrument and Methods in Physics Research B 1217,
321326.
6. Hazdra, P., Brand, K., and Vobecky, J. (2002) Nuclear Instrument and Methods in Physics Research
B 1192, 291300.
SOLAR ARRAY ARCING IN LEO
Abstract. It is often said that only the solar array or spacecraft surfaces that can
be reached by an arc plume are discharged in a solar array arc in LEO (low Earth
orbit). We present definitive results from ground test experiments done in the
National Plasma Interactions (N-PI) facility at the NASA Glenn Research Center
that this idea is mistaken. All structure surfaces in contact with the surrounding
plasma and connected to spacecraft ground are discharged, whether the arc plasma
can reach them or not. Implications for the strength and damaging effects of arcs
on LEO spacecraft are discussed, and mitigation techniques are proposed.
Key words: solar arrays, low Earth orbit, arcing, plasma interactions, arc plumes,
mitigation
1. Background
Modern solar arrays have areas of tens of square meters, and they operate with
bus voltages exceeding 100 V. Electrostatic discharges (arcs) are undesirable and
detrimental events for spacecraft function, and preventing these events and/or
mitigating their consequences are of primary importance for spacecraft designers.
There are two types of arcs that may occur in space. The first, the primary or trigger
arc, is a transient event that discharges some spacecraft capacitance. The second
type, the sustained arc, can occur between two solar array strings, and may be
powered by the arrays current generating capacity. In this paper, we consider only
the transient trigger or primary arcs.
Ground tests of small samples of large solar arrays have been used to provide
the necessary information regarding arc inception voltages and expected arc dam-
age for an entire array during its lifetime in space. However, the volume of the
space plasma and the size of the test arrays that may be simulated in ground tests
9
J.I. Kleiman (ed.), Protection of Materials and Structures from Space Environment, 919.
C 2006 Springer. Printed in the Netherlands.
10 D. C. FERGUSON ET AL.
is limited by the size of the test vacuum chamber, and this fact necessitates the
installation of additional capacitance between the sample and ground to simulate
the actual capacitance of a spacecraft and its solar array discharging through the
arc plasma. The magnitude of the capacitance to be added has been the subject
of discussions for many years (see for example, [1]). If the discharge of a space
solar array capacitance is caused by an arc plasma front propagating along the
array surface (see [2, p. 31]), this magnitude is limited to about 1 F, because the
array capacitance is approximately 0.25 Fm2 , and the propagation distance of
the dense arc plasma is less than a few meters under the conditions of a typical
LEO plasma.
On the other hand, if the entire array discharges through a current channel
created by an arc, this capacitance can even reach 103 F. The amplitude and
width of an arc current pulse are both increasing functions of the capacitance
discharged, and that is why the damage inflicted on the solar array by an arc
depends on the capacitance discharged in the arc. Is it the capacitance reached by
the dense arcplasma front, or is it the much larger capacitance of the coverglasses
of the entire array that is discharged?
The experiments described below confirm that the entire array capacitance
discharges through the arc current channel even under conditions when the arc
plasma front is prevented from propagating along the sample surface. Thus, the
proper value for an additional capacitor must be high (103 F) for ground tests
of arrays in order to properly simulate the damaging effects that may occur for
arcs on spacecraft with large arrays.
2. Experimental Setup
All of our tests were done in the National Plasma Interactions Facility (N-PI
Facility) at the NASA Glenn Research Center, Cleveland, Ohio. In our tests, the
LEO space plasma was simulated in a large vacuum chamber (2 m in diameter
and 3 m high) equipped with four oil diffusion vacuum pumps that provided a
background pressure about 0.5 Torr (66.6 Pa). One Kaufmann-type plasma
source generated a xenon plasma with an electron temperature of 11.3 eV, an
electron number density of (45) 105 cm3 , and a neutral gas pressure of about
50 Torr (6.67 mPa).
Two solar array samples (on fiberglass) were mounted on an aluminum sheet
and installed vertically in the middle of the chamber (figure 1). One sample (strings
1, 2, and 3) represented a silicon solar array with UVR coverglasses of 300 m
thickness that corresponds to a capacitance of 4344 pFstring1 . Another sam-
ple (strings 4, 5, and 6) had a capacitance of 7020 pFstring1 because of its
thinner coverglasses (150 m). The additional capacitor was chosen to have a ca-
pacitance of C = 0.03 F (10%) for the convenience of measurementssuch a
choice provided comparable currents in all branches of the bias circuit. However,
SOLAR ARRAY ARCING IN LEO 11
Figure 1. Two solar array samples are mounted on an aluminum sheet and installed vertically in a
large vacuum chamber. All dimensions are shown in mm
some measurements were done with a higher capacitance (0.25 F) to reveal the
dependence of the arc current pulse characteristics on the value of this capaci-
tance. Four current probes provided measurements of discharge currents flowing
in essential branches of the circuit (figure 2). For the second series of measure-
ments, a grounded aluminum plate was installed between the samples to prevent
the propagation of the arcplasma front from one sample to the other.
All four current pulse waveforms were registered by a four-channel digital os-
cilloscope and stored in a computer for further processing. Twenty events (arcs)
were observed for each configuration (positions of keys 14, and capacitance C).
That amounted to 260 files, one of which is shown in figure 3. Each file was used
to obtain the following data: (1)I p peak arc current; (2) 0.5 pulse width at
0.5 of the peak current value; (3)qi net electrical charge flowed through the
corresponding branch; and (4) tij time interval between current pulse peaks in
the different circuit branches.
The magnitude of the net electrical charge flowing through a branch was
calculated as
qi = Ii (t) dt (1)
12 D. C. FERGUSON ET AL.
Figure 2. Four current probes were used to measure the discharge currents in four different branches
of the circuit. The additional capacitance was chosen to be 0.03 F to obtain comparable current
magnitudes for all probes. A few measurements were done with C = 0.25 F
Then, the average and standard deviation over several measurements were calcu-
lated, and the resultant value was compared with the theoretically predicted value.
For example, the ratio of charges for string #1 (q1 ) and capacitor (q2 ) was
calculated by
q1 Cstring
= (2)
q2 C
Figure 3. One example of an arc pulse current sequence for C = 0.03 F, and for switches K1 and
K2 in the closed position. CP1 is the most positive current. The arc was initiated at time t = 0
SOLAR ARRAY ARCING IN LEO 13
It should be noted that the possible errors in the calculations of string ca-
pacitances could not be estimated properly because of unknown errors in the
corresponding geometrical and electrical parameters. However, the consistency
of all or our final results is a very convincing argument that the calculations of
array capacitances were done with an error of less than 10%. Also, the following
ratio:
qarc Cstring + C
= (3)
q2 C
was verified for all events when the experimental setup made it possible to do
so. This equality means that the array capacitance that discharged through the
arc current channel is independent of the distance between the arc site and other
strings, and the array discharged fully with or without an aluminum plate installed
between the two samples.
The results of our measurements and theoretical estimates are compiled in
table 1. Standard deviations (1 ) of the measurements are shown in parenthesis.
The numbers shown in table 1 demonstrate a very good agreement between
the measured parameters and their theoretical estimates. We believe that some
insignificant differences can be explained by our poor knowledge of the string
capacitances, possibly by a somewhat incomplete discharge of the panels, and
possibly by a somewhat inhomogeneous plasma potential distribution. However,
the considerably smaller-than-expected discharge of string #6 observed during
two different runs (run numbers 6 and 8 above) cannot be explained to date. These
results look particularly strange if one takes into account the very good agreement
between the measurements for strings #2/3 and their estimated values (run numbers
5 and 7 above), because these two runs were supposed to be symmetrical to each
other.
14 D. C. FERGUSON ET AL.
In the second stage of the experiment, an aluminum panel was installed between
the two samples to prevent the propagation of the arc plasma from the arc site to
the other sample (figure 1). Measurements were done with the same additional
capacitor (0.03 F, 10%) and with another capacitor (0.22 F, 5%) connected
in parallel with the first one. The results are shown in table 2.
It can be seen from the data in tables 1 and 2 that those strings that were
not arcing discharged fully in both caseswith or without the aluminum panel
between the samples. Thus, the mechanism of discharge of an entire array can only
be explained by an electron current flowing from the negatively charged conductor
(or semiconductor) to the surrounding plasma through the arcplasma conductive
channel. The positive charge of the coverglass is neutralizing by ambient plasma
electrons that are attracted by the positive potential of the coverglass.
The dependences of the arc current pulse width and amplitude on the net
capacitance were found from the experiments shown in table 1. However, the
narrow range of capacitances used (0.0420.064 F) and large deviations in
the measured values did not allow us to confirm (or to reject) any expected
square root dependence (figure 4). For a discussion of the square root dependence,
see [3].
For some measurements (shown in table 2), a bigger additional capacitor
(0.25 F) was used, and this provided the opportunity to verify the expected
dependence of the pulse width on the capacitance (figure 5). It turned out that this
dependence is slower than an expected (about 0.3 in power-law index, rather than
the expected 0.5 as Snyder depicted in [4].
One more interesting feature of the discharge process is a time delay between
the instant of the peaks in the arccurrent pulse and in the discharge current
of those strings not arcing (figure 3). We believe that this delay is caused by a
changing plasma potential during the discharge process (which corresponds to
the spacecraft potential for LEO orbit). In actuality, in the simple situation when
all keys (K1K4) are open, and the arc occurs on string #1, the relaxation current
SOLAR ARRAY ARCING IN LEO 15
Figure 4. Arc current amplitudes and pulse widths vs. net capacitance are shown for the experiments
without a conducting plate between the samples. Pulse width measurements are at top and peak
current measurements are at bottom. Units are as in the legend at top
Figure 5. In spite of large deviations, a square root dependence of pulse width on capacitance can
be excluded. The actual dependence is closer to a C 1/3 dependence
16 D. C. FERGUSON ET AL.
Figure 6. The theoretical time delay (in x units) between peaks of arc current and string discharge
current is similar to the observed one
If the arc current pulse is simulated by two exponents or by a Gaussian curve, the
solution of eq. (5) is shown in figure 6.
The measured time delay between current peaks in strings 1 and 6 was
1. with 280 V, no plate, 0.03 capacitor2.6 (0.8) s;
2. with 450 V, plate, 0.03 capacitor4.88 (1.68) s.
3. Conclusion
The results of our current experiments and their analysis confirm the necessity of
using a large additional capacitance (0.30.5 of the expected entire spacecraft solar
array capacitance) in ground tests in order to adequately simulate the consequences
of arcing on solar arrays in orbit. Spacecraft solar arrays that have very large
capacitances connected to spacecraft ground may be damaged by even transient
arcing events unless mitigation techniques are used. For example, a spacecraft
solar array with a 300 V power system and 10 F surface capacitance may be
subject to arcs of more than 1000 A peak strength and total energies greater
than 0.4 J, even if the arcs do not become sustained by the array power. Miller
[5] showed that even arcs of 40 A peak strength may severely damage solar cell
interconnects.
The design of a solar array must consider the plasma environment and interactions
with that environment. Arc prevention is extremely important. The following
techniques have been shown in ground and flight tests to prevent arcs or minimize
their damage:
SOLAR ARRAY ARCING IN LEO 17
1. If possible, use array string voltages of less than 55 V. No trigger arcs have
been seen on LEO arrays of less than about 55 V string voltage even under
simulated micrometeoroid bombardment. Solar arrays coming out of eclipse
will generate more voltage than when they operate at their max power point.
2. If solar array cell edges or interconnects are exposed to the LEO plasma and
string voltages are greater than 55 V, the strings should be laid out on the
substrate such that no two adjacent cells have a voltage difference of greater
than 40 V. Sometimes a leapfrog arrangement will be sufficient. In other high-
voltage arrays, the strings should be arranged parallel to each other. Serpentine
strings can be used to prevent the array width from becoming prohibitive. If
the string layout cannot be modified to prevent cells with more than 40 V
difference being adjacent to each other (anything less than about 1 cm may
be considered adjacent) then the total string voltage must be kept low enough
that the initial (trigger) arcs do not take place. The lowest known array trigger
arcing has occurred on thin-coverglass cells at about 75 V ([6], PASP Plus
results).
3. For array string voltages greater than about 75 V, trigger arcs in LEO can only be
completely prevented by encapsulating the cell or array edges so they do not see
the ambient plasma. If encapsulation is not possible, a thorough array bakeout
on-orbit (1 week at 100 C or more) may get rid of contaminants and prevent
trigger arcing up to about 300 V, or possibly more (see [7]). Recontamination
may occur on dirty spacecraft (spacecraft with excessive venting, cold gas
nozzles, etc.). Good encapsulation may prevent arcing up to 1000 V string
voltage.
4. Sustained (or continuous) arcs may occur whenever trigger arcs occur and
adjacent cells have more than 40 V potential differences. However, sustained
arcs, in addition to this voltage threshold, have a current threshold, below
which they will not occur. It is believed that the current threshold is greater
than about 0.5 A. If the current produced by each cell is above this threshold, a
single string may sustain arcs. If each cell is below this current threshold, then
isolating separate strings of solar cells from each other will prevent other strings
from feeding the arc site, and will prevent sustained arcs. This isolation can
be achieved by using blocking diodes in each string (EOS-AM1, now called
Terra, e.g., [8], for instance). Care must be taken that the power bus and/or other
components do not have the conditions necessary for sustained arcing. On the
Terra arrays, for instance, it was found that diodes used to block interstring
currents did not prevent the bus power traces from having sustained arcing
events. Covering all exposed bus conductors with Kapton insulation finally
solved the problem. Low-outgassing RTV may be used to cover bare conductors
as well.
5. RTV grout between adjacent solar cells and strings that have a high voltage with
respect to each other has been shown to effectively block sustained arcs between
18 D. C. FERGUSON ET AL.
cells and strings. The degree of coverage, etc., is important in determining the
final voltage threshold for sustained arcing.
6. Arrays of 300 V and greater string voltage must be fully encapsulated in order
to prevent arcing.
7. Finally, although design and construction are important in preventing trigger
arcs and sustained arcs, each new solar array implementation must be tested
in a simulated LEO plasma before it can be sure not to arc. This step must
not be omitted. The test bias voltage relative to the plasma should include the
maximum when the arrays come out of eclipse (or the highest potential expected
on the floating spacecraft). The interstring voltage should be at least as great
as that expected anywhere on the solar array on-orbit. Tests should ideally be
conducted at sample temperatures as low as the eclipse-egress temperature.
On the International Space Station (ISS), ground testing of the solar arrays showed
that the arrays did not arc at negative voltages less than about 250 V. This is
much less than the system operating voltage (160 V), so in the absence of auroral
charging, ISS is never expected to charge more than 160 V negative. And, if trigger
arcs do not occur, sustained arcs will not. However, ISS did take the precaution
of incorporating a system of PCUs (plasma contactor units) to prevent structure
charging by the solar arrays and to prevent dielectric breakdown of anodized
aluminum surfaces.
The LEO design guidelines [2] referenced above were based partly on ground
tests done for the SS/Loral satellite series and the EOS-AM1 spacecraft (now
called Terra). Mitigation techniques called out above were partly the result of
those tests, and therefore should be considered to be based on real spacecraft
experience in failure modes and failure investigations.
Kapton is a registered trademark of Dupont Co.
References
1. Levy, L., Reulet, R., Sarrail, D., and Migliorero, G. (1998) In 6th Spacecraft Charging Technology
Conference, Hanscom Air Force Base, MA, AFRL-VS-TR-20001578, pp. 4348.
2. Ferguson, D. C. and Hillard, G. B. (2003) Low Earth Orbit Spacecraft Charging Design Guide-
lines, NASA/TP2003-212287.
3. Galofaro, J., Vayner, B., Degroot, W., and Ferguson, D. (2002) The Role of Water Vapor and
Dissociative Recombination Processes in Solar Array Arc Initiation, NASA TM -2002-211328.
4. Purvis, C. K., Ferguson, D. C., Snyder, D. B., Grier, N. T., Staskus, J. V., and Roche, J. C. (1986)
Environmental Interactions for Space Station and Solar Array Design, Preliminary, NASA Lewis
Research Center, OH.
SOLAR ARRAY ARCING IN LEO 19
Abstract. The following paper presents multilevel protection mechanism for the
digital processing circuits, based on partially reconfigurable FPGA devices and
run-time reconfiguration of processing architectures in the FPGA device. This
approach assumes that if a hardware fault occurs in one of task processors, it can
be restored without interruption of any computing processes in the rest of FPGA
circuitry. The novel procedures of function restoration with and without perfor-
mance degradation are proposed. These procedures allow automatic restoration
of Application Specific Virtual Processors inside the FPGA device with mini-
mization of performance degradation when a fault occurs. Implementation of this
method can dramatically increase lifetime of the on-board computing platform
and at the same time reduce restoration time period.
1. Introduction
It is known fact [1, 2] that data-stream processing architecture can reach the
highest cost-effective parameters, if it reflects the Data Flow Graph (DFG) struc-
ture of the task. In [3] was proposed the approach where each DFG node called
SELF-RESTORATION AS SEU PROTECTION MECHANISM 23
Global Processing V
Routing Element (PEi) H
Lines C
Local routing
Interface
Element (IEj)
Tri-
state
Buffer
A B C D A B C D Y
+
+
X
X
%
Y
Global Routing
Lines (GRL)
Figure 2. ASVP on-chip architecture (micro-architecture) assembled from VHC EGO blocks in
accordance to task data-flow graph
global routing lines, will be connected to those lines but initially tri-stated at the
initial architecture loading state.
In [6] we proposed an approach where virtual hardware components are
LEGO-type building blocks that allow self-assembling of task optimized applica-
tion specific virtual processor (ASVP), inside the partially reconfigurable FPGA.
To illustrate this concept in figure 2the on-chip architecture (micro-architecture)
of Application Specific Virtual Processor assembled from virtual hardware com-
ponents is presented (figure 2(b)) according to task data-flow graph (figure 2(a)).
There are a few main reasons for hardware faults that can occur in FPGA-based
devices: (i) physical defects in wafer; (ii) hidden manufacturing defects; and (iii)
radiation effects. In our project, we considered faults occurring in small amount of
CLBs in the SRAM based on FPGA devices. Mostly, this type of faults occurs as a
result of radiation effects (SEUsingle event upset, SELsingle event lutch-up,
etc.) usual for aerospace applications [7] or applications for radiation intensive
environment (nuclear power stations, etc). As an alternative approach based on
protection of materials from radiation intensive environment, we proposed self-
restoration mechanism based on reassembling of damaged processing structures.
This mechanism utilize so-called scrubbing procedure [7], but allows functional
restoration to be much faster. In this paper we will give only a brief description of
proposed method. To restore ASVP we proposed a three-level procedure:
SELF-RESTORATION AS SEU PROTECTION MECHANISM 25
A B C D Y A B C D Y
+ +
X *
% %
Figure 3. VHC replacement using spare CLB-slots. In this example a fault is detected in the CLB-
column where Multiplier VHC was located. The dumm VHC (indicated as ) is loaded to the
faulty CLB-slot and then Multiplier VHC is loaded to one of spare CLB-slots
SELF-RESTORATION AS SEU PROTECTION MECHANISM 27
In this situation let us consider that all spare CLB-columns were already used or
were not reserved at all. We also should put in account the fact of hardware fault
in some CLB-slots, which means that one or few transistors in one or few logic
gates are damaged. A CLB-column consists of 20,000 up to 50,000 transistors
(depending on FPGA device specifics). Therefore, damaged logic gates can be
avoided by reloading other variants of virtual hardware componentRi, which
requires less logic gates. In result we have considered two possible solutions to
restore ASVP with performance degradation:
a) First approachMinimization of restoration timeassume reloading an-
other variant of the same VHC (e.g., Multiplier), which requires less area, to
the same CLB-slot. This approach gives the shortest restoration period and a
simplest restoration control scheme. However, this solution may not provide
minimum of performance degradation. For example, if a fault appears in a
CLB-slot where Multiplier is located it may be more efficient to relocate Mul-
tiplier as is to another slot and load the damaged slot by a smaller variant of
another VHC (e.g., Adder).
b) Second approachMinimization of performance degradationassume that
before functional restoration an optimization of ASVP architecture to the task
algorithm and data structure should be done. This procedure should select the
variant of ASVP architecture which satisfies new (after fault) resource limita-
tions and allows minimization of the performance degradation (e.g., latency,
data processing rate, etc.).
Firstly the approach which allows minimization of the restoration time was con-
sidered. The restoration procedure consists of the following steps:
1. Suspension of damaged ASVP pipeline.
2. Selection of one of VHC variants with reduced logic area and performance
(usually in this VHC some part is dummy and thus different VHC variants
has different location of the dummy component).
3. Loading configuration bit-stream of selected variant of VHC to the damaged
CLB-slot.
4. Continuation the data-stream processing on restored ASVP.
5. If hardware failure still affects data-stream processing the next VHC variant
with reduced area and performance should be selected from the VHC table.
6. Return to the point (iii) of presented procedure.
28 LEV KIRISCHIAN ET AL.
R1
Ri Ri Ri
1 2 3 4 5 6 7 8 9
1 2 3 4 5 6 7 8 9
Figure 4. Architecture configurations graph (arranged design space) for a ASVP (top tree) and
performance diagram associated with ASVP variants
SELF-RESTORATION AS SEU PROTECTION MECHANISM 29
5. Implementation
Figure 5. Architecture of reconfigurable parallel stream processor (RPSP) for multitask and multi-
mode workload
of I/O interface was equal to 7.2 Gbs1 (LVDS) and 8.5 Gbs1 (LVTTL). Cache
volume = 1 M 16 bits. Configuration bus bandwidth = 528 Mbs1 (33 MHz
16 bitcc1 ).
6. Experimental Results
To test and adjust ASVP restoration procedures a set of faulty virtual hardware
components (FVHCs) was developed for parallel video-stream processing tasks
such as: edge detection, brightness-to-colour conversion and image combination.
The experiment included
1. Initiation of ASVP and processing test-data stream.
2. Loading FVHC in selected CLB-slot at random moments of time.
3. Measurement of hardware fault detection time.
4. Measurement of scrubbing time till ASVP restoration.
5. Repetition of steps (ii)(iv) to initiate second level restoration.
6. Initiation of CLB-slot replacement to the spare CLB-slot.
7. Measurement of CLB-slot replacement time.
Results of the above experiments are summarized in table 1. We have measured
restoration periods for level 1 (VHC-scrubbing) and level 2 (CLB-slot replace-
ment). Based on this information we calculated the acceleration of the functional
restoration comparing with common approach when an entire FPGA device has
to be reprogrammed (see table 1).
These results show that restoration time directly depends on size of VHC.
Restoration time is minimal if VHC can be allocated in one CLB-column and
increases nearly linear when VHC requires 2, 3, or more CLB-columns. Another
result is that CLB-slot (column) scrubbing (1st level of restoration) works twice
faster than CLB-slot replacement. This effect appears because of the extra time
needed for a dummy VHC loading. Thus, for maximum acceleration (comparing
with the scrubbing of entire FPGA [8]) designers should develop a virtual hardware
components close to 1 CLB-column. This can dramatically increase the restoration
time (up to 70100 times depending on FPGA volume and organization), for the
reason that it was mentioned in [8] most of hardware faults in LEO and GEO are
caused by the SEU which can be restored by scrubbing procedure.
7. Summary
The major goal of the project was to find a mechanism that would be able to
restore FPGA-based computing platforms functionality quickly, when hardware
faults occur to minimize data corruption. Ideally, this restoration should not cause
performance degradation but it is assumed that if there is no other choice, the
system is better to decrease its performance characteristics rather then stop func-
tioning. Thus, a multilevel self-restoration mechanism was developed based on
the idea of self-assembled task optimized application specific virtual processors
(ASVP). These processors being just cores can be assembled on-chip from the
LEGO-type sub-coresvirtual hardware components (VHC). Furthermore, be-
cause ASVP is VHC-based devices they can be reassembled inside the FPGA
when hardware fault occurs in one of VHCs. Established on this idea special
procedures were developed for temporal faults caused by radiation (e.g., SEU) or
static faults (e.g., physical defects in the wafer). All procedures were implemented
on the reconfigurable platform based on the Xilinx Virtex FPGA (XCV-400 E).
It was shown and seen that these procedures can provide a much faster (in or-
ders) restoration of FPGA-based computing systems in comparison to a common
approach.
References
1. Micheli, G.D. (1994) Synthesis and Optimization of Digital Circuits, McGraw-Hill, Inc., New
York, 580 p.
2. Rixner, S. (2002) Stream Processor Architecture, Kluwer Academic Publishers, Nozwel, MA,
USA 120 p.
3. Kirischian, L. (2000) In Proceedings PARELEC 2000International Conference on Parallel Com-
puting in Electrical Engineering (IEEE #PR00759), Trois-Rivieres, Quebec, Canada, August
2000, pp. 100105.
4. XILINX Virtex II Platform FPGA Handbook VG002 v.1.0, 6 December, 2000.
5. XAPP151 v1.6. Virtex Series Configuration Architecture User Guide, Xilinx Inc., 2003.
6. Kirischian, L., Szajek, L., and Chayab, F. (2002) In Proceedings PARELEC 2002 Inter-
national Conference on Parallel Computing in Electrical Engineering, Warsaw, September
2002.
SELF-RESTORATION AS SEU PROTECTION MECHANISM 33
7. XAPP 216 v 1.0. Correcting Single Event Upsets through Virtex Partial Configuration, Xilinx
Inc., 1 June, 2000.
8. Carmichael, C. (2001) Triple Module Redundancy Design Techniques for Virtex FPGAs, Xilinx
Application Note XAPP197 (v 1.0), November 2001.
9. Abramovici, M., Breuer, M., and Friedman, A. (1994) Digital Systems Testing and Testable
Design, Wiley-IEEE Computer Society Press.
SYNERGISTIC EFFECT OF PROTONS AND ELECTRONS ON
RADIATION DAMAGE OF METHYL SILICONE RUBBER
1. Introduction
is introduced to discuss the additivity for the combined radiation effect of protons
and electrons:
P(p+e)
K add = (1)
Pp + Pe
where, K add is the additivity coefficient; P(p+e) is the property change due to
combined radiation of protons and electrons; Pp is the property change due to
proton radiation; Pe is the property change due to electron radiation. It can be
deduced that there is an additivity for the influence of the combined radiation
on material property when K add = 1, while the additiivity does not exist when
K add = 1.
2. Experimental
The silicone rubber with molecular mass of 68,000 was used as the experimen-
tal material, which was treated in vacuum, and then incorporated with 23 wt%
Si(OH)4 and 35 wt% Be2 Sn(OCOC11 H23 )2 . After mixing homogeneously, the
material was degassed for several minutes in vacuum, poured into a tetrafluo-
roethylene mould, and then vulcanized at room temperature.
The irradiation experiment was done in a system that can simultaneously
and separately simulate the radiation of electrons and protons with 30200
keV. In the test, the energy of protons and electrons was chosen as 180 keV,
the fluence of protons and electrons was 1016 cm2 , the vacuum in the sample
chamber was 105 Pa, and the surface temperatures of irradiated samples were
controlled at 10 5 C by a cold screen with liquid nitrogen. Sample thickness was
2 mm.
The sample surface morphology was evaluated using an OLYMPUS BH2-
UMA type optical microscope. The samples were coated with aluminium in vac-
uum for measuring the volume resistivity v , the dielectric coefficient and the
loss angle tangent tg . The v was measured using a ZC-36 type meter. The tests
for and tg were conducted using a TR-10C transformer bridge. The samples
were analyzed by a JES-FE3AX type electron spin resonance spectrometer made
by JEOL, Japan.
Figure 1(a) shows that the sample surface is smooth and flawless. After proton
irradiation, the surface colour of the silicone rubber darkened and aging cracks
SYNERGISTIC EFFECT OF PROTONS AND ELECTRONS 37
Figure 1. Change in surface morphology of silicone rubber before and after irradiation: (a) before
irradiation, (b) p = 1016 cm2 , (c) e = 1016 cm2 , (d) e = p = 1016 cm2
appeared, as shown in figure 1(b). Our previous work demonstrated that the aging
degradation of the silicone rubber surface occurred under the protons irradiation
with 180 keV to a fluence of 1016 cm2 . After the electron irradiation, the surface
colour of the silicone rubber was also darkened, and the arborescent discharge
stripes appeared on the surface, as seen in figure 1(c). Measurement of the surface
potential showed that the silicone rubber surface was negatively charged to 12 kV
under the electron irradiation. It is believed that the discharge is a main mode of
surface damage under the electron irradiation. After the combined irradiation by
protons and electrons, the surface colour of the silicone rubber was darkened
further and aging cracks and more arborescent discharge stripes were observed at
the same time, as shown in figure 1(d). There is a combined effect of the surface
damage modes for the two types of charged particle irradiations after the combined
radiation.
TABLE 1. The change in mass loss ratios and tensile fracture strength under different irradiation
conditions.
70
73
60
50
40
a.u.
57
30
16
20 15
2 14
1
10
0
0 10 20 30 40 50 60 70 80 90
Mass to charge ratio, m/z
Figure 2. Mass spectrum of atmosphere in the sample chamber during combined irradiation of
protons and electrons for silicone rubber
loss ratio for the combined irradiation is the greatest; the second for the proton
and the lowest for the electron irradiation. However, the additivity does not exist,
because K add = 1.18.
The in situ mass-spectrometric analysis showed that the micromolecule
gaseous products such as H2 (m/z:2.1, 1.1) and CH4 (m/z:16.2, 15.2, 14.2) were
formed during the proton, the electron and the combined irradiations [6, 7]. Also,
the gaseous product CH3 SiOCH3 (m/z:73.3, 57.3) was found during both the pro-
ton and the combined irradiations, as shown in figure 2. This could be attributed to
the protons participating in the destruction of macromolecular chains of silicone
rubber.
TABLE 2. Changes in electric properties of silicone rubber under different irradiation conditions
v 1013 tg /
Irradiation condition cm v /v % /% tg 103 tg %
Figure 3 presents the electron spin resonance (ESR) spectra for the silicone rubber
after the proton, the electron and the combined irradiation. In the figure, the g value
g=2.0036
g=2.0036 g=2.0036
50G
Figure 3. ESR spectra for silicone rubber after the proton, electron and combined irradiation: (a)
p = 1016 cm2 , (b) e = 1016 cm2 , (c) e = p = 1016 cm2
40 LIXIN ZHANG ET AL.
was equal to 2.0036, which represents the overlapped peak for the free radicals of
a a
!Si and !SiO . The fourfold split peak for the CH3 and the peak from ternary
split for the a
!SiCH2 do not appear, because of faster attenuation rates of these
free radicals. Some macrochains of the silicone rubber could break down due to
radiation. After the irradiations, these free radicals attenuate automatically. Figure
3 shows that the peak for the combined irradiation is the highest, implying that
the concentration of surviving free radicals is much higher than those for the
independent irradiation. Therefore, the decrease of the v is more obvious after
the combined irradiation.
The changes in the infrared spectrum of methyl silicone rubber before and after
the proton, the electron, and the combined irradiation are shown in figure 4.
Although no new peaks appear, the relative intensity of the peaks changes
after the irradiations. The peaks at the wave numbers of 800 cm1 and 1259 cm1
originate from the vibration of CH3 groups, and the overlapped peaks around the
800 cm1 might be related to the nonplanar vibration absorption of CH bonds. The
absorption peaks at 2962 cm1 and 2900 cm1 are due to the stretch vibration of
CH bonds, and the 1070 cm1 absorption peak was caused by the SiO stretching
vibration in the Si-O-Si structure. As seen in the figure, the height of all the peaks
decreases after the proton, the electrons and the combined irradiation, showing
that the degradation occurs for both the macromolecule main chain SiO bonds
and the lateral chains CH3 . Moreover, the decreasing extent for all the peaks is
larger for the combined irradiation, indicating that the combined one would lead
to more severe degradation effect on the methyl silicone rubber.
Figure 4. Changes in infrared spectrum of silicone rubber before and after the proton, electron,
and combined irradiation: (a) before irradiation, (b) p = 1016 cm2 , (c) e = 1016 cm2 , (d)
e = p = 1016 cm2 , (e) e = p = 2 1016 cm2
SYNERGISTIC EFFECT OF PROTONS AND ELECTRONS 41
4. Conclusions
The above results show that after the combined irradiation of protons and electrons,
the changes in the surface morphology for the silicone rubber show a combined
feature of the two independent radiation effects. The changes of tensile strength,
mass loss ratio, and electrical properties are not following simple additivity. The
combined radiation results in a more severe degradation effect on the methyl
silicone rubber, which is primarily responsible for the decrease in mechanical
properties and the increase in mass loss ratios. Under the combined radiation, the
formation of free radicals, due to the scission of molecular chains, leads to the
decrease in electrical properties of the methyl-silicone rubber.
References
1. Denkins, P., Badhwar, G., and Obot, V. (2001) Acta Astronautica 49(3), 313319.
2. Stevenson, I., David, L., Gauthier, C., Arambourg L., Davenas J., Vigier G. (2001) Polymer 42,
92879292.
3. Badhwar, G. D. and Oneill, P. M. (2001) Nuclear Instruments and Methods in Physics A 466,
464474.
4. Yibing, Z., Chunxiao, X., Wenyan, L., and Weiquan, F. (2003) European Space Agency, (Special
Publication) ESA SP 540, pp. 723725
5. Chang-Su, H., Bok-Hee, Y., and Sang-Youb, L. (2000) In Proceedings of the IEEE International
Conference on Properties and Applications of Dielectric Materials, Xian, China Vol. 1, 2000,
pp. 367370.
6. Zhang, L. X., Yang, S. Q., and He, S. Y. (2003) Chinese Journal of Polymer Science 21(5),
563567.
7. Zhang, L. X. and He, S. Y. (2004) Materials Chemistry and Physics 83, 255259.
INFLUENCE OF ELECTRON RADIATION ON OUTGASSING
OF SPACECRAFT MATERIALS
R. H. KHASSANCHINE,1 A. N. TIMOFEEV,1
A. N. GALYGIN,1 V. I. KOSTIUK,1 AND V. M. TSVELEV2
1
Joint-stock Company Kompozit, 4, Pionerskaya street, Korolev,
Moscow region, Russia
2
Lavochkin Association, Moscow, Russia
1. Introduction
Ci (x, t) 2 Ci (x, t) M
= Di i
rad
j C i (x, t) i C i (x, t) + Si (x, t)
rad rad
(1)
t x2 j=1
40
a 0.1
b
1
30
Ea (keV/electron)
2
E0=20; 30; 40; 50 keV
P (Gy*s 1)
3 0.01
20
4
10 1E-3
0
0 1 2 3 4 5 0 20 40 60 80 100
x (mg/sm2) x (m)
Figure 1. Distribution of the absorbed energy in ECOM-1: (a) for various values of electron energies
in laboratory experiments: 120 keV, 230 keV, 340 keV, 450 keV; (b) in material that is in service
in GSO
[4]. Curve 1(a) gives a distribution of the absorbed energy for various values of
electron energies in laboratory experiments, curve 1(b) gives a rough distribution
of the absorbed dose of electron radiation when material is in service in GEO.
It was found from the computational results of absorbed energy of electron
radiation in spacecraft coatings when in service that the VP source function can
be represented in the form
Sirad (x, t)i = Ai0 exp[0 (x + t h)] + Ai1 exp[1 (x + t h)] (4)
where Ai0 , Ai1 parameters depending on composition of material and electron
radiation spectrum. They are equal in values to the ratio of absorbed dose rate in
the near-surface layer of material to average energy required to generate an i-type
VP molecule in material being under exposure to low-energy and high-energy
spectrum parts of space electron radiation respectively;
0 , 1 effective coefficients of linear reduction of absorbed energy in space-
craft coatings from low-energy and high-energy spectrum parts of space electron
radiation respectively.
Having found the distribution Ci (x, t) of i-type VP in a sample from (1
3), dependences of VP mass in material and the flux through the unit surface of
material-vacuum boundary on time t are respectively determined from expressions
hvt
Msi (t) = S0 Ci (x, t)dx dFi (t)/dt = + ki airad Ci (h + t, t) (5)
0
When calculating, the parameters were chosen in such a way that results would
clearly show the outgassing processes described by the model.
Figure 2 shows influence of electron radiation on distribution of VP concentra-
tion in material at different model parameters. Plots 2(a) and 2(b) show changes of
VP concentration in material at equal values of D, k, , and temperatures in mate-
rial under thermal-vacuum action and when the thermal action is overlapped with
the electron radiation influence respectively ( i effective first-order reaction rate
with i-component.
M
i = i
rad
j + i
rad
)
m
) t (h x(
t (h
) x ( m )
0.8 0.15
0.6 0.10
0.4
0.05
0.2
0.0
t(
) h)
t (h m m)
) x( x (
radiation when there was no such VP in initial material (R = 0), plot 2(c) gives
the eventual case when radiation is responsible for the concentrations of some
types of VP to be reduced.
Figure 3 shows functions of mass loss Ms (t) and outgassing rate dF(t)/dt
from the unit surface of VP source under thermal vacuum action and influence of
electron radiation at various parameter values R = 1 (figure 3(a)), R = 0 (figure
3(b)) with other model parameters being equal. Outgassing rate is maximum at ini-
tial times in the first case and it grows up starting from zero value in the second case.
The mass loss function can be determined from
N
hvt
a b
R =1 dMsrad R =0 dMsrad
10
dMs
Ms, dF /dt
Ms, dF /dt
rad
dF/dt
1
rad
dF/dt
dF/dt
0.1
0 20 40 60 80 100 0 20 40 60 80 100
t (h) t (h)
Figure 3. Outgassing kinetics at different initial VP concentrations at electron radiation: (a) R = 1;
(b) R = 0 (D = 0.001, k = 0.01, = 106 , 0 = 0.1, 1 = 0.01, Ai0 = 0.01, Ai1 = 0.002)
The effect of electron radiation that influences the source material at different
remission coefficients kci on VP deposition kinetics is given in figure 4 showing
curves of VP mass deposited on the unit condensation surface Mci (t) by time t.
For large values of kci , the deposited mass of the VP initially grows due to the
significant outgassing rate, reaches maximum, and begins to decrease (figure 4(a)).
The VP mass chemically reacted on the condensation surface can be calculated
with the help of eqs. (6)(8).
Curves in figure 4(b) represent the mass loss Msi (t), the mass deposited onto
a unit of condensation surface Mci (t), and the outgassing rate dF(t)/dt functions
responding to a stepwise change of the model parameters resulted from electron
radiation for t0 = 40 h. Electron irradiation with time leads to the growth of VP
concentration in the near-surface layer due to radiolysis of material that sets off,
to some extent, the fall of the outgassing rate dF(t)/dt and leads to increase of
the mass loss Msi in VP source.
Figure 4. (a) VP deposition kinetics under thermal-vacuum action for different remission coeffi-
cients kc ; (b) responses of Msi (t), dF(t)/dt, and Mci (t) functions on electron radiation started at
t0 = 40 h
INFLUENCE OF ELECTRON RADIATION 49
Figure 5. Outgassing kinetics of ECOM-1 irradiated by 10-keV electrons of different fluxes: (1) =
0; (2) = 5 1015 cm2 , electron flux density 5 1011 cm2 s1 ; (3) = 5 1016 cm2 , electron
flux density 5 1011 cm2 s1 ; (4) = 5 1016 cm2 , electron flux density 5 1012 cm2 s1
4. Experimental Results
5. Conclusion
References
Key words: atomic oxygen, charging, low Earth orbit, synergy, polyimide
1. Introduction
2. Experimental Details
The atomic oxygen source used in this study was a laser detonation atomic oxygen
source. This type of source was developed by Physical Sciences Incorporation [6].
Details of the source used in this study were described elsewhere [7]. The PSI-
type atomic oxygen source was attached to the space environment simulation
facility at Kobe University (figure 1) [14]. The translational energy of atomic
oxygen beam was monitored by the time-of-flight (TOF) measurement system
consisting of a quadrupole mass spectrometer (QMS) and a multichannel scalar.
Mean translational energy of the atomic oxygen was calculated to be 5.0 eV, which
corresponds to the orbital impact velocity of atomic oxygen.
The atomic oxygen flux in a beam was measured by an Ag-coated QCM with an
accommodation coefficient of 0.62 [8]. The principle of measurement is explained
in following section. Since the reaction of atomic oxygen with Ag is a nonlinear
EFFECT OF SURFACE CHARGING ON THE EROSION RATE 53
X-ray
Source
DP-CMA
Polymer-coated QCM
Nozzle
PSV
Laser
QMS
Electron gun Au
Orifice Mirror
Figure 1. Atomic oxygen beam facility used in this study. Laser detonation atomic oxygen source
and high-energy electron gun are equipped
phenomenon, only the initial reaction, which gave a linear mass gain, was used to
calculate atomic oxygen flux [9].
The erosion rate of the polyimide film was measured by the resonant frequency
of QCM, which was coated by a thin film of polyimide. The change in mass of
the polyimide film was recorded during atomic oxygen exposure. The resonance
frequency of QCM is given by the following expression:
f = f 02 W/NA (1)
where, N is the frequency constant, A is the electrode area, is the density of
quartz, and f 0 is the resonant frequency. Since N , A, , f 0 are known factors, one
can calculate the mass change (W ) of the sensor crystal from the frequency shift
( f ) of the QCM. Resonant frequency was measured every 10 s with accuracy
of 0.1 Hz. For a 5 MHz AT-cut QCM sensor crystal used in this study, frequency
resolution of 0.1 Hz corresponds to mass resolution of 2 ng.
The QCM system used in the study was modified to apply bias voltages to the
front surface of a sensor crystal. Since a conventional QCM surface was grounded
in order to avoid ion sputtering effect during PECVD processes, an electrically
isolated manifold was used to float the QCM from the grounded vacuum chamber.
Bias voltage up to 1500 VDC was superimposed on to the driving voltage of QCM
sensor crystal (DC 8V) and the QCM holder. However, the computer interface of
54 MASAHITO TAGAWA ET AL.
the QCM driver was maintained grounded. This electric isolation was achieved
by a specially designed circuit.
The PMDA-ODA polyimide film was formed on a QCM sensor crystal by a
process described below. Precursor of PMDA-ODA polyimide (Semicofine SP-
510, Toray) was spin-coated on a QCM sensor crystal (12,000 rpm, 10 s). Two-
stage cure treatment (150 C, 1 h followed by 300 C, 1 h) in N2 atmosphere was
carried out to create PMDA-ODA structure. The PMDA-ODA film, thus formed,
has a thickness of approximately 0.1 m, and XPS spectrum of the film was similar
to that of commercially available polyimide (Kapton-H). Polysulfone film was also
spin-coated with a solution involving polysulfone pellet (Mw = 63,000) and N, N-
dimethylformamide as solvent. The spin-coated solution containing polysulfone
was dried in air at 70 C 1 h.
3. Results
Figure 2 shows the experimental setup to study the effect of simultaneous electron
beam irradiation on the atomic oxygen-induced erosion of polyimide. Electron
beam (acceleration voltage: 7 kV, filament current: 2.2 A) irradiated the sample
during atomic oxygen exposure. The experimental results for polyimide are shown
Figure 2. Photograph of the experimental set up for the e-beam experiment. A high-energy electron
gun is installed at the top flange of the reaction chamber. QCM was temporally attached to the source
chamber in order to measure atomic oxygen flux
EFFECT OF SURFACE CHARGING ON THE EROSION RATE 55
20
12
10
0
0 100 200 300 400 500
Elapsed Time (s)
Figure 3. Frequency shift of the polyimide-coated QCM under simultaneous exposure of atomic
oxygen and electron beam
Bias voltages from 0 to 1500 V (positive and negative) were applied to polyimide-
coated QCM, and the influence on atomic oxygen-induced mass loss phenomenon
-2
-4
Grounded
Floated
-6
-8
Figure 4. Frequency shift of the polyimide-coated QCM under electron beam exposure. Note that
atomic oxygen beam is turned off. No erosion is detected both for grounded and floated samples
was examined. Measurements were carried out during the increasing and decreas-
ing phases of bias voltages and the results were averaged. Figure 5 shows the rate
of resonant frequency change of polyimide-coated QCM during atomic oxygen
exposures with bias voltages from 0 to 1500 V. Figures 5(a) and 5(b) indicate the
results for negative and positive bias, respectively. It was clear that no significant
effect of bias voltages on atomic oxygen-induced erosion was observed. We have
also examined the polysulfone in the same experiment, however significant effect
was not observed. From the QCM experiments reported herein, we confirmed that
the bias voltage does not affect the reaction yield of atomic oxygen with poly-
imide. This experimental finding provided a conclusion that effect of charging
on polyimide is not necessary to consider when polyimide is used as a witness
sample for atomic oxygen fluence monitor.
4. Discussion
The effect of electron beam irradiation and bias voltage on atomic oxygen-induced
erosion of polysulfone was reported by King and Wilson [5].
They used a laser detonation atomic oxygen beam, which is the same type of
atomic oxygen source used in this study, and detected the reactive products of CO
and CO2 by quadrupole mass spectrometer during the experiment. They reported
no significant effect of electron beam irradiation when spin-coated polysulfone
sample was grounded, in contrast, significant increase in CO and CO2 signals
were detected when the sample was electrically floated. Since similar effect was
observed when sample was simply biased, they concluded that the increase in CO
EFFECT OF SURFACE CHARGING ON THE EROSION RATE 57
80
(a)
40
0V
500 V
20
1000 V
1500 V
0
0 50 100 150 200 250
Elapsed Time (s)
80
(b)
Frequency Shift (Hz)
60
40
0V
20 +500 V
+1000 V
+1500 V
0
0 50 100 150 200 250
Elapsed Time (s)
Figure 5. Frequency shift of the polyimide-coated QCM under oxygen beam exposure with bias
voltages of 01500 VDC: (a) neagative bias and (b) positive bias. Bias voltage of 0 V means that
the sample was grounded
and CO2 production yields are due to surface charging. Actually, they indicated
that CO2 signal increased when bias voltage was applied to the back plate of the
film.
However, the experimental results obtained here were inconsistent with Kings
report; i.e., mass loss phenomenon was not affected by electron beam irradiation
nor by application of bias voltages up to 1500 V (positive and negative). The origin
of this discrepancy would be explained by a surface ionization phenomenon. In
the Kings experiment, they detected CO and CO2 ions. These ions are ionized
in the electron bombardment ionizer at the QMS. Ionization cross section of
molecules in the electron bombardment ionizer has not been evaluated, but is
usually low (104 or even lower). On the other hand, ions originally contained
in the beam are detected with a cross section of almost 1. Kings analysis, where
CO and CO2 production yields become greater with bias voltage, is based on the
58 MASAHITO TAGAWA ET AL.
assumption that the amount of ions in the beam is unchanged by the application of
bias voltages. One of the mechanisms that explains two inconsistent experimental
results simultaneously is the effect of bias voltages onto surface ionization cross
section. Namely, if the fraction of ions in the reactive products (CO and CO2 )
becomes greater with bias voltages, the experimental results reported by King
and in this study can be explained. This is because the detection probability of
ion is much higher than that of molecule with QMS. Measuring of the surface
ionization cross section under application of bias voltages will verify the proposed
mechanism.
5. Conclusion
Acknowledgments
A part of this work was supported by the grant-in-aid of scientific research contract
no. 14350511 and 15560686 from the Ministry of Education, Culture, Sports,
Science, and Technology, Japan; Kawanishi memorial Shinmaywa Education
Foundation.
References
1. Tagawa, M., Yokota, K., Ohmae, N., and Kinoshita, H. (2002) Journal of Spacecraft and Rockets
39(3), 447451.
2. Yokota, K., Tagawa, M., and Ohmae, N. (2003) Journal of Spacecraft and Rockets 40(1), 143144.
3. Yokota, K., Tagawa, M., and Ohmae, N. (2002) Journal of Spacecraft and Rockets 39(1), 155156.
4. Yokota, K., Ohmae, N., and Tagawa, M. (2004) High Performance Polymers, Vol. 16, No. 2
(2004) pp. 221234.
EFFECT OF SURFACE CHARGING ON THE EROSION RATE 59
5. King, T. and Wilson, W. (1997) Synergistic effects of atomic oxygen with electrons, A Bound
Collection of Papers, AIAA Defence and Space Programs Conference and Exhibit, AIAA paper
No. 97-3901, pp. 1115.
6. Caledonia, G. E., Krech, R. H., and Green, D. B. (1987) AIAA Journal 25(1), 5963.
7. Tagawa, M., Muromoto, M., Hachiue, S., Yokota, K., Ohmae, N., Matsumoto, K., and Suzuki,
M. (2003) In Proceedings of the 10th European Space Mechanism and Tribology Conference,
ESA SP-524, pp. 311314.
8. Tagawa, M., Yokota, K., Kinoshita, H., and Ohmae, N. (2003) In Proceedings of the 9th In-
ternational Symposium on Materials in a Space, Noordwijk, The Netherlands, ESA SP-540,
pp. 247252.
9. Kinoshita, H., Tagawa, M., Yokota, K., and Ohmae, N. (2001) High Performance Polymers 13(4),
225234.
INFLUENCE OF SPACE ENVIRONMENT ON SPECTRAL OPTICAL
PROPERTIES OF THERMAL CONTROL COATINGS
V. M. PROSVIRIKOV,1 A. V. GRIGOREVSKIY,1
L. V. KISELEVA,1 A. P. ZELENKEVICH,1
AND V. M. TSVELEV2
1
Joint-stock Company Kompozit, Korolev,
Moscow region, Russia
2
Lavochkin Association, Khimki, Russia
1. Introduction
The solar radiation absorption coefficient (s ) and emissivity () are the major
parameters governing spacecraft surface temperature and its serviceability [1, 2].
It is known that s of white paint coatings increases whereas doesnt change
under the influence of environmental factors. The process of applying the paint
coatings to surfaces of various shapes is simple enough. The joint-stock company
Kompozit develops new paint coatings that can be used as thermal control
coatings (TCC) for prolonged exposures. To evaluate the effects of the space
environment, samples of coatings are subjected to accelerated tests, which include
both separate and combined influence of environmental factors, with the changes
of s being determined. Using the data obtained in such tests an analysis of
applicability of these coatings is carried on.
Studies of changes in reflectance spectrum of a coating due to the effect
of environmental factors are necessary to prove the legitimacy of simulation of
61
J.I. Kleiman (ed.), Protection of Materials and Structures from Space Environment, 6169.
C 2006 Springer. Printed in the Netherlands.
62 V. M. PROSVIRIKOV ET AL.
radiation conditions, to study the damage processes in TCC, and to support the
choice of coating degradation model that will be used in calculation of s to predict
the long-term serviceability of a TCC.
3. Equipment
Reflectance spectra of ECOM-1 samples were measured in air before and after
they have been exposed to environmental factors, separately and combined. The
INFLUENCE OF SPACE ENVIRONMENT 63
The ECOM-1 coating samples were illuminated by a xenon lamp that provides an
accelerating factor of 5. The samples were irradiated to a total of 2000 equivalent
solar hours (ESH), with measurements of s performed at intermediate points.
Figure 1 shows reflectance spectra of ECOM-1 before and after exposure. Main
changes in ECOM-1 reflectance spectrum were in the shortwave range near the
fundamental absorption line of ZnO (380 nm). Zone of irreversible changes ends
at about 800 nm wavelengths. The trends of coefficient changes in IR range, after
UV tests, followed those in the air bleaching process.
R,
(nm)
Figure 1. Reflectance spectrum of ECOM-1 TCC before and after UV exposure (2000 ESH)
64 V. M. PROSVIRIKOV ET AL.
R,
(nm)
Figure 2. ECOM-1 TCC reflectance spectra before and after irradiation with 45-keV electrons, flux
density = 5.0 1011 cm2 s1 up to fluence F = 1.5 1017 cm2
Irradiation of coatings was carried out with 45-keV electrons at flux densities of
5.0 1011 cm2 s1 up to the fluence of 1.5 1017 cm2 . According to calcu-
lations, the electron penetration into samples under study is about 15 m. While
irradiating, s was measured in intermediate points. ECOM-1 reflectance spec-
tra before and after electron irradiation are given in figure 2.
Changes in reflectance spectra of electron-irradiated sample are similar to re-
sults from UV irradiation and occur mostly in the shortwave range. The zone of
irreversible changes of the reflectance coefficient ends at about 900 nm. Figure 3
Figure 3. Bidirectional reflectances of radiation ( = 570 nm) from initial and postexposed ECOM-
1 TCC samples: E e = 45 keV, F = 1.5 1017 cm2 for incidence angle 70 with respect to surface
normal
INFLUENCE OF SPACE ENVIRONMENT 65
R,
(nm)
Figure 4. ECOM-1 TCC reflectance spectra before and after irradiation with protons of different
energies ( = 1.0 1011 cm2 s1 ): E p = 40 keV, F = 1.5 1016 cm2 ; E p = 150 keV, F = 1.2
1016 cm2 ; E p = 300 keV, F = 2.6 1016 cm2 ; E p = 500 keV, F = 2.0 1016 cm2
66 V. M. PROSVIRIKOV ET AL.
Figure 5. Bidirectional reflectances of radiation ( = 570 nm) from the initial and proton exposed
ECOM-1 TCC samples: E p = 40 keV, F = 1.5 1016 cm2 for incidence angle 70 with respect to
surface normal
8. Integrated Tests
R,
(nm)
Figure 6. ECOM-1 TCC reflectance spectra before and after integrated tests: K1E p = 30 keV,
Fp = 9.4 1015 cm2 ; E e = 10 keV, Fe = 9.4 1016 cm2 ; K2E p = 30 keV, Fp = 1.1 1016
cm2 ; E e = 10 keV, Fe = 2.8 1016 cm2 , UV 67.2 ESH; K3E p = 30 keV, Fp = 2.4 1016
cm2 ; E e = 10 keV, Fe = 3.6 1015 cm2
Figure 6 shows reflectance spectra before and after integrated tests conducted
at the above conditions. The trends in behavior of reflectance coefficients after tests
conducted at K1, K2, and K3 modes are different. The most significant changes
were in K3 irradiation mode up to maximum proton fluence of about 2.4 1016
cm2 . It is seen from figure 7 that beginning with fluences near 1.5 1016 cm2
Figure 7. Dependence of change of s on proton fluences for three irradiation modes: K1, K2,
and K3
68 V. M. PROSVIRIKOV ET AL.
F(cm2)
Figure 8. Dependence of s on proton fluence for K2 mode irradiation of ECOM series TCC
samples (UV = 2.7 ESR, E p = 30 keV, and Ee = 10 keV)
9. Conclusion
and their complexes. Absorption lines are shifted to longwave range area with an
increase of proton energy.
The most significant irreversible changes of optical properties of ECOM-1
TCC are produced by protons. In this case the shape of bidirectional reflectance
at sliding incidence angles of radiation changes appreciably.
When simulating the space environment, one should take into account, by
reason of nonadditivity, the influences of separate environmental factors and their
combined action.
ECOM-1P TCC is a prospective material showing good stability to environ-
mental factors.
References
1. Heat Transfer and Spacecraft Thermal Control. (1974) Edited by John W. Lucas; Translated from
English under edition of N.A. Anfimov. M.: Mir, pp. 2228.
2. Gurevich, M. M., Itsko, E. F., and Seredenko, M. M. (1984) L. Chemistry, 120, pp. 7985.
3. ASTM E 903-96, Standart Test Method for Solar Absorptance, Reflectance, and Transmittance of
Materials Using Integrating Spheres, American Society for Testing and Materials, Philadelphia,
USA.
4. ASTM E 490, Solar Constant and Air Mass Zero Solar Spectral Irradiance Tables. American
Society for Testing and Materials, Philadelphia, USA.
5. Soloviev, G. G. (1979) Science-Exploration Institute of Technical, Economical Problems,
NIITEChIM 9(159) pp. 139.
MITIGATION OF THRUSTER PLUME-INDUCED EROSION
OF ISS SENSITIVE HARDWARE
71
J.I. Kleiman (ed.), Protection of Materials and Structures from Space Environment, 7185.
C 2006 Springer. Printed in the Netherlands.
72 MITIGATION OF THRUSTER PLUME-INDUCED EROSION
1. Introduction
During International Space Station (ISS) operations, a variety of rocket engines are
used by the Space Station and its visiting vehicles (Space Shuttle Orbiter, Soyuz,
Progress, etc.) for attitude control, orbit reboost, and docking/separation. These
various engines are bipropellant thrusters using hypergolic components, either
monomethyl hydrazine (MMH) or unsymmetrical dimethyl hydrazine (UDMH)
as the fuel and nitrogen tetroxide (N2 O4 , or NTO) as the oxidizer. The exhaust
plumes (see figure 1) from these engines have been recognized as a potential
source of loads, heating, and contamination.
Laboratory studies have revealed the presence of unburned propellant in the
exhaust plume in the form of liquid particles [1, 2]. The origin of the particles,
which can range from 1100 microns in diameter, is commonly attributed to the
incomplete combustion. Although the induced external contamination from the
exhaust plume can come from the gaseous constituents, ISS sensitive surfaces do
not operate at temperatures low enough to promote condensation of the gas phase.
Hence, ISS plume-induced contamination is primarily due to drops of unburned
propellant in the plume. The gases in the exhaust plume accelerate these propellant
drops to high velocities (13 kms1 ) due to gas drag forces [1, 2, 9]. The effect of
these high-velocity drops impacting onto ISS sensitive surfaces, such as the solar
arrays and active radiators, is akin to the impact of micrometeoroid and orbital
debris (MM/OD) particles. Given that the flux of drops in thruster plumes is much
larger than the flux of MM/OD particles of comparable diameter [3] (see figure 2),
this erosion/pitting effect is of great concern to the ISS program.
1.00E+06
1.00E+05 PRCS Centerline Liquid Droplets (20m from nozzle)
1.00E+04 MM/OD
1.00E+03
1.00E+02
Number Density (#/cm^2-sec)
1.00E+01
1.00E+00
1.00E-01
1.00E-02
1.00E-03
1.00E-04
1.00E-05
1.00E-06
1.00E-07
1.00E-08
1.00E-09
1.00E-10
1.00E-11
0 10 20 30 40 50 60 70 80 90 100
Particle Diameter (um)
For many optically sensitive surfaces, special coatings are applied to enhance per-
formance or for environmental protection. For ISS sensitive surfaces, mechanical
74 MITIGATION OF THRUSTER PLUME-INDUCED EROSION
damage from thruster plume particle impacts has significant implications. The
Boeing ISS Environments Team began to examine the influence of thruster firings
on the degradation of ISS sensitive surfaces. Two sensitive surfaces that may see
a great number of thruster firings during ISS operations are the U.S. photovoltaic
(PV) solar arrays and the active thermal control system (ATCS) radiators.
Most ATCS radiators, which provide heat rejection for the core modules, are
coated with Z-93, a ceramic thermal control paint. Z-93 consists of a matrix of zinc
oxide (ZnO) dispersed in a potassium silicate binder (K2 Ox SiO2 with x = 35,
commonly referred to as waterglass). The minimum specified Z-93 paint thickness
for the ISS ATCS panels is 5 mils. A typical application of Z-93 is shown in
figure 5.
0.4
0.3
Absorptance
0.2
0.1
0 2 4 6 8
Z-93 Thickness (mil)
applied to the Kapton. While the SiOx coating performs no optical function, its
loss due to thruster plume erosion would allow AO erosion of the Kapton.
The Boeing ISS Environments Team developed an approach to modeling the me-
chanical erosion on surfaces due to the impact of particles in thruster plumes. A
Los Alamos National Laboratory Smooth Particle Hydrodynamics (SPH) code,
called SPHINX, was used to simulate thruster plume particle impact damage.
SPHINX has been tested and verified on a number of projectile impacts and astro-
physical problems [8]. In addition, the Boeing ISS Environments Team conducted
ground tests to verify SPHINX in the regime utilized for thruster particle im-
pact simulation. The results of this ground testing are being prepared for future
publication.
Boeing conducted parametric studies varying plume particle drop size, im-
pingement angle, and velocity via SPHINX simulations. Analysis of the SPHINX
output yielded a damage matrix for a specified impacted surface material. The
damage matrix information was coded inside Boeings NASAN-II contamina-
tion computer model [9]. NASAN-II is an integrated computer model, utilizing
inputted NASTRAN geometric models, view factor calculations, and transport
routines to assess a given thrusters effect on an ISS configuration, with results
available in tabular or graphics formats.
One result found during the parametric studies was that particle impacts at
angles greater than 75 with respect to the normal produce no damage. A synopsis
of these analyses is shown in figure 7. The application of this result to plume erosion
mitigation will be discussed in section 5.
All ISS vehicles which utilize thrusters must be considered as plume erosion
sources. The NASAN-II thruster plume erosion model was used to assess the
impact of thruster firing operations on the changing ISS configurations as radiators
and solar arrays are relocated and deployed. Results of these analyses were used
to determine if plume erosion mitigation would be required.
NASAN-II was used to conduct a parametric study of the thruster plume pitting
effect on the ATCS radiators during nominal ISS operations. The analyses showed
that the ATCS radiators Z-93 coating would be eroded but the total depth of
COURTNEY PANKOP ET AL. 77
Figure 7. SPHINX results for (a) 0 , (b) 45 , and (c) 75 impingement angles
erosion during nominal operations would be well under 1.0 mil. Since the loss
of 1.0 mil of Z-93 does not compromise the heat rejection optical properties (per
figure 6), the conclusion was made that the ATCS radiators were not adversely
impacted by thruster plume erosion.
Analysis of Z-93 ATCS radiators raised concern for ATCS radiators coated
in Silver Teflon (AgFEP). To prevent Atomic Oxygen (AO) erosion, AgFEP ra-
diators on ISS have a protective SiOx coating. If the SiOx were to be pitted by
thruster plume particle impacts, the Teflon layer would become exposed to AO. A
parametric study was conducted to determine possible degradation due to thruster
plume pitting and subsequent AO erosion on AgFEP radiators. Results showed
that although plume and AO erosion could degrade the AgFEP coating, the time
to effect would exceed the mean time between failures due to penetrant MM/OD
impacts. The Boeing ISS Thermal Team concluded that no mitigation would be
necessary to protect AgFEP radiators from thruster plume erosion.
The NASAN-II code was used to analyze the effect of plume particle impacts on
the U.S. Solar Arrays from all Russian and U.S. Vehicle thruster firings during
reboost, attitude control, and proximity operations (which includes visiting vehi-
cle approach and separation to ISS docking ports). Preliminary studies showed
that solar array coatings would be highly susceptible to plume erosion damage.
78 MITIGATION OF THRUSTER PLUME-INDUCED EROSION
P6 Solar Array
P4
Solar
Array
Orbiter
Progress (PMA2)
(SM Aft)
Soyuz
(FGB Nadir)
Progress
(DC1 Nadir)
P6 Solar Array
P4
Solar
Array
Figure 9. Predicted erosion on P6 solar array due to russian segment thruster firings from flight 4A
to 12A.1
Russian vehicles were also analyzed for later configurations when new solar
arrays will be deployed. The geometry of the ISS and visiting vehicle thrusters to
the new solar arrays changes considerably, resulting in new areas of concern.
For the P4 solar array, and all solar arrays deployed subsequently, thruster
firings from Progress/Soyuz during approach and separation to ISS are no longer
a plume erosion concern. The thruster pointing direction on these vehicles, though
severe for the P6 array, is advantageous for the P4 array. Progress vehicles docked
to the SM aft port, for ISS pitch and yaw control, are also in an advantageous
position to mitigate plume erosion. In this position, Progress thrusters are so far
aft that they do not have a centerline view to the P4 array. Progress docked to the
nadir port of the DC1 for ISS roll control will, however, pose a threat for plume
particle impacts.
For the ISS assembly complete configuration, analysis predicted that Progress
on DC1 nadir thruster firings would be the major Russian Segment contribu-
tor to pitting on the U.S. solar arrays. For this configuration, the analysis re-
sults show that some areas of the solar array wings would be 100% pitted (see
figure 10).
On the U.S. Segment, the Space Shuttle Orbiter was analyzed for plume erosion
impacts to U.S. solar arrays. The Orbiter thrusters will fire toward ISS during
approach and separation to the pressurized mating adapter 2 (PMA2) docking
port. At close ranges, a different set of side-pointing thrusters may also plume
solar arrays.
In addition, the Orbiter may perform several reboosts and maneuvers while
mated to ISS. These maneuvers, especially reboost, have the potential to cause
severe pitting damage to solar arrays given the Orbiter thruster configuration and
the high thruster firing time.
80 MITIGATION OF THRUSTER PLUME-INDUCED EROSION
Figure 10. Predicted solar array erosion for ISS thruster operations through assembly complete due
to Russian Segment thruster firings
Analysis results for both Russian Segment and Orbiter thruster firings indicated
that a mitigation technique would be required to protect solar arrays from thruster
plume pitting damage.
5. Mitigation
With predictions of solar array surface pitting, the next step was to determine the
extent of degradation to solar array performance resulting from erosion damage.
The data available to perform such an assessment was found to be inadequate.
A working group within the NASA ISS Program Office recommended a series
of ground-based tests to produce data detailing performance degradation to solar
cells due to surface pitting. The data from such tests, however, would not be
available in time to affect on-orbit solar array integration issues in the near term.
Therefore, the working group concluded that operational mitigation techniques
must be put in place now to protect ISS optical surfaces.
The Boeing ISS Environments Team was tasked to determine the angles to which
the solar arrays could rotate such that thruster plume erosion would be minimal.
Based on previous analyses which showed that particle impacts at angles greater
than 75 to normal produced no damage, the following mitigation condition were
initially derived: (1) the solar array must be rotated such that the plume impinge-
ment angle to a solar array surface is greater than 75 from the normal, and
(2) no thruster plume is allowed to contact the active side of the solar array.
COURTNEY PANKOP ET AL. 81
1.00E+05
1.00E+04 1.E-04
1.E-03
0.01
1.00E+03
0.1
1.00E+02 10
50
1.00E+01
1.00E+00
0 10 20 30 40 50 60 70 80 90
Angle From Plume Centerline (deg)
For the early ISS configuration, analyses found that the P6 solar array would receive
pitting damage during Progress, Soyuz, and Orbiter proximity operations. Existing
ISS Program Flight Rules for feathering the P6 solar array during Progress/Soyuz
approach and Orbiter approach and separation (put into place for other require-
ments) met requirements for plume erosion mitigation, so no further action was
required. For Progress/Soyuz separation, however, new flight rules were drafted to
82 MITIGATION OF THRUSTER PLUME-INDUCED EROSION
Figure 12. P4-2A allowable feathering angles for Progress on DC1 nadir roll control thruster firings
ensure that flight controllers would position the P6 solar array to mitigate plume
erosion prior to the separation operation. This flight rule is in effect for the current
ISS configuration and will remain in effect until the P6 solar array is relocated to
the main truss.
Analyses show that additional thrusters will become a plume erosion con-
cern when additional solar arrays are deployed. The thrusters of concern include
Progress on DC1 nadir thruster firings for ISS roll control, Orbiter thruster fir-
ings during approach and separation to PMA2, and Orbiter thruster firings during
mated ISS operations. Feathering angles to mitigate plume erosion must be defined
for each of these thruster firing events.
For Progress on DC1 nadir roll control firings, allowable solar array alpha/beta
angle pairs to mitigate erosion were defined per the mitigation criteria. These
alpha/beta pairs have been tabulated for inclusion into a flight rule to provide flight
controllers with the proper settings to assure solar array protection. A sample table
is shown for the P4 Solar Array in figure 12. It should be noted that allowable
solar array alpha/beta angle pairs are shown in gray.
Proximity operations have another factor adding complexity to the issue of
plume erosion. Solar arrays may need to be positioned to mitigate plume ero-
sion from the incoming (or departing) vehicles thruster firings. In addition, ISS
COURTNEY PANKOP ET AL. 83
Allowable
P4 2A Solar Array
Alpha/Beta Angle
Pair Combinations
Figure 13. P4-2A allowable feathering angles for Orbiter approach to ISS (with progress on SM
Aft and progress on DC1 Nadir attitude control)
6. Future Work
In upcoming years, ISS operations will include two new visiting vehicles: the Euro-
pean Space Agencys Automated Transfer Vehicle (ATV) as well as then Japanese
H-II Transfer Vehicle (HTV). ISS operational mitigation techniques for the U.S.
solar arrays during ATV and HTV proximity operations must be developed as well
as mitigation techniques for ATV mated operations. The Boeing ISS Environments
Team will perform analyses to define the allowable alpha/beta angles for the solar
arrays during these thruster firing events. The resulting requirements are to be
integrated with similar requirements for thermal and structural effects of thruster
plume impingement. Integrated requirements will be implemented through ISS
Program Flight Rules.
The Boeing ISS Environments Team will also continue updating previous
analyses as changes to the ISS configuration are made. In the near team, the
Service Module will be analyzed for unanticipated utilization during later ISS
assembly sequences. Russian Segment and U.S. vehicle analyses may also be
updated as new thruster firing data becomes available.
In addition to conducting plume erosion analyses, the Boeing ISS Environ-
ments Team will continue efforts to improve plume erosion modeling and predic-
tions. Ground-based tests, which will detail solar cell performance degradation
due to surface pitting, are planned to take place this year. Results will be incorpo-
rated into the NASAN-II plume erosion model.
7. Conclusions
Thruster plume particles can cause mechanical damage in the form of ero-
sion/pitting to ISS sensitive surfaces. In general, any spacecraft sensitive surfaces
COURTNEY PANKOP ET AL. 85
exposed to the exhaust plumes of thrusters on-orbit should either compensate for
thruster plume erosion via design or prepare mitigation techniques to minimize
thruster plume particle pitting.
In this paper, the development of the erosion model was outlined along with
predictions of damage due to thruster firings on selected ISS sensitive surfaces.
Analysis results indicate that plume erosion of ATCS radiator coatings would
not affect heat rejection properties. Characterization of thruster plume induced-
erosion/pitting of U.S. solar arrays due to nominal operations of ISS vehicle
thrusters was shown for an interim and final ISS assembly configuration. The
solar array analysis results show that plume erosion mitigation is required to pro-
tect solar array optical coatings.
A mitigation technique for ISS solar arrays, which minimizes thruster plume
erosion during current and future ISS assemblies, was presented. Implementation
of these mitigation techniques into ISS Program Fight Rules was discussed, as
well as plans for future analyses and ground-based testing to characterize plume
erosion effects.
The Boeing ISS Environments Team pursues these activities to ensure a known
environment around the ISS so as to guarantee the success of the ISS as a platform
for scientific experiments in low Earth orbit.
References
1. Trinks, H. (1987) Experimental Investigation of the Exhaust Plume Flow Fields of Various Small
Bipropellant and Monopropellant Thrusters, AIAA Paper 871607, June 1987.
2. Rebrov, S., and Gerasimov, Y.; Investigation of the Contamination Properties of Bipropellant
Thrusters, AIAA 2001-2818, 35th AIAA Thermophysics Conference, Anaheim, CA, 1114
June 2001.
3. Theall, J. (2000) JSC Space Science Branch, Private communication, 27 November 2000.
4. Koontz, S., Melendez, O., Zolensky, M., and Soares, C. (1996) SPIFEX Contamination Studies,
JSC-27399, SPIE May 1996, 411.
5. Soares, C. E. and Mikatarian, R. R. (2002) In International Symposium on Optical Science and
Technology, SPIE 4774-20, Seattle, July 2002, 213.
6. Orr, W. (2000) PIC Camera Lens Test Analysis, Lockheed Martin Science, Engineering, Test, and
Analysis Report to NASA JSC EM2, 16 August 2000.
7. Daneman, B., and Thomlinson, H. (1993) Advancements in Long-Life Thermal Control Coatings
for Low Earth Orbit Applications, McDonnell Douglas Aerospace, #93222.
8. Wingate, C. and Stellingwerf, R. (1993) Smooth Particle Hydrodynamics: The SPHINX and
SPHC Codes, Technical Report LA-UR-9301938, Los Alamos National Laboratory, January
1993.
9. Alred, J., Boeder, P., Mikatarian, R., Pankop, C., and Schmidl, W. (2003) In Proceedings of the 9th
International Symposium on Materials in a Space Environment, Noordwijk, The Netherlands,
1620 June 2003.
DEGRADATION OF THERMAL CONTROL COATINGS UNDER
INFLUENCE OF PROTON IRRADIATION
1. Experimental Studies
Tests of thermal control coating (TCC) samples under impact of protons with
energies up to 500 keV and fluences from 1014 to 2 1016 cm2 have been car-
ried out on the KG-500 accelerator in Skobeltsyn Institute of Nuclear Physics
Moscow State University (SINP MSU). The accelerator provides energy stabil-
ity of 0.1%. The beam current of protons was set not to exceed 1 A. At the
maximum energy of protons (0.5 MeV), the power at the TCC sample was less
than 0.5 W, and the temperature of the sample during the irradiation did not ex-
ceed 100 C. The proton fluence was measured with accuracy of at least 1%. The
measurement of the TCC optical absorption coefficient s was done with the pho-
tometer built in the vacuum chamber. The high value of the proton fluence (2
1016 cm2 ) was chosen to determine the parameters in function s = f (F, E)
used in physical and mathematical model of the TCC optical degradation [1
3]. It was shown [4] that protons produce the strongest irreversible changes in
87
J.I. Kleiman (ed.), Protection of Materials and Structures from Space Environment, 8791.
C 2006 Springer. Printed in the Netherlands.
88 DEGRADATION OF THERMAL CONTROL COATINGS
optical thermal properties that could be associated with processes of lattice defect
generation.
2. Test Results
As
0.6
1
0.5
2
0.4
0.3
0.2
0.1 3
0 4
10 14 5 . 1015 1016 1.5 . 1016 F, cm2
Figure 1. Change of the TCC s Coefficient after the proton irradiation with energy 300 keV:
(1) ECOM-1, (2) TR-SO-TSM, (3) SS-1, (4) ECOM-2
L. S. NOVIKOV ET AL. 89
As
0.6 1
0.5 2
0.4 3
0.3
0.2
0.1
0
15 16
0 5.0 . 10 1.0 . 10 1.5 . 1016 2.0 . 10
16
F, cm2
Figure 2. Change of the ECOM-1 coating As coefficient as function of the proton fluence with
energies (1) 150, (2) 300, and (3) 500 keV
radiation defects localization depth, where the influence of the defects on the s
value decreases.
The parameters of the physical and mathematical model describing the process of
optical degradation of the TCC due to radiation are determined from the experi-
mental data obtained during irradiation of the coatings by monoenergetic beams.
These parameters are used in calculation of the forecast of the s coefficient
change under impact of corpuscular radiation with continuous energy spectrum
in spacecraft orbit. In the given work, the model was chosen that describes the
As
0.5 3
2
0.4 1
0.3
0.2
0.1
0
0 5.0 . 10
15
1.0 . 1016 1.5 . 1016 2.0.1016
F, cm2
Figure 3. Change of the TR-SO-TSM coating As coefficient as function of the proton fluence with
energies (1) 150, (2) 300, and (3) 500 keV
90 DEGRADATION OF THERMAL CONTROL COATINGS
As
0.7
1
0.6
0.5 2
0.4
0.3
0.2
0.1
0
0 100 200 300 400 500
E, keV
Figure 4. Dependence of the As coefficient of (1) ECOM-1 and (2) TR-SO-TSM coatings on the
proton energy at a fluence 2 1016 cm2
4. Conclusions
r Tests of various types of TCC under impact of protons with energies 150
500 keV with fluence up to 2 1016 cm2 revealed that theAsvalue variation
TCC type a B
does not exceed 0.05 for black coating ECOM-2 and 0.1 for grey SS-1 coating.
The variation is 0.30.6 for white coatings ECOM-1 and TR-SO-TSM.
r Radiation resistance of the ECOM-1 enamel coating is higher than that of the
TR-SO-TSN ceramic coating.
r The As value functions of the proton fluence and energy obtained for the ECOM-
1 coating and TR-SO-TSN coating were used the determine the predicted vari-
ations of the As value during the spacecraft flight in GSO up to 10 years.
r ECOM-2 coating and TR-SO-TSM coating are recommended as the most
preferable for TCC on spacecrafts in GSO.
References
1. Mikhailov, M. M., Dvoretsky, M. I., and Krutikov, V. N. (1982) Space Technology and Material
Research, Science, Moscow, pp. 95100.
2. Titov, V. I. and Tarasov, Y. I. (1984) Journal of Physics and Chemistry 58(5), 12121214.
3. Vasiliev, V. N., Grigorevskiy, A. V., and Gordeev, Y. P. (2002) In Proceedings of the 6th Inter-
national Space Conference on Protection of Materials and Structures from Space Environment,
Toronto, Canada, 2002, pp. 543550.
4. Prosvizikov, V. M., Grigorievskiy, A. V., Kiseleva, L. V., Zelenkevich, A. P., and Tsvelev, V. M.
(2004) In Proceedings of the 7th International Space Conference on Space Materials, Toronto,
Canada, Springer, 6169.
5. Garret, H. B., Schwank, D. C., and De Forest, S. E. (1981) Ions Planetary and Space Science
29(10), 10451060.
MITIGATION OF DAMAGE TO THE INTERNATIONAL SPACE
STATION (ISS) FROM WATER DUMPS
Abstract. The International Space Station (ISS) and Orbiter dump water over-
board. This water is from the ISS condensate system, and from the Orbiters fuel
cell (supply side) and wastewater (urine and condensate) systems. When water
is dumped into a vacuum, some of it flashes into a vapor. The expanding vapor
bursts the stream into vapor, and small and large liquid/ice particles. The large liq-
uid/ice particles are approximately 2 mm in diameter and have nominal velocities
of approximately 31 fts1 (U.S. Lab) and 50 fts1 (Orbiter). As these liquid/ice
particles impact, they may cause mechanical damage due to erosion/pitting of
sensitive surfaces, including solar array or radiator surfaces. Solar arrays are of
particular concern because of the thin optical coatings on the surface of the cells.
The thickness of these coatings is 43300 A. Damage to these coatings can cause
degradation of the cells optical characteristics that can potentially reduce perfor-
mance and shorten the life of the cells. To mitigate damage from water dumps, the
characteristics of the water dumps were studied and the results used to develop
the constraints needed to mitigate damage to ISS hardware from U.S. Lab and
Orbiter water dumps.
Key words: water dump, vent, purge, impact, damage, erosion, pitting
1. Introduction
The International Space Station (ISS) and Orbiter dump water overboard into
space. The phenomena of releasing water into a vacuum has been studied for
many years [13]. Figure 1 shows a schematic of water dumping into a vacuum.
It is known that as the liquid exits the nozzle it freezes. Then the expanding gas
bubbles burst the stream into vapor, and small and large liquid/ice particles that
can travel in various directions.
When water is dumped overboard, there is a concern that direct contact from
the liquid/ice particles on ISS hardware can cause mechanical damage to sensitive
93
J.I. Kleiman (ed.), Protection of Materials and Structures from Space Environment, 93105.
C 2006 Springer. Printed in the Netherlands.
94 MITIGATION OF DAMAGE TO THE INTERNATIONAL SPACE STATION
Ice Particles
Figure 1. Schematic of water dumping into a vacuum. The outer layer of water freezes as the inner
core of liquid boils and forms bubbles. The expanding bubbles burst the stream into liquid/ice
particles
Condensate is produced aboard the ISS by the thermal heating and cooling system
and stored in a tank that must eventually be emptied. The tank is fill restricted to
100 lbm (60.35% of capacity) to prevent hard-filling the tank and to maximize
the tank bellows life. The U.S. Lab condensate system dumps the excess water at
cabin pressure (14.7 psi) and with a mass flow rate that is approximately 14.6 gs1 .
The condensate system dumps the excess water overboard through nozzles
located on the U.S. Lab forward end-cone. These nozzles are located on the
WILLIAM SCHMIDL ET AL. 95
port-zenith side, and starboard-nadir sides, and have a diameter of 1.4 mm. Each
nozzle is preheated to 120 C and heated while dumping to keep ice from forming
close to the nozzle exit that could result in clogging the nozzle, forming an icicle,
or diverting the dump stream. Figure 2 shows the relative position of the conden-
sate dump nozzles and their direction. The plume from the nozzles is directed
37.4 from the zenith and nadir directions.
The flight experiment SDTO 16004-A, performed on 7 September 2001, was
to video a condensate dump using the ISS robotic arm, the space station remote
manipulator system (SSRMS) and to measure the characteristics of the U.S. Lab
water dump [46]. The images were captured using the SSRMS cameras of both
the port and starboard dumps.
The SDTO measurements included the plume cone angle, velocity of the large
liquid/ice particles, and the duration of each of the dump phases. Figure 3 shows the
distinct dump phases that were observed in the SDTO video. The phases observed
were a startup, steady state, and shutdown phase. The startup phase occurs when the
line valve is opened and the liquid first moves into the lower pressure environment.
The steady state phase occurs for the majority of the dumping duration. The liquid
is confined to a narrow directed cone. It lasts for the remainder of the dump,
excluding the startup and shutdown phases. The shutdown phase occurs when gas
becomes entrained in the condensate water being dumped. The shutdown phase
was also divided into an initial shutdown phase and a sputtering shutdown phase.
The initial shutdown phase occurs when gas becomes entrained in the condensate
96 MITIGATION OF DAMAGE TO THE INTERNATIONAL SPACE STATION
Figure 3. U.S. Lab water dump video captured during flight experiment SDTO 16004-A. Three
phases were observed during the U.S. Lab condensate water dump: a startup, steady state, and shut-
down phase. The startup phase lasts approximately 36 s, the steady state phase lasted approximately
564 s, and the shutdown phase lasted approximately 54 s
tank and the valve is still open. The shutdown sputtering phase occurs after the
valve is closed.
The plume cone angles were measured from the centerline for each of the
dump phases. It was observed that the plume had a core region where the majority
of the liquid/ice particles were concentrated; however, some particles were also
observed outside the core region. For SDTO 16004-A, approximately 145 particles
were ejected outside the core region.
In figure 3, it can be observed that the steady state phase has a tight cone.
The startup and shutdown phases, at the beginning and end of the dump event,
have large cone angles. This is due to the bursting of the stream. The plume cone
angle for the steady state phase was measured to be 10 . However, for engineering
margin, the plume cone angle for the steady state phase was defined to be 20 . For
the startup and shutdown phases, the plume cone angle was defined to be 60 .
Figure 4 shows the angular distribution function that was developed for each
of the phases observed during SDTO 16004-A. For the startup, shutdown initial,
and shutdown sputtering phases, the distribution is flat to account for the larger
plume cone angle that was observed. Figure 4(b) shows the angular distribution
function for the steady state phase. It can be seen that the majority of large particles
is concentrated near the plume centerline out to 10 . This region is defined as the
impact zone. The ISS External Contamination Team defined the cone out to 20
from the plume centerline as the impact zone with engineering margin.
Figures 5(a) and 5(b) shows the field of views from the port and starboard U.S.
Lab condensate water dump nozzles for ISS assembly complete. It can be seen
that if the solar arrays are not rotated and feathered for the water dump, they could
rotate directly into the impact zone with engineering margin of the water dump.
For each vent phase of SDTO 16004-A, the liquid/ice particle velocities were
measured both outside and inside the core region of the plume. For the startup
phase, velocities measured outside the 60 core region ranged from 5.8 to 15 fts1
(measured velocities including the core region ranged from 2.8 to 30.6 fts1 ). For
WILLIAM SCHMIDL ET AL. 97
1 1
Angular Distribution
Total
0.6 0.6 Total
Large Particle
0.5 0.5
Large Particle
Gas
0.4 0.4 Gas
Gas Small Particle
0.3 0.3
Small Particle
0.2 0.2
Gas
0.1
Small Particle 0.1 Small Particle
0 0
0 10 20 30 40 50 60 70 80 90 0 10 20 30 40 50 60 70 80 90
Angle (degrees) Angle (degrees)
(a) (b)
1 1
Total Shutdown Initial Phase Total Shutdown Sputtering Phase
0.9 Angular Distribution vs Angle (degrees) 0.9 Angular Distribution vs Angle (degrees)
0.8 0.8
Angular Distribution
Total
0.6 0.6 Large Particle
Large Particle
0.5 0.5 Gas
Gas
0.4 0.4 Small Particle
Gas Small Particle
Gas
0.3 0.3
0.2 0.2
0.1 0.1
Small Particle Small Particle
0 0
0 10 20 30 40 50 60 70 80 90 0 10 20 30 40 50 60 70 80 90
Angle (degrees) Angle (degrees)
(c) (d)
Figure 4. U.S. Lab water dump cumulative mass fraction distribution: (a) startup phase, (b) steady
state phase, (c) shutdown initial phase, and (d) shutdown sputtering phase
the steady state phase (core region within 20 ), from average flow rate and nozzle
diameter, the average velocity was determined to be 31 fts1 . For the initial
shutdown phase, measured velocities outside the core region ranged from 2.0
to 3.3 fts1 (measured velocities including the core region ranged from 2.0 to
8.0 fts1 ). For the shutdown sputtering phase, measured velocities outside the 60
core region ranged from 2.0 to 4.6 fts1 (measured velocities including the core
region ranged from 1.5 to 9.1 fts1 ).
The phase durations for the SDTO 16004-A dump, as shown in figure 3, were
determined from the video by the NASA Imaging Science and Analysis Group
(IS&AG). The startup phase took approximately 36 s, the initial shutdown phase
took approximately 22 s, and the shutdown sputtering phase took approximately
32 s. The steady state phase lasted for the remainder of the dump, which for this
experiment (SDTO 16004-A), lasted approximately 564 s.
The Orbiter dumps both waste water (urine and condensate) and supply water.
Supply water is generated by the Orbiters fuel cells and is extremely pure, so it is
98 MITIGATION OF DAMAGE TO THE INTERNATIONAL SPACE STATION
(a)
Impact Zone
(b)
Impact Zone
not a molecular deposition concern. The condensate water is collected from the
Orbiter cabin and has approximately 0.004% nonvolatile residues in the water.
So, it does not leave a significant deposit. Pure urine contains approximately 4%
residue and is a deposition concern. To protect the solar cells from urine deposits,
the solar arrays are feathered so that only the backside is hit. The feathering angles
defined to protect against erosion/pitting will also mitigate deposition concerns.
Flight rules are currently in place to minimize urine dumps. To minimize
urine dumps, urine and condensate water are separated, and the condensate water
WILLIAM SCHMIDL ET AL. 99
Figure 6. Orbiter water dump for ISS assembly complete. The Orbiter dumps water out to the port
side of ISS
is stored in contingency water containers (CWCs). Only urine in excess of the tank
capacity is dumped overboard. All urine dumping is to be discontinued when the
Japanese hardware goes up, because it is in the direct line of sight of the Orbiter
vent nozzles.
Water released from the Orbiter is dumped through two nozzles located close
to each other on the Orbiters port side and are directed to the port side of ISS.
One nozzle is for supply side water and the other is for wastewater. The nozzles
are 1.4 mm in diameter. Water released from each nozzle is dumped at 31 psia at
23.7 gs1 .
The concern for Orbiter water dumps is with sensitive hardware on the port
side of ISS. Figure 6 shows a schematic of the direction of the Orbiter dump for
ISS assembly complete.
Figure 7 shows an image of the Orbiter dumping water. The inset figure was
captured from a video of the Orbiter water dump.
The Boeing external contamination team developed a model for the Orbiter
water dump [7]. Figure 8 shows the plume distribution for that model. It can be
seen that the majority of large particles is concentrated near the centerline of the
plume out to 10 . This region is defined as the impact zone. The ISS External
Contamination Team defined the cone out to 20 from the plume centerline as the
impact zone with engineering margin.
100 MITIGATION OF DAMAGE TO THE INTERNATIONAL SPACE STATION
Figure 7. Image of Orbiter water dump. The inset image was captured from a video of the Orbiter
water dump
0.8
0.7
Angular Distribution
0.6 Total
0.4 Gas
0.2
0.1
0
0 10 20 30 40 50 60 70 80 90
Angle (degrees)
Figure 8. Boeing two-phase water dump plume model for Orbiter. The region from the plume
centerline out to 10 is defined as the impact zone. The region out to 20 is defined as the impact
zone with engineering margin
WILLIAM SCHMIDL ET AL. 101
Impact Zone
Figure 9 shows the field of view from the Orbiter dump nozzles for ISS as-
sembly complete. Sensitive surfaces of concern are the Japanese Aerospace Ex-
ploration Agency (JAXA) payload sites and the P4/P6 radiators and solar arrays.
The JAXA payload sites can be seen just below the center of the plume. It can be
seen that if the solar array is not feathered, it will be impacted broadside. In this
configuration, the radiator is rotated out of the field of view. As will be shown
later, the solar arrays should be feathered so that the impacts occur on the backside
of the array and at a shallow (15 ) angle.
The Arnold Engineering Development Center (AEDC) measured the charac-
teristics of the Orbiter water dump. These measured characteristics include the
Orbiter water dump plume cone angle, the composition of the plume (the fractions
of vapor/small particles/large particles), and the velocity of the large particles. The
results from the AEDC ground tests have been compared with flight images and
an Orbiter dump model was developed [7]. The results show that the velocity of
the particles is nominally 50 fts1 and the plume cone angle is 10 .
3. Damage Mitigation
to the surface of less than 15 (or 75 from the surface normal) did not show any
damage to the solar cells.
Using these results as a starting point, the constraints for liquid/ice particle
impacts were developed. Although, the results did not show any damage to the
solar cell, the UVE protective coating is thin, 43,300 A. If the coating is damaged,
it degrades the performance and lifetime of the solar cell. To be conservative, it
was determined that the operational constraint will be to not allow impacts onto
the active side.
The backside of the solar array has a 1300 A SIOx coating to protect the kapton
backing from atomic oxygen erosion. Below the kapton, there are additional layers
that if lost will not affect the performance of the solar cell. However, to minimize
damage to the solar array, the operational constraint developed for the backside of
the solar array is that impacts will be at a shallow angle, less than 15 to the surface.
In addition, to minimize the number of impacts, an operational constraint was
developed so that the solar arrays will be rotated to remain outside the impact
zone with engineering margin (the 20 half cone angle cone around the plume
centerline).
To mitigate damage to the solar array photovoltaic radiators (PVR), an op-
erational constraint was developed to keep the radiators away from the plume
centerline. The radiators operate cold and cannot be feathered. In addition, the
effect of liquid/ice particle impacts on the radiators has not been well defined yet.
The constraint is to keep a solar array between the impact zone with engineering
margin and the radiator. This constraint will allow the radiators no closer than 50
to 60 from the plume centerline.
Figure 10 shows the field of view from the port side U.S. Lab condensate water
nozzle for ISS assembly complete. The solar arrays are rotated outside of the
impact zone with engineering margin and feathered so that impacts from the
water dump occur on the backside of the arrays.
An example of the feathering angles defined, based on the defined constraints,
to feather the solar array wings (SAWs) so that impacts from the liquid/ice particles
occur on the backside of the arrays at a shallow angle is shown in figure 11.
Figure 12 shows the field of view from the Orbiter dump nozzle for a solar array
rotated out of the high impact zone and with the solar array feathered so that
impacts are on the backside of the array at a shallow angle. The solar array is
feathered so that there are no impacts on the active side of the array and so that
impacts on the backside of the array are at a shallow angle.
WILLIAM SCHMIDL ET AL. 103
Impact Zone
Figure 11. Example U.S. Lab port side condensate water dump allowable solar array feathering
angles for the P4-4A solar array wing (SAW) for beta gimbal assembly (BGA) beta rotations from
0 to 238
104 MITIGATION OF DAMAGE TO THE INTERNATIONAL SPACE STATION
Impact Zone
4. Summary
The International Space Station (ISS) and Orbiter both dump water overboard into
space. The dumped stream bursts into liquid/ice particles. The large liquid/ice
particles are approximately 2 mm in diameter and have nominal velocities of
approximately 31 fts1 (U.S. Lab) and 50 fts1 (Orbiter). As these liquid/ice
particles impact, they can cause mechanical damage due to erosion/pitting of
sensitive surfaces, such as the coatings used on the solar array or radiator surfaces.
Solar arrays are of particular concern because of the thin optical coatings on the
surface of the cells. Damage to these coatings can cause degradation of the cells
performance and operational lifetime.
To mitigate damage from water dumps, the characteristics of the water dumps
were studied. The results were used to develop the constraints needed to mitigate
damage to ISS hardware from the U.S. Lab and Orbiter water dumps. The results
of these studies show that the ISS solar arrays can be parked at select angles during
water dump operations to preclude damage to the solar array and radiator surfaces.
References
5. Schmidl, W. D., Alred, J. W., Mikatarian, R. R., Soares, C., Miles, E., Howorth, W., Mishina, L.,
and Murtazin, R. (2003a) In 33rd AIAA Fluid Dynamics Conference and Exhibit, AIAA paper
2003-4268, Orlando, FL, 2326 June 2003.
6. Schmidl, W. D., Alred, J. W., Mikatarian, R. R., Soares, C., Miles, E., Howorth, W., Mishina,
L., and Murtazin, R. (2003b) In 9th International Symposium on Materials in Space, ESTEC,
Noordwijk, Netherlands, June 2003, 423430.
7. Alred, J. W., Smith, L. N., Wang, K. C., Lumpkin, F. E., and Fitzgerald, S. M. (1998) In
Proceedings of AIAA/ASME Joint Thermophysics and Heat Transfer Conference, Albuquerque,
NM, June 1518, 1998, 982588.
8. Alred, J., Boeder, P., Mikatarian, R., Pankop, C., and Schmidl, W. (2003) In 9th International
Symposium on Materials in Space, ESTEC, Noordwijk, Netherlands, June 2003, 19.
9. Wingate, C. and Stellingwerf, R. (1993) Smooth Particle Hydrodynamics: The SPHINX and
SPHC Codes, Technical Report LA-UR-9301938, Los Alamos National Laboratory, January
1993.
10. Stellingwerf, R. and Wingate, C. (1993) International Journal of Impact Engineering 14, 707
718.
INVESTIGATION OF SYNERGISTIC EFFECTS OF PROTON
AND ELECTRON RADIATION ON THE DYEING OF
OPTICAL QUARTZ GLASS
Abstract. Optical materials have been widely used in spacecraft. Many optical
elements exposed in open space, such as mirrors of telescopes, fused quartz win-
dows, lenses and cover plates on solar cells, are affected by various space en-
vironmental factors. These include high vacuum, thermal cycling, irradiation by
charged particles, and solar electromagnetic irradiation. It is important to investi-
gate the interaction of optical materials with space environmental factors.
The synergistic effects of proton and electron radiation on optical properties of
quartz glass for space applications were studied using a complex simulator for the
space environment named KIFK. The energy of protons and electrons ranged
from 60 to 180 keV, which is within the energy range of the Earths radiation
belt. The changes in transmission of quartz glass under radiation were analyzed.
Synergistic effects were found during combined irradiation. The synergistic effect
of complex radiation indicates at the necessity of developing a complex space
environment test.
Key words: optical quartz glass, proton and electron radiation, synergistic effect,
transmission
1. Introduction
With increasing human activity in space, the reliability and lifetime of spacecraft
is becoming an important restriction. Space environment affects the reliability and
lifetime of a spacecraft. Great effort has gone into studying the space environment
effects on spacecraft among which one can mention the space environment and
effects (SEE) plan of USA. Research of space effects has two approaches, i.e.,
space exposure testing and ground based simulation testing. Although the two ap-
proaches do not conflict with each other, the ground based simulation testing is a
more economic and feasible method available to most researchers, rather than more
107
J.I. Kleiman (ed.), Protection of Materials and Structures from Space Environment, 107113.
C 2006 Springer. Printed in the Netherlands.
108 SYNERGISTIC EFFECTS OF PROTON AND ELECTRON RADIATION
The primary effect of charged particles on optical quartz glass is dyeing. In ref-
erence [2] the radiation induced dyeing of optical quartz glass by single particle
was investigated. Figure 1 shows the optical spectral transmission of JGS3 quartz
glass before and after radiation by protons and electrons.
Radiation by protons and electrons results in dyeing of JGS3 optical quartz
glass within the ultraviolet to visual wavelength range, while it has little influ-
ence within the infrared range. The critical dose of proton radiation that induces
dyeing is c = 5 1014 partcm2 . The transmission is first decreased within the
200250 nm of ultraviolet zone and then the variation in transmission is shifted
to longer wavelengths with increasing radiation dose. Transmission decreases lin-
early with increasing proton radiation for doses less than 2 1016 partcm2 . The
critical dose for electron radiation induced dyeing is c = 5 1014 partcm2
with similar variation in dyeing of that under proton radiation. However, the
electron induced decrease in transmission is higher than that under proton ra-
diation with the same energy and dose. A critical dose at 60 keV of simultaneous
HAI LIU ET AL. 109
100 100
1
1
90 90
2
Transmission(%)
Transmission(%)
2
3 Ep=180keV
80 3 Ep=80keV 80 1- before radiation
1- before radiation 4 15 2
2- =
10 part/cm
2- =
2 1
15
0part/cm
2 5 3- =
2
15
10part/cm
2
70 4 16 2
70 16 2
3- =
10 part/cm 4- =
10 part/cm
16 2 16 2
4- =
2 1 0part/cm 5- =
2 10part/cm
60 60
50 50
200 300 400 500 600 700 800 900 1000 200 300 400 500 600 700 800 900 1000
Wavelength(nm) Wavelength(nm)
100 100
e n
t re te tx h
e e
r
1 1
90 90
Transmission, %
2
2
Transmission, %
3
3
4
80 Ee=100keV 80
1- before radiation
15 2
2- =
10 /cm
70 4
3- =
2 1
15
0/cm
2
70
4- =
16
10 /cm
2 Ee=140keV
1- before radiation
60
60 2- =
5
14
10/cm
2
15 2
3- =
10 /cm
15 2
50 4- =
5 10/cm
200 300 400 500 600 700 800 50
200 300 400 500 600 700 800
Wavelength, nm
Wavelength, nm
Figure 1. Optical spectral transmission of JGS3 quartz glass before and after radiation by protons
and electrons
radiation is pc = ec = 1014 partcm2 that decreases with increasing particle
energy.
Under proton and electron irradiation, the variations in optical spectral prop-
erties of JGS3 quartz glass share the common characteristic of separated radiation
by the two types of particles. The simultaneous radiation affects the optical spec-
tral properties of quartz glass in the 200800 nm range with little influence in
the infrared zone. The critical dose of simultaneous radiation to change the spec-
trum for JGS3 quartz glass is lower than that found for separate radiation by each
single particle, that is, effect on properties has close relationship with particle en-
ergy. With increasing energy, the critical radiation dose shows a decreasing trend.
Under the simultaneous radiation, there are two distinct absorption bands within
the 430450 nm range and at 540 nm, as well as including the strong absorp-
tion band within 200250 nm that exist simultaneously when the critical dose is
reached. The absorption band at 540 nm has no ability to recover. Under a certain
energy and radiation dose, the decrease in transmission caused by simultaneous
radiation by both type of particles is larger than found under radiation by each
type of particles.
110 SYNERGISTIC EFFECTS OF PROTON AND ELECTRON RADIATION
16
Change in transmittance, %
15 2
1- combined irradiation, p= e=10 part/cm
2- sum of curres 3 and 4
15 2
12 3- electron irradiation e=10 part/cm
15 2
4- proton irradiation p=10 part/cm
1
8
2
4
3
4
0
200 300 400 500 600 700 800
Wavelength, nm
25 35
Change in transmittance,%
Change in transmittance, %
15 2 16 2
1- combined irradiation, p= e=2*10 part/cm 1- combined irradiation, p= e=10 part/cm
2- sum of curres 3 and 4 30 2- sum of curres 3 and 4
20 15 2 16 2
3- electron irradiation e=2*10 part/cm 3- electron irradiation e=10 /cm
15 2 25 16 2
4- proton irradiation p=2*10 part/cm 4- proton irradiation p=10 /cm
15
20
2
1 15
10
2 1
10 4
5 3
5 3
4
0 0
200 300 400 500 600 700 800 200 300 400 500 600 700 800
Wavelength, nm Wavelength, nm
Figure 2. Synergistic effect of simultaneous irradiation of quartz glass by protons and electrons with
energy of 100 keV (a) = 1015 partcm2 ; (b) = 2 1015 partcm2 ; (c) = 1016 partcm2
Under the proton and electron radiation, the main radiation induced absorption
bands in quartz glass are at 203215 and 220240 nm, corresponding to the E
center which is the main radiation defect [35]. Here we are going to analyze the
mechanism of synergistic effects of simultaneous radiation by protons and elec-
trons and explain the experimental data using figure 3. Assuming the particle flux
irradiates the specimen from left to right, Rp and Re are the shot penetration dis-
tances of protons and electrons, respectively. Within the shot range, simultaneously
applied protons and electrons will act with the outer electron shells of the atom
resulting in the atom ionization and formation of secondary electrons and a cavity.
Since the energy of the incident protons and electrons is far higher than ionization
energy of atoms in target material, the ionization is a complicated cascade pro-
cess. One incident particle will interact with many atoms in sequence to produce
multi-ionization until it looses its energy. On the other hand, the ionized electron
sample
Beams
p++e-
Rp
Re
Figure 3. Schematic presentation of the material irradiated simultaneously by protons and electrons.
Rp penetration depth of protons; Re penetration depth of electrons
112 SYNERGISTIC EFFECTS OF PROTON AND ELECTRON RADIATION
5. Conclusions
dose is exceeded, the synergistic effect has less affect on the transmission, with
the total transmittance being less than the sum result of the two single factors. The
existence of synergistic effects justifies the approach of simultaneous irradiation
space environmental testing.
Hydrogen has particular effects during formation of color centers. Although
no hydrogen is contained in JGS3 quartz glass, the implanted proton reacts with
some bonds to produce disturbances or transformations for color center of E 1 ,
reducing the amount of E 1 center and nonbridging oxygen cavities and improving
the stability of radiation resistance for quartz glass within the ultraviolet range.
The particular effect of hydrogen is the main reason for the difference in re-
sults obtained from single proton irradiation and simultaneous protonelectron
irradiation.
The existence of synergistic effects in simultaneous irradiation indicates that
it is difficult obtaining convincing results in any simultaneous realistic conditions
using results from irradiation by single factors. This proves the necessity to develop
complex space environment simulation systems.
References
1. Abraimov, V. V., Shiqin, Y., Shiyu, H., Dezhuang, Y., Kolybaev, L. K., and Verhovtseva, E.
T. (2000) In the Fifth Sino-Russian-Ukrainian Symposium on Space Science and Technology
Held Jointly with the First International forum on Astronautics and Aeronautics Symposium
Proceedings, Harbin Institute of Technology, Harbin, P.R.China, 2000, pp. 706713.
2. Liu, H., Geng, H., He, S., Yang, S., Yang, D., Abraimov, V. V., and Wang, H. (2002) In Sixth Inter-
national Space Conference on Protection of Materials and Structures from Space Environment,
Toronto, Canada, 13 May 2002, pp. 8190.
3. Hosono, H., Kawazoe, H., Oyoshi, K., and Tanaka, S. (1994) Journal of Non-Crystalline Solids
179, 3950.
4. Griscom, D. L. (1980) Physical Review B 22(9), 41924201.
5. Marshall, C. D., Speth, J. A., and Payne, S. A. (1997) Journal of Non-Crystalline Solids 212,
5973.
6. Belostotskyi, V. I. (1998) Journal of Inorganic Materials 34(6), 738741. (Russian Translation)
THE ROLE OF ABNORMAL ELECTRON FLUXES WITH ENERGY
<1 MeV IN THE SURFACE CHARGING DOSE OF SPACECRAFT
1. Introduction
In the present study we continue to analyze the influence of low energy charged
particles (mostly electrons with energy up to 1 MeV) on surface layers of materials
and film coverings. These particles are regularly registered in the inner radiation
belt of the Earth by different spacecraft. We analyze electron fluxes data obtained
from ACTIVE (INTERCOSMOS-24) satellite experiment (E e < 500 keV, orbit
altitude H = 5002500 km, orbit inclination I = 81 ), SPRUT-VI experiment
onboard MIR station (E e = 0.31.0 MeV, orbit altitude H = 350 km, orbit in-
clination I = 51.6 ) and SAMPEX satellite experiment (E e > 150 keV, orbit
altitude H = 520670 km, orbit inclination I = 82 ). These electrons were
measured at geomagnetic index L < 2. The electron flux data were used for
calculations of surface dose (thickness of material layer up to 1020 m).
The distributions of surface dose and electron spectra for different altitudes for
115
J.I. Kleiman (ed.), Protection of Materials and Structures from Space Environment, 115122.
C 2006 Springer. Printed in the Netherlands.
116 O. R. GRIGORYAN ET AL.
2. Instrumentation
In this paper a detailed analysis of electron flux data obtained onboard ACTIVE
(INTERCOSMOS-24) satellite is presented. Electrons were registered by single
Si-surface barrier detectors. The diameter and the thickness of each detector were
8 mm and 300 m, correspondingly. The electron detectors geometric factor was
0.03 cm2 .sr. There were three pairs of detectors that measured electron fluxes at
different angles (99 , 69 , 39 ) with respect to zenith axis of the satellite. All de-
tectors were protected by a Mylar foil to stop protons with energy E p < 500 keV.
The period between subsequent electron flux measurements was 10 s. The pecu-
liarities of the satellite orbit allowed analysis of electron distribution in a wide
range of altitudes. The ACTIVE satellite was launched into a 5002500 km with
an inclination of 81 . The detectors installed on the ACTIVE satellite measured
electrons in seven energy channels from 30 to 500 keV.
3. Results
Electron fluxes under the inner radiation belt were observed in different experi-
ments since 1980s [15]. Previous experiments revealed the existence of electrons
at L = 1.21.8. Figure 1 presents the experimental data obtained in four different
experiments a follows:
A. SPRUT-VI experiment onboard MIR station, altitude H 350 km, electron
energy Ee > 75 keV [1],
B. CORONAS-I satellite experiment, altitude H 500 km, electron energy E e >
0,5 MeV [2],
C. OHZORA satellite experiment, altitude H = 350850 km, electron energy
E e = 0.193.2 MeV,
D. ACTIVE satellite experiment, altitude H = 5002500 km, electron energy
E e = 30500 keV.
Shaded areas correspond to the flux level of approximately 100 (cm2 .s.sr)1 .
A comparison of the data from mentioned experiments shows that locations of
observed electron precipitation zones are similar. These zones have temporal and
spatial stability.
THE ROLE OF ABNORMAL ELECTRON FLUXES 117
by electron interaction with the lightning generated waves as they first cross the
magnetic equatorial plane (at relatively low L) and before they are reflected. The
differences would be due to the latitude of the lightning, to the density gradients
in the ionosphere and the dependence of precipitation flux on L value.
Figure 4 presents the time history of the electron intensity for each energy channel
on the ACTIVE satellite. Two narrow peaks at L region from 1.2 to 1.8 with
maximums at L = 1.5 and 1.78 are observed.
Figure 4. The time history of the electron flux intensity at ACTIVE satellite
THE ROLE OF ABNORMAL ELECTRON FLUXES 119
From the results of experimental data processing it can be suggested that the
longitudinal boundaries and shape of electron flux registration are stable in time
and space for the different year experiments. This also proves the existence of
a constantly operating mechanism. Electrons are registered at low and middle
latitudes at L = 1.21.8 with maximum intensities at L-values around 1.31.4
and 1.71.8.
The next step was the investigation of the dependence of the electron flux
distribution on altitude. The wide altitude range allowed comparing zones of
electron registration and electron spectra at different altitudes. Figure 5 presents the
electron flux distribution in geographical coordinates at L = 1.21.8 for different
altitudes. These maps were constructed using ACTIVE satellite data.
The position of electron flux precipitation at selected L-values and the depen-
dence of these zones on altitude (500, 700, 900, 1100, and 1300 km) are shown,
not taking into account all experimental data at L > 2 and in the region of South
Atlantic Anomaly. The value of electron energy is 30500 keV (every section of
the figure is plotted for whole energy range). The following observations could be
made analyzing figure 5: (a) zones of electron flux registration under the inner ra-
diation belt exist constantly in an altitude interval from 350 km (see previous figure
based on other satellite experiments) up to 1300 km; (b) electrons are observed
at L = 1.21.8 both in northern and in southern hemispheres; (c) the longitu-
dinal size of the zones practically does not depend on altitude up to 1300 km,
120 O. R. GRIGORYAN ET AL.
Figure 6. Electron spectra at (a) L = 1.21.5 and (b) L = 1.61.9 plotted for different altitudes
from 500 km with an increment of 100 km
at higher altitudes the electron flux becomes larger and it is difficult to locate
one.
Figure 6 presents electron spectra at different altitudes from 500 to 1000
km in two L-zones; L = 1.21.5 (figure 6(a)) and L = 1.61.9 (figure 6(b)).
As can be seen, the electron flux value increases with the altitude (10 times
more at 1000 km altitude than at 500 km). Figure 7 compares results obtained in
ACTIVE satellite experiment with other experimental data and with AE8 MIN and
AE8 MAX models. Electron spectra obtained onboard MIR station (SPRUT-VI
experiment), SAMPEX satellite and ACTIVE satellite experiments are presented.
These satellites and the MIR station were launched to different orbits, so there
is experimental data for a wide altitude range. SPRUT-VI data give the minimal
value of electron flux, because the data were obtained of 350 km altitude. At 1000
km altitude the electron flux has the value of AE8 model flux.
Based on the data above, the observed radiation doze values were computed for the
ACTIVE satellite orbit for two altitude values (500 and 1000 km). The results of
computations of the annual absorbed radiation dose in Si as a function of thickness
of an infinite flat Al shield for electron fluxes for 500 km (bottom) and 1000 km
altitude (top) are presented in figure 8.
As can be seen from figure 8, the contribution of the abnormal electron
fluxes under the inner radiation belt at L = 1.21.9 in the absorbed dose increases
with altitude and it is important to take it into account.
THE ROLE OF ABNORMAL ELECTRON FLUXES 121
Figure 8. Absorbed electron dose for various spectra of electron flux at 500 and 1000 km
6. Conclusions
Spacecraft were exposed to impact of charged particles at low and middle latitudes
(L-shell value from 1.2 to 1.8). The contribution of abnormal electron forma-
tions increase with the altitude and are ten times more at 1000 km comparing with
the value at 500 km. For more correct estimation of absorbed radiation dose value,
the modified AP-8 and AE-8 models for low energy values are necessary.
References
1. Grachev, Y. A., Grigoryan, O. R., Novikov, L. S., and Tchurilo, I. V. (2003) In J. Kleiman and
Z. Iskanderova (eds.), Proceedings of ICPMSE-6, 2002, Toronto, Canada. Kluwer Academic
Publishers, 2003, pp. 123130.
2. Biryukov, A., Grigoryan, O., Kuznetsov, S., Ryaboshapka, A., and Ryabukha, S. (1996) Advance
in Space Research 17, 189.
3. Bashkirov, V. F., Denisov, Y. I., Gotseluk, Y. V., Kuznetsov, S. N., Myagkova, I. N., and Sinyakov,
A. V. (1999) Radiation Measurements 30, 537.
4. Seltzer, S. M. (1979) IEEE Transactions on Nuclear Science NS26, 2160.
5. Blake, J. B., Inan, U. S., Walt, M., Bell, T. F., Bortnik, J., Chenette, D. L., and Christian, H. J.
(2001) Journal of Geophysical Research 106, 733743.
VACUUM ULTRAVIOLET RADIATION EFFECTS ON DC93-500
SILICONE FILM
1. Introduction
2. Experimental Methods
Dow Corning (DC) 93-500 Space Grade Encapsulant two-part silicone elastomer
was cast and cured by ENTECH Inc. into a film of approximately 150 m thickness
following a proprietary process for mixing, casting, and curing via a proven tem-
perature/time schedule in an environment free of cure inhibitors. Samples tested
were cut from this stock film.
Tensile testing was conducted using a DDL Inc. Model 200Q electromechanical
test system. Tensile test specimens were punched using a die fabricated to the
specifications defined in American Society for Testing and Materials (ASTM)
126 JOYCE A. DEVER ET AL.
Standard D-638 for type V tensile specimens [10]. Whereas the tensile die pro-
duces an overall sample length of 63.5 mm, samples for these experiments were
trimmed to an overall length of 44.5 mm to better accommodate the VUV ex-
posure chamber dimensions. The trimmed length was simply the excess material
that would fall outside the grips of the tensile tester. Sample width is 9.52 mm
with a narrow section of 3.18 mm width and 7.62 mm length. The initial grip
separation distance was 25.4 mm. Samples were tested using a speed of 63.5
mmmin1 and load vs. displacement data were obtained. From the load vs. dis-
placement data, ultimate tensile strength (UTS) and elongation at failure were
obtained.
The facility used for vacuum ultraviolet exposure has been described in detail
elsewhere [6]. The facility uses a cryogenic vacuum pumping system. The test
chamber contains four individual VUV exposure areas separated by water-cooled
copper walls to minimize cross interactions between compartments. Each expo-
sure area contained a 30-watt VUV deuterium lamp with a magnesium fluoride
end-window (Hamamatsu model L7293) which provided broad spectrum VUV
with a lower cut-off wavelength of 115 nm. Sample stages, one per area, were mo-
tor controlled so that the intensity of VUV light could be adjusted by changing the
distance between the VUV light source and the sample. Each exposure area was
equipped with a cesium iodide (CsI) phototube, calibrated to a NIST-measured
deuterium source, to make measurements of lamp intensity (in the wavelength
range of sensitivity of the detector, 115200 nm) at the sample distance. Lamp in-
tensity measurements were used to determine the number of VUV suns, defined
as the ratio of lamp intensity to air mass zero solar intensity the 115200 nm wave-
length range. For reference, the air mass zero solar intensity in the 115200 nm
wavelength range is 1.073E-5 Wcm2 [7]. Multiplying the number of suns by test
exposure hours provides the number of equivalent sun hours, or ESH, which is
the equivalent time for direct air mass zero solar exposure. Additionally, multiply-
ing the lamp intensity (in units of Wcm2 ) by test exposure duration (in seconds)
provides incident energy fluence (in units of Jcm2 ). These exposure conditions
will be reported with material degradation results.
Prior to exposure, samples were installed in the facility and the chamber was
brought to high vacuum. An operating pressure of approximately 4 106 Torr
was achieved within 24 h of establishing high vacuum prior to commencing ul-
traviolet exposure. The test chamber was brought to atmosphere using nitrogen in
order to remove or replace samples. Separate from the sample exposure tests, ther-
mocouple temperature measurements were made which indicated that the VUV
lamps did not cause the temperature of the sample stage at the testing distances
to rise above room temperature.
A comparison of the air mass zero solar irradiance spectrum and the deuterium
lamp irradiance spectrum is shown in figure 1 [12]. The lamp irradiance spectrum
was obtained by the National Institute of Standards and Technology using a dis-
tance of 25.4 cm from the source to the detector. For the results described in this
128 JOYCE A. DEVER ET AL.
1 E-03
1 E-04
Irradiance (W/cm2 nm)
1 E-05
1 E-06
1 E-07
1 E-10
100 200 300 400
Wavelength (nm)
Figure 1. Air mass zero solar irradiance compared to deuterium VUV lamp irradiance at a 25.4 cm
source-to-detector measurement distance
paper, the VUV exposure facility used distances that were significantly greater
than 25.4 cm, between 50 and 92 cm, so the absolute irradiance values shown
in figure 1 are much higher than would be used for typical test conditions. How-
ever, from figure 1 it is evident that the spectral shapes are very different between
the air mass zero sun and the deuterium lamp. For example, the deuterium lamp
shows a peak around 160 nm that is absent in the solar spectrum. Additionally,
beyond approximately 170 nm, the deuterium lamp irradiance is much less than
that of the sun. Test conditions using this facility for wavelength dependence and
intensity dependence studies for silicone DC93-500 degradation are described in
the sections below.
Source to
sample Exposure Avg. Cumulative Avg. Cumulative
Test distance duration intensity incident energy number of equivalent
area (cm) (h) (Wcm2 ) fluence (Jcm2 ) suns sun hours
UV light is at the lamp output window. Because the VUV lamps are located on
ports that can be isolated from the main sample chamber, lamp maintenance was
conducted while the samples remained under high vacuum conditions. Whereas
table 2 indicates the time-average intensity and number of suns, there is an intensity
variation across the sample area, which is approximately 15% of the measured
value.
Figure 2 shows the total transmittance spectra of 152 m thick DC93-500 films
that were exposed to VUV from beneath various narrow bandpass filters, where
test conditions are fully described in table 1. Data in figure 2 are shown only for
samples exposed to VUV for 408 h, because spectra for samples exposed for 74 h
were difficult to distinguish from the spectrum of the pristine sample. The legend
for figure 2 indicates the filter peak wavelength and equivalent space exposure
in units of ESH for each wavelength band from the exposure conditions listed in
table 1. The most significant changes to the spectra are in the UV wavelengths
below 300 nm as shown in figure 2(a). Significant degradation was observed
for all four exposure wavelength ranges. This indicates that degradation is not
exclusively caused by exposure to an individual wavelength or narrow wavelength
band. However, data can be further analyzed to establish whether any of these
wavelength bands is more effective than others in causing degradation. Table 3
shows data for transmittance at 250 nm for all exposed samples and a pristine
VACUUM ULTRAVIOLET RADIATION EFFECTS 131
90 a
b
80
Transmittance (%)
c
d
e
70
a: Pristine
60 b: 188 nm, 30 ESH
40
200 210 220 230 240 250 260 270 280 290 300
Wavelength (nm)
(a)
100
90
Transmittance (%)
80
a: Pristine
50
200 700 1200 1700 2200
Wavelength (nm)
(b)
Figure 2. Effect of VUV exposure of various wavelength ranges on spectral transmittance of 152 m
DC93-500 silicone film for (a) the wavelength range of 200300 nm where the most significant
degradation occurred and (b) the overall spectral measurement range of 2102500 nm
sample obtained from the spectral data shown in figure 2. Figure 3 shows a plot of
transmittance at 250 nm, a wavelength where significant degradation is evident, vs.
incident energy fluence (listed in table 1) for all exposures. It is evident from this
figure that the VUV filters providing longer VUV wavelengths cause a higher rate
132 JOYCE A. DEVER ET AL.
Equivalent sun
Filter peak VUV exposure Cumulative hours in filter Transmittance
wavelength wavelength incident energy wavelength measured at
(nm) range (nm) fluence (Jcm2 ) range 250 nm
of transmittance degradation than the shorter VUV wavelengths. The most likely
explanation for this is that the longer wavelengths are more deeply penetrating
into DC93-500 and, therefore, affect more of the bulk of the material. Depth
of penetration of VUV radiation relative to these observations will be further
discussed in section 3.2.
100
80
Transmittance
60
151 nm filter
170 nm filter
40
182 nm filter
188 nm filter
151 nm filter
20 170 nm filter
182 nm filter
188 nm filter
0
0 1 2 3 4 5 6
1000
100
Penetration depth, m
10
0.1
0.01
125 150 175 200 225 250 275
Wavelength, nm
Figure 5 shows the total transmittance spectra of 152 m thick DC93-500 films
exposed to VUV of various intensity levels and for various equivalent sun hours.
134 JOYCE A. DEVER ET AL.
100
(a)
90
80
(b)
70
Transmittance (%)
(c)
60
(d)
(e)
50 (a) Pristine
(f)
40 (b) 200 ESH, 1.7 suns
(g)
(c) 370 ESH, 3.2 suns
30
(d) 570 ESH, 1.5 suns
20 (e) 700 ESH, 6.0 suns
(f) 1120 ESH, 3.0 suns
10
(g) 1930 ESH, 5.5 suns
0
200 300 400
Wavelength (nm)
(a)
100
90
80
70
Transmittance (%)
60
50 (a) Pristine
(b) 200 ESH, 1.7 suns
40
(c) 370 ESH, 3.2 suns
0
200 400 600 800 1000 1200 1400 1600 1800 2000 2200 2400
Wavelength (nm)
(b)
Samples were exposed to the unfiltered deuterium lamp output, wavelengths above
115 nm, and output is reported in the 115200 nm band, which contains the
majority of lamp output as shown in figure 1. It is evident that the most significant
degradation occurred in the ultraviolet wavelengths, below 400 nm, as is shown in
VACUUM ULTRAVIOLET RADIATION EFFECTS 135
figure 5(a). It is also evident that degradation increases with increasing equivalent
sun hours of exposure.
Table 4 and figure 6 show transmittance at 300 nm (a wavelength at which
significant transmittance degradation of DC93-500 is evident, based on figure 5)
as a function of exposure, represented by thousands of ESH, for various intensities
(number of suns). In the figure, symbols indicate the measured data points, and
lines indicate exponential decay curve fits. In addition to examining data for
individual intensities, data were considered all together, independent of intensity,
as one of the curves shown in figure 6.
100
80
Transmittance, %
60
All Data
1.5 suns
40 3.0 suns
5.5 suns
All Data
20 1.5 suns
3.0 suns
5.5 suns
0
0.0 0.5 1.0 1.5 2.0
Thousands of ESH
When transmittance data for all intensities are considered together, the best
curve fit equation is given in the format shown in eq. (3):
= y0 + a exp(kx) (3)
where is transmittance (%), x is exposure duration (thousands of equivalent sun
hours), and y0 , a, and k are constants. Eq. (3) represents an exponential decay
reaching an asymptotic value, which is evident from examination of figure 5. We
can further determine the expression for transmittance of unexposed materials, or
0 , by setting x = 0:
0 = y0 + a (4)
Using the expression for a as a function of y0 and 0 from eq. (4), and by
expressing x (thousands of equivalent sun hours) as the product of equivalent
suns, S, and thousands of exposure hours, h, eq. (3) can be rewritten as
= y0 + (0 y0 ) exp(k Sh) (5)
In order to determine whether the rate of transmittance decay shows a dependence
upon S, the intensity (expressed as number of equivalent suns), eq. (5) is solved
for k, the constant in the exponential expression, to give:
1 y0
k= ln (6)
Sh 0 y0
The curve fit for all measured transmittance data at 300 nm (figure 6) produces
a value of y0 = 46.7, which is the asymptote being approached by the decay
in transmittance and which is assumed to be constant, independent of intensity.
If we examine transmittance data at 300 nm for all intensities and plot k vs.
S, it is possible to determine whether k is constant, which would mean it is
independent of intensity, or whether it varies as a function of intensity, indicating
intensity dependence. The values for S and h are obtained from the VUV exposure
conditions (table 2), and and 0 values are the measured transmittance data
(shown in table 4). The plot of k vs. S is shown in figure 7. Based on the data
shown in figure 7, it is evident that there is no statistically significant dependence
upon exposure intensity between approximately 1.5 and 5.5 VUV suns.
Table 5 and figures 8(a) and 8(b) show mechanical properties of 152 m DC93-
500 silicone films as a function of exposure (equivalent sun hours). Both ultimate
tensile strength (figure 8(a)) and elongation at failure (figure 8(b)) decrease with
increasing exposure and indicate the approach of an asymptotic value near the
2000 ESH exposure level. An exponential decay curve fit is shown in each figure.
Based on these data, especially because of similar degradation for similar ESH,
regardless of intensity, there is no clear trend indicating an intensity dependence
upon the rate of mechanical properties degradation.
VACUUM ULTRAVIOLET RADIATION EFFECTS 137
2.5
1.5
k
0.5
0
0 1 2 3 4 5 6
S (Number of Suns)
Figure 7. Coefficient, k, described by eq. (6), as a function of VUV exposure intensity for DC93-500
film samples
4. Conclusions
DC93-500 silicone has been found to undergo degradation in optical and mechan-
ical properties upon exposure to a laboratory deuterium lamp providing VUV
radiation of wavelengths greater than 115 nm. In one experiment, samples of
DC93-500 films were exposed to narrow bands of VUV radiation (20 nm) in
wavelength ranges between approximately 140 and 200 nm by using a broad
spectrum deuterium lamp as the VUV source and narrow bandpass filters over
10.0
7.0
6.0
UTS (MPa)
5.0
4.0
3.0
2.0
1.0
0.0
0 200 400 600 800 1000 1200 1400 1600 1800 2000
VUV Exposure (ESH)
(a)
160.0
5.5 suns
140.0 3 suns
1.5 suns
S i 1
Elongation at Failure (%)
120.0
100.0
80.0
60.0
40.0
20.0
0.0
0 200 400 600 800 1000 1200 1400 1600 1800 2000
VUV Exposure (ESH)
(b)
Figure 8. Mechanical properties of (a) tensile strength and (b) elongation at failure for 152 m
DC93-500 silicone as a function of VUV exposure equivalent sun hours
the DC93-500 samples. Results indicated that each wavelength range used for the
exposures produced degradation in DC 93-500 optical properties. However, degra-
dation per incident energy fluence indicated the highest rate of degradation for
samples exposed from beneath filters which included wavelengths above 185 nm.
Vacuum ultraviolet ellipsometric optical measurements were made on DC93-500
silicone to determine the depth of penetration of vacuum ultraviolet light as a func-
tion of wavelength. Data showed that vacuum ultraviolet of wavelengths below
VACUUM ULTRAVIOLET RADIATION EFFECTS 139
Acknowledgments
The authors gratefully acknowledge the technical support of Frank Lam and James
Mazor (Akima), Michael DePauw (NASA), Scott Panko and Edward Sechkar
(QSS Group, Inc.), Michael Piszczor (NASA) and Mark ONeill, Don Spears,
and A. J. McDanal (ENTECH, Inc.).
References
1. Dow Corning Corp. (2001) Product Information for Dow Corning Space-Grade Silicone
Sealants, 2001.
2. ONeill, M. J., Piszczor, M. F., Eskenazi, M. I., McDanal, A. J., George, P. J., Botke, M. M.,
Brandhorst, H. W., Edwards, D. L., and Hoppe, D. T. (2003), In International Symposium on
Optical Science and Technology, SPIEs 48th Annual Meeting, SPIE Paper no. 517917, August
2003.
3. Edwards, D. L., Finckenor, M. M., ONeil, M., and McDanal, A. J. (2000) In 8th International
Symposium on Materials in a Space Environment, Arcachon, France, 2000.
140 JOYCE A. DEVER ET AL.
4. Dever, J., Messer, R., Powers, C., Townsend, J., and Wooldridge, E. (2001) High Performance
Polymers 13(3), S391S399.
5. Dever, J., Semmel, C., Edwards, D., Messer, R., Peters, W., Carter, A., and Puckett, D. (2002)Ra-
diation Durability of Candidate Polymer Films for the Next Generation Space Telescope Sun-
shield, AIAA 2002-1564, April 2002.
6. Dever, J. and McCracken, C. (2004) High Performance Polymers 16(2), 289302.
7. American Society for Testing and Materials (Reapproved 1992). Solar Constant and Air Mass
Zero Solar Spectral Irradiance Tables, ASTM-E 490-73a.
8. Dever, J. A. (1994) Flight-vehicle materials, structures, and dynamicsAssessment and
Future Directions, Vol. 2, Advanced Metallics, Metal-Matrix and Polymer-Matrix Composites,
American Society of Mechanical Engineers, New York, 1994, pp. 422433.
9. Adams, M. R. (1993) The Degradation of Polymeric Spacecraft Materials by Far-UV Radiation
and Atomic Oxygen, UMI Dissertation Services, Ann Arbor, MI, 1993, p. 138.
10. American Society for Testing and Materials. (1995) Standard Test Method for Tensile Properties
of Plastics, ASTM D 63895.
11. Herzinger, C. M., Snyder, P. G., Johs, B., and Woollam, J. A. (1995) Journal of Applied Physics
77(4), 17151724.
12. Dever, J. A., Pietromica, A. J., Stueber, T. J., Sechkar, E. A., and Messer, R. K. (2001) Simulated
Space Vacuum Ultraviolet (VUV) Exposure Testing for Polymer Films, AIAA Paper no. 2001
1054, American Institute of Aeronautics and Astronautics, January 2001.
ENHANCEMENT OF ATOMIC OXYGEN-INDUCED EROSION OF
SPACECRAFT POLYMERIC MATERIALS BY SIMULTANEOUS
ULTRAVIOLET EXPOSURE
Key words: atomic oxygen, ultraviolet, low Earth orbit, synergy, space environ-
ment effect
1. Introduction
There are many environmental factors in low Earth orbit (LEO) such as micro-
gravity, thermal cycling, plasma environment (ions and electrons), ultraviolet,
radiation, neutral gas, high-energy charged particles, and space debris. However,
it has been recognized that atomic oxygen is one of the most important hazards to
Corresponding authors
141
J.I. Kleiman (ed.), Protection of Materials and Structures from Space Environment, 141152.
C 2006 Springer. Printed in the Netherlands.
142 KUMIKO YOKOTA ET AL.
2. Experimental Details
The samples used in this experiment were low-density polyethylene (LDPE) and
polyimide films. Both of the films were spin-coated on a QCM sensor crystal. The
polyethylene solution containing 0.3 g of LDPE (average molecular weight: 6500)
in 40 ml xylene was prepared for polyethylene film. The pyromelliticdianhydride
(PMDA)-oxydianiline (ODA) polyimide, which was supplied by Toray Industries
Inc. (Semicofine SP-510), was used as a polyimide sample. Precursor of PMDA-
ODA polyimide was spin-coated on a QCM sensor crystal and then annealed
at 150 C and at 300 C. Details of the sample preparation are reported in [6].
The polyimide film, thus prepared, showed an XPS spectrum similar to that of
Kapton-H film, which is commercially available polyimide film.
ENHANCEMENT OF ATOMIC OXYGEN-INDUCED EROSION 143
XPS
Temperature-controlled
Rotatable QCM
Nozzle
AO
PSV
Laser
QMS
Au Mirror
TOF System
UV Source Sample
(excimer 172 nm)
Figure 1. A schematic drawing of the space environmental simulation facility using a laser detona-
tion atomic oxygen beam source
25
0 deg.
10 deg.
20 20 deg.
30 deg.
0
0 50 100 150 200 250
Figure 2. Resonant frequency shift of the polyethylene-coated QCM during atomic oxygen expo-
sures at impingement angles from 0 to 90 . Atomic oxygen flux: 5.9 1014 atomscm2
In this study, the relative intensity of atomic oxygen and ultraviolet was adjusted
by rotating QCM sample along the axis that is perpendicular to the axes of both the
atomic oxygen and the ultraviolet. However, in this experimental configuration,
the incident angle of atomic oxygen at polymer surface was changed when relative
atomic oxygen/ultraviolet intensity was adjusted. Therefore, it was necessary to
investigate the effect of incident angle of atomic oxygen and ultraviolet on the
polymer erosion prior to the quantitative analysis of synergistic effect of atomic
oxygen and ultraviolet. Figure 2 shows the frequency shift of polyethylene-coated
QCM during atomic oxygen exposures at incident angles from 0 to 90 . The
incident angle of 0 means that atomic oxygen hit the sample surface at normal
incidence. As can be seen in figure 2, a good linear relationship between the
frequency shift and exposure time was observed at all incident angles. The slope
of the frequency shift, i.e., the mass loss rate of polyethylene film, was calculated
by a least square fit and plotted against each incident angle. The result is shown in
figure 3. It is observed that the erosion rate decreases with increasing incident angle
and follows a cosine distribution, which is indicated by the solid line. The data
point at 80 did not fit the cosine curve because the QCM holder partially blocked
atomic oxygen beam. The experimental result on incident angle dependence of
atomic oxygen-induced erosion of polyethylene showed a similar tendency with
those reported on polyimide film [8]. The fact that the incident angle dependence of
the erosion rate follows a cosine law indicates that the erosion rate of polyethylene
is proportional to the effective flux of atomic oxygen; namely, the reaction yield
of atomic oxygen with polyethylene is constant at all incident angles.
ENHANCEMENT OF ATOMIC OXYGEN-INDUCED EROSION 145
0.07
0.06
0.04
0.03
0.02
0.01
0.00
0 10 20 30 40 50 60 70 80 90 100
Figure 4. The atomic force microscope image of atomic oxyge-exposed polyethylene surface.
Atomic oxygen fluence: 3.0 1018 atomscm2
146 KUMIKO YOKOTA ET AL.
0 to 60 leads to the reduction of effective incident energy from 5.0 to 1.25 eV; it
is proportional to the cosine square of the incident angle. The activation energies
of gasification reaction of polymers in this incident energy range are expected as
low as 103 eV [9]. It was, thus, concluded that the effect of activation energies
in the incident angle dependence of erosion rates could be negligibly small. From
the experimental results shown here, it is concluded that the flux control of atomic
oxygen can be achieved by tilting the sample with respect to the beam axis.
0.10
0.08 UV, AO
Erosion Rate (Hz/s)
q
0.06
0.04
0.02
0.00
0 10 20 30 40 50 60 70 80 90
AO
q
90
UV
Figure 6. The set up of AO beam and UV: The axes of AO and UV were crossed at 90 degree, so
that relative intensity of AO and UV can be adjusted
0.10 0.10
(a) (b)
0.08 0.08
Erosion Rate (Hz/s)
0.06 0.06
0.04 0.04
0.02 0.02
0.00 0.00
0 10 20 30 40 50 60 70 80 90 0 10 20 30 40 50 60 70 80 90
Figure 7. Erosion rate of the polyimide-coated (a) and polyethylene-coated (b) QCM during only
atomic oxygen () and atomic oxygen & ultraviolet exposures (). (a) AO: 6.8 1014 atomscm2 /s,
UV: 0.40 mW/cm2 , (b) AO: 6.0 1014 atomscm2 /s, UV: 0.55 mW/cm2
148 KUMIKO YOKOTA ET AL.
0
0 1 2 3
Figure 8. The relationship between the relative intensity of UV/AO and the normalized erosion rate
of polyethylene and polyimide
As reported elsewhere [11], the synergistic effect of atomic oxygen and 172 nm
ultraviolet on polyimide was obvious when ultraviolet intensity is high compared
with atomic oxygen; (UV/AO ratio >1 1014 mJatom1 ). However, this syn-
ergistic effect was not observed in the range of UV/AO ratio < =1015 mJatom1 ,
as demonstrated in figure 7(a); atomic oxygen-induced erosion data with ultravi-
olet (empty circle) and without ultraviolet (solid circle) are identical. In contrast,
polyethylene shows clear effect of simultaneous ultraviolet irradiation even in the
range of UV/AO ratio < =1015 mJatom1 as shown in figure 7(b); i.e., the ero-
sion rates with ultraviolet exposure (empty circle) are 30100% greater than those
without ultraviolet exposure (solid circle).
From the result of figure 7, the relationship between the relative intensity of
UV/AO and flux normalized erosion rate of polyethylene was replotted in fig-
ure 8. The ordinate was normalized erosion rate, namely that means the erosion
rate in the case of atomic oxygen and ultraviolet exposures was divided by those
in the case of only atomic oxygen exposure. The erosion rate of atomic oxygen-
induced polyethylene was enhanced 150180% at the relative intensity of UV/AO
0.52.6 1015 mJatom1 by simultaneous ultraviolet exposure. It was observed
that the erosion rate increased 300% when the ultraviolet intensity was high (not
shown). In the case of polyimide, the similar effect by simultaneous ultravio-
let exposure was observed at relative intensities one order greater than that of
polyethylene [11]. This finding suggests that polyimide is a better material as a
witness sample for measuring atomic oxygen fluence in LEO since synergistic
effect was only in high UV/AO conditions.
ENHANCEMENT OF ATOMIC OXYGEN-INDUCED EROSION 149
0.025
Residual
VUV only
0.020 least square fits
0.015
0.010
0.005
0.000
0 10 20 30 40 50 60 70 80 90 100
Figure 9. Erosion Rate of polyethylene-coated QCM under VUV exposure (), and the contribution
of VUV in the synergistic effect (). Note that the abscissa represents the incident angle of atomic
oxygen
in the present experimental condition. Since the absolute reaction rate is greater
in atomic oxygen-induced erosion, it was considered that the overall erosion was
rate-limited by atomic oxygen.
In the previous section, the importance of the oxygen atoms adsorbed at the
polyethylene surface on the ultraviolet-induced erosion was suggested. XPS was
used for analyzing the oxygen adsorbed polyethylene surface. The experiment was
carried out with atomic oxygen fluence of 3.0 1018 atomscm2 . The sample was
exposed to 172 nm ultraviolet radiation (0.45 mWcm2 ) for 40 min in vacuum.
Figure 10 indicates the C1s core level XPS spectra of atomic oxygen-exposed
polyethylene before and after ultraviolet exposure. As clearly observed in figure
10, atomic oxygen-exposed polyethylene showed a high-energy shoulder at 288.5
eV besides the main peak at 284.8 eV. This shoulder is contributed by carboxyl
groups. The high-energy peak disappeared under the ultraviolet irradiation. Simi-
lar results were obtained from polyimide samples. Table 1 shows the composition
of the polyethylene and polyimide surfaces analyzed by XPS. After atomic oxygen
exposure, the oxygen composition increased to 27% for polyethylene and to 33%
for polyimide. Increase in oxygen composition corresponds to surface oxidation
forming carboxyl groups. However, the high-energy shoulder disappeared after
ultraviolet exposure, and the surface composition recovered to normal values. This
spectral change in XPS is explained by the photodissociation of carboxyl group
created by exposure to atomic oxygen [12]. Similar conclusion was also obtained
in polyimide [12]. It was, thus, confirmed by XPS that ultraviolet irradiation
promotes the decomposition of carboxyl group formed by atomic oxygen expo-
sure at polyethylene and polyimide surfaces. This photochemical reaction is the
2500
(a
(a)
2000
(b)
1500
(c)
c/s
1000
500
Figure 10. C1S XPS spectra of atomic oxygen-exposed polyethylene before and after ultraviolet
exposure. (a): Pristine, (b): AO exposed, (c): AOUV exposed
ENHANCEMENT OF ATOMIC OXYGEN-INDUCED EROSION 151
TABLE 1. The composition of polyethylene and polyimide before/after atomic oxygen and
ultraviolet exposure
Polyethylene Polyimide
Samples C O C O N
Polyethylene: AO fluence: 3.0 1018 atoms/cm2 , UV intensity: 0.45 mW/cm2 for 40 min.
Polyimide: AO fluence: 9.6 1017 atoms/cm2 , UV intensity: 0.6 mW/cm2 for 4 hr.
4. Conclusion
Acknowledgments
This study was partially supported by the Grant-in-Aid for Scientific Research
from the Ministry of Education, Culture, Sports, Science, and Technology, Japan
152 KUMIKO YOKOTA ET AL.
under contact no. 13750842, 14350511, and 15560686; and the Space Utiliza-
tion Promotion from the Japan Space Forum. Financial support from Kawanishi
Memorial Shinmaywa Education Foundation is also acknowledged. The authors
express their acknowledgments to S. Seikyu and K. Maeda of Kobe University
for their help with this experiment.
References
1. Standard Practices for Ground Laboratory Atomic Oxygen Interaction Evaluation of Materials
for Space Applications, ASTM Designation, E2089-00.
2. Atomic Oxygen Effects Measurements for Shuttle Missions STS-8 and 41-G, NASA TM-100459,
1988.
3. Rutledge, S. K., Banks, B. A., and Kitral, M. (1998) A Comparison of Space and Ground Based
Facility Environmental Effects for FEP Teflon, NASA TM 207918.
4. Brinza, D. E., Chung, S. Y., Minton, K. T., and Liang, R. H. (1994) Final Report on the NASA/JPL
Evaluation of Oxygen Interactions with Materials-3 (EOIM-3), JPL Publication, pp. 9431.
5. Koontz, S. L., Leger, L. J., Albyn, K., Cross, J. (1990) Journal of Spacecrafts and Rockets 27,
346348.
6. Kinoshita, H., Tagawa, M., Umeno, M., and Ohmae, N. (1998) Transaction of the Japan Society
for Aeronautical and Space Science 41(132), 9499.
7. Tagawa, M., Yokota, K., Ohmae, N., and Kinoshita, H. (2002) Journal of Spacecraft and Rockets
39(3), 447451.
8. Yokota, K., Tagawa, M., and Ohmae, N. (2002) Journal of Spacecraft and Rockets 39(1), 155
156.
9. Tagawa, M., Yokota, K., Kida, T., and Ohmae, N. (2003) In J. Kleiman and Z. Iskanderova (eds.),
Protection of Materials and Structures from Space Environment Space Technology Proceedings,
Vol. 5, Kluwer Academic Publishers, Dordrecht, pp. 391400.
10. http://www1.ushio.co.jp/tech/
11. Yokota, K., Ohmae, N., and Tagawa, M. (2004) High Performance Polymers 16, 221234.
12. Scnabel, W. (1992) Polymer Degradation-Principles Practical Applications, Carl Hanser Verlag,
Munich.
GROUND SIMULATION OF HYPERVELOCITY SPACE DEBRIS
IMPACTS ON POLYMERS
1. Introduction
Many of the satellites nowadays are being launched into low Earth orbit (LEO),
ranging from 200 to 700 km. The LEO space environment possesses many
obstacles to a successive spacecraft mission. These obstacles include ambi-
ent space conditions such as ultrahigh vacuum, as well as man-made obstacles
such as spacecraft debris. In LEO, spacecrafts are subjected to various destruc-
tive environmental components, such as ionizing radiation (electrons, protons),
153
J.I. Kleiman (ed.), Protection of Materials and Structures from Space Environment, 153165.
C 2006 Springer. Printed in the Netherlands.
154 R. VERKER ET AL.
vacuum ultraviolet (VUV) radiation, hyperthermal atomic oxygen (AO), and other
factors such as extreme temperature variations, micrometeoroids, and orbital
debris.
Due to either singular or synergistic interactions with these space components,
structural materialsin particular polymer-based materialssuffer a relatively
rapid erosion (mass loss), structure modification, and surface roughening, leading
to irreversible degradation of their physical characteristics (optical, thermal, elec-
trical, and mechanical) [1, 2]. Therefore, a careful selection of surface materials,
namely polymer films and paints, is required [3].
Micrometeoroids originate naturally from planetary or asteroidal collisions
and cometary ejecta [4]. The large debris population at LEO altitudes com-
prises the waste products of spacecraft operations. Typical velocities of debris
particles range from few kilometers per second up to 16 kms1 , making these
particles a threat to spacecrafts. The debris issue must be quantified over the pro-
jected lifetime of a space system to determine the life expectancy of exposed
systems and to quantify necessary shielding requirements [5]. Artificial space de-
bris consists of large objects such as spent satellites and rockets, and mostly of
small objects such as aluminum oxide fuel particles, paint chips, and fragmentation
objects from collisions of these bodies in orbit [4].
The recovery of several spacecrafts in the last decade offers information con-
cerning the directionality of the LEO meteoroids and space debris fluxes [6]. Such
recovered spacecrafts and spacecrafts parts include one of the Hubble space tele-
scope solar arrays (retrieved in 1993), the European retrievable carrierEURECA
(retrieved in 1993), and the long duration exposure facilityLDEF (retrieved in
1990 after 69 months in LEO).
Spacecraft debris impact damage can degrade the performance of exposed
spacecraft materials and, in some cases, destroy a satellites ability to perform or
complete its mission [3]. The Hubble space telescope solar array, for example,
suffered impacts at ultrahigh velocities ranging from 2.9 to 11.5 kms1 from
particles 7 to 98 m in diameter [7]. Particles traveling at ultrahigh velocities
generate temperatures in the range of 5000 K and pressures of several megabars
when they collide with a surface [8].
Accumulation of impacts over the large surface area of solar panels leads, in
some cases, to degradation in efficiency [9]. Impacts into metals form craters,
which have diameters averaging about 5 times the impact diameter. These craters
are of concern because they can prevent impacted components from operating. In
the case of composites, if a complete penetration occurs, this can lead to further
breakdown of the composite during subsequent exposure to AO or VUV. Debris
impacts into polymer films occurs quite often, since they are used extensively
onboard spacecrafts, mainly as thermal blankets. Mostly, these materials are thin
laminated layers; thus, the impacts cause delamination of these layers into many
times the diameter of the crater [3].
GROUND SIMULATION OF HYPERVELOCITY SPACE DEBRIS IMPACTS 155
2. Experimental
The laser driven flyer (LDF) method was used for simulating space hypervelocity
debris with dimensions ranging from 10 to 100s m and velocities of up to
2900 ms1 . LDF is attractive as an acceleration technique for debris simulation
due to its relative simplicity, relatively low cost, ease of incorporation into a
vacuum facility, and high shot rate capability [8, 10, 11].
Figure 1 shows a schematic diagram of the LDF process. In this method,
a high-intensity laser beam is shot into a metal foil (325 m-thick) tenaciously
bonded to a glass substrate (hereafter referred as target). The beam passes through
the glass and hits the aluminumglass interface. At the interface, high-pressure
Plasma
(Shock wave)
Laser beam
Glass/Aluminum
laminate
plasma is formed at the range of giga-Pascal (GPa). The instantaneous high pres-
sure generates a compressive shock wave that propagates into the aluminum faster
than the speed of sound. As the shock wave reaches the aluminum surface and
the laser pulse ends, two (tensile) rarefaction waves are generated propagating
one toward the other. When the rarefaction waves, and their tension pressure is
higher than the spall pressure, a spall is produced. A pressure gradient between
the plasma pressure on one side and the vacuum zero pressure on the other side
causes the spalled layer acceleration which results in an aluminum layer to fly away
at ultrahigh velocity. The aluminum flyers in this work were accelerated toward
polymer samples located at a selectable distance of 212 mm from the laminate
structure. The whole system was placed inside a vacuum chamber operating at a
base pressure of 20 mTorr.
Soreqs LDF system is using a Titanium:Sapphire laser operating under a
single-pulse mode at 810 nm wavelength. The length of each pulse is 300 ps,
and the energy of the pulse can be controlled within the range of 250750 mJ.
The experiments were carried out with single laser shots. After each shot, a new
sample was positioned and a new unexposed area of aluminumglass target was
placed into the laser beam path.
5000 Theoretical
Measured
Velocity (m/s)
4000
3000
2000
1000
200 300 400 500 600 700 800
Laser pulse energy (mJ)
Figure 2. Flyer measured and theoretical velocities as a function of the laser pulse energy
GROUND SIMULATION OF HYPERVELOCITY SPACE DEBRIS IMPACTS 157
0.008
0.007
0.005
0.004
0.003
0.002
0.001
0.000
6 6 6 5 5
0.0 3.0x10 6.0x10 9.0x10 1.2x10 1.5x10
Time (sec)
(a) (b)
Figure 3. (a) Schematic description of the LDF velocity measurements, and (b) a scope display
during a velocity measurement experiment
hydrodynamic coefficient H = 1.0. This value means that the whole laser pulse
energy is transferred into flyer kinetic energy. The hydrodynamic coefficient is
defined as
E KF V2
H = = M2 (1)
E LP VT
where E KF is the flyer kinetic energy, E LP is the laser pulse energy, and VM and
VT are the measured and theoretical velocities, respectively.
According to the theoretical and measured velocities presented in figure 2 and
eq. (1), the Soreqs LDF system hydrodynamic coefficient was calculated to be
H = 0.23. The flyer velocity was measured using the system shown schematically
in figure 3(a).
A continuous He:Ne laser beam was set orthogonal to the flyers trajectory,
so that the beam crossed the flyer path twice in the presence of a prism. The
two parallel beams were set at a known distance of 13 mm from each other.
A photodiode attached to a scope received the continuous laser signal. As the
flyers crossed the continuous lasers path, two peaks were detected by the scope,
allowing the velocity calculation. Figure 3(b) shows a typical scope display with
peaks time difference of 6 s obtained at laser pulse energy of 650 mJ, indicating
a flyer velocity of 2000 ms1 .
100 m
Figure 4. A typical ESEM image of an impact-induced crater on a glass surface. Impact conditions
were 250 mJ pulse energy, 300 ps pulse length, 100 mTorr system pressure, and flyer velocity of
1400 ms1
mTorr, and pulse length of 300 ps. The flyer velocity was 1400 ms1 . Sample
morphology was characterized using an environmental scanning electron micro-
scope (ESEM, model Quanta 200 from FEI), which allows characterization of
nonconductive samples (e.g., Kapton and glass) without the need for a conductive
coating. Figure 4 shows a typical SEM image of an impact-induced crater on a
glass surface.
Flyer velocities and crater diameters, as determined from the ESEM images,
were applied into the conchoidal cracking diameter equation [7]:
0.5
Dco = 5 104 S P0.71 dP1.13 vp0.754 (2)
where Dco is the concentric cracking region diameter, s and p are the sample and
flyer densities, respectively, dp is the plate diameter, and vp is the plate velocity.
The flyer dimensions were calculated to be 2329 m in diameter. This calculation
was conducted only for a velocity of 1400 ms1 .
In practice, each laser shot at the target produced several major flyers being
part of a cloud of such flyer fragments, all traveling at ultrahigh velocity. In order
to evaluate the clouds dimensions, a 12 m-thick aluminum foil was used as the
impacted sample. The cloud of flyers punctured the foil, creating a hole of 1.5 mm
in diameter. The diameter of the hole resembles the dimensions of the cloud.
The morphology of the fractures created by the debris impacts was analyzed using
the same ESEM.
GROUND SIMULATION OF HYPERVELOCITY SPACE DEBRIS IMPACTS 159
3. Materials
Figure 5 demonstrates the effect of flyer velocity on the extent and nature of
damage developed in impacted 25 m-thick Kapton films. The fractures were
created using final flyer velocities VF of 1400 (figure 5(a)), 1650 (figure 5(b)),
Figure 5. ESEM images of 25 m-thick Kapton films impacted by debris at velocities of (a) 1400
ms1 , (b) 1650 ms1 , (c) 1730 ms1 , and (d) 2900 ms1 , respectively
160 R. VERKER ET AL.
1730 (figure 5(c)), and 2900 (figure 5(d)) ms1 . All ESEM images were taken
from the impact exit side.
The volcano-like puncture sites obtained at the lowest velocity (figure 5(a))
may indicate ductile rupture of the polymer. At this relatively low velocity, only few
flyers could penetrate the film and create the punctures. At a higher flyer velocity of
1650 ms1 (figure 5(b)), ductile rupture is still dominant, but some cracks begin to
form around these volcano-like punctures. A further increase in the flyer velocity
resulted in radial cracking around the central impact zone (figure 5c). The results
indicate also that all flyers in the cloud had sufficient energy to penetrate the Kapton
film. At the highest tested velocity of 2900 ms1 , these radial cracks completely
developed into a brittle fracture of the polymer (figure 5(d)). The transition from
ductile to brittle fracture may be expected because such transitions are strongly
dependent on the strain rate. As the flyer velocity increased, the strain rate also
increased. Brittle fractures are associated with less energy absorbance compared
to ductile fractures [12]. At relatively low strain rates (figures 5(a) and 5(b)) the
kinetic energy lost by the flyers was transferred into pronounced deformation
energy. At relatively high strain rates (figures 5(c) and 5(d)), on the other hand,
the successive kinetic energy was transformed into crack propagation energy and
the associated formation of new surfaces.
The effect of film thickness on the extent and nature of damage introduced into the
Kapton film is demonstrated in figure 6. Laser driven flyers with velocity of 1730
Figure 6. Impacts at velocity of 1730 ms1 into (a) 25 m, (b) 50 m, and (c) 125 m-thick Kapton
films. Note the different scale bar in (c)
GROUND SIMULATION OF HYPERVELOCITY SPACE DEBRIS IMPACTS 161
ms1 were shot against 25, 50, and 125 m-thick Kapton films. All images in
figure 6 were taken from the impacted sample exit side. It is evident that as the film
thickness increases, a transition from brittle (figure 6(a)) to ductile (figure 6(c))
fracture occurs and the overall extent of damage caused by the impact is reduced
too. The 25 m-thick Kapton film (figure 6(a)) experienced significant damage
with radial brittle-like cracks emanating from a central impact zone. The 50 m-
thick Kapton film (figure 6(b)) experienced less damage, lacking any radial cracks;
only few punctures were noticed. The least damage was introduced into the 125
m-thick Kapton film (figure 6(c)); only a single penetration zone is observed,
exhibiting a volcano-like puncture.
The following argument may be given to explain the results aforementioned.
As the Kapton film becomes thicker, its ability to absorb energy and slow the
flyer increases. Consequently, the strain rate associated with the impact process is
reduced. For the 25 m-thick film, the strain rate is high enough to catalyze the
formation of brittle radial cracks. For the 50 m-thick film, intermediate strain
rates probably existed, leading to a semiductile fracture of the polymer. In this
case, no sufficient energy was left to allow radial cracking. Finally, in the case
of the 125 m-thick film, only a single puncture was formedmost likely by a
single flyer. The strain rate under which this process took place was low enough
to enable ductile fracture.
Close examination of the sample impacted at the highest velocity of 2900 ms1
is shown in figure 7 at a higher magnification. This figure shows also the fracture
surface morphology (fractography) around the circumference of the penetrating
hole. It is clear that this morphology changes significantly, indicating the possible
involvement of different mechanisms of fracture.
The characteristics of fracture morphology seem surprising at firstwhile the
central penetration zone that experienced the highest strain rate failed in a fairly
ductile manner, the radial cracks that formed subsequently under lower strain
rates (i.e., as a secondary process) exhibit a more brittle fracture morphology.
We believe that this behavior may be explained in terms of a high temperature
gradient that is established within the polymer sample as the flyer hits its surface
and penetrates through the film. It is well known that ductilebrittle transitions
depend strongly on the local temperature. Whereas a ductile fracture is expected
above the glass transition temperature, Tg , below this temperature brittle fractures
are most likely to occur. It should be noted that in Kapton, a second-order transition
occurs within the temperature range of 360410 C, which is assumed to be the
glass transition temperature [13]. This temperature dependence of fracture mode
also reminds the deformation map suggested by Spaepen for metallic glasses [14].
Ultrahigh velocity impacts generate temperatures in the range of 17276727 C
162 R. VERKER ET AL.
Figure 7. (a) A 25 m-thick Kapton film impacted at hypervelocity of 2900 ms1 . Two distinct
modes of fracture are evident: (b) a fairly brittle fracture in the radial cracking region, and (c) a
more ductile fracture around the central penetration region
and shock pressures of 30100 GPa when striking ceramic surfaces [9, 15]. High-
density polyethylene projectiles shot at an average velocity of 5 kms1 were
observed to generate temperatures of up to 7927 C in the case of head-on impacts
on aluminum targets [16].
Hence, the following model can explain the phenomenon shown in figure 7.
Due to the high temperature generated at the penetration zone and despite the
ultrahigh strain rate involved in impact, the Kapton film exhibits a fairly ductile
fracture in this zone. On the other hand, the significantly lower temperatures of
T < Tg are not sufficient to compensate for the still high strain rates, thus the
secondary cracks far from the impact point exhibit brittle fracture characteristics.
In order to support the above model, an attempt was made to estimate the mean
temperature developed within a sample at the time of penetration. Figure 8 shows
a 25 m-thick Kapton film impacted at an entrance flyer velocity Vi of 2200
ms1 . The flyer velocity at the exit, Vo , was also measured and found to be
1540 ms1 .
The sample was divided into three regions: (a) the penetration hole, at which a
piece of Kapton was sheared-off, (b) a region where ductile fracture was observed,
and (c) a region where brittle fracture was noticed.
GROUND SIMULATION OF HYPERVELOCITY SPACE DEBRIS IMPACTS 163
Figure 8. A 25 m-thick Kapton film impacted at a hypervelocity of 2200 ms1 . The sample is
divided schematically into three regions: (a) the penetration hole, at which a piece of Kapton was
sheared-off, (b) a region where ductile fracture was observed, and (c) a region where brittle fracture
was noticed
where PS is the shearing force, h is the film thickness, l is the length of cut,
UTS is the ultimate tensile strength of the material being sheared, and E Sh is the
shearing energy. For h = 25 m, l = 3.5 103 m, UTS = 2.31 108 Nm2 ,
and PS = 17 N, we get E sh = 0.2 mJ. E C was evaluated according to
E C = 2 a h (S + P ) (7)
where a is the total length of cracks which were formed by the impact, s is
the specific surface energy, and p is the plastic deformation energy. Using the
measured value a = 0.37 cm, and the typical values s = 5.8 104 Ncm 1
and p = 0.5 Ncm1 , we find E C = 9.3 103 mJ.
Now, by substituting these values into eq. (3), we find the amount of energy
transformed into heat, E H = 30 mJ. Using the common expression for adiabatic
heat transfer equilibrium process
E H = m Cp T (8)
where m is the mass of Kapton confined within region (b) in figure 8 (m =
5.7 105 g), CP is the Kaptons specific heat (CP = 1.09 J(gK)1 ), and T is
the temperature difference, a mean temperature increase of T = 920 C was cal-
culated. This calculation thus shows that a flyer penetrating at a velocity of 2200
ms1 causes a significant temperature increase in the penetration zone (figure 8,
region (b)), raising the local temperature to as high as approximately 947 C.
This estimated temperature is much higher than the glass transition temperature
of Kapton, supporting the suggested model and explaining the ductile-like frac-
ture within the penetration zone. This large temperature gradient also predicts a
temperature lower than Tg in region (c), where the high strain rate results in a
brittle-like fracture.
A laser driven flyer (LDF) system was developed at Soreq NRC for simulation of
space debris impacts. The system produces a cloud of flyers with diameters of up
to 30 m and measured velocities of 14002900 ms1 . The effect of simulated
hypervelocity debris impacts on space-qualified Kapton films was studied for
different film thickness (25125 m) and impact velocities.
As the Kapton film thickness was increased from 25 to 125 m at a fixed
impact velocity, a transition from brittle to ductile fracture was observed. At a
constant thickness of 25 m, low impact velocities of 14001600 ms1 resulted
in a ductile-like fracture. Increasing the flyer velocity resulted in cracks formation
and a fracture that was mostly brittle-like. At impact velocities higher than 1700
ms1 , the central impact region (which is exposed to the highest strain rate) was
characterized by a ductile-like fracture, while remote radial crack regions were
GROUND SIMULATION OF HYPERVELOCITY SPACE DEBRIS IMPACTS 165
Acknowledgment
References
1. Grossman, E., Gouzman, I., Viel-Inguimbert, V., and Diguirard, M. (2003) Journal of Spacecraft
and Rockets 40, 110.
2. Houdayer, A., Cerny, G., Klemberg-Sapieha, J.E., Czeremuszkin, G., and Wertheimer, M.R.
(1997) Nuclear Instruments and Methods in Physics Research B 131, 335.
3. Silverman, E. M. (1995) Space Enviromental Effects on Spacecraft, LEO Material Selection
Guide, NASA Contractor Report 4661, Langley Research Center, pp. 41.
4. Tennyson, R. C. and Shortliffe, G. (1997) In A. Paillous (ed.) Proceedings of the 7th International
Symposium on Materials in Space Environment, Toulouse, France, ESA Publication, ESTEC
Noordwijk, The Netherlands, 1620 June 1997, p. 485.
5. Hastings, D. and Garrett, H. (1996) Spacecraft-Environment Interactions, Cambridge University
Press, Cambridge, U.K., pp. 4599.
6. Miao, J. P. W. (2001) Stark Planetary and Space Science 49, 927.
7. Paul, K. G., Igenbergs, E. B., and Berthoud, L. (1997) International Journal of Impact Engi-
neering 20, 627.
8. Stein, C., Roybal, R., and Tlomak, P. (2000) In E. Werling (ed.) Proceedings of the 8th Interna-
tional Symposium on Materials in Space Environment, Arcachon, France, CNES Publication,
Toulouse, France, 59 June 2000.
9. Medina, D.F., Wright, L., and Campbell, M. (2001) Advances in Space Research 28, 1347.
10. Tighe, A., Gabriel, S., and Van Eesbeek, M. (2000) In E. Werling (ed.) Proceedings of the
8th International Symposium on Materials in Space Environment, Arcachon, France, CNES
Publication, Toulouse, France, 59 June 2000.
11. Roybal, R., Tlomak, P., Stein, C., and Stokes, H. (1999) International Journal of Impact Engi-
neering 23, 811.
12. Callister, W. D. (2000) Materials Science and Engineering, John Wiley and Sons Inc., New-York,
pp. 124134.
13. Du-Pont Inc. Technical bulletin, http://www.dupont.com/kapton/general/H-38492-2.pdf.
14. Spaepen, F. (1977) Acta Metallurgica 25, 407.
15. Akins, J.A. (2003) Dynamic Compression of Minerals in the MgOFeO-SiO2 System, Ph.D.
thesis, California Institute of Technology, CA.
16. Ramjaun, D., Shinohara, M., Kato, I., and Takayama, K. (2001) In Proceedings of the 23rd
International Symposium on Shock Waves (ISSW23), Fort Worth, TX, 2227 July 2001.
TESTING OF SPACECRAFT MATERIALS FOR LONG
FLIGHTS IN LOW EARTH ORBIT
1. Introduction
One of the principal damaging factors of space in low Earth orbits (LEO) is the
atomic oxygen (AO). In prolonged exposures of materials on the outside surfaces
of spacecraft, their resistance to AO attack is of major importance. Since polymer-
based materials are widely used on spacecrafts and most of them are susceptible
to AO attack, their protection by various means, including protective coatings, is
necessary.
The long-term LEO flight simulation requires irradiation of materials with
high-fluence AO up to 1022 1023 cm2 . Simulator beam intensities do not exceed
1017 cm2 s1 (usually 1015 1016 cm2 s1 ) [1] that results in practically unac-
ceptable test duration. The reduction of test duration is achieved by an increase
of the particle beam energy within the limits of conservation of the interaction
mechanism with the test material [2].
167
J.I. Kleiman (ed.), Protection of Materials and Structures from Space Environment, 167174.
C 2006 Springer. Printed in the Netherlands.
168 L. S. NOVIKOV ET AL.
In our work, the accelerated simulation material tests were carried out in
oxygen plasma beams, formed by a plasma accelerator, with the atomic oxygen
particle energy of 20 eV. The changes of weight, mechanical and optical properties
of prospective materials were investigated at equivalent AO fluences up to 3.5
1022 cm2 .
2. Test Materials
The experiments were carried out in the plasma accelerator [3] in oxygen plasma
beams consisting of ions, atoms, and molecules of oxygen with mean velocity of
16 kms1 (mean atom energy of 20 eV) and a flux density of (2.53.5) 1016
cm2 s1 . Due to the dissociation of fast molecules and neutralization of the ions
TESTING OF SPACECRAFT MATERIALS 169
at collision with the surface, only the atoms with mean velocity of 16 kms1
impact the material. The AO equivalent fluence was evaluated using the witness
film mass loss in an assumption of the erosion yield Y = 4.4 1024 gatom1
O. The fluence is equal to a fluence of a fictitious 5 eV-AO beam that initiates
the same polyimide mass loss. We followed the standard technique of AO fluence
determination in ground-based facilities [4].
When material properties were studied the changes of weight and thickness
were measured, and the reflection spectra in the range 0.22.5 m were registered.
The stressstrain properties were determined on a tensile test machine.
4. Experimental Results
The specific mass loss and thickness of the black reinforced plastic samples un-
coated and with the coating as a function of the AO equivalent fluence are given
in figure 1. As can be seen, the mass loss and thickness of the uncoated sam-
ples are linear functions of the AO fluence. At the same time, the mass loss of
the coated sample is lower by a factor of 6.5, and thickness change is negligibly
small.
Reactivity of the unprotected black reinforced plastic consisting of carbon
fibers in epoxy matrix, measured in the tests, was 1.2 1024 cm3 per O atom. It
corresponds to the flight data known for the composite components: 1 1024 cm3
per O atom for carbon and 1.7 1024 cm3 per O atom for epoxy [3] that confirms
that the plasma accelerator conditions can imitate adequately the LEO conditions.
The polymer binder on the coated samples was etched away after plasma impact
80 1.2
no coating
70 1
Specific mass loss,
with coating
Thickness, mm
60
0.8
50
mg/cm2
40 0.6
30
0.4
20 no coating
0.2
10 with coating
0 0
0 1 2 3 4 0 1 2 3 4
AO equivalent fluence, 1022 1/cm2 AO equivalent fluence, 1022 1/cm2
Figure 1. Specific mass loss and thickness of samples of the black reinforced plastic uncoated and
with the coating as functions of the AO equivalent fluence
170 L. S. NOVIKOV ET AL.
1400
1200
1000
Load, N 800
600
400
200
0
0 10 20 30 40
Elongation, %
Figure 2. Tensile stressstrain diagrams of the sennit PARML samples before and after the AO
exposure
exposing the fibers that were also partially destroyed with an observed increase in
the surface roughness [5].
The picture is similar to observations in flight experiments onboard the space
shuttle spacecraft, Salute-6 and Mir space stations [6].
The results of measurements of specific mass loss of the Terlon fabric samples
and the sennit PARML samples show, that the Terlon fabric is close to polyimide
in AO resistivity, as the materials mass losses are close to 5.4 and 5.6 mgcm2
accordingly.
A complete etching of the polyimide fibers in a significant part of bundles
in the sennit PARML was observed. The tensile stressstrain diagrams of the
sennit PARML samples before and after the AO exposure to an equivalent flu-
ence of 3.7 1021 cm2 are given in figure 2. The breaking load reduction by
a factor of 5, and a drop of relative elongation at rupture by a factor of 1.7 was
observed.
After AO exposure the destruction of seam strings and Terlon fabric was
observed on the fragment of the Terlon screen with the seam. The tensile stress
strain diagrams of reference and AO exposed samples of the shield fragments
are presented in figure 3. When loading up to approximately 20-fold working
load identical elongations were observed in both samples. By further increasing
the loading, the exposed sample fails that manifests itself as a sharp break in
the curve and an increase in elongation is observed without an increase or even
on reduction of loading. The elongation on rupture of the exposed sample also
decreases by a factor of 2.
TESTING OF SPACECRAFT MATERIALS 171
800
600
Load, N
400
200 1
2
0
0 2 4 6 8 10 12 14
Elongation, mm
Figure 3. Tensile stressstrain diagrams of reference (1) and AO exposed (2) samples of the shield
fragments
In figure 4, the dependences of the mass loss of the polyimide films protected by
deposited silica layers and by silicone coatings of two varnish types of AO fluence
are presented. We quantify the AO protection efficiency as a ratio of bare film
mass loss dm 0 to the protected film one dm i under equal fluence increment dF, i.e.
= dm 0 /dm i |d F = const.
The protection efficiency is greatest for silica layer ( = 830) and decreases up
to = 250430 for varnish coverings [6]. The exposure to AO causes formation
of microcracks in the silica layers. On the contrary, no cracking occurs in the
silicone coatings.
Thermo-optical characteristics of the samples changed very little. For the silica
coating an insignificant and uniform in wavelength range reflectance reduction is
observed that can be associated with scattering from the grid of microcracks. Thus,
the solar absorptance s increases from 0.366 to 0.380. The reflectance increase
is typical for varnish coatings, especially appreciable in long wavelength part of
the spectrum.
60
Mass loss, mg/cm2
50
40
30
20
10 silicon 3 oxide silicon 5
0
0 5 10 15
AO equivalent fluence, 1020 1/cm2
Figure 4. Dependences of the mass loss of the polyimide films with protection by silica layers
deposited and silicone coatings of two varnish types on AO fluence
172 L. S. NOVIKOV ET AL.
The paint test results have shown different AO resistance for various types of
enamels [7].
Color change and significant mass losses were observed for the epoxy enam-
els. The erosion yields of the color epoxy enamels (0.30.5 1024 gatom1 O)
are much lower than that of the binder and of the witness polymers
KO-811 blue
0.15
Reflection coefficient
0.1 after
before
0.05
0
300 400 500 600 700
Wavelength, nm
EP-140 blue
0.35
0.3
Reflection coefficient
0.25
0.2 after
0.15 before
0.1
0.05
0
300 400 500 600 700
Wavelength, nm
after
0.15 before
0.1
0.05
0
300 400 500 600 700
Wavelength, nm
Figure 5. Visible spectrum reflectance for three coatings after oxygen plasma exposure with AO
equivalent fluence 1.4 1021 cm2
TESTING OF SPACECRAFT MATERIALS 173
5. Conclusions
It was demonstrated that oxygen plasma beams can be used in accelerated tests
of carbon-based and polymeric materials (with the exception of fluorinated hy-
drocarbons) and inorganic coatings to evaluate their resistance to the AO impact
during the simulation of the long flights in LEO.
The resistance of prospective spacecraft materials: polyimide films, synthetic
Terlon fabric, sennit PARML, black reinforced plastic, polymeric paints, and
thermo-control coatings under oxygen plasma beams simulating the AO fluxes
in LEO was investigated. For the unprotected materials sharp fall of mechani-
cal properties (manifesting in lower failure loads and relative elongation at rup-
ture) was observed. Optical characteristics deteriorated as well. Application of
protective coatings had shown to reduce the degradation of mechanical and op-
tical properties. The protection efficiency is the greatest for coatings containing
silicon.
References
1. Kleiman, J., Iskanderova, Z., Gudimenko, Y., and Horodetsky, S. (2003) In Proceedings of the
9th International Symposium on Materials in Space Environment, ISMSE-9, Nordwijk, 2003,
pp. 313324.
2. Rutledge, S.K., Banks, B.A., Dever, J., and Savage, W. (2000) In Proceedings of the 5th Interna-
tional Conference on Protection of Materials and Structures from the LEO Space Environment,
ICPMSE-5, Arcachon, France ESA Publishing, Noordwijk, The Netherlands, 2000.
3. Akishin, A. I., Novikov, L. S., and Chernik, V. N. (2000) In L.S. Novikov and M.I. Panasyuk
(eds.), New High Technologies and Technics, Vol. 17, Moscow, ENTSITEH, p. 100 (in Russian).
4. Protocol for Atomic Oxygen Testing of Materials in Ground-Based Facilities, JPL Publication,
pp. 9517.
174 L. S. NOVIKOV ET AL.
5. Novikov, L. S., Chernik, V. N., Babaevskij, P. G., Kozlov, N. A., Chalyh, A. E., Balashova, E. V.,
and Smirnova, T. N. (2001) Perspektivny materialy 5, 2026 (in Russian) (Journal of Advanced
Materials, Cambridge Interscience Publication).
6. Chernik, V., Naumov, S., Demidov, S., Sokolova, S., and Svechkin, V. (2000) In 5th Interna-
tional Conference on Protection of Materials and Structures from the LEO Space Environment,
ICPMSE-5, Arcachon, France, 2000.
7. Chernik, V. N., Naumov, S. F., Sokolova, S. P., Gerasimova, T. I., Kurilyonok, A. O., Poruchikova,
Y. V., and Novikova, V. A. (2003) In Proceedings of the 9th International Symposium on Mate-
rials in Space Environment, ESTEC, Noordwijk, 2003, pp. 281285.
M/OD IMPACTS ON THE MULTIPURPOSE LOGISTICS MODULE
Abstract. The International Space Station (ISS) program has manifested a Mul-
tipurpose Logistics Module (MPLM) on five launch packages since 2001, with
the next flight scheduled for the STS-114 mission. The MPLM has been deployed
each time on the nadir docking port of node 1 by the space shuttle for a period
of 56 days, and then returned to Earth for refurbishment prior to the next resup-
ply mission. MPLM flight module 1 (Leonardo) and flight module 2 (Raffaello)
have accumulated about 700 h of low Earth orbit (LEO) exposure time on the
ISS. Through five missions, there have been two perforations of the aluminum
outer wall of the MPLM shielding and 24 craters due to meteoroid/orbital debris
impact.
This paper will document the results of an ongoing postflight inspection cam-
paign that identifies hypervelocity impact (HVI) damage to the meteoroid and
debris protection system (MDPS) of the MPLM through observations, data col-
lection and analysis. Residual projectile materials are collected at each impact site
and subjected to scanning electron microscope/energy dispersive X-ray spectro-
metric analysis to identify the elemental composition of the impactor and presumed
source (meteoroid or orbital debris) of the impact damage. The observed impact
damage exhibits a marked directionality on the module, with the majority of dam-
ages occurring to the forward half of the MPLM. Postflight predictions of damage
to MPLM from the BUMPER code, which is utilized by the aerospace community
in meteoroid/orbital debris risk assessments, are compared to the observed dam-
ages. Results from BUMPER very clearly show the directionality of the expected
impact damage as was observed in the postflight MPLM inspections. Observations
and analytical data from MPLM demonstrate the meteoroid/orbital debris impact
hazards that the ISS, shuttle, and other spacecraft face in low Earth orbit envi-
ronment. Implications of this work on the design of adequate meteoroid/orbital
debris protection for future vehicles will be provided.
175
J.I. Kleiman (ed.), Protection of Materials and Structures from Space Environment, 175191.
C 2006 Springer. Printed in the Netherlands.
176 JAMES L. HYDE ET AL.
1. Introduction
1.1. PURPOSE
One of the primary reasons for tracking on-orbit damage is the fact that NASA has
used observation data from returned spacecraft surfaces to calibrate the ORDEM
2000 engineering model of the low Earth orbit (LEO) environment [1]. Postflight
inspections of shuttle surfaces such as the Crew Module Windows, Payload Bay
Door Radiators has produced a database with thousands of entries [2]. The long
duration exposure facility (LDEF), retrieved by the shuttle in 1990 after nearly 6
years on orbit, contributed a rich set of data [3] that has been used in the formulation
of the orbital debris environment.
Postflight inspections also provide an opportunity to validate meteoroid/orbital
debris (M/OD) threat assessment codes such as NASAs BUMPER-II [4, 5] or
ESABASE/Debris. The ability to compare predictions with observations is a valu-
able component in the development life cycle of M/OD analysis software.
The performance and adequacy of meteoroid/debris protection systems can be
demonstrated with postflight damage inspections. The flight history of the MPLM
has established the protection capability of the shielding system.
1.2. OBJECTIVE
This paper will document the results of the five postflight inspections that have
been performed on the MPLM, providing data on damage sizes and sources of
the impact. The results will also include an estimate of the equivalent diameter
of the impacting particle. Due to the nature of hypervelocity impacts, projectile
shape and impact angle data are not readily discernable.
2. MPLM Overview
2.1. DESCRIPTION
Three flight units, built by Alenia Spazio for the Italian Space Agency (ASI) as
part of an agreement with NASA, were named after famous Italian engineers:
Leonardo da Vinci, Donato di Niccolo, di Betto Bardi, and Raffaello Sanzio.
The MPLM functions as a cargo container for ISS outfitting and resupply mis-
sions and as a pressurized module while attached to the ISS. The modules can be
configured as active, when a payload requires power, data or cooling resources.
M/OD IMPACTS ON THE MULTIPURPOSE LOGISTICS MODUL 177
The MPLM can also be manifested in a passive configuration when the contents
require no power, data or cooling resources (dry cargo). Each module has a design
life of 10 years or 25 flights. Each module has an overall length of about 6.2 m
and a diameter of about 4.5 m [6], giving a surface area of approximately 100 m2 .
2.2. INSTALATION
MPLM modules are carried into orbit and placed on the ISS by the shuttle arm as
shown in Figure 1. At the end of a mission, the module is restowed in the payload
bay of the orbiter and returned to Earth.
There have been five MPLM missions to date, with a cumulative docked exposure
time of 701 h. Three missions used flight module (FM) 1 and two missions used
FM2. Details of the missions are shown in figure 4. The modules were always
MDPS FRGF
PANEL
MAIN
TRUNNION
SUPPORT
BRACKET
FOR FLUID
PDA
STABILIZER
TRUNNION
FRGF
MAIN
TRUNNION
MDPS
PANEL
mated with the ISS at the nadir docking port on node 1. This placed the port edge
of the MPLM (relative to its stowed position in the shuttle) in the ram direction
when docked to the ISS.
The predominant flight attitude of the ISS when the shuttle and MPLM are
attached is a positive pitch bias of about 23 , as shown in figure 5, the front of the
station is angled upwards 23 .
3. Impact Survey
The following section gives details of the impact features observed during post-
flight processing. Figure 6 illustrates the measurement terminology that was used
to characterize the impacts.
Spall
Lip
Crater
Depth
3.1. PROCEDURE
As part of the standard postflight processing workflow, all MDPS panels are in-
spected by technicians at a processing facility at the Kennedy Space Center (KSC).
They are tasked with the documenting defects on external and internal surfaces. In-
stances of scratches, scuffs, ding, and rub marks have all been recorded, along with
hypervelocity impact craters. Specialists from the Johnson Space Center with hy-
pervelocity impact (HVI) inspection experience travel to KSC and independently
examine as much as possible of the exterior for HVI damage.
When an impact site is identified, the location is recorded and a graduated
handheld magnifier is used to measure dimensions of the impact features (crater,
lip, and spall diameters). For impacts into monolithic aluminum MDPS surfaces,
a soft wooden probe is used to collect potential impactor residue from the crater
or hole. Dental mold impressions are another technique that is used when surface
damage needs to be sampled. Finally, when soft goods such as beta cloth are
impacted, an adhesive agent is sometimes used to aid in data collection. Crater
depth is obtained back at the lab by measuring the effected region of the sampling
device.
The figure above illustrates the inspection access limitations at the space station
processing facility (SSPF). The mounting fixture has to be rotated at least once to
allow a complete inspection of all MDPS panels.
There were three hypervelocity impact sites observed after the first MPLM mis-
sion. One of the impacts produced a perforation of an MDPS bumper. The impact
site, designated 1.1.1, had a 1.44 mm diameter hole with a 2.45 mm diameter lip.
There was no discernable damage to the MLI or to the pressure wall of the MPLM.
In an attempt to determine the origin of each impactor, samples collected from
the three impact sites and subjected to a scanning electron microscope (SEM) and
energy dispersive X-ray (EDX) analysis. The results are shown in the following
table. Analysis indicated that the impactor that produced the bumper perforation
was a particle of spacecraft paint. There was no discernable impact residue in the
sample from site 1.1.2.
When the particle type is known, an equivalent diameter can be estimated
using eq. (1) with an assumed impact velocity of 9 kms1 and an impact angle of
45 .
D = [(P/5.24 H 1/4 (B /P )1/2 (S/Vn )2/3 ]18/19 (1)
IMP A C T D E TA IL S Estimated
Impact Location Diam (mm) Depth SEM/EDXA Particle
Number Map ID Type Region Material Spall Lip Crater Hole (mm) Results Dia. (mm)
1.1.1 A6 MDPS bumper port/lower 0.8mm Al 6061-T6 --- 2.45 --- 1.44 --- Orbital Debris: Paint 0.46
1
1.1.2 B5 MDPS bumper port/lower 0.8mm Al 6061-T6 --- 1.25 0.78 --- 0.51 Unknown ---
1.1.3 D5 MDPS bumper port/lower 0.8mm Al 6061-T6 --- 0.88 0.54 --- 0.3 Meteoroid 0.19
where
Destimated particle diameter (cm)
Pcrater depth (cm)
H Brinell hardness of bumper (MDPS = Al6061-T6, 95)
B density of bumper (MDPS = Al6061-T6, 2.713 gcc1 )
P density of projectile, 2.5 gcc1 (for spacecraft paint)
Sspeed of sound in target (Al6061-T6 MDPS bumper, 5.1 kms1 )
Vn normal component of impact velocity
(assuming a velocity of 9 kms1 and an impact angle of 45 , 6.36 kms1 )
Note: in the cases where a bumper perforation occurs, the value of P is extrapolated
from the crater diameter.
IMP A C T D E TA IL S Estimated
Impact Location Diam (mm) Depth SEM/EDXA Particle
Number Map ID Type Region Material Spall Lip Crater Hole (mm) Results Dia. (mm)
2
1.2.1 D2 MDPS bumper port/upper 0.8mm Al 6061-T6 --- 2.80 2.0 --- 0.9 Unknown
2
1.2.2 B12 MDPS bumper stbd/upper 0.8mm Al 6061-T6 --- 1.80 1.5 --- 0.75 Unknown
1.2.3 B11 MDPS bumper stbd/upper 0.8mm Al 6061-T6 --- 0.85 0.8 --- 0.55 Orbital Debris: SS 0.14
There were three hypervelocity impact craters observed on the aluminum MDPS
bumpers after the third mission (Fig. 12). No perforations were detected. SEM/
EDX analysis of site 1.2.3 indicated that the crater was caused by a particle of
stainless steel. The other two sites were not available for sampling.
Eight impacts were observed after this mission, four on the MDPS bumpers and
one each on a grapple fixture, trunnion brace, and scuff plate (Fig. 13). Impact
2.2.8 was the site of the second bumper perforation. The impact produced a 1.2 mm
diameter hole with a lip diameter of 1.8 mm. As with the first perforation, there was
no visible damage to the underlying structure or components. SEM/EDX analysis
indicated that the impactor was a particle of stainless steel. Samples were taken
at all eight impact sites. Two impacts were shown to be from orbital debris and
two were meteoroids. The remaining four impact samples revealed no discernable
impact source.
Postflight inspections for the third mission of FM1 (Leonardo) recorded 12 impact
sites (Fig. 16). Ten occurred on the MDPS panels, another in the material between
IMP A C T D E TA IL S Estimated
Impact Location Diam (mm) Depth SEM/EDXA Particle
Number Map ID Type Region Material Spall Lip Crater Hole (mm) Results Dia. (mm)
2.2.1 LA4 scuff plate port/lower yellow Sheldahl tape 1.4 --- 0.45 --- 0.15 Meteoriod: Fe,Ni,S 0.14
1
2.2.2 A6 MDPS bumper port/lower 0.8mm Al 6061-T6 --- 0.70 0.55 --- 0.25 Unknown ---
1
2.2.3 FC9 MDPS bumper fwd cone 0.8mm Al 6061-T6 --- 0.60 0.4 --- 0.50 Unknown ---
2.2.4 LA4 trunnion brace port Betacloth --- --- 0.31 --- 0.10 Orbital Debris: Al 0.10
2.2.5 D10 MDPS bumper stbd/upper 0.8mm Al 6061-T6 --- 0.45 0.35 --- 0.50 Meteoroid 0.24
1
2.2.6 A2 grapple fixture port Al 6061-T6 2.1 1.05 0.85 --- 0.95 Unknown ---
1
2.2.7 A12 MDPS bumper stbd/upper 0.8mm Al 6061-T6 --- --- 0.45 --- 0.50 Unknown ---
2.2.8 D12 MDPS bumper stbd/upper 0.8mm Al 6061-T6 --- 1.80 --- 1.20 --- Orbital Debris: SS 0.19
IMP A C T D E TA IL S Estimated
Impact Location Diam (mm) Depth SEM/EDXA Particle
Number Map ID Type Region Material Spall Lip Crater Hole (mm) Results Dia. (mm)
1.3.1 D6 MDPS bumper port/lower 0.8mm Al 6061-T6 --- 0.60 0.45 --- 0.60 Meteoroid 0.28
1
1.3.2 D6 MDPS bumper port/lower 0.8mm Al 6061-T6 --- 0.70 0.55 --- 0.60 Unknown
1
1.3.3 D6 MDPS bumper port/lower 0.8mm Al 6061-T6 --- 0.68 0.5 --- 0.45 Unknown
1
1.3.4 D5 MDPS bumper port/lower 0.8mm Al 6061-T6 --- 0.43 --- --- --- Unknown
1
1.3.5 C6 MDPS bumper port/lower 0.8mm Al 6061-T6 --- 0.40 --- --- --- Unknown
1.3.6 C4 MDPS bumper port/lower 0.8mm Al 6061-T6 --- 0.63 0.5 --- 0.35 Orbital Debris: Paint 0.23
1.3.7 A2 grapple fixture port Al 6061-T6 4.25 1.00 0.82 --- 0.65 Orbital Debris: SS 0.20
1
1.3.8 B2 MDPS bumper port/upper 0.8mm Al 6061-T6 --- 0.70 0.55 --- 0.5 Unknown ---
1.3.9 B2 MDPS bumper port/upper 0.8mm Al 6061-T6 --- 0.48 0.4 --- 0.25 Orbital Debris: Paint 0.17
1.3.10 3-2 intercostal port/upper 0.8mm Al 6061-T6 --- 0.80 0.65 --- 0.45 Meteoroid 0.21
1
1.3.11 B2 MDPS bumper port/upper 0.8mm Al 6061-T6 --- 0.25 --- --- --- Unknown ---
1
1.3.12 C12 MDPS bumper stbd/upper 0.8mm Al 6061-T6 --- 0.45 --- --- --- Unknown ---
panels and one impact was observed on the port grapple fixture. Samples were
obtained at all 12 impact sites. SEM/EDX analysis did not reveal impact residue
in seven of the samples. Of the five remaining, three were determined to be orbital
debris.
4. Summary
Through five MPLM missions, 26 hypervelocity impact sites have been observed
during postflight inspections. Although more than 80% of the impacts occurred
on the MDPS panels, other parts of the module also sustained impact damage.
Samples were obtained at 24 impact sites. SEM/EDX analysis was performed on
each sample and 12 yielded results that could be interpreted as either meteoroid
or orbital debris in nature. Of the known impact sources, seven were determined
to be orbital debris and five were meteoroid. The table in figure 17 provides a
summary of the 12 impacts. The estimated diameter of the particle that caused
the damage is shown in the last column of the table, sorted from least to greatest.
An analysis of the expected number of MDPS outer wall perforations was per-
formed using the BUMPER code, the meteoroid/orbital debris risk assessment
software tool used by NASA to determine risk for the International Space Station
[7] and the space shuttle [8]. The intent was to determine if BUMPER code predic-
tions for the number of MDPS perforations came close to matching observations.
Another type of prediction vs. observation activity that can be performed with
the BUMPER code is comparison of observed impact locations to the predicted
distribution on the MDPS.
186 JAMES L. HYDE ET AL.
1 2
3 4
Using attitude information from the as-flown timeline in addition to the orbital
and environmental parameters shown in the table below (Fig. 19), an analysis was
performed with the BUMPER code to determine if the number of observed MDPS
outer wall perforations agrees with predictions.
10 days 18 hours
Mission Duration
258 hours
Altitude 400 km
Orbit Inclination 51.6
Flight Year 2001
bias -XLV +ZVV = 0, 113.5, 0 130 hrs
LVLH Attitudes - RPY
bias +ZLV +XVV = 0, 327, 0 5 hrs
Orbiter PYR Euler Sequence
bias -XLV +YVV = 90, 90, 23.5 5 hrs
(Exposure Hrs | % Duration)
-ZLV -XVV = 0, 180, 0 118 hrs
Orbiter Finite Element Model ISS 5A.1 mated with orbiter, MPLM on Node 1
Orbital Debris Density constant, 2.8 g/cc
Orbital Debris Environment ORDEM2000
Meteoroid Density variable, 0.5 - 2.0 g/cc
Meteoroid Enviroment SSP30425 Rev. B
Meteoroid Velocity Distribution variable, SSP30425
Meteoroid Showers none
Space
Velocity
Space
Space Velocity
Velocity
Figure 20. BUMPER finite element models showing typical MPLM mission attitudes
Figure 21. Expected number of MDPS outer wall perforations for a 258 h mission
188 JAMES L. HYDE ET AL.
Figures 22 and 23 show the numbering scheme for the MDPS panels. The top,
bottom, forward, and aft designations refer to the position of the module
when it is stowed in the shuttle payload bay.
Figure 22 above illustrates the 12 rows or bays of MDPS panels that are
referenced in the analysis results table in figure 21.
Figures 2426 show the relative distribution of MDPS outer wall perforation risk
from meteoroid and orbital debris impacts. The highest risk of bumper perforation
can be seen in the red bands 30 70 off of the velocity vector on the port and
starboard sides. It can be seen in figures 24 and 25 that the MDPS panels on the
Figure 26. MDPS impact risk distributionforward view (ISS and Orbiter removed for clarity)
end cones have considerably lower risk than the cylinder region. In figure 26, the
meteoroid and orbital debris flux shadowing effect from the docked Orbiter can
be seen in the forward view of the risk distribution.
The correlation between observed and predicted impacts is shown in figure 27,
where the values from the impact risk plots in figures 24 through 26 are mapped
to the MDPS cylinder panels and observed impact locations. Of the 25 observed
impacts in the cylinder region, 16 (64%) occurred in the top 2 high-risk areas.
6. Discussion
This paper details the status of an ongoing inspection and analysis campaign for
the MPLM. More flights are planned in the future and postflight inspections will
be needed. One item for investigation is the determination of MDPS hole size at
onset of significant MLI degradation/rear wall damage. Information of this type
could aid in postflight serviceability requirements.
The aluminum composition of the MDPS outer wall makes it difficult for the
SEM/EDX analysis to discern aluminum projectile residue in the midst of the com-
mon background material. Many of the impacts in the database that were assigned
to the unknown category are probably aluminum in nature. Since aluminum is
one of the more common sources of orbital debris impacts, this discrepancy needs
to be addressed.
M/OD IMPACTS ON THE MULTIPURPOSE LOGISTICS MODUL 191
Figure 27. MDPS cylinder region risk distribution with locations of observed impacts (dotted line
indicates leading edge of module when docked at ISS)
References
1. Liou, J. C., Matney, M. J., Anz-Meador, P. D., Kessler D., Jansen M., Theall, J. R. (2002) The
New NASA Orbital Debris Engineering Model ORDEM2000, NASA/TP 2002-210780, May
2002.
2. Hyde, J. L. Christiansen, E. L., Bernhard, R. P., Kerr J. H., Lear, D. M. (2001) A History of
Meteoroid and Orbital Debris Impacts on the Space Shuttle, Proc. Third European Conference
on Space Debris, ESA SP-473, October 2001, pp. 191196
3. See, T. H., Allbrooks, M. A., Atkinson, D. R., Simon, C. G. and Zolensky, M. (1990) Mete-
oroid and Debris Impact Features Documented on the Long Duration Exposure Facility, NASA
documents published by the Johnson Space Center JSC-24608, August 1990.
4. Hyde, J. L. (2000) As-Flown Shuttle Orbiter Meteoroid/Orbital Debris Shield Assessment: Phase
1Shuttle/Mir Missions, NASA documents published by the Johnson Space Center JSC-28768,
January 2000.
5. Hyde, J. L. (2000) As-Flown Shuttle Orbiter Meteoroid/Orbital Debris Assessment: Phase 2
Flights, NASA documents published by the Johnson Space Center JSC-29070, September 2000.
6. MPLM Interface Definition Document (ISS-MPLM-IDD-006), December 2000.
7. Prior, T. G. (2003) International Space Station Meteoroid & Orbital Debris Integrated Threat
Assessment #10, NASA documents published by the Johnson Space Center Revision C (JSC-
29951), April 2003.
8. Hyde, J. L. and Christiansen, E. L. (2001) Space Shuttle Meteoroid & Orbital Debris Threat
Assessment Handbook, NASA documents published by the Johnson Space Center JSC-29581,
December 2001.
9. Christiansen, E. L. (1991) Shield Sizing Equations, NASA inter-department communication
SN-90-131, October 1991.
FUEL OXIDIZER REACTION PRODUCTS (FORP)
CONTAMINATION OF SERVICE MODULE AND RELEASE OF
N-NITROSODIMETHYLAMINE IN A HUMID ENVIRONMENT
FROM CREW EVA SUITS CONTAMINATED WITH FORP
Abstract. The U.S. Control Moment Gyros (CMGs) maintain the International
Space Station (ISS) vehicle attitude by compensating for disturbances. It is pre-
ferred to use CMGs, instead of attitude control thruster firings. However, prior
to an extravehicular activity (EVA) on the Russian Segment (RS), the Docking
Compartment (DC1) must be depressurized, as it is used as an airlock. When
the DC1 is depressurized, the CMGs margin of momentum is insufficient to
compensate for the disturbance and the service module (SM) attitude control
thrusters need to fire to desaturate the CMGs. The SM roll control thruster fir-
ings induce fuel oxidizer reaction products (FORP) contamination of the adjacent
SM surfaces. One of the components present in FORP is the potent carcinogen
N-nitrosodimethylamine (NDMA). Since the EVA crewmembers often enter the
area surrounding the thrusters for tasks on the aft end of the SM and when trans-
lating to other areas of the Russian Segment, the presence of FORP contamination
is a concern. This paper will discuss FORP contamination of the SM surfaces, the
potential release of NDMA in a humid environment from crew EVA suits, if they
happen to be contaminated with FORP, and the toxicological risk associated with
the NDMA release.
193
J.I. Kleiman (ed.), Protection of Materials and Structures from Space Environment, 193208.
C 2006 Springer. Printed in the Netherlands.
194 WILLIAM SCHMIDL ET AL.
1. Introduction
The U.S. Control Moment Gyros (CMGs) maintain the International Space
Station (ISS) vehicle attitude by compensating for disturbances. However, when
the docking compartment (DC1), the location of the Russian airlock, of the Inter-
national Space Station (ISS) is depressurized for extravehicular activities (EVAs),
the service module (SM) attitude control thrusters have to fire because the CMGs
have an insufficient margin of momentum to compensate for the disturbance and
must be desaturated.
Thruster firings produce fuel oxidizer reaction products (FORP) that can con-
taminate adjacent surfaces. For EVAs on the aft end of the service module (SM)
of the Russian Segment (RS), there is a concern that when the EVA crewmem-
ber translates around the FORP contaminated area they could inadvertently brush
against the FORP and transfer some of it to their suit.
FORP is composed of both volatile and nonvolatile components. How
fast the volatile components leave varies. One of the components present in
FORP that represents a toxicological risk to the crew is the potent carcinogen
N-nitrosodimethylamine (NDMA). Although NDMA is volatile, it does persist
long enough to be a concern. In addition, when dried FORP is reintroduced into a
humid environment, like the ISS cabin, NDMA can reform from the components
remaining in the FORP. So the concern is that when FORP (on the suit) is brought
back into the humid environment of the ISS cabin, it can release NDMA into the
atmosphere.
2. Background
Discoloration around the SM zenith roll thrusters was observed during the ISS
flight 5A Orbiter fly around of the ISS, as shown in the image in figure 1. In the
image, the pitch thrusters are closest to the aft end of the SM (right side of the
image). The roll thrusters are to the left of the pitch thrusters and raised above
the surface of the SM. The discoloration (brown color) can be clearly seen close
to the roll thrusters in the inset zoom image. The EVA handrails can be seen in
the inset image close to the roll thrusters and on either side of the pitch thrusters.
Contamination has also been observed around the SM nadir roll thrusters.
Figure 2 shows the relative position and direction of the SM attitude control
thrusters. It can be seen that the roll thrusters plume centerline is directed 47
away from the surface normal (lower diagram), while the pitch and yaw thrusters
plume centerline is directed 13 away from the surface normal (top diagram). In
addition, the pitch and yaw thrusters are recessed below the SM surface, while the
roll thrusters are elevated above the surface to provide the roll control component.
So the plumes from the roll thrusters are more likely to contaminate the adjacent
FUEL OXIDIZER REACTION PRODUCTS 195
Figure 1. Fight 5A observation of darkening near service module (SM) zenith roll thrusters. The
inset image shows an enlarged image of the zenith roll and pitch thrusters. The brownish discoloration
is visible near the role thrusters
SM surfaces than the plumes from either the pitch or yaw thrusters. The ISS Ex-
ternal Contamination team has concluded that the discoloration is plume contam-
ination due to the thrusters firings. In such areas, it must be assumed that FORP is
present.
Roll Thrusters 13
Pitch Thrusters
Yaw Thrusters
Roll Thrusters
Figure 2. Service module (SM) attitude control thrusters position and direction. The pitch and yaw
thrusters point away from the vehicle surface (13 from the normal). The roll thrusters point 47
from the normal. contamination from the roll thrusters is expected on the adjacent SM surfaces
196 WILLIAM SCHMIDL ET AL.
Figure 3. (a) Service module (SM) Gas Dynamic Protection Unit (GZU) for the SM roll thrusters
prior to flight and installation; (b) SM roll thrusters prior to installation of the GZU; and (c) SM roll
thrusters (right side of the image) and pitch thrusters (left side of the image) prior to installation of
the GZU
The image in figure 1 was acquired before shields, gas dynamic protection
devices (GZUs), were installed on the thrusters in January 2002. Rocket and
Space Corporation Energia (RSC-E) has designed the GZUs to constrain the
thruster plume and limit contamination of the surrounding surfaces. Figure 3(a)
shows the GZU for one of the roll thrusters prior to flight. It fits over the top of
the roll thrusters. The handle on the top of the GZU is used during installation
and is not used for translating during EVAs, as it becomes contaminated. The
brackets on the sides of the GZU are used to install the GZU on fittings that were
preinstalled on the SM surface prior to flight. Figure 3(b) shows an image of roll
thrusters prior to installation of the GZU. Figure 3(c) shows the other side of the
roll thrusters (right side of the image) and the corresponding pitch thrusters (left
side of the image). It should be noted that the GZUs used for the pitch and yaw
thrusters are different than the ones used for the roll thrusters because they are
recessed below the SM surface.
Images have been taken at regular intervals from the DC1 window of the SM
nadir attitude control thrusters and the Russian Kromka experiment. Figure 4
shows one of these images. Kromka is an experiment to measure how well the
GZUs are performing and to test how well some material samples perform in
space [3]. The Kromka experiment is the material tray visible in the middle of the
image. It has some small material samples mounted on it. In front of the Kromka
are the pitch thrusters with its GZU installed. In front of the pitch thrusters are
FUEL OXIDIZER REACTION PRODUCTS 197
GZU for SM
roll thrusters
Figure 4. Image taken from Docking Compartment 1 (DC1) window. Kromka is visible in the
middle of the image
the roll thrusters with its GZU installed. It can be seen that the GZUs are heavily
contaminated compared to the preflight GZU in figure 3(a). Figure 5 shows a
diagram of the nadir side of the SM and the relative position of the Kromka
experiment and the thrusters. The Kromka experiment is visible near the SM
pitch thrusters. EVA handrails can be seen on either side of the pitch thrusters.
3. Eva Constraints
Due to the presence of FORP on the SM surfaces adjacent to the roll thrusters,
additional EVA constraints were required to be implemented in these areas. The
constraints were initially established through a nonconformance report (NCR)
that discussed the removal of the Kromka 1-0 experiment and installation of the
Kromka 1-1 experiment and in subsequent ISS Program Safety Review Panel
(SRP) discussions. The constraints were later confirmed in the protocol from a
Figure 5. Nadir side of service module (SM). The Kromka experiment is visible near the SM pitch
thrusters
198 WILLIAM SCHMIDL ET AL.
4.1. INTRODUCTION
A test program was setup at the NASA White Sands Test Facility to obtain the
data that would be needed to determine what EVA constraints would be required
to mitigate the risk of an EVA crewmember inadvertently contacting a FORP
contaminated surface and bringing the FORP back into the humid ISS cabin. The
program included tests to determine the evaporation rate of FORP on the zenith
(25 C, hot) and nadir (40 C, cold) sides of the ISS, the evaporation rate of
NDMA within the FORP on the zenith and nadir sides of the ISS, the quantity of
NDMA that would be released in a humid pressurized environment from the dried
FORP, and the rate at which NDMA reforms when dried FORP is introduced into
a humid environment.
Two groups of tests were performed. The first group of tests were performed
during Calendar Year 2003. Results from this group of tests are designated CY
2003. Results from these tests included 100 h evaporation rate data for FORP for
both the zenith and nadir cases, NDMA evaporation rate data for the corresponding
zenith and nadir cases, and NDMA formation rate data. Based on these results,
the time to remain outside the keep-out zone (KOZ) after the last SM thrusters
fired was reduced from 2.5 to 2 h.
Additional tests were requested to determine if the time to remain outside
of the KOZ could be further reduced from 2 to 1 h. This was the series of tests
performed during early Calendar Year 2004 and designated CY 2004. Results
from these tests included 1 to 6 h evaporation rate data for FORP for the zenith
FUEL OXIDIZER REACTION PRODUCTS 199
and nadir cases, NDMA evaporation rate data for the corresponding zenith and
nadir cases, and NDMA formation rate data.
For each group of tests, a batch of FORP was generated using a permeation
technique developed at the NASA White Sands Test Facility (WSTF) [2]. In this
technique, separate Unsymmetrical Dimethylhydrazine (UDMH) and NO2 gas
streams are concentrated in a small controlled area. The batch of FORP needed
for the tests was prepared over a couple of weeks. For the formation test, a sample
of FORP from each batch was evaporated for 5 days at 25 C in a vacuum to
generate a sample of dried FORP. Since the FORP was prepared over a long
duration, the composition of the two batches of FORP varied. The composition of
the FORP batches used in the tests is shown in table 1. For the CY 2003 FORP,
only the composition of the evaporated sample was measured. When comparing
the two evaporated samples, it can be seen that the CY 2004 FORP has a higher
concentration of dimethylammonium, nitrates, and nitrites, 7.9% vs. 13%, 36%
vs. 48%, and 0.08% vs. 0.6% respectively, than the CY 2003 FORP. The higher
concentration of dimethylammonium, nitrates, and nitrites in the CY 2004 FORP
likely explains why it has more mass remaining after the 5 day evaporation, 14%
vs. 10%, than the CY 2003 FORP.
The nitrite levels of the CY 2004 FORP before evaporation and after evapora-
tion, 20% of the initial mass and 0.6% of the 14% mass remaining, indicate that
the nitrite concentration is decreasing. A lower nitrite level significantly decreases
the NDMA formation rate. This was seen in the results that will be discussed later.
The concentration of dimethylammonium is also higher in the CY 2004 dried
FORP, 13% vs. 7.9%, than in the CY 2003 FORP. The higher concentration of
dimethylammonium and nitrites indicates that a higher NDMA formation rate
200 WILLIAM SCHMIDL ET AL.
would be expected for the CY 2004 FORP than the CY 2003 FORP. This was seen
in the results.
The technique used to measure the evaporation rate for FORP and NDMA and the
NDMA formation rate is discussed in the WSTF report Evaporation rate study
and NDMA formation from UDMH/NO2 Reaction Product, WSTF-IR-0188-
001-03 (2003) [1].
g NDMA/g FORP(t=0)[2003(-40C)]
g NDMA/g FORP(t=0)[2004(-40C)]20
10000
1000
g NDMA/g FORP(t=0)
100
10
1
0 1 2 3 4 5 6 7
Time (hrs)
Figure 6. Concentration of NDMA relative to the initial mass of FORP (g NDMA/g FORP) for
the 40 C case (nadir, cold case) decreases rapidly with time
the CY 2003 tests. It can be seen that the CY 2004 FORP starts (time = 0) with
a higher concentration of NDMA than the CY 2003 FORP. However, by the time
it reaches the 2 and 4 h data points, the concentration of NDMA relative to the
initial amount of FORP is comparable and the CY 2004 and CY 2003 data points
overlap.
The results in figure 7 are for the zenith (25 C, hot) case. It can be seen
that the NDMA concentration drops more rapidly for the 25 C case than for the
40 C case. After 1 h the concentration has dropped approximately 2 orders of
magnitude, compared with the 1 order of magnitude for the 40 C case. It can
also be seen that by 1 h the concentration drop has reached a plateau.
g NDMA/g FORP(t=0)[2003(25C)]
g NDMA/g FORP(t=0)[2004(25C)])
10000
1000
g NDMA/g FORP(t=0)
100
10
1
0 1 2 3 4 5 6 7
Time (hrs)
Figure 7. Concentration of NDMA relative to the initial mass of FORP (g NDMA/g FORP) for
the 25 C case (zenith, hot case) decreases rapidly with time. Within 1 h the concentration level has
reached a plateau
10
1
0 1 2 3 4 5 6 7
Time (hrs)
Figure 8. FORP weight (%) vs. time for the 40 C case (nadir, cold case)
FUEL OXIDIZER REACTION PRODUCTS 203
3 pts 2 pts
FORP Weight %
10
2 pts 2 pts
1
0 1 2 3 4 5 6 7
Time (hrs)
Figure 9. FORP weight (%) vs. time for the 25 C case (zenith, hot case). FORP volatilizes rapidly
to a stable mass that persists for a longer period of time
the higher mass remaining is likely due to the higher concentration of nitrites and
nitrates.
1.00E+04
1.00E+03
g NDMA / g FORP
1.00E+02
2004 g NDMA/g FORP (Nitriteadded)
2004 g NDMA/g FORP (Nonitriteadded)
2003 g NDMA/g FORP (Nitriteadded)
1.00E+01
2003 g NDMA/g FORP (Nonitriteadded)
1.00E+00
1.00E-01
0 1 2 3 4 5 6 7
Time (hrs)
Figure 10. NDMA concentration (g NDMA/g FORP) vs. time measured during the NDMA for-
mation test. It can be seen that the NDMA forms rapidly
added. It can be seen that when no nitrite is present the NDMA formation rate is
very low. For the CY 2003 FORP, no NDMA formation was detected. For the CY
2004, a low rate of 100 g NDMA/g FORP was measured.
The results in figure 10 also show that for the CY 2003 FORP, the NDMA
formation rate is 1800 g NDMA/g FORP present after 2 h and that the formation
rate for the CY 2004 FORP was higher, 18400 g NDMA/g FORP present after
2 h. One cause for the higher formation rate is the higher Dimethylammonium
concentration in the CY 2004 FORP. The mass of CY 2004 FORP remaining
after the 5 days of evaporation was also higher. This might indicate that there
were other components present in the FORP that might have resulted in a higher
NDMA formation rate.
5. Methodology
Figure 11. Zenith side of service module (SM). Gas Dynamic Protection Units (GZUs) constrain
the thruster plume. The distance from the roll thruster to the closest SM surface outside the GZUs
is 3.2 in (8 cm)
firing times were obtained from the ISS Programs Guidance, Navigation and
Control group (GN&C). The value of 45 s was used because it is the longest
thrusters firing time that has been observed during the previous EVAs.
The next step is to calculate how much FORP would remain 1 h after the SM
thrusters have fired. This was determined using the WSTF laboratory evapora-
tion test data for the Nadir case. The conservative value of 36% FORP remaining
used. Figure 11 shows a map of the distribution of FORP from the thruster re-
maining on the SM thruster based on the plume model and laboratory data. It
can be seen that the FORP concentration drops rapidly with distance from the
thrusters.
Next, to be conservative, it is assumed that all the FORP in a 100 cm2 area
is transferred to the suit by the inadvertent swipe. Based on the amount of FORP
transferred, the amount of NDMA that would be released from this FORP inside
the cabin in the first 2 h was calculated using the rate of 400 g NDMA/g FORP
present. This value was obtained from the WSTF data based on the evaporation
rate of NDMA for the zenith hot case. The amount of NDMA that would be
formed from the dried FORP in the first 2 h was also calculated based on the more
conservative CY 2004 rate, 18400 g NDMA/g FORP present.
Figure 12 shows that since the GZUs were installed, the closest point on the
SM surface that can be touched is 3.2 in (8 cm) from the edge of the thruster.
The diagram is of the zenith side of the SM. The figure on the left shows the roll
thrusters with the GZUs installed. The figure on the right shows the distance from
the edge of the roll thruster to the closest SM surface outside of the GZUs. The
roll thruster diameter is 1.8 in. The distance from the roll thruster centerline to
the closest SM surface is 5 in (12.7 cm).
The predicted concentrations of NDMA with distance from the SM thrusters
that would be released by an inadvertent swipe into the DC1 and ISS compartments
206 WILLIAM SCHMIDL ET AL.
Figure 12. Map of the FORP predicted to remain after 1 h on the surface around the SM roll thrusters
after a 45 s roll thruster firing. It can be seen that the FORP concentration drops rapidly with distance
from the thruster
are shown in tables 2 and 3. It can be seen that the concentration drops off rapidly.
Also the closest point is at the GZU itself, which is defined as a no touch area
by flight rule. This data was given to the Toxicology group for assessment.
6. Toxicological Assessment
The quantity of FORP present and NDMA released into the ISS from an inad-
vertent swipe by an EVA crewmember of the contaminated area around the SM
TABLE 2. NDMA concentration predicted to be released in the cabin for FORP transferred to the
EVA suit one hour after the roll thruster firing at different distances from the roll thruster for the
40 C case (Nadir)
NDMA
NDMA released concentration (ppb)
Distance from FORP present in pressurized
roll jet (m) (gcm2 ) environment DC1 ISS
TABLE 3. NDMA concentration predicted to be released in the cabin for FORP transferred to the
EVA suit one hour after the roll thruster firing at different distances from the roll thruster for the
25 C case (Zenith)
NDMA
NDMA released concentration (ppb)
Distance from FORP present in pressurized
roll jet (m) (gcm2 ) environment (g) DC1 ISS
roll thrusters was calculated and provided to the NASA Toxicology group for an
assessment. This data is shown in tables 2 and 3.
Acute toxicity has not been reported at concentrations below 10 ppm. There-
fore, it can be concluded that at concentrations below 1 ppm, it is very unlikely that
NDMA will produce any acute toxic reactions. So, the NASA Toxicology Group
concluded that for the concentration levels expected, it is unlikely that NDMA
will produce acute toxicity.
The NASA Toxicology Group also found that the highest calculated risk from
the projected NDMA concentrations is less than 8.46 105 (40 C, distance
0.08 m from the thrusters). The NASA Toxicology Group, with the concurrence of
the National Research Council Spacecraft Maximum Allowable Concentrations
(SMAC) Subcommittee, accepts a cancer risk of 1/10,000 (i.e., 104 ) in deriving
SMACs on carcinogenic compounds, such as benzene.
7. Summary
Prior to an extravehicular activity (EVA) on the Russian Segment (RS), the Dock-
ing Compartment (DC1) must be depressurized, as it is used as an airlock. When
the DC1 is depressurized, the CMGs margin of momentum is insufficient to
compensate for the disturbance and the Service Module (SM) attitude control
thrusters need to fire to desaturate the CMGs. The thruster firings result in FORP
contamination of the adjacent SM surfaces.
The FORP contamination of the SM surfaces, the release of NDMA in a
humid environment from crew EVA suits, if they happen to be contaminated
with FORP, and the toxicological risk associated with the NDMA release were
calculated. It was determined that the FORP and NDMA evaporate rapidly and
that their concentration drops off rapidly with distance from the thrusters. The
208 WILLIAM SCHMIDL ET AL.
FORP remaining after 1 h for the Nadir case was found to be 36% of the initial
mass. For the Zenith case the FORP remaining after 1 h was 22% of the initial
mass.
The NASA Toxicology Group found that the highest calculated risk from
the projected NDMA concentrations is less than 8.46 105 (40 C, distance
0.08 m from the thrusters). The NASA Toxicology Group, with the concurrence of
the National Research Council Spacecraft Maximum Allowable Concentrations
(SMAC) Subcommittee, accepts a cancer risk of 1/10,000 (i.e., 104 ) in deriving
SMACs on carcinogenic compounds, such as benzene.
Based on these results the time to remain outside the 1 m KOZ could be reduced
from 2 to 1 h.
Acknowledgments
The authors gratefully acknowledge the ISS Program office for supporting this
study, and the NASA White Sands Test Facility (WSTF) and its personnel for
conducting the study.
References
1. Buchanan, V. D. and Dee, L. A. (2003) White Sands Test Facility (WSTF) Investigative Report,
WSTF-IR-0188-001-03.
2. Dee, L. A., Davidson, V. D., and Baker, D. L. (2000) Protocol External Contamination Technical
Interchange Meeting Fuel/Oxidizer Reaction Products (FORP) Plume Induced Contamination,
1526 April 2002.
3. Naumov, S. F., Gerasimov, Y. I., Sokolova, S. P., Rebrov, S. G., Gerasimova, T. I., Kalistratova,
O. V., Prokofyev, M. A., Grigorevsky, A. V., Prosvirikov, V. M., Buryak, A. K., and Chernik,
V. N. (2003) In 9th International Symposium on Materials in a Space Environment, ESTEC,
Noordwijk, The Netherlands, 1620 June 2003 .
EFFECT OF VACUUM THERMOCYCLING ON PROPERTIES OF
UNIDIRECTIONAL M40J/AG-80 COMPOSITES
1. Introduction
Also, under high vacuum, the mass loss of the epoxy matrix composites could
take place. Therefore, it is important to study the behavior of the composites
under vacuum thermocycling, which might influence the performance of space-
craft in orbit [35]. The AG-80 resin is a new type of thermosetting matrix for
advanced polymetric composites, which exhibits a unique resistance to wet and
heat [6], showing a good prospect for the usage in spacecraft. The aim of this
study was to examine the effect of vacuum thermocycling on the unidirectional
M40J/AG-80 composite, in order to provide the basic information for its fur-
ther development and application in spacecraft. The AG-80 resin is a type of
tetraglycidyl diaminodiphenyl methane (TGDDM) with diaminodiphenyl sulfone
(DDS).
2. Experimental
A commercial grade of AG-80 resin was used as the matrix material, which
was supplied by Shanghai Institute of Synthetic Resins in China. The chemical
structure is shown in figure 1. Its epoxy value is approximately 0.80, and the curing
agent is diaminodiphenyl sulfone (DDS). The M40J carbon fibers were chosen as
the reinforcement for the composites, which were supplied by TORAY Company
in Japan. The prepreg tape for the M40J/AG-80 composite specimens was laid
on a mould board and placed in an autoclave for curing. The curing cycles were:
heating to 130 C from room temperature with the heating rate 1.52.0 Cmin1 ,
and holding for 4060 min at 130 C under the pressure of 0.60.7 Mpa; and then
heating to 180 C with the rate 1.52.0 Cmin1 and holding for 120 min; finally,
cooling to room temperature in the autoclave. The volume fraction of M40J fibers
in the manufactured composites was approximately 60%.
The vacuum thermocycling regime is shown in figure 2. Before and after
the vacuum thermocycling, the bending strength and modulus of the M40J/AG-
80 composites were measured using a MTS 810 type test machine. The mass
loss ratios for the specimens after the thermocycling were characterized by a
high precision microbalance with a sensitivity of 105 g. The change in surface
morphology of specimens due to the vacuum thermocycling was observed using
an atomic force microscope (AFM) of Nanoscope IIIa Dimension 3100 type.
H H
H2 C C CH2 CH2 C CH2
O O
N CH2 N
H H
H2C C CH2 CH2 C CH2
O O
200
(140C,10min)
105Pa
Temperature/C
100
100
200 (180C,10min)
0 10 20 30 40 50 60 70
Time/min
Figure 2. Temperature vs. time curve for the vacuum thermocycling
The laid direction of fibers was parallel to the observed surface of specimens.
In order to reveal the effect of thermocycling on the chemical structure of the
AG-80 matrix, the dynamic mechanics thermal analysis (DMA) was performed
using a Rheometric Scientific DMTA Vtype spectrometer by means of a standard
three-point mode. The loading frequency was 1 Hz, the temperature ranged from
130 to +300 C and the heating rate was chosen as 5 Cmin1 . The size of
DMA samples was 40 7 1 mm3 , and the longitudinal direction of samples
was parallel to the laid direction of the carbon fibers.
Figures 3(a) and 3(b) show the changes in bending strength and modulus with
vacuum-thermal cycles for the M40J/AG-80 composites, respectively. The varia-
tion trends for the bending strength and modulus are similar. With increasing the
1800 215
(a) (b)
Bending modulus/GPa
105Pa 105Pa
210
Bend strength/MPa
93 413K 93 413K
1700
205
200
1600
195
1500 190
185
1400
0 50 100 150 200 250 300 350 0 50 100 150 200 250 300
Thermal cycles Thermal cycles
Figure 3. The bending strength and (a) modulus vs. (b) vacuum-thermal cycles for M40J/AG-80
composites
212 YU GAO ET AL.
0.4
0.3
Weight loss/%
5
10 Pa
0.2 93 413K
0.1
0.0
0 40 80 120 160
Thermal cycles
Figure 4. The mass loss ratio vs. vacuum-thermal cycles for M40J/AG-80 composites
vacuum-thermal cycles, both the bending strength and modulus increase firstly
and begin to decrease after approximately 40 cycles. After 97 cycles, the decrease
tends to level off.
Figure 5 shows the change in the surface morphology after vacuum-thermal cy-
cling for the M40J/AG-80 composites, in which the carbon fibers, the epoxy
matrix, and the interfacial layers are indicated by A, B, and C respectively. Before
the vacuum thermocycling, the bonding of the matrix with carbon fibers is in good
EFFECT OF VACUUM THERMOCYCLING 213
Figure 5. AFM micrographs showing the change in surface morphology after vacuum-thermal
cycling for M40J/AG-80 composites: (a) 0 cycles; (b) 40 cycles; (c) 274 cycles
Figure 6 shows the change of the loss tangent (tan ) vs. temperature spectrum
after vacuum thermocycling for the M40J/AG-80 composites. The spectrum can
be divided into two portions, as shown in figures 6(a) and 6(b), respectively.
Figure 6(a) shows the change of the peaks for the primary or -transition. The
temperature at which the -transition peak appears is corresponding to the glass
transition temperature (Tg ). It can be seen that with increasing the vacuum-thermal
cycles, the Tg value decreases slightly. After 40 cycles, the peak height for the -
transition drops obviously (almost by 18%), implying that the cross-linking extent
increases for the M40J/AG-80 composites. The increase in the cross-linking extent
demonstrates that the vacuum thermocycling would lead to a post curing effect on
the epoxy matrix. However, when the thermal cycles increase from 40 to 274, the
214 YU GAO ET AL.
0.045
0.30 (a) (b) original state
0.040 vacuum thermo-cycling
for 40 cycles
0.035 vacuum thermo-cycling
0.25
for 274 cycles
0.030
Y
tan
tan
0.20 0.025
Figure 6. The tan vs. temperature spectra for the M40J/AG-80 composites after vacuum thermo-
cycling for various cycles
peak height change a little, indicating that the post curing does not take place any
more.
Figure 6(b) shows the changes of the peaks for the secondary transitions,
including the and ones. The peak height decreases with increasing the
thermal cycles, and disappears after 274 cycles. The transition occurring in
the temperature range of 10100 C can be related to the movement of benzene
rings, which would weaken with increasing the cross-linking extent. The change
in the peak height could also demonstrate the post curing effect during vacuum
thermocycling.
Also, as can be seen in figure 6(b), the -transition takes place in the tem-
perature range of 12050 C, which can be related to the movement of the chain
segment CH2 CHOH(R)CH2 . Such a chain segment could be formed by break-
ing the epoxy rings. The peak signals are stronger for the vacuum thermocycling
of 40 cycles than those before the thermocycling. This phenomenon indicates that
the amount of the CH2 CHOHCH2 segments in the epoxy matrix is larger in the
former state than in the latter one, and thus the post curing would occur due to the
thermocycling.
4. Conclusions
References
Abstract. The damage characteristics of Zr41 Ti14 Cu12.5 Ni10 Be22.5 bulk metallic
glass under planar shock wave have been investigated by firing aluminum projec-
tiles using a two-stage light gas gun. The SEM results show that radial and sym-
metric cracks formed on the shocked plane of the sample at the impacted location
by aluminum projectile at the velocity of 2.7 kms1 . Parallel shear crack/bands in
the sublayer under the shocked plane were formed. For a better understanding of
the response features under shock wave, hypervelocity impact tests with conven-
tional sphere aluminum projectiles were carried out. Besides the same adiabatic
shear crack/bands and crack propagations, craters were formed and lamination
cracks occurred.
1. Introduction
Multicomponent Zr41 Ti14 Cu12.5 Ni10 Be22.5 bulk metallic glass (BMG) with ex-
tremely high glass forming ability, one of the most widely studied BMGs [1],
has recently gained considerable attention due to its low density, high strength
and fracture toughness, good corrosion and wear resistance, excellent ductibil-
ity, and uniquely dynamical deformation characteristics such as adiabatic shear
banding. In recent years, this specific alloy has already made commercial applica-
tion potential such as headway in golf clubs. BMG is also a likely candidate as a
217
J.I. Kleiman (ed.), Protection of Materials and Structures from Space Environment, 217223.
C 2006 Springer. Printed in the Netherlands.
218 C. YANG ET AL.
space vehicle material. The BMG used for the spacecrafts might face a danger of
collision by space debris and micrometeoroids. In order to protect the space vehi-
cles against the impact of space debris and micrometeoroids, related knowledge
of high strain rate response must be obtained to evaluate the dynamic properties
of the BMG impacted by hypervelocity projectiles. Extensive research on damage
behavior of Zr41 Ti14 Cu12.5 Ni10 Be22.5 BMG under the uniaxial compression [24],
tensile tests [5, 6], and shocking loading by a powder gun loading system [7] has
been carried out in the past. However, these studies paid particular emphasis to low
and medium strain rate response of the BMG. In this paper, we report the damage
characteristics of a Zr41 Ti14 Cu12.5 Ni10 Be22.5 BMG under the planar shock wave
treatment and the impact of hypervelocity projectiles by accelerating aluminum
projectiles using two-stage light gas gun.
2. Experimental Procedure
The ingots of the alloy with a nominal composition of Zr41 Ti14 Cu12.5 Ni10 Be22.5
were firstly prepared from a mixture of pure elements in an arc-melting furnace
under Ti-gettered Argon atmosphere. The purity of Zr, Ti, Cu, Ni, and Be were
99.999, 99.9, 99.5, 99, and 99.5% respectively. The ingots were crushed and
remelted in a quartz tube, and then were quenched into water to obtain a BMG
rod. The structure of the BMG was identified by X-ray diffraction (XRD) to be
fully amorphous and no crystalline phase was detected. Small cylinders 10 mm
long, cut out from the alloy rod, were utilized as impacted target. Targets were
recovered using a soft recovery device, as shown in figure 1. The amorphous
BMG cylinder was embedded into a copper cylinder with a diameter of 40 mm.
Figure 1. Schematic diagram of the recovery device for high speed impact [(1) steel cylinder;
(2) steel cylinder tube; (3) aluminum cylinder; (4) polyethylene cylinder; (5) copper cylinder; (6).
Zr41 Ti14 Cu12.5 Ni10 Be22.5 BMG target]
DAMAGE CHARACTERISTICS OF Zr41 Ti14 Cu12.5 Ni10 Be22.5 BMG 219
A steel cylinder was mounted tightly beneath the copper cylinder and aluminum
cylinder was mounted tightly beneath the steel cylinder. The polyethylene cylin-
der was mounted tightly beneath the aluminum cylinder. Finally, the assembly of
these cylinders was inserted into a steel cylinder with 40 mm inside diameter and
120 mm outer diameter. The end facing the direction of the cylinder assembly
was machined as shown in figure 1. The other end was fixed by bolts. The BMG
cylinder samples of diameter 22 mm were impacted by the hypervelocity spher-
ical aluminum projectiles of 5 mm in diameter at speeds of about 2.6, 3.2, and
3.7 kms1 , respectively. Though these speeds are not very high, they can cause
a catastrophic failure of protected layers for space vehicles [8]. As compared
with the damage behavior for hypervelocity impact, planar shock experiment was
carried out using the same recovery device except that the BMG sample with a
diameter of 18 mm was covered by a copper plate with a thickness of 2 mm. In the
planar shock experiment, the cylinder aluminum projectile was fired at a speed of
2.7 kms1 . The projectile was designed to be 22 mm in diameter and 2 mm thick
in order to avoid the influence of side rarefaction wave and reflected rarefaction
wave on the sample. The axis of projectile is coincident with the steel cylinder
assembly. The projectiles were fired by two-stage light gas gun. The velocities of
the projectiles were measured by the electromagnetic induction method. The spec-
imens were cut from the impacted recovered targets with an electric spark-cutting
machine, and were grinded and polished mechanically after cutting. The morphol-
ogy of the craters and microstructure changes in the region near the crater and
deformation damage characteristics of cross section such as adiabatic shear bands
and microcracks were observed by XL30 S-FEG scanning electron microscope
(SEM).
Figure 2 shows the craters that were formed in the Zr41 Ti14 Cu12.5 Ni10 Be22.5 BMG
target after the hypervelocity impact with velocities of 2.6, 3.2, 3.7 kms1 , re-
spectively. The samples projectile 22 mm in diameter and a thickness of 10 mm
belong to the medium thick target, in which the impact wave can interact with the
side and back faces, resulting in significant effect on the impact process. It can be
seen that shape of the craters is almost spherical coronary without an obvious outer
projecting fringe in the mouth of the crater, which is extruded from the crater
by high pressure and commonly observed in the target impacted by hypervelocity
projectiles. This may be related to the high strength and low ductility of the BMG
220 C. YANG ET AL.
a b
Figure 2. Cross-section morphologies of damage craters formed in the Zr41 Ti14 Cu12.5 Ni10 Be22.5
BMGs medium thick target: (a) sample A: v = 2.6 kms1 ; (b) sample B: v = 3.2 kms1 ;
(c) sample C: v = 3.7 kms1
materials. The shape of the crater becomes spherical from spherical coronary with
the increase of projectile velocity.
The resulting dimensions of the sample after the planar shock loading with an
aluminum flyer at a velocity of 2.7 kms1 were 20.5 mm in diameter and 7.3 mm
thickness compared to 17.5 mm in diameter and 10 mm thickness for the orig-
inal sample. The upper surface facing the flyer exhibits an annular crack and
many densely radial cracks initiating from a central point. The reason for this
phenomenon is that the central part of aluminum flyer collided firstly with the
central part of the BMG sample because the flyer was hindered by air in the target
housing. The central part of the flyer deforms during the flight. The compres-
sive shock wave therefore acts initially on the center point of the BMG target,
thus initiating the failure in this part first. Then, radial cracks propagate from the
point under the action of strong shock wave. Annular cracks formed naturally
by interaction between the wave (tensile wave) reflected from the cylinder copper
wall and compressive shock wave. The recovered aluminum flyer with a prominent
DAMAGE CHARACTERISTICS OF Zr41 Ti14 Cu12.5 Ni10 Be22.5 BMG 221
Figure 3. SEM micrographs showing cross-section microstructures of Zr41 Ti14 Cu12.5 Ni10 Be22.5
BMG under planar shock loading of aluminum flyer with a speed of 2.7 kms1
feature formed along the direction of flight demonstrates experimentally the above
discussed phenomenon.
Figure 3 shows a SEM microstructure micrograph of Zr41 Ti14 Cu12.5 Ni10 Be22.5
BMG target hit by aluminum flyer with a speed of 2.7 km/s. Several fracture planes
(shear cracks/bands) 2 mm apart and approximately parallel to each other were
seen in the upper part of the sample (that are not shown in figure 3 due to limitation
of the view field of SEM). Another kind of shear cracks/bands with an angle of
40 between the fracture plane and horizontal direction result from the interaction
between the shock wave reflected from the cylinder copper wall and compressive
shock wave. This kind of cracks/bands cannot appear in usually uniaxial compres-
sive and tensile tests because the exterior walls of the cylinder sample has not been
restrained under these conditions. In our case, as discussed above, the sample was
reduced in thickness and expanded in diameter after the impact. Once the shearing
occurs along the direction shown by the left arrow, it results in the formation of
primarily adiabatic shear band and induces a new secondary shear band along the
direction shown by the left arrow. Thus a triangular region with an apex angle
of 115 is extruded from the sample because of convergence between a reflected
shock compressive wave along the radial direction of the sample from the interior
wall of the recovery copper cylinder and a compressive shock wave along the pro-
jectile trajectory. The upper two shear cracks/bands has already intersected each
other, while the subjacent two havent. The zigzag-shaped cracks (indicated by
right arrow) on the micrograph of the BMG suggest that the spalling of the BMG is
characterized by brittle damage resulting from the shear localization even though
the sample is under tension during decompression process. Shear cracks/bands
propagate and end at a point where several secondary radial microcracks (shear
222 C. YANG ET AL.
Figure 4. Typical SEM micrographs of shear bands/cracks observed on the cross section of the
recovered specimen after shock loading with a velocity of v = 3.7 kms1
sizes of these shear bands/cracks increase gradually with the increase of the impact
velocity.
4. Conclusions
Zr41 Ti14 Cu12.5 Ni10 Be22.5 BMG glasses under shock loading by two-stage light
gas gun revealed shear fracture cracks/bands formation. Radial symmetric cracks
were formed at the shocked plane. Shear crack/bands parallel to each other on the
part close to the shocked plane were formed readily on the cross section of the
recovered sample. The depth of crater in the target increases linearly with increase
of the projectile velocity in the range of experimental velocity. Edge sides shear
band/crack, together with lamination shear band/crack has been observed on the
recovered samples impacted by hypervelocity projectiles. Based on these results,
it can be concluded that BMG can be used as a candidate for a space vehicle
material that resist the impact of small space debris.
Acknowledgment
We wish to thank the National Natural Science Foundation for its support of this
research through Grant 50171077, 50171059, and 50325103.
References
1. Tang, X. P., Geyer, U., Busch, R., Johnson, W. L., and Wu, Y. (1999) Nature 402, 160162.
2. Subhash, G., Dowding, R. J., and Kecskes, L. J. (2002) Materials Science and Engineering A334,
3340.
3. Gilbert, C. J., Ager III, J. W., Schroeder, V., and Ritchie, R. O. (1999) Applied Physics Letters
74, 3809.
4. Wright, W. J., Schwarz, R. B., and Nix, W. D. (2001) Materials Science and Engineering A
319321, 229232.
5. Li, J. X., Shan, G. B., Gao, K. W., Qiao, L. J., and Chu, W. Y. (2003) Materials Science and
Engineering A 354, 337343.
6. Flores, K. M. and Dauskardt, R. H. (2001) Materials Science and Engineering A 319321,
511515.
7. Zhuang, S. M., Lu, J., and Ravichandran, G. (2002) Applied Physics Letters 80, 4522.
8. Zukas, J. A. (1989) Impact Dynamics (translated into Chinese by Zhiyun Zhang et al.), Publishing
Company of Military Industry, P. R. China, p. 182.
EFFECT OF VUV RADIATION ON PROPERTIES AND CHEMICAL
STRUCTURE OF POLYETHYLENE TEREPHTHALATE FILM
1. Introduction
Polymeric materials have been widely employed in spacecraft for parabolic anten-
nas, solar wings, and thermal control systems [13]. Their performance in space
will directly influence the reliability and lifetime of spacecraft. Vacuum ultraviolet
(VUV) radiation is one of the major environment factors causing the degradation
of polymers. Although the VUV energy percentage in the total spectrum of the
Sun is very limited, its photon energy is high enough to break most chemical
bonds in the polymers [4]. Therefore, it is important to study the effect of VUV
radiation on the chemical structure and properties of polymeric materials, such as
the polyethylene terephthalate (PET) film. Most published work about the effects
of VUV on polymeric films was performed using various VUV lamps, with which
it could be difficult to obtain a VUV spectrum similar to the Sun. A gas-jet type
VUV source was used in this study, which could give a spectrum similar to the Sun
in the wavelength range of 5200 nm [4]. The aim of this work was to examine the
changes in chemical structure and properties of the PET film under VUV exposure.
225
J.I. Kleiman (ed.), Protection of Materials and Structures from Space Environment, 225232.
C 2006 Springer. Printed in the Netherlands.
226 GUIRONG PENG ET AL.
O O
C C O CH2 CH2 O
2. Experimental
The PET film was 60 m thick. Figure 1 shows the chemical structure of the PET
film. The film was annealed at 70 C for 3.5 h, rinsed with analytically pure acetone
and ethanol, and dried in a desiccator for more than 48 h at ambient temperature
before the VUV exposure.
The VUV source used in this study operates with supersonic argon gas, which
is excited by a high-energy electron beam to give the VUV wavelengths ranged
from 5 to 200 nm. The VUV intensity of 0.24 wm2 was acquired at a distance of
70 cm from the source, corresponding to 10 times the VUV solar constant (VUV
suns). In this study, the gas-jet pressure was 5 atm, the electron-beam energy
was 1000 eV and the electric current was 15 mA. The vacuum before and after
the injection of argon gas into the chamber was 105 and 103 Pa, respectively.
The samples were located at 72, 34, and 22 cm away from the VUV source,
corresponding to the 10, 40, and 100 VUV suns respectively.
The tensile tests were carried out at room temperature after the VUV exposure.
The engaged surface area of the film samples was 20 5 mm2 , and the crosshead
speed was 2.4 mmmin1 . The transmittance of the VUV irradiated film sam-
ples in the wavelength range of 2003200 nm was measured with a UV-3101PC
scanning spectrophotometer. The X-ray photoelectron spectroscopy (XPS) was
performed using a VG-ESCALAB Mark-2 type spectrometer with Mg K source.
The vacuum in the chamber was 106 Pa, and the pass energy 20 eV. IR spec-
tra were acquired using a Perkin Elmer System 2000 Fourier Transform Infrared
(FTIR) spectrophotometer. The resolution of the spectrophotometer was 1 cm1 .
The test wave numbers were in the range of 4004000 cm1 . The electron-spin
resonance analysis (ESR) was performed at room temperature in a JES-FE3AX
type spectrometer.
Figure 2 shows the changes in tensile strength and elongation for the PET film
irradiated with various doses of VUV. It can be seen that both the strength and
elongation decrease slightly with increasing the VUV dose.
EFFECT OF VUV RADIATION 227
240
120
(a) (b)
200
90
f , MPa
, %
160 60
100 VUV suns
30 100 VUV suns
120
0
0 2000 4000 6000 0 2000 4000 6000
Irradiation dose, esh Irradiation dose, esh
Figure 2. (a) The tensile strength f and (b) elongation vs. VUV dose for PET film
Figure 3(a) shows the change in spectral absorbance A for the PET film
after VUV irradiation. The absorbance limit is at the wavelength of 322 nm and
changes little with VUV dose.
The A increases with the irradiation dose. Figure 3(b) illustrates the effect of
VUV intensity on A . The A in the near-ultraviolet region changes remarkably
with the VUV intensity; but less in the visible to near-infrared regions. Moreover,
under the same VUV dose, the A in the near-ultraviolet region increases less
for the PET film irradiated with 100 VUV suns than that under 40 VUV suns (see
the curves 3 and 4), but more than that under 10 VUV suns (see the curves 1 and
2). Figure 4 shows the change in A at 322 and 600 nm with VUV dose. With
increasing the dose, the A322 increases rapidly and then tends to level off, while
A600 changes almost linearly.
Figure 5 shows the C1s spectra of PET film irradiated with 100 VUV suns and
various doses. Before the irradiation, the C1s spectrum would be composed of
0.5
0.6 (a)
(b)
0.4
0.5 4 4, 1600 esh @ 40 VUV suns
3, 2940 esh
0.4
0.3 3, 1600 esh @ 100 VUV suns
3 2, 1600 esh 3 2, 680 esh @ 100 VUV suns
A
Figure 3. The change in absorbance A vs. (a) VUV dose and (b) intensity for PET film
228 GUIRONG PENG ET AL.
0.6
100 VUV suns A 322
0.5
0.4
A
0.3
0.2
A 600
0.1
0.0
0 600 1200 1800 2400 3000
Irradiation dose, esh
Figure 4. The change in absorbance at 322 nm (A322 ) and 600 nm (A600 ) vs. VUV dose for PET
film irradiated with 100 VUV suns
three peaks with binding energy at 284.6, 286.3, and 288.4 eV in turn, as shown
in figure 5(a). The three peaks might be related to the carbon in the benzene rings,
the carbon singly bound with oxygen in the ethylene, and that in the carbonyl
groups, respectively [56]. After the VUV irradiation, the C1s spectra could be
still characterized with the three peaks, as shown in figures 5(b) and 5(c). Table 1
shows the change in the corresponding area ratio for the three characteristic peaks
(c)
2940 esh
(b)
CPS
680 esh
(a)
unirradiated
Figure 5. C1s spectra of PET film irradiated with 100 VUV suns for (a) 0, (b) 680, and (c) 2940 esh
EFFECT OF VUV RADIATION 229
TABLE 1. The change in area ratios of the fitting peaks for C1s
spectra of PET film after 100 VUV suns irradiation
with VUV dose. With increasing the irradiation dose, the area ratio for the peak
at 284.6 eV increases, but those at 286.3 and 288.4 eV decrease obviously.
Figure 6 shows the change in O1s spectrum after VUV irradiation. Before the
exposure, the O1s spectrum is fitted into two Gauss peaks with the binding energy
at 531.8 and 533.3 eV, respectively. The former could be related to the oxygen in
the carbonyl groups, and the latter to the oxygen singly bound with carbon [56].
After the irradiation, the intensities of the two peaks decrease. From the above
XPS analysis, it is believed that the single bonds of carbon with oxygen (CO) in
the PET molecules can be broken under the VUV radiation, and further resulting
in decarbonylation.
The FTIR spectra for the PET film before and after VUV irradiation are shown
in figure 7. The absorption peaks at 3627 and 3545 cm1 are related to the vibra-
tion of hydroxy1 groups [78]. After the VUV irradiation, these two peaks are
(b)
680 esh
CPS
(a)
0 esh
Figure 6. O1s spectra of PET film irradiated with 100 VUV suns for (a) 0, (b) 680, and (c) 2940 esh
230 GUIRONG PENG ET AL.
50
80 unirradiated (b)
100 VUV suns, 2940 esh
40
60
30
40
T,%
T,%
(a) 20
20
10
unirrradiated
0
100 VUV suns, 2940 esh
0
3600 3200 2800 2400 1800 1700 1600 1500
1 1
Wave number, cm Wave number, cm
Figure 7. FTIR spectra for PET film before and after irradiation with 100 VUV suns for 2940 esh
in the wave number ranges of (a) 37002300 and (b) 18001480 cm1
enhanced, indicating that the content of the end hydroxy1 groups might be in-
creased to some extent. In addition, several new absorption peaks appeared in the
region from 1680 to 1530 cm1 . These peaks originate from the change in the sub-
stitute groups in the benzene rings and the formation of ethylene groups such as
CH=CH [7].
Figure 8 shows the ESR spectra of PET film after VUV irradiation. The g value
is 2.004 and the line width 9G. On the basis of the ESR analysis, the free radical
concentration in the PET film samples could be estimated using the formula in
reference [9], as shown in figure 9. It is shown that the free radical concentration
increases rapidly and then tends to level off with the increase of the VUV dose.
This trend is similar to that of A322 vs. dose in figure 4, demonstrating that the
change of optical properties has a relation with the free radicals.
Figure 8. ESR spectra for PET film irradiated with 100 VUV suns for (a) 680 and (b) 1600 esh
EFFECT OF VUV RADIATION 231
25
100 VUV suns
4. Conclusion
References
1. Nakayama, Y., Imagawa, K., Tagshira, M., Nakai, M., Kudoh, H., Sugimoto, M., Kasai, N., and
Seguchi, T. (2001) Journal of High Performance Polymers 13(3), 433451.
2. Dever, J. A., Pietromica, A. J., Stueber, T. J., Sechkar, E. A., Messer, R. K. (2002) Simulated
Space Vacuum Ultraviolet (VUV) Exposure Testing for Polymer Films, NASA Report TM-
2002-211337.
3. Iwata, M., Tohyama, F., Ohnishi, A., and Hirosawa, H. (2001) Journal of Spacecraft and Rockets
38(4), 504509.
4. Verkhovtseva, E. T., Yaremenko, V. I., and Telepnev, V. D. (1997) In The Proceedings of 7th
International Symposium on Materials in Space Environment, Toulouse, France, 1620 June
1997, p. 119.
5. Beamson, G. and Briggs, D. (1992) High Resolution XPS of Organic Polymers: The Scienta
ESCA300 Database, John Wiley and Sons, Chichester, UK.
6. Ektessabi, U. and Yamaguchi, K. (2000) Thin Slid Films 377378, 793797.
232 GUIRONG PENG ET AL.
7. Grossetete, T., Rivaton, A., Gardette, J. L., Hoyle, C. E., Ziemer, M., Fagerburg, D. R., and
Clauberg, H. (2000) Polymer 41(10), 35413554.
8. Miyake, A. (1959) Journal of Polymer Science 38, 497512.
9. Yin, J. and Mo, Z. (2001) Modern Polymer Physics, Science Publisher, Beijing, 743829, 679
680.
STATUS OF SOLAR SAIL MATERIAL CHARACTERIZATION AT
NASAS MARSHALL SPACE FLIGHT CENTER
Abstract. Near term solar sail propelled science missions are targeting the La-
grange point 1 (L1) as well as locations sunward of L1 as destinations. These near
term missions include the Solar Polar Imager [1] and the L1 Diamond [2]. The
Environmental Effects Group at NASAs Marshall Space Flight Center (MSFC)
continues to actively characterize solar sail materials in preparation for these near
term solar sail missions. Previous investigations indicated that space environmen-
tal effects on sail material thermo-optical properties were minimal and would not
significantly affect the propulsion efficiency of the sail [35]. These investigations
also indicated that the sail material mechanical stability degrades with increasing
radiation exposure. This paper will further quantify the effects of space environ-
mental exposure on the mechanical and thermo-optical properties of candidate sail
materials. Candidate sail materials for these missions include Aluminum coated
MylarTM , TeonexTM , and CP1 (Colorless Polyimide).
Experimental data will be presented on the response of sail material to charged
particle radiation that indicates change in the mechanical and thermo-optical
properties. Thermo-optical property data will also be presented indicating the
effects of long-term Near Ultraviolet (NUV) exposure. Charged particle radiation
with subsequent Hypervelocity Impact (HVI) results will be presented, indicat-
ing that sail material damage is primarily limited to the diameter of the incident
projectile.
1. Introduction
The non-Keplerian mission opportunities offered by solar sail propulsion are driv-
ing many science missions to require the capabilities of solar sail propulsion. This
233
J.I. Kleiman (ed.), Protection of Materials and Structures from Space Environment, 233246.
C 2006 Springer. Printed in the Netherlands.
234 DAVID L. EDWARDS ET AL.
renewed interest in solar sails has resulted in escalating development of solar sail
materials, structures, and deployment mechanisms. Many challenges confront sail
material technologists and among these is the development of low areal density
sails. Present state-of-the-art sails have an areal density ranging between 5 and 15
gm2 .
Sail materials investigated in this effort are on the order of 3 gm2 . The far-
term goal of sail material manufacturers is a sail material on the orderof 0.01
gm2 [6].
As reported by McInnes [7], a solar sail is a thin membrane that uses the
momentum carried by photons to propel spacecraft. These photons originate from
the sun or can be beamed onto the sail with a laser. If the sail is a good reflector,
the momentum transferred to the sail can be almost doubled.
Since the momentum carried by a single photon is extremely small, the surface
area of a sail must be large to produce a reasonable acceleration, as indicated by
a = (2Ap t )/m. (1)
where,
anonrelativistic sail acceleration (ms2 ),
Asurface area (m2 ),
pt incident radiation pressure (Nm2 ),
mmass (kg).
The radiation pressure ( pt ) varies as the inverse square of the distance from the
Sun as shown by:
pt = 4.56 106 [(1 + R)/rAU
2
] (2)
where,
Rsurface reflectivity (0 < R < 1),
rAU distance to the Sun in astronomical units (AU).
A perfectly reflective sail (R = 1) at a distance of 1 AU from the Sun experiences
a light pressure of 9.1 Nm2 . Substituting eq. (2) into (1) and solving for the
acceleration gives:
a = 9.12 106 [(1 + R)A/mr 2AU ] (3)
For example, a 1000 kg spacecraft with a perfectly reflective (R = 1) sail
area of 106 m2 at a distance of 1 AU from the Sun experiences an acceleration of
approximately 1.8 cms2 .
As shown in eq. (3), the reflectivity, R, of a sail material is a primary factor in
sail performance. A factor not specifically referenced in eq. (3) is the mechanical
stability of the film. This property is implied in the variable defining the area, A.
Area loss can occur through numerous mechanisms including rips, delaminating
STATUS OF SOLAR SAIL MATERIAL CHARACTERIZATION 235
Sail Reflectivity
1.5
Acceleration (cm/s^2)
0
1 0.9 0.8 0.7 0.6 0.5
Ref lectivity
Figure 1. Sail characteristic acceleration dependence on sail material reflectivity and sail area
Calculations were made assuming an initial sail area of 1 106 m2
The Pelletron test system provides the unique capability to irradiate material to
a simultaneous or sequential exposure to an emulated space environment, and
perform in-vacuum reflectance measurements.
This facility possesses the capability to irradiate materials to: high energy
electrons (1002.5 MeV), low energy electrons (1100 keV), protons (40700
keV) Vacuum Ultraviolet (VUV) (photon radiation over the wavelength range from
121 to 200 nm) and Near Ultraviolet (NUV) (photon radiation over the wavelength
range from 200 to 400 nm). The exposure coverage is nominally a 7.6 cm 7.6
cm (3 inch 3 inch) area. In-vacuum reflectance measurements can be obtained
on two 1-inch diameter coupons or on one 1-inch 2-inch rectangular sample.
The reflectance measurements are taken over the wavelength range from 250 to
2500 nm. The solar absorptance ( s ) is calculated from this spectral reflectance
data. The Pelletron system is shown in figure 2.
Protons Test
Chamber
VUV
High Energy
Electrons
Reflectance
Low Energy Measurement
Electrons
Figure 2. Photograph of the Pelletron combined space environmental effects exposure system
STATUS OF SOLAR SAIL MATERIAL CHARACTERIZATION 237
photodiodes located at each end of the flight tube. Sail materials, investigated for
this work, were approximately 15.2 cm 15.2 cm (6 inch 6 inch) and mounted
in a frame. Test coupons can be as large as 20.3 cm 25.4 cm (8 inch 10 inch).
The MSFC MLGG is shown in figure 4.
3. Experimental Procedure
The charged particle exposure activity focuses on the determination of the me-
chanical and thermo-optical property response of sail material to various types
of exposure conditions (e.g., electron, proton, combined electron and proton).
The objective of this work is to determine the radiation tolerance of the candi-
date sail materials for each type of charged particle radiation. Each candidate
STATUS OF SOLAR SAIL MATERIAL CHARACTERIZATION 239
10
10
Rads
9
10
8
10
0 0.5 1 1.5 2 2.5 3 3.5 4
Microns
Figure 5. Flat dose profile of 50 keV electrons and 700 keV protons in aluminized Mylar
T +R+A=1 (4)
R =1 A (5)
Thermal emittance is measured using the TEMP 2000 and the Laboratory
Portable Infrared Reflectometer (LPIR) instruments. The TEMP 2000 measures
thermal emittance over an integrated wavelength range from 2.5 to 20 m,
while the LPIR provide spectral data over the wavelength range from 2 to
20 m.
Penetration Area
Damage Area
Figure 6. Penetration site of HVI of a 1.0 mm diameter projectile with velocity of 7 kms1 . The
photograph shows the penetration area and the surrounding damage area induced by the impact. Sail
material was Al/Mylar
to as Beta factors. These Beta factors are defined in eqs. (6) and (7).
(Area of Penetration)
Beta Penetration = (6)
(Cross-Sectional Area of Projectile)
(Damage Area)
Beta Damage Area = (7)
(Cross-Sectional Area of Projectile)
4. Results
Preliminary results on the effects of charged particle exposure have yielded some
interesting, but not unexpected results. The mechanical properties, including, ul-
timate stress and ultimate strain, decrease with increasing radiation dose. Stress-
strain data was obtained for each exposure condition. Three to five tensile coupons
of each material were exposed to each exposure condition. The results, to date,
are shown in figures 7 and 8.
In addition to mechanical properties, thermo-optical properties are measured,
when practical. Figure 9 shows the effect of electron radiation on the reflectance
property of aluminized Mylar as a function of increasing dose. The solar absorp-
tance values, shown in table 1, indicate little change with radiation exposure.
The spectral data, shown in figure 9, shows a slight decrease in reflection with
increased radiation exposure. This reflection loss occurs in the low wavelength
region of the reflectance spectrum. Table 1 details the thermo-optical properties
obtained to date for materials exposed to long term Near Ultraviolet Radiation
and other materials exposed to high doses of electron radiation.
242 DAVID L. EDWARDS ET AL.
5
10
4
10
3
10
2 3 4 5
0.001 0.01 0.1 1 10 10 10 10 10
Dose (Mrad)
Figure 7. Relationship between ultimate strength of candidate sail materials and radiation dose
The data, shown in table 1, indicates the solar absorptance of the Aluminum
surface of the sail material has not significantly changed with either UV exposure
or electron exposure. The back surface (e.g., polymer surface) emittance has also
remained relatively stable, showing, at most, a 4% change after 1 Grad of electron
exposure.
0.1
Strain to Failure
0.001
2 3 4 5
0.001 0.01 0.1 1 10 10 10 10 10
Dose (Mrad)
Figure 8. Relationship between strain to failure of candidate sail materials and radiation dose
STATUS OF SOLAR SAIL MATERIAL CHARACTERIZATION 243
UV ESH Alpha Emmitance Reflectance Alpha Emmitance Reflectance Alpha Emmitance Reflectance
The Hypervelocity Impact (HVI) testing was designed to determine the ef-
fects of HVI over the functional lifetime of the sail material. Impact testing was
performed at dose levels of 100 Mrad and 1 Grad. Sail material specimens were
first exposed to electron radiation, then subjected to HVI. Beta factors for pen-
etration and damage areas for the candidate materials Al/ Teonex and Al/Mylar
after radiation exposure and HVI characterization are shown in figures 10 and 11.
These materials were exposed to radiation doses of 100 Mrad and 1 Grad and then
subjected to HVI with projectile of 3 kms1 and diameters of 0.4 and 1.0 mm.
The complete data set for tests completed to date is shown in table 2.
0.98
0.96 Pre-Exposure
100 Mrad
0.94 1 Grad
0.92
Ref lectance
0.9
0.88
0.86
0.84
0.82
200 300 400 500 600 700 800 900 1000
Wavelength (nm)
Figure 9. Reflectance of aluminized Mylar solar sail material as a function of electron radiation
dosage
244 DAVID L. EDWARDS ET AL.
1.6
1.5
1.4
1.1
0.9
2 3
0.001 0.01 0.1 1 10 10 10
Dose (Mrad)
Figure 10. Beta factor (penetration area) relationship with radiation dose in candidate sail
material
This data does not constitute a statistically significant data set. This data does
consistently indicate the damage area is no larger than a factor of 2 of the projectile
diameter. The tests also indicated that the impacts did not result in rip propagation.
The data indicates little change in hypervelocity impact response with increased
radiation exposure of these materials.
1.8
1.7
1.6
Beta Damage
1.5
1.2
1.1
2 3
0.001 0.01 0.1 1 10 10 10
Dose (Mrada)
Figure 11. Beta factor (damage area) relationship with radiation dose in candidate sail material
STATUS OF SOLAR SAIL MATERIAL CHARACTERIZATION 245
5. Future Plans
Upcoming testing on the candidate solar sail material will focus on mission-
specific applications. Test plans are being developed to expose the material to
emulated Solar Polar Imager (SPI) and L1 Diamond radiation environments. The
mechanical and thermo-optical properties will be evaluated over mission lifetime
dose exposures. Hypervelocity impact tests will be conducted on sail materials ex-
posed to mission middle and end-of-life radiation doses. The goal of this research
is to select the optimum material for these specialized missions.
References
3. Edwards, D., Hubbs, W., Gray, P. Wertz, G., Hoppe, D., Nehls, M., Semmel, C., Albarado, T., and
Hollerman, W. (2003) In Proceedings of the 9th International Symposium on Material in a Space
Environment, Noordwijk, The Netherlands, ESA Publications, Noordwijk, The Netherlands
June 2003, pp. 1620.
4. Albarado, T., Hollerman, W., Edwards, D., Hubbs, W., and Semmel, C. (2003) In Proceedings
of ISEC 2003: 2003 International Solar Energy Conference, Hawaii, 1518 March 2003.
5. Edwards, D., Hubbs, W., Stanaland, T., Hollerman, A., and Altstatt, R. (2002) In Proceedings
of SPIE Photonics for Space Environments VIII, Vol. 4823, 2002.
6. Tech ICP-Solar Sails: Potential Mission Applications. www. inspacepropulsion.com/tech/
sails missionapps.html
7. McInnes, C. R. (1999) Solar Sailing: Technology, Dynamics, and Mission Applications, Praxis
Publishing, Chichester, UK, pp. 150.
8. ASTM E-490-00a. (2000) Standard Solar Constant and Air Mass Zero solar Spectral Irradiance
Tables, 200.
9. Halbleib, J. A., Kensek, R. P., Mehlhorn, T. A., Valdez, G. D., Seltzer, S. M., Berger, M. J. ITS
Version 3.0: The Integrated TIGER Series of Coupled Electron/Photon Monte Carlo Transport
Codes, SAND91-1634, (March 1992).
10. Ziegler, J. F. (2000) The Stopping and Range of Ions in Matter, SRIM-2000.40, www.SRIM.org.
ATOMIC OXYGEN DURABILITY EVALUATION OF A UV CURABLE
CERAMER PROTECTIVE COATING
Abstract. The exposure of most silicones to atomic oxygen in low Earth orbit
(LEO) results in the oxidative loss of methyl groups with a gradual conversion
to oxides of silicon. Typically, there is surface shrinkage of the oxidized silicone
protective coatings which leads to cracking of the partially oxidized brittle sur-
face. Such cracks widen, branch and can propagate with continued atomic oxygen
exposure ultimately allowing atomic oxygen to reach any hydrocarbon polymers
under the silicone coating. A need exists for a paintable silicone coating that is
free from such surface cracking and can be effectively used for protection of
polymers and composites in LEO. A new type of silicone-based protective coat-
ing holding such potential was evaluated for atomic oxygen durability in an RF
atomic oxygen plasma exposure facility. The coating consisted of a UV curable
inorganic/organic hybrid coating, known as a ceramer, which was fabricated us-
ing a methyl substituted polysiloxane binder and nanophase silicon-oxo-clusters
derived from sol-gel precursors. The polysiloxane was functionalized with a cy-
cloaliphatic epoxide in order to be cured at ambient temperature via a cationic
UV induced curing mechanism. Alkoxy silane groups were also grafted onto the
polysiloxane chain, through hydrosilation, in order to form a network with the
incorporated silicon-oxo-clusters. The prepared polymer was characterized by 1 H
and 29 Si NMR, FT-IR, and electrospray ionization mass spectroscopy. The pa-
per will present the results of atomic oxygen protection ability of thin ceramer
coatings on Kapton H as evaluated over a range of atomic oxygen fluence levels.
247
J.I. Kleiman (ed.), Protection of Materials and Structures from Space Environment, 247263.
C 2006 Springer. Printed in the Netherlands.
248 BRUCE A. BANKS ET AL.
1. Introduction
Silicones are one of the few polymers that can be applied by painting or spray-
ing over composite or other organic spacecraft materials which have afforded
reasonable protection from low Earth orbital (LEO) atomic oxygen attack. The
gradual oxidation of the silicones in LEO results in an oxidized silicone surface
which becomes a silicate or silica [13]. This surface conversion from silicone to
silica also tends to produce tensile stresses in the surface of the oxidized silicone.
An increase in the surface microhardness also results due to the atomic oxygen con-
version reaction with the silicone [4]. A variety of approaches have been or are now
being explored to identify silicones, silicone copolymers or silicone-hydrocarbon
blends that could provide flexibility as well as atomic oxygen protection [58].
Results to date indicate that hybrid polymers composed of inorganic and organic
polymers hold potential to survive LEO atomic oxygen attack. The silicones which
are dominated by a oxygen-to-silicon ratio of 1.5 have shown greater resistance to
atomic oxygen attack than the silicones with a ratio of 1.0. Polyhedral oligomeric
silsesquioxane (POSS) contains covalently bonded reactive functionalities appro-
priate for polymerization or grafting. It can be blended or copolymerized with
many aerospace polymers and is being considered for atomic oxygen durability
[8]. The resistance to atomic oxygen attack of silicone blended or copolymerized
polymers has been dependent not only on the oxygen-to-silicone ratio but the
fractional fill of the silicone. The challenge to make functional use of such blends
has been to find an adequately silicone-filled polymer that contains the appropriate
protective silicone such that it has mission dependent properties that are accept-
able. Examples of some of these are atomic oxygen durability, volatility, optical,
thermal, mechanical and ease of application.
Because of their ability to provide atomic oxygen protection, thermal sta-
bility, flexibility, and stability; polysiloxanes are an attractive candidate solution
to achieving ideal protection from the elements of space. However, this is just
part of the solution. The vacuum ultraviolet (VUV) radiation and high energy
particles can still damage and degrade the composite material. Therefore, to in-
corporate protection from those components as well, ceramer coatings; which are
inorganic/organic hybrid materials, can be utilized.
Ceramers are part ceramic (inorganic) and part polymer (organic) and can
offer protection from atomic oxygen as well as UV radiation and high energy
particles via the in situ fabrication of nanophase silicon-oxo-clusters [9, 10]. The
silicon-oxo-clusters are formed through a series of hydrolysis and condensation
reactions between sol-gel precursors.
The intention of a ceramer approach is to acquire a synergistic effect between
the inorganic and organic phases on a nanoscale through the use of phase coupling
agents, which for this system are alkoxy silanes pendant from the polysiloxane
chain. There is confirmation of a synergy between the phases and this approach
affords a uniformly distributed nanophase within a continuous organic phase [11].
ATOMIC OXYGEN DURABILITY 249
O Atomic Oxygen
O OH O O
O O O O
O O O
OH High Energy Particle
O DUV
HO O O O OH
Metal Oxide
Si Si Si Si
Dissipated Heat
O O O O O
SiO2 Layer
Coating
Composite
Substrate
Figure 1. Depiction of the formation and function of protective silicon oxide layer and silicon-oxo-
clusters
Once the coating is exposed to atomic oxygen, a protective layer of silicon oxide
is formed and, with the incorporation of silicon-oxo-clusters, the coating should
protect the composite material against further atomic oxygen erosion, high energy
particles, and VUV radiation. Figure 1 is an overall diagram of the coatings
function [12].
This paper summarizes an investigation of the high fluence atomic oxygen
durability of ultraviolet radiation curable ceramer protective coating consisting
of methyl substituted polysiloxane and compares the results with the commonly
used silicone coating DC93-500.
2.1. SYNTHESIS
2.1.1. Experimental
Materials. Octamethylcyclotetrasiloxane,1,3,5,7-tetramethylcyclotetrasil-oxane,
1,1,3,3-tetramethyldisiloxane, and vinyl triethoxysilane were purchased from
Gelest, Inc. and used as supplied. Wilkinsons catalyst (chlorotris (triph-
enylphosphine) rhodium(I), 99.99%), Amberlyst 15 ion-exchange resin, and
4-vinyl-1-cyclohexene 1,2-epoxide were purchased from Aldrich and used as
supplied. Toluene, supplied by Aldrich Chemical Co., was distilled in order to
eliminate any impurities. Irgacure 250 was supplied by Ciba Specialty Chemicals
and used as received. Air sensitive materials were transferred and weighed in an
inert atmosphere dry box under argon.
O Rh Catalyst
N2 Blanket
H3C == CHCH3Si(OEt)3
toluene (30 g) was added. The reaction was held at 75 C by means of an oil
bath and mechanically stirred under a nitrogen blanket. The disappearance of the
Si-H functionality was monitored by FT-IR and using the disappearance of the
peak at 2160 cm1 as an indication that the reaction was complete. Any solvent
and unreacted starting materials were removed under vacuum (35 mm Hg). Cy-
cloaliphatic epoxide and alkoxy silane functionalization was confirmed/analyzed
through 1 HNMR, FT-IR analysis, and titration.
2.2. STRUCTURE
Once cured the coating should form a strong interlocking network consisting
of a cross linked polysiloxane phase with interconnected silicon-oxo-clusters
(figure 3). The silicon-oxo-clusters will be connected to the polysiloxane backbone
though hydrolysis and condensation reactions with the tethered alkoxy silane.
The polysiloxane was diluted with dry toluene (25% w/w) in order to reduce
the viscosity. Sol-gel precursor (5% w/w) and photo initiator (3% w/w) was also
252 BRUCE A. BANKS ET AL.
3.1.1. Instruments
Viscosity measurements were taken on an AR 500 Rheometer (Thermal Analysis)
equipped with a cup and bob sample holder and operated at 21.1 C. Pencil hard-
ness tests were conducted according to ASTM method D3363-00. Taber scratch
test was performed using a Taber Shear/Scratch Tester model 502 and conform-
ing to ASTM method G171-03. Taber Abrasion tests were conducted on a Taber
Industries 5130 Abraser using a CS-10 test wheel. Taber abrasion studies corre-
sponded to ASTM method D552-93a. Thermogravimetric analysis was performed
on a TGA Q 500 (Thermal Analysis).
X-ray photoelectron spectroscopy was completed on a Kratos Model ES3000
with a non-monochromatic 120 W Al K-Alpha radiation source under 108 Torr.
Scanning electron microscopy was performed on a Hitachi S-2150 operating at
15 kV. Atomic force microscopy was performed on a multimode scanning probe
microscope (Digital Instruments) using the tapping mode.
1 cm
Figure 4. Sample holder to prevent curling of 2.54 cm diameter samples when they were exposed
to atomic oxygen
The abrasion and scratch resistance of the cured coating were studied to determine
how susceptible it is to physical damage. The Taber abrasion and scratch tests
yielded a value of 183 wear cycles per mil and a scratch value of 50 g. Both of
these values are low and show that the coating has poor abrasion resistance. These
values were cross-checked using the pencil hardness test, which gave a value of
2B/B. This value is also low and this trend could be a result of the very low glass
transition temperature of the coating, which is approximately 130 C. Such a
low glass transition temperature makes the coating soft and vulnerable to damage.
Varying the pendant group to raise the glass transition of the coating could be a
potential answer in improving the abrasion resistance.
Thermal gravimetric analysis was performed in order to observe the thermal
stability of the cross-linked polysiloxanes. Irreversible changes to the cross-linked
structure of silicone polymers unavoidably occur at high temperatures due to chain
scission or oxidative cross-linking [14]. In an inert atmosphere, depolymerization
occurs with the loss of volatile products, mostly low molecular weight cyclic
oligomers; but is often catalyzed by traces of acids, bases, water, or residual catalyst
used in the polymers original production [15]. Typically, depolymerization occurs
near 400 C for reasonably pure polydimethylsiloxane [16].
Thermal gravimetric analysis of the cured coatings illustrates the loss of small
molecular weight oligomers in the early stages of the analysis (figure 5). As
100
80
Weight (%)
60
40
20
0
0 200 400 600 800 1000
Temperature (C)
Figure 6. AFM images of crosslinked methyl polysiloxane substituted with 5% sol-gel precursor
expected, the depolymerization occurred near 400 C for the sample tested. The
range of molecular weights give way to the multiple slopes the curve exhibits.
It is also important to note that the sample generated a small amount of residue
(13%), which can be attributed to the silicon-oxo-clusters formed during the
polymerization process and high molecular weight chains that may not have com-
pletely volatized/degraded.
One of the most important aspects of the coating is the presence of the silicon-
oxo-clusters. By utilizing the AFMs tapping mode, it will be possible to detect
hard (silicon-oxo-clusters) and soft (polymer) regions within the cross-linked
polymer network. These clusters provide additional protection against high-energy
particles and deep UV light. Figure 6 is an AFM image of a sample with 5% (w/w)
sol-gel precursor added prior to casting. The silicon-oxo-clusters are clearly visi-
ble in the subjected sample. The average size for the methyl substituted polysilox-
ane nano phase is 125 nm. Figure 6 shows a more disperse and uniformly sized
nanophase, which could be attributed to the small size of the pendant methyl group
allowing more freedom to the growing nano clusters.
Conductivity tests were performed on the cross-linked coatings. The coatings
showed no signs of electrical conductivity, which is expected due to their insulating
nature [14].
Photographs of the silicone coated Kapton H samples and silicone coated fused
silica samples after two different levels of atomic oxygen exposure are shown in
figure 7.
256 BRUCE A. BANKS ET AL.
(a)
(b)
(c)
Figure 7. Photographs of silicone coated Kapton H and silicone coated fused silica after atomic
oxygen exposure to moderate and high fluence levels (a) Ceramer coated samples at a Kapton
effective fluence of 2.22 1021 atoms/cm2 ; (b) DC93-500 coated samples at a Kapton effective
fluence of 2.22 1021 atoms/cm2 ; (c) Ceramer coated samples at a Kapton effective fluence of
1.38 1022 atoms/cm2 ; (d) DC93-500 coated samples at a Kapton effective fluence of 1.38 1022
atoms/cm2
ATOMIC OXYGEN DURABILITY 257
(d)
Figure 7. (Continued)
As can be seen in figure 7, atomic oxygen exposure of the ceramer and DC93-
500 provide excellent protection for moderate (2.22 1021 atomscm2 ) fluence
levels. The ceramer coating appears to be a significant improvement at moderate
fluence levels in that there is no sign of microcracking as occurs for DC93-500.
However, at high fluences (1.38 1022 atomscm2 ), both the ceramer and DC93-
500 develop microcracks. Unlike the DC93-500, the ceramer tends to detach from
its substrate causing greater coating shrinkage due to atomic oxygen attack on both
surfaces of the coating. This coating shrinkage and opening of the cracks allows
atomic oxygen to attack the underlying Kapton. Thus, the ceramer coating is better
for moderate fluences and could potentially be used for coating optical polymers
such as Fresnel lenses for concentrators over solar cells. In such applications,
protective coatings that form microcracks would not be suitable due to loss in
specular transmittance.
The mass loss of coated Kapton as a function of atomic oxygen fluence for
both the ceramer and DC93-500 coatings is shown in figure 8.
As can be seen from figure 8, the ceramer coating as well as the DC93-
500 coating does provide significant atomic oxygen protection for most fluences.
However, at fluences above 1 1022 atomscm2 , the ceramer coating develops
apertures in it due to microcracking and the rate of oxidation of the underlying
Kapton greatly increases. There did not seem to be significant differences in atomic
oxygen protection resulting from the laboratory dust abrasion of the coatings. This
is thought to be due to the very shallow scratches resulting without any going all
the way through the coatings.
Atomic oxygen exposure of the ceramer coatings causes an increase in optical
absorptance, and therefore a reduction in transmittance for wavelengths <800 nm
with little change in reflectance, as shown in figure 9.
258 BRUCE A. BANKS ET AL.
100
90
80
Mass lost (mg) 70
60
50
40
30
20
10
0
0 5E+21 1E+22 1.5E+22
Ceramer - scratched
2 Ceramer - unscratched
Fluence (atoms/cm ) DC - scratched
DC - unscratched
untreated kapton
Figure 8. Mass loss of coated Kapton as a function of atomic oxygen fluence for both the ceramer
and DC93-500 coatings
Similar results were observed for DC93-500 and the ceramer as shown in
figure 10. However, if one considers applications such as protective coatings for
Fresnel concentrator solar cell arrays there is an important distinction between the
two coatings for moderate fluences.
The specular transmittance degradation caused by atomic oxygen exposure is
in part due to increased absorption for wavelengths <800 nm as well as microc-
racking, as shown in figure 11.
As can be seen by examining figures 11a and 7 (a and b), the absence (or
presence) of microcracking at a fluence level of 2.22 1021 atomscm2 makes
a significant impact on specularly transmitted light. The ceramer coating does
allow much greater specular transmittance than the DC93-500 for fluences up to
2.22 1021 atomscm2 .
X-ray photoelectron spectroscopy was performed in order to confirm the pres-
ence of a protective oxide layer (figure 12). It is important to note that sputtering
was not performed during the analysis. This ensures that only the surface of the
samples was analyzed. The initial XPS spectrum shows high amounts of both sili-
con and oxygen, which is expected since these elements are present in the polymer
backbone. After atomic oxygen exposure, the oxygen peak increased while the
silicon peaks decreased. This is anticipated due to the protective oxide layer pos-
sessing a high amount of oxygen compared to silicon. The oxide layer should be
composed of silicon atoms whose valences are filled by oxygen atoms, yielding
a Si-O4 network. The presence of carbon after exposure to atomic oxygen is due
to impurities on the surface of the film such as dust, dirt, etc. and is, therefore,
always present [12].
Cross contamination tests performed separately on the ceramer and DC93-
500 coating using fused silica witness slides adjoining the silicone samples
ATOMIC OXYGEN DURABILITY 259
110
Absorptance (%) 90
70
50
30
10
10
0 500 1000 1500 2000 2500
Wavelength (nm)
original
2.22E=21 atoms/cm2
1.38E+22 atoms/cm2
a. Total absorptance
100
Transmittance (%)
80
60
40
20
0
0 500 1000 1500 2000 2500
Wavelength (nm)
original
2.22E+21 atoms/cm2
1.38E+22 atoms/cm2
b. Total transmittance
50
40
Reflectance (%)
30
20
10
10
0 500 1000 1500 2000 2500
Wavelength (nm) original
2.22E+21 atoms/cm2
1.38E+22 atoms/cm2
c. Total reflectance
Figure 9. Effects of atomic oxygen on total optical properties for ceramer coated fused silica
260 BRUCE A. BANKS ET AL.
100
Absorptance (%) 80
60
40
20
0
20
0 500 1000 1500 2000 2500
Wavelength (nm) original
2.22E+21 atoms/cm2
1.38E+22 atoms/cm2
a. Total absorptance
100
90
80
Transmittance (%)
70
60
50
40
30
20
10
0
0 500 1000 1500 2000 2500
Wavelength (nm) original
2.22E+21 atoms/cm2
1.38E+22 atoms/cm2
b. Total transmittance
50
45
40
Reflectance (%)
35
30
25
20
15
10
5
0
0 500 1000 1500 2000 2500
Wavelength (nm) original
2.22E+21 atoms/cm2
1.38E+22 atoms/cm2
c. Total reflectance
Figure 10. Effects of atomic oxygen on total optical properties for DC93-500 coated fused silica
100
Transmittance (%)
80
60
40
20
0
20
0 500 1000 1500 2000 2500 3000
Wavelength (nm) original
2.22 E+21 atoms/cm2
1.38 E+22 atoms/cm2
a. Ceramer
120
Transmittance (%)
100
80
60
40
20
0
0 500 1000 1500 2000 2500
Wavelength (nm) original
2.22 E+21 atoms/cm2
1.38 E+22 atoms/cm2
b. DC93-500
Figure 11. Effects of atomic oxygen on specular transmittance of ceramer and DC93-500 coated
fused silica
9000
Si C O
8000
7000
6000
Intensity (arb. units)
5000
4000
3000
2000
1000
0
0 100 200 300 400 500 600 700
Energy (eV)
Before AO Exposure After AO Exposure
Figure 12. XPS spectrum of the cross-linked methyl substituted polysiloxane before and after atomic
oxygen exposure
262 BRUCE A. BANKS ET AL.
indicated that there was transport of short chain silicones to the fused silica that
resulted in a silica deposit 857 Angstroms for a Kapton effective fluence of 1
1022 atomscm2 for the ceramer, but no measurable contamination from the
DC93-500. This is probably due to the lack of vacuum stripping of the ceramer
whereas the DC93-500 is vacuum stripped. A no sample test of the facility did,
in fact, result in no deposit of oxidized silicone.
5. Conclusion
Atomic oxygen exposure of ceramer and DC93-500 silicone coated Kapton H and
fused silica slides indicates that the ceramer coating has superiority over DC93-
500 coatings for moderate (up to 2.22 1021 atomscm2 ) Kapton effective atomic
oxygen fluences. The ceramer coatings at this fluence resulted in low mass loss of
coated Kapton samples and did not show evidence of the extended microcracking
that occurred for DC93-500. This results in superior specular light transmittance
for the ceramer coatings. This may allow its use as an atomic oxygen protective
coating over silicone Fresnel concentrators for solar arrays. At high fluence levels
(1.38 1022 atomscm2 ), the ceramer coating develops microcracks that result
in detachment of the coating causing exposure of the underlying Kapton which
does not occur for DC93-500. Thus at high fluences DC93-500 would be a better
choice of protective coating.
The ceramer coating produced cross contamination of silica on witness slides
which suggests that the ceramer should be vacuum stripped to prevent the transport
of short chain silicones that can result in contamination.
Acknowledgments
The authors gratefully acknowledge Justin Tokash and Dr. Rex Ramsier of the
University of Akrons Physics Department for their XPS contributions.
References
1. Banks, B. A., Dever J. A., Gebauer L., and Hill, C. M. (1991) Paper presented at the 1st LDEF
Post-Retrieval Symposium, Kissimmee, Florida, 28 June 1991.
2. Banks, B. A., Rutledge, S. K., de Groh, K. K., Mirtich, M. J., Gebauer, L., Olle, R., and Hill,
C. M. (1991) Paper presented at the 5th International Symposium on Materials in a Space
Environment, Cannes-Mandelieu, France, 1620 September 1991.
3. Banks, B., Rutledge, S., Sechkar, E., Stueber, T., Snyder, A., Hatas, C., and Brinker, D. (2000)
In Proceedings of the 8th International Symposium on Materials in a Space Environment and
the 5th International Conference on Protection of Materials and Structures from the LEO Space
ATOMIC OXYGEN DURABILITY 263
Environment cosponsored by the CNES, Integrity Testing Laboratory, ESA, ONERA and the
Canadian Space Agency, Arcachon, France, 49 June 2000.
4. de Groh, K., Banks, B., and Ma, D. (2004) In 7th International Conference on Protection of
Materials and Structures form Space Environment, Toronto, Canada, 1013 May 2004.
5. Rutledge, S., Cooper, J., and Olle, R. (1990) In Proceedings of the Space Operations Applications
and Research Symposium, NASA CP-3103, Albuquerque, NM, 2628 June 1990.
6. Hung, C. (1994) Reaction and Protection of Wire Insulators in Atomic Oxygen Environment,
NASA TM106767, 1994.
7. Zhang, C., Babonneau, F., Bonhomme, C., Laine, R. M., Soles, C. L., Hristov, H. A., and Yee,
A. F. (1998) Journal of American Chemical Society 120, 83808391.
8. Brunsvold, A. L., Minton, T. K., Gouzman, I., Grossman, E., and Gonzalez, R. I. (2004) Journal
of High Performance Polymers, V. 16, No. 2, pp. 303318.
9. Soucek, M. D. and Tuman, S. J. (1996) Journal of Coatings Technology 68(854), 73.
10. Tuman, S. J., Chamberlain, D., Scholsky, K. M., and Soucek, M. D. (1996) Progress in Organic
Coatings 28, 251.
11. Wold, C. R. and Soucek, M. D. (1997) Journal of Coatings Technology 70, 43.
12. Dworak, D. P. and Soucek, M. D. (2003) Progress in Organic Coatings 47, 448.
13. ASTM E 2089-00, Standard Practices for Ground Laboratory Atomic Oxygen Interaction
Evaluation of Materials for Space Applications, June 2000, American Society for Testing and
Materials, Philadelphia, USA.
14. Smith, A. L. (1991) The Analytical Chemistry of Silicones, Wiley, New York.
15. Kang D. W., Rajendran, G. P., and Zeldin, M. (1986) Journal of Polymer Science, Part A:
Polymer Chemistry 24, 1085.
16. Grassie, N. and Macfarlane, I. G. (1978) European Polymer Journal 14, 875.
CERMET THERMAL CONVERSION COATINGS FOR SPACE
APPLICATIONS
Abstract. Cermet films of gold and aluminum oxide were sputter deposited onto
silicon substrates. These films consisted of nanometer-sized gold grains embed-
ded in a matrix of aluminum oxide, and were designed to maximally absorb visible
radiation and minimally emit infrared radiation. Ex situ and in-situ spectroscopic
ellipsometry were used to characterize film optical constants from vacuum ul-
traviolet to middle infrared. The optical constants, for cermet films with gold
volume fractions ranging from 0 to 1, were used in predictive optical models.
Thermal performance was optimized using a new algorithm based on energy bal-
ances. This algorithm weights the parameter fit in predictive optical models to
an ideal reflectance spectrum. This weighting factor emphasizes fitting in more
significant spectral ranges and deemphasizes fitting in less significant wavelength
ranges. Absorptivity, emissivity, and their ratio were calculated for optimized film
structures.
1. Introduction
Cermet films contain metallic and ceramic materials. Because they can be tailored
to have a wide range of optical, electrical, mechanical, and chemical properties,
cermet films have many potential applications [1, 2]. In the present work, the
focus was using spectroscopic ellipsometry to engineer optical coating thickness
and concentration for optimal photothermal energy conversion [3]. These cermet
films absorb energy in the visible range and reflect infrared (IR) radiation. The
desired films should have maximum absorptivity in the visible spectral range and
minimum emissivity in the infrared. One application for these films is to provide
energy for a Stirling engine by absorbing solar energy in the space environment
[4]. A Stirling engine converts heat into mechanical energy. Another application
is photothermal energy conversion for terrestrial heating applications [2].
265
J.I. Kleiman (ed.), Protection of Materials and Structures from Space Environment, 265276.
C 2006 Springer. Printed in the Netherlands.
266 BENJAMIN W. WOODS ET AL.
3.0 1.0
2.5
Irradiance (W/m /nm) 0.8
2.0
0.6
Reflectance
2
1.5
0.4
1.0
AM0 Solar Irradiance
450C Blackbody Irradiance
Ideal Reflectance 0.2
0.5
0.0 0.0
0 1000 2000 3000 4000 5000
Wavelength (nm)
Figure 1. AM0 Solar irradiance spectrum, 450 C blackbody spectrum, and ideal photothermal film
reflectance spectrum
2. Experimental
All films were sputter deposited on flat, polished silicon wafer substrates. Depo-
sition occurred in a Kurt J. Lesker Co. vacuum chamber with base and deposition
pressures being 7 105 and 4 101 Pa, respectively. The chamber had a
rotating platten, positioned by a stepper motor under computer control, as shown
in figure 3. Cermet films were deposited via a cyclic process of: positioning the
substrate above one of two sputter guns, dwelling for a chosen time period, po-
sitioning the substrate above the other sputter gun, and dwelling for a different
Alumina
Alumina Cermet B
time period. The sputter guns were fitted with conical metal chimneys to prevent
overspray. In other words, the chimneys prevented gold from being deposited on
a substrate positioned above the aluminum target and vice versa. These chimneys
also served as inlets for argon and oxygen. Gold was DC sputtered with a con-
trolled power of 15W. A low wattage was chosen to minimize the gold deposition
rate. The idea was to have a low deposition rate and short deposition times to
minimize grain size.
Initially, alumina was reactively sputtered using RF power. This was later re-
placed by pulsed-DC power. For RF alumina deposition, an RFX 600 power supply
was used with a controlled bias of 700V which resulted in 250300W power out-
put. For DC alumina deposition, an MDX 1K power supply was put in series
with a Sparc-le V pulse generator. DC power was set at 150W. The DC pulse
generator was purchased to add stability and reproducibility compared to the
RF deposition process. During deposition, oxygen gas was introduced through
the chimney above an aluminum target at rates between 0.8 and 1 sccm. Be-
fore cermet deposition, each silicon wafer substrate was coated with 20 nm of
chromium (to promote adhesion) followed by an optically thick (120 nm) gold
layer.
A set of nine cermet samples were made with gold compositions of 10%, 20%,
30%, . . . 90% (by volume) along with samples of pure gold and pure alumina.
Composition estimation was based on the deposition rates of alumina and gold.
Composition was controlled by the amount of time a substrate spent above a
sputter gun. Since each cermet layer was composed of many sublayers of gold and
alumina, an additional constraint was put on the deposition process. Each sublayer
of gold was deposited with a thickness of 1 nm. This constraint was chosen so that
gold grains would be small and roughly spherical. TEM and AFM micrographs
of two cermet films are shown in figure 4. We believe that the dark regions in
the TEM micrograph are gold which would give a gold grain size of 5 nm and
smaller.
CERMET THERMAL CONVERSION COATINGS 269
Figure 4. Left) AFM micrograph of 45% gold 55% alumina cermet film (500 500 nm) Right)
TEM micrograph of 30% gold 70% alumina cermet film (20 nm scalebar)
One goal of the present work was to find the best means to compare performance of
films used for photothermal energy conversion. Currently, absorptivity, emissivity,
and their ratio are the most commonly used figures of merit. Absorptivity and
emissivity are defined by
A() [1 R(, )] d
= 0 (4)
0 A() d
270 BENJAMIN W. WOODS ET AL.
and
B(, T ) [1 R(, )] d
= 0
(5)
0 B(, T ) d
respectively, where A() is the solar irradiance spectrum, B(, T ) is the blackbody
spectrum, and R(, ) is spectral reflectance. Angles of incidence are near normal
for all calculations. In practice, the range of wavelengths for integration is set by
the intensity of the blackbody and solar irradiance spectra. Thus, a range is chosen
which includes all wavelengths for which the blackbody and solar spectra have
significant magnitude. In this work, the range of integration is 0.230 microns.
Absorptivity and emissivity determine how efficiently a film absorbs and emits
radiation. Since high solar absorption and low blackbody emission are desired it is
logical to create a single figure of merit /. Unfortunately, / is more sensitive
to changes in than . It is also difficult to use the two parameters separately
because one must choose if high solar absorption or low infrared emission is
more important. When there is significant overlap between the solar irradiance
and blackbody spectra, there is an additional complication. The ideal spectral
reflectance is zero for wavelengths where the solar irradiance is more intense
and 100% for wavelengths where the blackbody spectrum is more intense. This
implies that an ideal film with this spectral reflectance could not have 100%
absorptivity or 0% emissivity. To visualize this, notice the area beneath the AM0
curve and at wavelengths higher than 1700 nm in figure 1. None of this area could
contribute to the numerator in the calculation of in eq. (4). In fact, for the AM0
solar spectrum and 450 C blackbody the absorptivity calculated from the ideal
reflectance spectrum is 90%.
We define a new figure of merit that incorporates the solar and blackbody spec-
tra and spectral reflectance. The parameter, photothermal conversion efficiency,
PCE, is defined by
[X A() B(, T )] [1 Rreal (, )] d
PCE = 0 (6)
0
[X A() B(, T )] [1 Rideal (, )] d
where X is the solar concentration factor, A() is the solar irradiance spectrum,
B(, T ) is the blackbody spectrum, Rreal (, ) is the measured spectral reflectance,
and Rideal (, ) is the ideal spectral reflectance. Ideal spectral reflectance, for
absorbers, is zero for wavelengths where solar irradiance is greater than blackbody
and 100% for the opposite case. The solar concentration factor accounts for the
use of lenses or mirrors to enhance solar irradiance on the film. In this work,
the solar concentration factor is assumed to be unity unless otherwise specified.
PCE is a ratio of the net power absorbed divided by the net power available. It
incorporates both absorption and emission. It has meaningful values ranging from
0 to 1.
CERMET THERMAL CONVERSION COATINGS 271
2.2
2.0 14
100% gold
1.8 80% gold
12
1.6 60% gold
1.4
10 40% gold
20% gold
1.2 8
100% alumina
1.0
20% gold 6
0.8
40% gold
0.6 60% gold 4
0.4 80% gold
100% gold 2
0.2
0.0 0
200 400 600 800 1000 1200 1400 1600 1800 200 200 400 600 800 1000 1200 1400 1600 1800 200
Wavelength (nm) Wavelength (nm)
Figure 5. Optical constants for cermet films. The extinction coefficient of pure alumina is not shown
because it is effectively zero over this range
It is possible to calculate a negative value for PCE, which simply means that the
film is a better emitter than absorber. In the event, that PCE is negative, inverting the
ideal spectral reflectance so that 100% becomes 0 and vice versa, then recalculating
PCE will yield an emission efficiency instead of an absorption efficiency.
Ellipsometric data (in-situ and ex-situ) for all 11 samples were analyzed. In-
situ ellipsometric data were optically modeled to yield preliminary thicknesses and
optical constants for each sample. These thicknesses and optical constants were
used to set up optical models for the ex-situ data. In the end, oscillator models
were used to characterize optical constants for each sample for the spectral range
between (0.2 and 30 microns). These optical constants, shown in figure 5, were
subsequently used to predict the optimum thicknesses for cermet films, given a
solar irradiance spectrum and a blackbody temperature.
To find the optimal structures (with respect to PCE) in the modeling software,
a weighting factor was created to emphasize fitting in parts of the spectrum where
solar irradiance or blackbody emission is more intense. This weighting factor is
calculated at each wavelength with the equation
1
i = (7)
(X Ai Bi )i
where X is the solar concentration factor (unity unless specified), Ai and Bi are the
solar irradiance and blackbody spectra, respectively. i accounts for the variable
spacing of wavelengths of the measured solar irradiance spectrum. This makes
a function of the areas beneath A and B, and not the absolute values of A and B
for a given wavelength. The optical analysis software seeks to minimize a mean
squared error, MSE, defined by
N ideal 2
1 Ri Rimod
MSE = (8)
2N M i=1 i
272 BENJAMIN W. WOODS ET AL.
TABLE 1. Calculated figures of merit as a function of solar and blackbody spectra. All simulated
films were optimized with respect to PCE. The simulated reflectance was derived from the optical
constants and layer thicknesses using Fresnel equations
where R mod is the reflectance predicted by the optical model, R ideal is the re-
flectance value from the ideal reflectance file, N is the total number of wavelengths,
M is the number of variables in the model (a small number compared to N ), and
is a weighting factor. Combining eqs. (7 and 8), the MSE calculation becomes
1 N
ideal
2
MSE = Ri Rimod [X Ai Bi ] i (9)
2N M i=1
Data points which correspond to wavelengths where there is a large difference
between the solar irradiance and blackbody spectra are weighted most heavily
because they make the largest contribution to the number being minimized, MSE.
Ideally, analysis software would optimize by maximizing PCE directly. However,
the optimization described above was easy to use with existing software. The
reflectance spectra, generated from the fit optical model, yielded the largest PCE
for a given set of conditions (solar and blackbody irradiance). This means that
attempts were made to modify variables from the values found by the software, but
none of these modifications made the film structure perform any better with respect
to PCE. Surprisingly, the optimal structure (with respect to PCE) for the AM0 solar
spectrum and 450 C blackbody was a single cermet layer with 20% gold (77 nm
thick) atop an optically thick gold layer, as shown in table 1. One might expect
that the optimal structure would have an antireflection layer, but this is not the
case. The optimal structure for the AM1.5 solar spectrum at the same blackbody
temperature has the same structure, but the thickness of the cermet layer was 82 nm,
see table 1. Note that models with multiple layers including antireflective coatings
were tested. If additional protection from the space environment is desired, a 20
nm layer of alumina can be deposited on top of the cermet without significantly
changing PCE.
Optimizations were performed giving PCE for gold films with a single cermet
layer on gold, and secondly for a cermet layer with an alumina layer on gold for
two blackbody temperatures with the AM0 solar spectrum. These simulations
were based on the optical constants obtained from deposited samples. Results of
these simulations are summarized in figure 6. Notice that the optimal structures
at both temperatures have a 30% gold cermet with an alumina over-layer. Since
CERMET THERMAL CONVERSION COATINGS 273
1.0
0.8
0.7
0.6
0.5
0.4
the optimized structures described in the previous paragraph were optimized for
a higher temperature the alumina over-layer did not enhance performance. This is
because the alumina layer enhances the absorption of light for wavelengths before
the reflectance transition (2001700 nm), but it also may increase blackbody
emission for higher wavelengths. For high temperatures, the software puts more
emphasis on maximizing infrared reflectance than minimizing visible reflectance
because the area beneath the blackbody curve is larger than that of the solar
irradiance curve.
A set of simulations was done to show how a photothermal film structure would
perform if the actual operating temperature was not the same as the blackbody tem-
perature assumed for the design of the film structure. First, three multilayer optical
models were constructed. Each model was optimized (with respect to PCE) for
the AM0 solar spectrum and a different temperature. From each optimized optical
model, a reflectance spectrum was generated. PCE was calculated for the optical
models over a range of temperatures using the generated reflectance spectra. The
results of these simulations are shown in figure 7. Notice how the conversion
efficiency of a given structure asymptotically approaches a maximum value as
operating temperature decreases. This means that there is a temperature range for
which the efficiency doesnt change significantly. Figure 7 also indicates that the
film structures designed for lower operating temperatures have higher maximum
efficiency values than structures designed for higher operating temperatures.
274 BENJAMIN W. WOODS ET AL.
1.0
Optimized for 50C
Photothermal Conversion Efficiency, PCE
Optimized for 250C
Optimized for 450C
0.8
0.6
0.4
0.2
0.0
0 100 200 300 400 500 600 700
Two samples were deposited which consisted of single cermet layers on gold.
The first sample had a cermet layer with 20% gold and about 80 nm thick. This
sample was deposited because the software indicated that it was the best structure
for AM0 and 450 C blackbody spectra. The second sample had a cermet layer
with 30% gold about 100 nm thick. It was not optimized for any solar or blackbody
spectrum. Figures of merit were calculated for the reflectance generated by optical
models and measured reflectance.
These values are shown in table 2. All reflectance measurements were taken
at near normal incidence and at room temperature. Temperature dependence was
introduced using Plancks equation for a blackbody. AM0 is the relevant solar
irradiance for all calculations. Notice that model values and measured values are
close, if not exact, for PCE. The differences in , , and their ratio for modeled and
measured values are understandable when one considers that none of the models
were optimized for or .
4. Conclusion
In this work, cermet films made from gold and alumina were deposited and opti-
cally characterized with spectroscopic ellipsometry over the spectral range from
0.2 to 30 microns. Once the optical constants were determined, optical models
CERMET THERMAL CONVERSION COATINGS 275
Sample PCE /
were set up to simulate and predict performance of cermet film structures. A new
figure of merit, PCE, was used to quantify the net power absorbed with respect
to the net amount of power available for absorption. A weighting factor was in-
troduced so that optical analysis software would optimize film structures based
on PCE. Simulations were performed to show how various cermet film structures
would perform given blackbody temperatures and solar irradiance spectra.
PCE decreases as blackbody temperature increases. Thus a new optimization
problem is defined: A high blackbody temperature is desired to enhance heat
transfer from a photothermal film to the Stirling engine is efficient as well as
increase the Carnot efficiency of the Stirling engine. A low blackbody temperature
is desired because photothermal conversion efficiency and net available power (the
denominator in eq. (6) are higher. Thus, total energy conversion efficiency must
be determined for each specific application.
The predictive capability demonstrated in this paper should be helpful in this
process.
Acknowledgments
This research was generously supported by NASA Glenn Research Center, grant
number NAG3-2219.
References
1. Kreibig, U. and Vollmer, M. (1995) Optical Properties of Metal Clusters, Springer, Berlin.
2. Wackelgard, E., Niklasson, G. A., and Granqvist, C. G. (2001) Jeffrey Gordon (ed.) Published by
James & James (Science Publishers) Ltd, London. Solar Energy: The State of the Art, London.
3. Adsten, M., Joerger, R., Jarrendahl, K., and Wackelgrrd, E. (2000) Solar Energy 68, 325.
4. Jaworske, D. A. and Shumway, D. A. (2003) Space Technology & Applications International
Forum, STAIF-2003, Albuquerque, NM, February 2003, pp. 6570.
276 BENJAMIN W. WOODS ET AL.
5. Marco, J., Bhojaraj, H., and Hulyal, R. (2003) In Proceedings from the 9th International Sym-
posium on Materials in a Space Environment, Noordwijk, The Netherlands, 1620 June 2003,
p. 359.
6. Duvignacq, C., Hespel, L., Roze, C., and Girasole, T. (2003) In Proceedings from the 9th
International Symposium on Materials in a Space Environment, Noordwijk, The Netherlands,
1620 June 2003, p. 399.
7. Itawa, M., Ohnishi, A., Saito, R., and Ohkubo, K. (1997) In Proceedings of the 7th International
Symposium on Materials in a Space Environment, Toulouse, France, 1620 June 1997, p. 59.
8. Marco, J., Paillous, A., and Gourmelon, G. (1994) In Proceedings of the 6th International
Symposium of Materials in a Space Environment, ESTEC, Noordwijk, The Netherlands, 1923
September 1994, p. 77.
9. Tenditnyi, V. A., Smirnov-Vasilev, K. G., Yevkin, I. V., and Mironovich, V. V. (1994) In Pro-
ceedings of the 6th International Symposium of Materials in a Space Environment, ESTEC,
Noordwijk, The Netherlands, 1923 September 1994, p. 113.
10. Hidden, G., Vega, J. M., Remaury, S., and Nabarra, P. (2003) In Proceedings from the 9th
International Symposium on Materials in a Space Environment, Noordwijk, The Netherlands,
1620 June 2003, p. 49.
11. Triolo, J., Ackerman, N., Neuberger, D., and Harris, G. (1997) In Proceedings of the 7th Inter-
national Symposium on Materials in a Space Environment, Toulouse, France, 1620 June 1997,
p. 411.
12. Guillamon, J. C. (1997) In Proceedings of the 7th International Symposium on Materials in a
Space Environment, Toulouse, France, 1620 June 1997, p. 427.
13. Granqvist, C. G. (1989) Spectrally Selective Surfaces for Heating and Cooling Applications
SPIE, Bellingham, 1989.
14. Halliday, D. and Resnick, R. (1981) Fundamentals of Physics, 2nd ed., Wiley, New York,
p. 777.
15. Wehrli, C. Extraterrestrial Solar Spectrum, Publication no. 615, Physikalisch-
Meteorologisches Observatorium + World Radiation Center (PMO/WRC) Davos Dorf,
Switzerland, July 1985. Also available at time of publishing from http://rredc.nrel.gov/solar/
spectra/am0/wehrli1985.new.html.
16. Neckel, H. and Labs, D. (1984) Solar Physics 90, 205.
17. Ritchie, I. T. and Window, B. (1977) Applied Optics 16, 1439.
18. Trotter, D. M. and Sievers, A. J. (1980) Applied Optics 19, 711.
19. Zhang, Q. C. and Mills, D. R. (1992) Journal of Applied Physics 72, 3013.
20. Zhang, Q. C. (2000) Solar Energy Matererials in Solar Cells 62, 63.
21. Harding, G. L. and Craig, S. (1981) Solar Energy Materials 5, 149.
22. Farooq, M., Green, A. A., and Hutchins, M. G. (1998) Solar Energy Materials in Solar Cells
54, 67.
MULTIFUNCTION SMART COATINGS FOR SPACE APPLICATIONS
Abstract. This paper describes a new multifunction smart coating that can pro-
vide atomic oxygen (AO) and electrostatic discharge (ESD) protection, while also
improving the thermal control of space structures. The methodology is based on
a passive thin-film structure employing VOn transition metal oxides that exhibit a
metal to insulator transition. The coating, depending on its formulation, can pro-
vide a variable heat-transfer/emitter structure that operates passively in response
to changes in the temperature of the space structure, by dynamically varying the
ratio of solar absorptance () to thermal emittance (). This enhances self-heating
of the structure at lower temperatures and cooling through thermal radiation at
elevated temperatures. Work is currently underway to apply this coating to various
polymers and membranes to improve their performance in space.
In the space environment, such as low Earth orbit (LEO), the coating will
be subject to various stresses including VUV radiation and AO. Atomic oxygen
testing in a simulated environment at CSA indicated no resolvable change in
the morphology or thickness of the coatings. The thermo-optic characteristics
after AO exposure were similar to the as deposited films. Additional long-term
radiation exposure at the Centre National dEtudes SpatialesFrance (CNES),
equivalent to three years in a geostationary orbit (GEO) environment, resulted in
a change in the coating and of less than 0.002.
1. Introduction
radiation and reactive atomic oxygen (AO) that are present to a greater, or lesser
extent, in various orbits. These environmental factors can significantly modify
the performance of the outer layers of a spacecraft, and for certain materials,
especially various polymers, cause significant material erosion and degradation
of their thermo-optic characteristics. The effects of the space environment on
various materials has been examined through various studies including NASAs
Long Duration Exposure Facility (LDEF) set of experiments [1]. In particular,
Kapton polyimide and Teflon FEP are widely employed in space as part of thermal
blankets and for membrane antennas but suffer significant erosion due to the VUV
AO that is present, especially in the low Earth orbit (LEO) environment [2].
The outer spacecraft coverings control the thermal balance between the space-
craft, incident solar and terrestrial radiation, and thermal emittance to dark space.
The efficient thermal control of spacecraft structures is an important issue that
impacts directly on the performance and longevity of attached and internal sub-
systems. While spacecraft can be subjected to external temperature swings from
about 150 to +150 C; the corresponding internal temperature must be regulated
over a nominal range, typically 1030 C. The tighter the temperature control of
the spacecraft; the better the performance ratings and the lifetime of the spacecraft
subsystems and payloads.
Thermal control systems typically require heaters with significant power re-
quirements to maintain payload temperatures at acceptable levels during cold
swings to compensate for the radiator heat loss to dark space. Traditional dynamic
thermal-control systems have employed various forms of mechanical louvers to
regulate heat dissipation into deep space. However, these are bulky (3.3 kgm2 ),
subject to wear, require additional power, and are costly, exceeding $100K US
m2 [3]. With the trend towards smaller, highly functional satellites; there is
a need for cost-effective smart radiator devices (SRD) that are lightweight and
reliable.
The main smart material systems that have been considered for thermal control
applications include V1xy Mx N y On [5], where M and N are dopants such as W,
Ni, Cr, Ti, and Mo; ceramic tiles based La1x Mx MnOn perovskite materials (Mx
can be Ca, Sr, or Ba [6]); and electrochromic active devices based on WO3 [7, 8].
The alloying or inclusion of M and N dopants is used to modify the characteristics
of the material. In this paper, the authors focus mainly on the VOn based smart
material system.
VOn exhibits one of the largest observed variations in electrical and optical
characteristics due to the metal-insulator transition. The transition temperature
increases with the oxygen content; varying from 126 K for VO, to 140K for
V2 O3 , and to 341 K for VO2 [9, 10]. The metal-insulator transition in Vx On is
associated with a change in structure from a tetragonal rutile structure with metallic
characteristics above the transition temperature, to a monoclinic structure with
insulator-like characteristics below the transition temperature. It is also feasible
to induce the metal-insulator transition in VO2 actively using an applied electric
field. Measurements by Stefanovich et al. [11] of the switching characteristics in
the electrical current using the VO2 in a FET configuration indicate that the metal-
insulator transition in VO2 is related to a critical electron density that induces the
transition to a metallic state, similar to the classical Mott metal-insulator transition.
The Mott criteria for the transition is given by
(n c )1/3 H 0.25,
where n c is the critical electron density and H is the Bohr radius. The value of
the critical density was found experimentally [11] to be about 3 1018 cm3 , in
good agreement with theoretical predictions. Since the driving mechanism is an
electric field effect, the power requirements to induce the switching are moderate,
comparable to the operation of a CMOS device.
The broad-band optical transmittance characteristics of a 250 nm thick VO2
coating deposited on c-Si, as measured by FT-IR [5], are shown in figure 1 for the
metallic (95 C) and insulating (29 C) states. The VO2 was prepared by reactive
laser ablation [5]. This provided high transmittance in the insulating state that
extends from below 2 m in the NIR to beyond 15 m in the infrared. The relatively
low optical absorption by the VO2 in this insulating state suggests a relatively clean
band-gap, free of defect states. In the metallic state, the optical transmittance of the
VO2 /Si decreased close to 0%, remaining below 0.5% into the FIR. The nominal
transition temperature for undoped crystalline VO2 is about 68 C.
The corresponding switching in the electrical characteristics of the crystalline
VO2 is shown in figure 2. There is a three order of magnitude decrease in the
resistivity of the VO2 due to the transition to a metallic state above 70 C over a
transition region that is about 10 wide.
La1x Mx MnOn is a class of double-exchange (DE) ferromagnets [6]. The
double-exchange refers to a double electron exchange between neighbouring Mn
280 ROMAN V. KRUZELECKY ET AL.
100
90
Transmittance (%)
80
70
60 29C
50
40
30
20 95C
10
0
2 4 6 8 10 12 14 16 18 20
Wavelength (micrometers)
Figure 1. Optical transmittance of undoped VO2 deposited onto Si in the insulating (29 C) and
metallic (95 C) states (near T = 0%)
and O. This affects the Mn-O-Mn bond angle, and hence, the crystal structure. The
DE ferromagnets differ from the vanadium oxide family in that the DE class of
materials is metallic at lower temperatures and semiconducting above the transition
temperature, over the appropriate compositions. The large change in properties
occurs near x = 0.3 in composition with LaMnO3 behaving as an insulator at one
extreme of the compositional variation. Bulk samples are typically prepared by
sintering a powder mixture of the appropriate composition at high temperatures
above 1200 C for about 60 h. This can be employed to produce tiles of the material.
Electrochromic structures based on WO3 [7, 8] employ an active switching
mechanism involving the electric-field induced motion of group 1 ions, such as
H+ or Li+ , from a storage reservoir through an ion conducting layer to the WO3
layer. The resulting reactions convert the WO3 , which is IR transparent, to HWO3
(or LiWO3 ) complexes that are metallic and reflective in the IR. This entails a
multilayer structure; consisting of a polymer that acts as a storage reservoir for the
Electrical resistance ()
105
104
103
20 40 60 80 100
Temperature (C)
Figure 2. Variation of the electrical resistance of VO2 deposited onto quartz at 500 C as measured
using the four-point probe method
MULTIFUNCTION SMART COATINGS FOR SPACE APPLICATIONS 281
2. Experimental Results
This work focused on the growth of VOn on Kapton and Al for various thermal-
control applications in space. VOn films were also prepared on crystalline Si (c-Si)
and quartz to enable various thin-film characterization. Growth on Si facilitated
IR characterization of the film thermo-optical characteristics; while growth on
quartz enabled characterization of the electrical characteristics and the optical
transmittance in the VIS/NIR spectral range.
Film characterization included
Undoped and doped VO2 films were prepared by reactive pulsed laserablation us-
ing a special target and a XeCl excimer laser, as shown in figure 3. The depositions
were performed in a controlled gas-phase background consisting of an O2 .Ar1
gas mixture at total pressures of about 100 mTorr. The substrate temperature was
systematically varied from 200 to 350 C for growth on Kapton, and from 300 to
500 C for growth on Al. The film thickness ranged from 0.1 to about 1.0 m. The
trial test coupons were 25 mm O.D. in size for the coated area.
X-ray diffraction (XRD) was employed to study the structure and stoichiometry
of the resulting coatings. Deposition conditions were established for the formation
of stoichiometric VO2 . At the higher concentrations of O2 in the gas mixture and
higher gas-phase pressures, a mixed-phase structure containing VO2 and V2 O5
was observed by XRD. As the total pressure was reduced to below 100 mTorr,
and the concentration of O2 in the mixture was reduced to about 2.5%, films that
were mainly VO2 were obtained. For film growth on crystalline substrates such
as quartz and c-Si, a distinguishable diffraction peak associated with a crystalline
282 ROMAN V. KRUZELECKY ET AL.
VO2/Si
VO2
400000 TS=520C
200000
Intensity (Counts)
0
400000 10 20 30 40 50 60
TS=420C
200000
0
10 20 30 40 50 60
400000
TS=330C
200000
0
10 20 30 40 50 60
2 (degree)
Figure 4. XRD analysis of the deposition temperature effect on the vanadium oxide structure and
stoichiometry, as deposited onto c-Si
MULTIFUNCTION SMART COATINGS FOR SPACE APPLICATIONS 283
(a) (b)
Al VO2/Aluminum 250
700
kapton substrate
600
200
Al
500
150
Intensity
Intensity
400
Al
300 100
200 Al
VO2 50
100
0 0
10 20 30 40 50 60 70 80 10 20 30 40 50 60
2 (degree)
Figure 5. X-ray diffraction spectra of VOn deposited onto (a) polished Al at 520 C and (b) Kapton
at 300 C
0.9
0.8
Transmittance
0.7
0.6
0.5
0.4
500 550 600 650 700 750 800 850
Wavelength (nm)
Figure 7. VIS Transmittance of a 100 nm thick VOn deposited onto Glass at 150 C
(a) (b)
AM0 Spectra
2200
Spectral Irradiance W*m2*micron1
2000
1800
1600
1400
1200
1000
800
600
400
200
0
0 0.5 1 1.5 2 2.5 3 3.5 4
Wavelength micrometers
ASTM E-490 Wehrli WMO
Figure 8. (a) AM0 spectrum (after NASA [12]) and (b) Optical simulation of VOn -based VIS
selective reflector on Kapton
MULTIFUNCTION SMART COATINGS FOR SPACE APPLICATIONS 285
1.00E+07
1.00E+06
1.00E+05
30 40 50 60 70 80 90 100 110 120
Temperature (C)
Figure 9. Sheet resistance (Ohm/) versus temperature for 150 nm thick VOn films deposited
onto 5 mil Kapton at various deposition temperatures (250 CSSH04; 300 CSSH03; 350 C
SSH05)
The sheet resistivity characteristics of 150 nm thick VOn films deposited onto
Kapton were examined using a commercial 4-point probe. Figure 9 shows the
variation of the sheet resistivity with temperature for VOn films deposited between
250 (K02) and 350 C (K05) on Kapton. Relative to the resistivity characteristics
of the crystalline VO2 deposited on quartz, as shown in figure 2, the electrical
characteristics of the VOn /Kapton did not exhibit a significant transition near
70 C. Lowering the deposition temperature tended to increase the resultant sample
resistivity (VO2-SSH-K04), as shown by figure 9. This may be due to a more
amorphous structure as the deposition temperature is lowered, as suggested by
the XRD measurements, that results in greater carrier scattering and trapping.
Typically, the room-temperature resistivity of the amorphous VOn deposited near
300 C on Kapton was about an order of magnitude larger than that of crystalline
VO2. Since background scattering by the Kapton obscures the XRD measurements,
286 ROMAN V. KRUZELECKY ET AL.
additional XPS measurements are planned to clarify the stoichiometry of the VOn
deposited at lower substrate temperatures onto the Kapton.
Broad-band sample emissivity data was provided using the measurement set-up
shown in figure 10. The samples were placed on a polished Al pad that could be
heated or cooled using a proportional-integral controller. The thermally-induced
IR emission from the sample was collected and focused onto a pyroelectric detector
using a system of ZnSe lenses. An optical chopper and signal lock-in amplifier
were used to extend the signal measurement range and to reduce background
effects. An f = 25 mm lens allowed signal collection from a small section of
the sample surface to reduce background effects. For the initial measurement set-
up, the pyroelectric detector was used without a window, providing full spectral
measurements of the total emitted thermal signal.
The thermal emissivity of VOn on 5 mil Kapton HN was estimated using a
polished Al backing to compensate for the IR transmission of the Kapton. Figure
11 shows a graph of the measured signal detected by the pyroelectric detector
versus the surface temperature, as determined using a miniature thermister, for
a 150 nm thick VOn sample deposited at 300 C. Relative to a blackbody refer-
ence standard, the measurements suggest a net emissivity of about 0.9 for the
VOn /Kapton combination. The metal-insulator transition expected near 68 C in
undoped crystalline VO2 was not resolvable.
Figure 12 shows the measured pyroelectric detector signal for an engineered
VO2 sample on an Al substrate (ISRD-010), that was structured for tuneable
0
20 30 40 50 60 70 80 90 100
Calibrated Temperature (C)
blackbody Polished AL VO2_SSH-K-01
Figure 11. Thermal emission of a 150 nm VOn film on Kapton HN (a type of Kapton provided by
Dupont)
5.0
4.5
Pyroelectric Detector Signal (V)
4.0
3.5
3.0
2.5
2.0
1.5
1.0
0.5
0.0
20.0 30.0 40.0 50.0 60.0 70.0 80.0 90.0 100.0
0.7
0.6
0.5
Calibrated IR Emissivity
0.4
0.3
0.2
0.1
0
20 30 40 50 60 70 80 90 100
Surface Temperature (C)
Figure 13. Esimated emissivity for VOn /Al sample (ISRD-10) versus temperature
2.5. DOPING
Ti Doping 0.0 0% 0% 1%
Transition Temperature 68 C 30 C 20 C About 0 C
3. Discussion
Two space applications are currently being developed by MPB for the VOn material
system:
1. multifunction protective coating on polymers such as Kapton.
2. thin-film tunable emittance coating on Al for thermal radiator tiles,
As a coating on polymers such as Kapton, the VOn structure can offer several
functionalities:
r solar absorptance below 0.2, with high VIS reflectance,
r IR thermal emittance exceeding 0.8,
r Rs(T ) < 109 / for ESD protection to below 100 C,
r AO protection.
Work is currently ongoing to optimize the low-temperature deposition process
for VOn on Kapton to also provide some tunability of the net thermal emissivity.
The goal is a tunability in the net emissivity of about 0.20.3 between the high and
low-temperature values to facilitate dynamic temperature control of membrane
structures.
Spacecraft employing passive thermal radiators can require substantial heating
at lower temperatures to compensate for the high thermal exchange to dark space.
Active systems using mechanical louvers are subject to wear and bear a consid-
erable mass penalty. MPB is developing a thin-film smart radiator device (SRD)
based on its VOn technology that is applied directly to the spacecraft Al radiator,
resulting in minimal added mass (<100 gmm2 ), long-term reliable operation
with a target of 15 years, and robust integration with the spacecraft structure.
Advanced nano-engineering concepts are being employed to structure the VOn on
Al in order to optimize the emissivity tunability. The design optimization and test-
ing is currently ongoing. As previously shown in figure 12, broadband emissivity
tunability (420 m) exceeding 0.45 has been achieved.
MULTIFUNCTION SMART COATINGS FOR SPACE APPLICATIONS 291
The MPB thin-film SRD functions to control the differential heat exchange
between the spacecraft and vacuum by dynamically controlling the heat return
to the spacecraft, the heat transfer to the external radiating surface, and the heat
exchange to dark space. Figure 14 provides a schematic of the operation of the SRD
below (low-temperature case), and above (high-temperature case) a preselected
thermal-switch point.
When the spacecraft is cold, heat loss into space is minimized ( < 0.3) to stabi-
lize the spacecraft temperature. Otherwise, this heat loss needs to be compensated
using internal heaters with an additional power penalty. When the spacecraft is
hot, maximal heat is dynamically transferred to the radiator and the heat radia-
tion to dark space is maximized ( > 0.65), while heat return to the spaceship is
minimized.
The proposed dynamic SRD significantly reduces the heat loss at lower tem-
peratures to enable spacecraft designers to avoid or reduce internal heating require-
ments. Nominally, the performance of space systems must be derated to account
for the effects of the large temperature swings. The SRD can also assist to reduce
292 ROMAN V. KRUZELECKY ET AL.
the peak-to-peak temperature swings of the spacecraft. This would enable existing
space systems to operate in temperature ranges of maximum performance.
4. Conclusion
High-quality VO2 films have been achieved on Si, quartz and Al substrates by
hybrid reactive pulsed-laser ablation. The metal-insulator transition temperature
can be tailored from below 20 to above 70 C through doping. Below this tran-
sition temperature, the VO2 films are insulating in nature and display high optical
transmittance over a broad spectral range that extends from the visible to the far
infrared. Above the transition temperature, the films exhibit a sharp drop in optical
transmission and a corresponding sharp increase in their electrical conductivity
and reflectivity. The VO2 -based smart coatings can be employed in a variety of
ways to provide very flexible thermal management of spacecraft.
The deposition of VOn onto Kapton at relatively low temperatures of 150
and 300 C results in an amorphous structure with a weakened metal/insulator
transition.
Adhesion to Kapton is relatively good, with no observable flaking or crack-
ing. The low-temperature deposition (<350 C) results in an amorphous structure
that exhibits higher resistivities and a lower conduction activation energy than the
crystalline VO2 deposited at 520 C, as used for the thin-film SRD applications.
VOn -based structures are being developed for Kapton that provide an emissivity
exceeding 0.85 and an / > 2 for the thermal control of membrane structures.
The VOn coating also exhibits excellent stability under exposure to AO and radi-
ation as encountered in LEO and GEO environments. The thin-film structure can
provide several functionalities for polymers including AO and ESD protection,
thermal control and selective rejection of incident solar radiation. Work is ongoing
to facilitate dynamic variation of the / passively in response to the temperature
of the membrane.
Both active and passive SRD devices using the VOn on Al are being developed
for the dynamic thermal control of spacecraft. Passively, the VO2 can be used as a
combined variable-conductance switch/variable-emittance radiator. This enhances
the net dynamic variation in emissivity that is possible for a passive SRD device.
Combining these effects has provided experimentally measured net variations
in IR emissivity from below 0.25 at lower temperatures to above 0.65 at higher
temperatures for a passive device on Al. Still greater high-temperature emissivities
may be possible by further optimizing the structure and doping of the VO2 to
enhance IR optical absorption in the metallic state.
The proposed thin-film thermal-control methodology has significant advan-
tages over competitive technologies in terms of weight, cost, structural sim-
plicity with no mechanical components, integration with the space structure,
MULTIFUNCTION SMART COATINGS FOR SPACE APPLICATIONS 293
compatibility with the space environment and dynamic thermal emittance perfor-
mance. The proposed passive thin-film SRD employs vacuum-compatible layers
that have low solar absorptance and good AO resistance. Stability in the pres-
ence of ions and protons should also be good due to the strong stoichiometric
bonding. An added benefit is the ability to deposit the thin-film SRD on various
qualified space materials such as Al and Kapton to minimize the overall mass of
the spacecraft thermal control system.
Acknowledgments
References
1. Whitaker, A. F. and Gregory, J. (1993) LDEF Materials Results for Spacecraft Applications,
Reports: N-94-31012; NASA-CP3257; M742; NAS1.55:3257; CONF-9210467.
2. Silverman, E. M. (1995) Space Environmental Effects on Spacecraft: LEO Materials Selection
Guide, NASA CR-4661.
3. Tennyson, R. C. and Morison, W. D. (1990) In Proceedings TMS Symposium on Space Envi-
ronmental Effects on Materials, Anaheim, CA, February 1990.
4. Gilmore, D. G. (1994) Satellite Thermal Control Handbook, The Aerospace Corporation Press,
El Segundo, CA.
5. Mahan, B. A. (1975) University Chemistry, 3rd ed., Addison-Wesley Publishing, Reading, MA,
USA, pp. 685743.
6. Kruzelecky, R. V., Haddad, E., Jamroz, W., Soltani, M., Chaker, M., Nikanpour, D., and Jiang,
X. X. (2003) In the 33nd International Conference on Environmental System 33-IES, Paper
03ICES-242, Vancouver, Canada, July 2003.
7. Shimakawa, Y., Yoshitake, T., Kubo, Y., Machida, T., Shinagawa, K., Okamoto, A., Nakamura,
Y., Ochi, A., Tachikawa, S., and Ohnishi, A. (2002) Applied Physics Letters 80, 48644866.
8. Franke, E., Neumann, H., Schubert, M., Trimble, C. L., Yan, L., Woollam, J. A. (2002) Surface
and Coating Technology 151152, 285288.
9. Douglas, D. T., Swanso, R., Osiander, J., Champion, J., Darrin, A. G., Biter, W., and Chan-
drasekha, P. (2002) In M. El-Glenk (ed.) Space Technology and Applications International
Forum (STAIF)-2002, CP 608, American Institute of Physics, pp. 204210.
10. Griffiths, C. H. and Eastwood, H. K. (1974) Journal of Applied Physics 45, 22012206.
11. Egorov, F. A., Yu. Sh. Temirov, Dvoryankin, V. F., Potapov, V. T., and Sokolovskii, A. A. (1991)
Sov. Tech. Phys. Lett. 17, 295296.
12. Stefanovich, G., Pergament, A., and Stefanovich, D. (2000) Journal of Physics. Condensed
Matter 12, 88378845.
13. ASTM-E-490. Version SOLAR2000, Model AM0 (Solar irradiance, Air mass zero, distance of
1 A.U.) rredc.nrel.gov/solar/standards/am0/wehrli1985.new.html.
EFFECTS OF SPACE ENVIRONMENT EXPOSURE ON THE
BLOCKING FORCE OF SILICONE ADHESIVE
Abstract. An issue has been identified with the diode tape used on the Interna-
tional Space Station (ISS) solar arrays to provide a high emittance surface for the
underlying diodes. The diode tape consists of silicone pressure sensitive adhesive
(Dow Corning QC-7725) with a protective Kapton over-layer to prevent sticking
prior to solar array deployment. On-orbit, the Kapton over-layer will erode under
exposure to atomic oxygen (AO). Under additional AO exposure, the underlying
exposed silicone adhesive is expected to convert to a glass like silicate.
Conventional thinking suggests that at some point, the Kapton will be eroded,
exposing pure silicone adhesive. The silicone will lose its adhesive characteristics
after some time as it converts to a silicate due to AO exposure. This presents a
problem because the current operational plan to retract ISS solar array P6 and leave
it stored under load for a long duration (6 months or more) may be incompatible
with the timing of blocking minimization for the silicone adhesive. Previous test-
ing by Lockheed-Martin Space Systems (LMSS) characterized silicone blocking
following exposure to atomic oxygen (AO) in an electron cyclotron resonance
(ECR) source facility. LMSS blocking test results, combined with analysis results
of predicted ISS on-orbit AO exposure levels indicate the silicone blocking forces
resulting from retraction and storage of the P6 4B solar array for an extended
period of time may reach unacceptable levels. That is, the silicone may not be
sufficiently converted to a silicate.
The LMSS ECR source test results and their application to ISS on-orbit con-
ditions are believed to be conservative. To address this conservatism, the Envi-
ronment Effects Group at Marshall Space Flight Center, under direction from
the ISS Program Office Environments Team, performed environments exposure
testing with a high energy 5 eV AO beam along with near ultraviolet (NUV)
295
J.I. Kleiman (ed.), Protection of Materials and Structures from Space Environment, 295307.
C 2006 Springer. Printed in the Netherlands.
296 PAUL BOEDER ET AL.
radiation and ionizing radiation to more accurately simulate the actual on-orbit
environment. Diode tape samples were exposed to each environment constituent
individually, put under preload for seven days and then the resulting blocking force
was measured using a tensile machine. Additional samples were exposed to AO,
NUV and electrons in series and then put under long term (three to ten months)
preload to determine the effect of preload duration on the resulting blocking force
of the silicone-to-silicone bond.
In this paper, we present the results of laboratory tests to determine the effect of
space environment exposure (5 eV atomic oxygen, ultraviolet radiation, and elec-
tron radiation) on the blocking force of silicone adhesive in a silicone-to-silicone
bond. Test results indicate that high energy AO, ultraviolet radiation and electron
ionizing radiation exposure all reduce the blocking force for a silicone-to-silicone
bond. AO exposure produces the most significant reduction in blocking force.
Key words: atomic oxygen, ultraviolet radiation, electron radiation, silicone ad-
hesive, blocking
1. Introduction
In this paper, we present the results of laboratory tests to determine the effect
of space environment exposure on the blocking force of silicone adhesive in a
silicone-to-silicone bond. Environment exposure includes both individual and
combined exposure to 5 eV atomic oxygen, ultraviolet radiation, and energetic
electron ionizing radiation. Test results available to date and presented herein
show that long-term exposure to each of the environment constituents reduces the
adhesive properties of a silicone adhesive. Section 2 describes the background
and reasoning for performing the testing. Section 3 describes the testing program.
Section 4 summarizes the test results and their implications. Conclusions are given
in section 5.
2. Background
MD-944 diode tape used on the International Space Station (ISS) U.S. solar array
panels provides a high emittance surface for the underlying diodes in the solar array
panel circuit. MD-944 diode tape consists of a pressure sensitive silicone adhesive
(Dow Corning QC-7725) with a protective Kapton over-layer to prevent sticking
prior to solar array deployment. On-orbit, the Kapton over-layer will erode under
exposure to atomic oxygen (AO). Once exposed to AO, the underlying silicone
adhesive is expected to ultimately convert to a silicate with no residual adhesive
properties.
EFFECTS OF SPACE ENVIRONMENT EXPOSURE 297
[3]. To explore this question, the Environment Effects Group at Marshall Space
Flight Center, under direction from the ISS Program Office Environments Team,
performed environmental exposure testing on samples of MD-944 diode tape.
The MD-944 diode tape Combined Effects Test Program is being conducted by
the Environment Effects Group at the Marshall Space Flight Center [4]. The test
program simulates exposure to the various components of the space environment,
singly and in combination. The test program calls for samples of the MD-944
diode tape to be exposed to the following:
1. 5 eV atomic oxygen with concurrent vacuum UV radiation;
2. Near UV radiation (NUV);
3. Electron radiation (250 keV energy);
4. Thermal energy atomic oxygen (<0.1 eV);
5. 5 eV atomic oxygen, NUV radiation, and 250 keV electrons combined in series.
Test items 13 have been completed while items 4 and 5 are still in progress at
the time this paper was prepared. A number of control samples were also prepared
to serve as a baseline for testing results. All exposure testing involves exposing
the silicone adhesive component of the diode tape to the environment constituent
of interest. This was accomplished by mounting the diode tape on an aluminum
substrate (block) so that the Kapton over-layer of the tape was bonded to the
aluminum substrate. A pressure sensitive adhesive was used to form the Kapton
to aluminum bond. Figure 2 shows the sample configuration for the exposure
portion of the test. Figure 3 shows a representative tape sample prior to exposure.
Test exposure levels are based upon our best estimate of environmental ex-
posure levels on ISS at the time the test program was put in place (July 2003).
The on-orbit exposure period covers the time span from P6 array deployment in
December of 2000 to April 2003. The estimated total AO fluence incident on the
P6 solar array port panel diode tape is 2 1021 AOcm2 . The total ultraviolet
radiation exposure is estimated at >12,000 ESH. The total surface electron dose is
estimated at 3.3 106 rads. The thermal cycle range for the deployed solar array
is 174+286 F.
Ultraviolet and electron radiation exposure levels will have continued to in-
crease between April 2003 and the present. However, only a minor increase (10
20%) in the total AO fluence exposure is expected for the P6 solar array diode
tape. This is because ISS has been flying in a minimum drag configuration in
order to minimize propellant consumption following the Columbia accident. The
minimum drag configuration places the solar array panels edge on to the velocity
vector and minimizes the MD-944 diode tape exposure to AO.
The MSFC Atomic Oxygen Beam Facility (AOBF) produces a flux of 5 eV atomic
oxygen atoms of approximately 5 1015 atomscm2 s1 . AO fluence is deter-
mined from mass loss of Kapton, with corroboration to beam current, duty cycle,
and time of exposure. During AO production, vacuum UV radiation is produced,
primarily at 130 nm. The VUV dose is dependent on the time in chamber more
than the fluence. A sample receiving 1 1021 atomscm2 would get 820900
equivalent sun hours of VUV. Samples of the diode tape with the silicone adhe-
sive side up were exposed to AO fluences ranging from 1.3 1018 to 1.5 1021
atomscm2 .
A Mercury-Xenon NUV lamp was used as the source of the NUV radiation. The
samples were mounted in a vacuum chamber and illuminated from outside of the
300 PAUL BOEDER ET AL.
chamber through a port in the side of the chamber. The plate to which the sample
holder was mounted was water cooled and the system operating temperature was
2225 C. During the exposure, the chamber was evacuated to a pressure of 106
Torr or less. The lamp was characterized with a radiometer prior to sample ex-
posure, and the output was monitored throughout the test. Samples of diode tape
were exposed to 500 equivalent sun-hours (ESH), 1000 ESH, and 2200 ESH.
The electron radiation energy and fluences for this test were based on radiation
environment predictions from the MSFC Environments Group and dose-depth
profiling for a 2-mil adhesive layer. The electrons were produced by a PelletronR
accelerator capable of generating 200 keV to 2.5 MeV energy electrons. Samples
of diode tape were exposed to 250 keV electrons to a fluence of 1.75 1013 ,
3.90 1013 , and 1.15 1014 e-cm2 .
250 keV electrons produce uniform dose-depth profiles as they are energetic
enough to penetrate all the way through 2 mils of adhesive. 1.75 1013 e-cm2
produces a calculated dose of 0.6 106 rads at a depth of 2 mils in the silicone
adhesive but under-doses the front of the adhesive (02 mils). 3.9 1013 e-cm2
produces a calculated dose of 1.3 106 rads at a depth of 1 mil in the silicone but
under-doses the adhesive at lesser depths and overdoses the adhesive at greater
depth. 1.15 1014 e-cm2 produces a calculated dose of 3.3 106 rads for the
surface of the adhesive but overdoses the adhesive at depth.
3.4. THERMAL ENERGY ATOMIC OXYGEN (< 0.1 EV) EXPOSURE TEST
The MSFC Atomic Oxygen Drift Tube Beam Facility produces a flux of <0.1 eV
atomic oxygen atoms of approximately 2 1016 atomscm2 s1 . During AO pro-
duction, a minimal amount of vacuum UV radiation is produced. This test had not
been completed at the time this paper was prepared. Diode tape samples will be ex-
posed to AO fluences comparable to the fluence levels used in the 5 eV AOBF tests.
The combined exposure test involves exposing samples to 5 eV AO, NUV radiation
and 250 keV electron radiation in the order indicated. This test was still in progress
at the time this paper was prepared.
the tape. After processing the samples, two substrates are pressed together using
the exposed silicone adhesive layer of the MD-944 tape (silicone to silicone bond);
and placed under 2.5 psi preload for a predetermined interval. The majority of
the test samples were placed under preload for 7 days prior to performing blocking
tests. Some of the combined effects test samples have been placed under longer
duration preloads of 3, 6, and 10 months to simulate the effects of long term solar
array storage. After preloading, the bond between the substrates is failed by a
tensile loading of the bond to quantify the degradation of the silicone adhesive.
Figure 4 shows the test set up for the blocking force tensile loading test. Figure 5
shows two samples bonded together and mounted in the tensile test machine prior
to initiating a tensile test.
A small number of samples were set aside so that x-ray photo-electron spec-
troscopy (XPS) and fourier transform infrared (FTIR) spectroscopy analyses could
Figure 5. Two samples mounted in the tensile test machine prior to test
302 PAUL BOEDER ET AL.
be performed. FTIR analyses were performed using the Horizontal Attenuated To-
tal Reflectance (HATR) technique. The analysis of the tape samples was done by
applying the silicone adhesive layer of the tape to the surface of a BBBB ATR
crystal, installed in an FTIR spectrometer. One hundred scans, at 4-cm1 res-
olution, were made to obtain the spectra reported herein. The spectra collected
were evaluated against spectra of the silicone adhesive and the Kapton backing of
the tape taken as reference spectra or control samples. The depth of penetration
into the sample is typically one to two microns or less. Variables such as sample
composition, angle of penetration into the sample, and amount of surface contact
with the sample have a large influence on the depth of sample penetration and the
internal reflectance sampling beam. XPS studies had not yet been completed at
the time this paper was prepared.
Test results are presented herein for NUV, energetic electron ionizing radiation, and
5 eV AO exposures. Test results are presented as a function of the tensile stress
force (pounds per square inchpsi) required to separate two silicone samples
bonded together after exposure to various levels of each environment constituent.
All samples for which results are presented herein were placed under a nominal
2.5 psi preload for seven days prior to the tensile test being performed. Figure 6
140
7-day preload at room temperature & 2.5 psi
Individual Data Points
Bonding area = 1" x 0.5"
Mean
120
Random Variation of Mean (95%)
100
Stress at Max. Load (psi)
3 pts
3 pts
80 8 pts
60
40
3 pts
20
0
0 500 1000 1500 2000 2500
NUV Exposure (ESH)
80
Individual Data Points
70 8 pts Mean
Random Variation of Mean (95%)
3 pts
60
Stress at Max. Load (psi)
3 pts
50 6 pts
40
30
20
10
7-day preload at room temperature & 2.5 psi
Bonding area = 1" x 0.5"
0
0.0E+00 2.0E+13 4.0E +13 6.0E +13 8.0E+13 1.0E+14 1.2E+14
250 KeV Electron Exposure (e/cm^2)
shows the change in tensile stress required to separate two samples exposed to
various levels of NUV. Figure 7 shows the change in tensile stress as a function of
energetic electron ionizing radiation exposure, while figure 8 shows similar data
as a function of 5 eV AO fluence exposure levels.
Figure 6 indicates a 67% mean reduction in blocking force after 2200 NUV
equivalent sun hours (ESH). We do not know why there is such a large spread in the
NUV exposure tensile test results shown in figure 6. Lamp calibration measure-
ments made at the NUV sample fixture location showed a standard deviation of
0.03 for nine measurements. Any systematic errors with the preloading or tensile
test phase of the test should also be apparent in the other data sets for electron
radiation and AO. Figure 7 shows a 30% mean reduction in blocking force after
exposure to a 250 keV electron fluence of 1.15 1014 electronscm2 . We spec-
ulate that the reason why NUV produced a greater reduction in blocking force
as compared to electron radiation is that the NUV radiation is concentrated near
the surface of the sample while the electron radiation is spread throughout the
material. Figure 8 shows a 100% reduction in blocking force after exposure to a
5 eV AO fluence of 1021 atomscm2 . The samples exposed to this AO fluence
did not exhibit any blocking force following a 7-day preload at 2.5 psi.
Test results are also presented for selected FTIR analyses. FTIR analyses were
performed on control samples as well as samples exposed to various levels of
environment constituents. Figure 9 shows the absorbance profile of an unexposed
304 PAUL BOEDER ET AL.
80
3 pts
Individual Data Points
8 pts Mean
70
Random Variation of Mean (95%)
60
7-day preload at room temperature & 2.5 psi
Bonding area = 1" x 0.5"
Stress at Max. Load (psi)
50
40
30 3 pts
20
10 3 pts 3 pts
0
0
1.E+17 1.E+18 1.E+19 1.E+20 1.E+21 1.E+22
5 eV Atomic Oxygen Kapton Effective Fluence (ao/cm^2)
2.8
2.6
2.4
2.2
2.0
1.8
Absorbance
1.6
1.4
1.2
1.0
0.8
0.6
0.4
0.2
0.0
4000 3500 3000 2500 2000 1500 1000
Wavenumbers (cm-1)
0.9
0.8
0.7
0.6
Absorbance
0.5
0.4
0.3
0.2
0.1
0.0
Figure 10. Sample # 95 (BF4a) exposed to 5 eV AO atoms to a fluence of 9.7 1019 atomscm2
The absorbance peak at 1250 cm1 in figures 9 and 10 indicates the presence
of the silicone adhesive. This peak is significantly reduced in figure 11 indicating
the silicone adhesive has been altered by the AO and near-UV exposure. None of
the FTIR analyses detected a significant absorbance band at 1400 cm1 , which
0.055
0.050
0.045
0.040
0.035
Absorbance
0.030
0.025
0.020
0.015
0.010
0.005
0.000
0.005
3500 3000 2500 2000 1500 1000
Wavenumbers (cm-1)
Figure 11. Sample # 163 (BF1a) exposed to 5 eV AO atoms to a fluence of 1.5 1021 atomscm2
followed by 2200-ESH near-UV
306 PAUL BOEDER ET AL.
Figure 12. Diode tape sample following exposure to a total 5 eV AO Kapton effective fluence of
1.48 1021 atomscm2
would indicate the presence of a silicate or aliphatic silicone (Si-O-R). The broad
absorbance band at 11101100 cm1 in figures 9 through 11 also does not provide
a clear indication of the conversion of the silicone adhesive to a silicate (SiOx ). We
do not know the depth of the expected silicone/silicate interface so it is difficult
to say whether we are looking at the AO effected layer or that we have penetrated
into the bulk silicone material with our FTIR analysis. We expect planned XPS
analysis to resolve any uncertainty in our FTIR analysis results. The peak at 3000
cm1 in figures 911 is a carbon to hydrogen bond (CH).
Figures 12 shows a sample that has been exposed to 5 eV AO to a total
Kapton effective fluence of 1.48 1021 atomscm2 . This exposed sample may
be compared to the unexposed sample shown in figure 3. No obvious surface
features are apparent on the unexposed sample in figure 3, while the post AO
exposure sample in figure 12 appears to have formed a glass-like surface coating
with fine cracks. FTIR analyses of exposed samples do not show evidence of SiOx
formation, but it is unclear whether the FTIR analysis is measuring the properties
of the AO affected layer or of the bulk silicone material. XPS depth profile studies
are planned to characterize the elemental composition of the diode tape materials
before and after exposure to AO.
As previously mentioned, the LMSS atomic oxygen blocking test results ini-
tially indicated the silicone blocking forces resulting from retraction and storage
of the P6 solar array for an extended period of time may reach unacceptable lev-
els. However, more recent analyses conducted at NASA Glenn Research Center
to determine the silicone effective fluence correlation for the LMSS tests indicate
that the Kapton effective fluence in the ECR facility needs to be 2.64 times higher
than in low Earth orbit to replicate the equivalent damage induced in silicone [3]
The silicone effective fluence correlation is based on changes in surface hard-
ness of silicone, whereas Kapton effective fluence is based on mass loss. These
new results imply that the LMSS atomic oxygen tests based on Kapton effective
EFFECTS OF SPACE ENVIRONMENT EXPOSURE 307
fluence were conservative, and therefore the LMSS tests are more consistent with
the results reported here.
5. Conclusions
Testing has been performed by the MSFC Space Environment Effects Group to
characterize the effect of environmental exposure on the blocking force of silicone
adhesive. Testing indicates that long-term exposure to each of the environment
constituents (AO, NUV, and electron ionizing radiation) will reduce the adhesion
properties of Dow Corning QC-7725 silicone adhesive. AO is the most efficient
in this respect, while NUV and electron ionizing radiation will also reduce the
adhesive properties of the silicone to a lesser extent. Contrary to expectations,
FTIR analyses do not show any evidence of silicate formation following exposure
to AO. Additional test plans are in place to
1. Perform XPS scans of control and exposed samples to determine elemental
composition as a function of depth in the sample,
2. Complete combined effects exposure testing (5 eV AO, NUV, and electrons in
series) followed by long term preloads of 3, 6, and 10 months in duration.
3. Complete sample exposures to low energy AO (<0.1 eV) in the MSFC Atomic
Oxygen Drift Tube facility.
Acknowledgments
References
1. Introduction
Titanium alloys have been widely used in aerospace industries because of their
high strength-to-weight ratio and excellent corrosion resistance [1]. However,
their poor wear resistance prevents the applications of titanium alloy as structural
materials [2]. The poor tribological properties of titanium alloy can be attributed
to low resistance to plastic shearing and low protection exerted by the surface
oxide [3]. For these reasons, many surface modification treatments are under
continuous development to improve their tribological behavior. Various surface
treatments have been widely studied such as, anodizing and MoS2 codeposition
[3], ion implantation [4], plasma nitriding [5], laser nitriding alloying [69], dia-
mond coating [10], plasma spray [11], and physical vapor deposition [12]. How-
ever, in order to optimize the surface treatment it is necessary to understand the
309
J.I. Kleiman (ed.), Protection of Materials and Structures from Space Environment, 309316.
C 2006 Springer. Printed in the Netherlands.
310 YONG LIU ET AL.
2. Experimental
The material used in this study was Ti-6Al-4V alloy (chemical composition in
weight percent, 5.56.5% Al, 3.54.5% V, 0.03% Fe, 0.15% O, balance Ti). The
material received in the form of hot rolled rod was annealed at 760 C for 60
minutes and cooled in the furnace to obtain an average hardness of 345 HV (35
HRC). The counter material was steel GCr15 (around 1% carbon, 1.5% chromium
in weight percent) quenched in water at 840 C and then followed by tempering at
150 C to have the hardness of 62 HRC.
Wear tests were carried out using a pin-on-disc scheme, in which the wear
couple was composed of a Ti-6Al-4V alloy pin and a steel disc. The pin was
cylindrical in shape, 9 mm in diameter, and 20 mm long, with a spherical end with
a radius of 40 mm. The pin was sliding against the steel disc (70 mm in diameter
and 10 mm thick). The surfaces of the pin and the disc were both grinded to
obtain roughness Ra of about 0.42 and 3.20 m, respectively. The wear tests were
performed at loads of 30 N and within a sliding velocity range of 0.21.2 ms1 .
The sliding distance was 1000 m. The wear results were characterized by the mass
loss of the pin using an electronic balance with the precision of 105 g.
The wear teats were carried out under the environmental pressure of 105 Pa using
a UTI TV 100 type apparatus for vacuum friction and wear (made in Ukraine). The
worn surface was prepared for SEM observation by ultrasonic washing in acetone
and drying. A Hitachi S-570 type scanning electron microscope was used.
DRY SLIDING WEAR OF Ti-6Al-4V ALLOY 311
TABLE 1. Friction coefficient of Ti-6Al-4V under 30 N and various sliding velocities at low
temperature in vacuum
Table 1 lists friction coefficient for Ti-6Al-4V alloy against GCr15 steel under 30
N and different sliding velocities in vacuum. The friction coefficient at low tem-
perature ranged from 0.33 to 0.56, which is lower than that at ambient temperature
under the same friction condition (within 0.680.84).
The measurement results indicate that the friction coefficients of Ti-6Al-4V
against steel at ambient temperature and low temperature in vacuum show a non-
linear variation with sliding velocity, implying the complexity of frictional contact
and difficulty in describing the friction using an uniform process. In addition, the
friction coefficient at low temperature is obviously lower than that at ambient
temperature, suggesting that the decrease in temperature will reduce the friction
coefficient for Ti-6Al-4V.
Figure 1 shows the variation of wear rate of Ti-6Al-4V with different sliding
velocities under 30 N at 203 and 300 K in vacuum. The wear rate of Ti-6Al-4V
Figure 1. Wear rate of Ti-6Al-4V alloy as a function of the sliding velocities under 30 N at low
temperature in vacuum
312 YONG LIU ET AL.
Figure 2. SEM micrographs showing morphologies of the worn surfaces of Ti-6Al-4V alloy under
0.2 ms1 and 30 N in vacuum at low temperature
is increased with increasing sliding velocity both at low and high temperature. At
ambient temperature, wear rate shows a near-linear variation with sliding velocity,
while it is increased nonlinearly with sliding velocity at low temperature. The rate
of wear at low temperature is different within the different sliding velocity ranges,
i.e., it is slow between 0.2 and 0.6 ms1 , fast within 0.60.8 ms1 , and little
changed between 0.8 and 1.2 ms1 . The calculation based on figure 1 indicates
that the wear rate is higher at ambient temperature than that at low temperature.
Figure 2 shows the surface morphologies of Ti-6Al-4V worn at low temper-
ature in vacuum. The main characteristics of the worn surface are the plough
grooves and sliding traces that are similar to those at ambient temperature in
vacuum. However, there are differences in characteristics of surfaces worn at
lower sliding velocities under low temperature and vacuum condition comparing
DRY SLIDING WEAR OF Ti-6Al-4V ALLOY 313
to these in high temperature vacuum environment. Cracks and black foreign ad-
hered residue can be found in figure 2(a). Figure 2(b) shows the edge morphology
due to plastic deformation caused by squeezing of surface during friction. Some
adhesion traces and layered steps caused by squeezing deformation can be ob-
served in figures 2(c) and 2(d), respectively. The analysis of figure 2 implies that
the worn surface of Ti-6Al-4V after wear at lower sliding velocities under low
temperature vacuum condition is characterized by plastic deformation, fracture,
and adhesion.
No cracks can be observed on the surface worn at higher sliding velocities
comparing to that at lower velocities. Adhesion traces and layered steps also
can be found on the surface worn at 1.2 ms1 under low temperature vacuum
condition. It is proposed that the worn surface of Ti-6Al-4V after wear at higher
sliding velocities under low temperature vacuum condition is characterized by
plastic deformation and adhesion.
Table 2 lists the average microhardness of Ti-6Al-4V surface worn under 30
N and different sliding velocities at low temperature in vacuum. The average
microhardness of the surface worn at low temperature is higher than that of the
original surface (345 HV). Figure 3 shows XRD spectrum for Ti-6Al-4V wear
under 0.2 ms1 and 30 N at low temperature in vacuum. The worn surface layer
of Ti-6Al-4V contains -Ti and -Fe phase. It implies that wear took place on
Figure 3. XRD spectrum for the worn surface of Ti-6Al-4V alloy in vacuum at low temperature
314 YONG LIU ET AL.
Figure 4. Temperature of the worn surface of Ti-6Al-4V alloy specimens as a function of the sliding
time under 30 N and various sliding velocities in vacuum at low temperature
GCr15 steel during friction. The debris of GCr15 steel will remain on Ti-6Al-
4V surface. The existence of -Fe in the worn surface layer of Ti-6Al-4V will
result in increase in microhardness. In addition, the hardening caused by plastic
deformation during friction also affects the increase in hardness.
Figure 4 shows the variation of surface temperature of Ti-6Al-4V with time
during sliding at different sliding velocities at low temperature in vacuum. The
surface temperature is increased linearly with increasing time. Increase in sliding
velocity will lead to an increase in heating rate of the surface. For example, the
surface temperature under 1.2 ms1 rises from original 203 to 235 K only after
3 min of sliding. The above analysis on morphology and XRD of the worn sur-
face of Ti-6Al-4V against GCr15 steel shows that, the wear of Ti-6Al-4V at low
temperature in vacuum is characterized by typical adhesion and plough with ma-
terial transfer. This implies that the friction process is mainly carried out between
Ti-6Al-4V and the transfer material on GCr15 steel. XRD spectrum indicates
that there is mutual material transfer between Ti-6Al-4V and GCr15 steel. At
low temperature in vacuum, the surface hardness of Ti-6Al-4V is increased lead-
ing to the wear of GCr15 steel. The formed Fe debris embedded in Ti-6Al-4V
surface results in an increase in hardness of Ti-6Al-4V surface. Comparing the
wear results at ambient temperature in vacuum, the difference in contacting con-
dition between sliding surfaces will change the trend of variation in wear rate
with sliding velocity, leading to a nonlinear increase of wear rate with increasing
sliding velocity at low temperature in vacuum.
With the surface temperature increasing with increasing sliding velocity [9],
the extent of temperature rise under lower velocities is lower than that under
higher velocities. On the other hand, the brittleness of Ti-6Al-4V is decreased with
DRY SLIDING WEAR OF Ti-6Al-4V ALLOY 315
decreasing temperature. Thus, cracks are easily formed at lower sliding velocities.
Increase in velocity will reduce the trend of fracture with fewer cracks existing on
the worn surface.
4. Conclusion
1. The friction coefficient for Ti-6Al-4V alloy against GCr15 steel in vacuum
ranged within 0.680.84 at ambient temperature and 0.350.55 at low temper-
ature.
2. The wear rate of Ti-6Al-4V alloy increased with increasing the sliding velocity
in vacuum, with the trend being higher at ambient temperature than that at low
temperature.
3. The surface of Ti-6Al-4V alloy worn at low temperature in vacuum was char-
acterized by increased adhesion. Under the low sliding velocity, microcracks
formed on the worn surface indicating of fracture due to low temperature, while
no such microcracks were observed under the higher sliding velocity.
4. The XRD analysis showed that the debris contain Fe, implying that the wear
took place for the steel disc although the original hardness of the steel was
higher than that of the Ti-6Al-4V specimen.
5. The surface temperature increased linearly with increasing time. A higher slid-
ing velocity would create an increased rate in temperature rise.
Acknowledgment
The authors are grateful to professor G. D. Gumulya for his interest in and en-
couragement of this work. This research is supported by National Basis Research
Foundation of China.
References
1. Wan, J. Y., Ge, A. M., and Zhou, Y. B. (1987) Titanium Alloys in the Aeronautic Application,
Science and Techology Press, Shanghai, China, pp. 545.
2. Budinski, K. G. (1991) Wear 151, 203217.
3. Molinari, A., Straffelini, G., Tesi, B., and Bacci, T. (1997) Wear 208, 105112.
4. Weisheit, A. and Mordike, B. L. (1995) Surface Engineering 11(2), 160161.
5. Ignatiev, M., Kovalev, E., Melekhin, I., Smurov, I. Y., and Sturlese, S. (1993) Wear 166, 233236.
6. Yerramsreddy, S. and Bahadur, S. (1992) Wear 157, 245262.
7. Ji, H. B., Xia, L. F., Ma, X. X., Sun, Y., and Sun, M. (1999) Tribology International 32, 265.
8. Wilson, A., Matthews, A., Housden, J., Turne, R., and Garside, B. (1993) Surface and Coatings
Technology 62, 600602.
316 YONG LIU ET AL.
9. Herr, W., Matthes, B., Broszeit, E., Meyer, M., and Suchentrunk, R. (1993) Surface and Coatings
Technology 57, 7780.
10. Bloyce, A., Qi, P.-Y., Dong, H., and Bell, T. (1998) Surface and Coatings Technology 107,
125132.
11. Dong, H. and Li, X. Y. (2000) Materials Science and Engineering A280, 303310.
12. Kragelski, I. V., Lyubarski, I. M., and Guslyakov, M. I. (1973) Friction and Wear in Vacuum,
Mashinostroenie, Moscow.
13. Vlandman, U., Luedtke, W. D., and Ringer, M. (1992) Wear 153(1), 330.
14. Miyoshi, K., and Pepper, S. V. (1992) Properties Data for Opening Galileos Partially Unfurled
Antenna, NASA, Cleveland, OH, NASA-TM-105355.
15. Sliney, H. E., Lukaszewicz, V., and DellaCorte, C. (1994) The Tribology of PS212 Coatings and
PM212 Composites for the Lubrication of Titanium 6Al-4V Components of a Stirling Engine
Space Power System, NASA-TM-106462, NASA, Cleveland, OH.
EROSION OF KAPTON H BY HYPERTHERMAL ATOMIC OXYGEN:
DEPENDENCE ON O-ATOM FLUENCE AND SURFACE
TEMPERATURE
Abstract. Organic polymers are susceptible to erosion from reaction with am-
bient atomic oxygen in low Earth orbit. We have investigated the linearity of the
O-atom fluence dependence of Kapton H erosion and the dependence of Kapton
H erosion yield on surface temperature. Sample exposures were performed with
a pulsed beam containing hyperthermal O atoms that were generated with a laser
detonation source. After exposure, samples were removed from the chamber in
which the exposures were done, and postexposure analyses were performed: etch
depth (profilometry) and surface topography (atomic force microscopy). A sys-
tematic set of exposures, which eroded room-temperature Kapton H from 1.4 to
25 m, showed that the erosion yield of Kapton H is linearly dependent on O-atom
fluence. This result helps validate the use of Kapton H mass loss (or erosion depth)
as a linear measure of the O-atom fluence of a materials exposure. The erosion of
Kapton H was strongly temperature dependent. At lower temperatures (<100 C),
the erosion yield appeared to be independent of sample temperature. But above
100 C, the erosion yield exhibited an Arrhenius-like temperature dependence,
with an apparent activation energy of 0.31 eV. These observations suggest that
O-atom-induced erosion of Kapton H proceeds through direct, nonthermal, gas-
surface reactions, and through reactions that depend on the surface temperature.
1. Introduction
2. Experimental Details
The sample exposures were conducted in the source region of a molecular beam
apparatus [3, 5]. A schematic diagram of the experimental arrangement is shown
in figure 1. The pulsed, hyperthermal, O-containing beam was produced with
a laser detonation source that is similar to the original laser-detonation source
SOURCE
SAMPLE
PULSED NOZZLE CHAMBER
MOUNT COLLIMATING
VALVE APERTURE
MIRROR
IONIZER QUADRUPOLE
SKIMMER
MASS FILTER
CO2 LASER
MAIN SCATTERING CHAMBER
Figure 1. Schematic diagram of the beam source, sample mount location, main scattering chamber,
and rotatable mass spectrometer detector
320 DEANNA M. BUCZALA AND TIMOTHY K. MINTON
designed by Physical Sciences Inc. [14]. The source produces neutral beams (ion
content 1%) containing 6070% O atoms and 3040% O2 molecules traveling
at velocities in the vicinity of 8 kms1 . The average O-atom translational energy
used in the experiments reported here was approximately 5 eV, and the energy
width (full width at half maximum) was approximately 2.5 eV. Characterization
of the beam has been done to determine that the hyperthermal O atoms are in their
ground electronic state, O(3P) [15]. In another similar study, we have concluded
that the O2 molecules in the beam are in the ground O2 (3 g ) state [16]. These
results establish that the exposure environment used in the laboratory subjects
materials to ground-state O and O2 , as do space exposures.
The samples were exposed to the hyperthermal beam at a distance of 40
cm from the apex of the conical nozzle source. A small portion of the O-atom
beam passed through several apertures and regions of differential pumping before
entering a mass spectrometer, which was used to characterize the ratio of O to O2
in the beam and the velocity distributions of these species. The Kapton-equivalent
O-atom fluence of an exposure was determined by measuring the erosion depth of
a Kapton H reference sample that is assumed to have an erosion yield of 3.00
1024 cm3 O-atom1 . The Kapton-equivalent fluence measurements led to the
conclusion that the average O-atom flux of one beam pulse at the samples was
1.67 1015 O atomscm2 pulse1 .
The samples were placed in a sample mount (see figure 2) such that the beam
impingement angle was a few degrees off normal. The entire mount was exposed
to the beam at a repetition rate of 2 pulses per second. Kapton H samples 1.22
cm diameter and 125-m thick were placed in the sample mount, with a stainless
steel wire mesh placed over them. The mesh allowed the erosion depth between
EROSION OF KAPTON H BY HYPERTHERMAL ATOMIC OXYGEN 321
the exposed and unexposed areas of the samples to be measured. The sample
mount contains nine locations for samples. The Kapton H reference sample was
placed in position 5. The reference sample was thermally isolated from the rest of
the sample mount and was maintained at 296 K. The temperature of the samples
at the other eight positions can be elevated from room temperature to 573 K by
resistive heaters embedded in the sample mount. The specific temperatures used
were 298, 333, 373, 423, 498, and 573 K.
Kapton H samples were always placed in positions 4 and 7 of the sample mount.
In order to ensure uniform temperature of these thermally insulating samples, thin
films of Kapton H were bonded to silicon substrates with a thermally conductive,
silver-filled polyimide adhesive (Ablebond 701). This adhesive was cured in
vacuum, with a maximum curing temperature of 573 K. When mounted in the
sample mount, the back sides of the silicon substrates were in intimate contact
with metal surfaces that were at the equilibrium temperature of the heated mount.
All samples used for studying the temperature dependence of the erosion yield
were exposed to 100,000 pulses of the hyperthermal beam. Many additional expo-
sures (for other purposes) were carried out over a two-year period, and the Kapton
H reference samples for all the exposures during this period were used to study the
dependence of Kapton H erosion depth on exposure duration. After exposure, sam-
ples were removed from the chamber and examined by profilometry and atomic
force microscopy (AFM). The wire mesh allowed the measurement of many steps
on exposed samples with the use of a Dektak-3 (Veeco Metrology Group) surface
profiler. The average erosion depth and corresponding standard deviation for a
given sample were calculated from measurements of 4045 different step heights.
It has previously been shown that the determination of the erosion yield of Kapton
may be accomplished with equivalent results by mass loss with a microbalance
or by step height measurements with a profilometer [17]. AFM measurements of
surface roughness of the exposed regions of sample surfaces were obtained with
a nanoscope IIIa tapping mode atomic force microscope (Digital Instruments).
14
12
Erosion Depth/m 10
0
0 50 100 150 200 250 300
Exposure Duration/1000 beam pulses
20
18
16
Erosion Depth/m
14
12
10
2
250 300 350 400 450 500 550 600
Temperature/Kelvin
Figure 4. Erosion depth of Kapton H as a function of temperature. The r points represent sample
position #7, the points represent sample position #4, and the points represent the Kapton H
reference sample (maintained at 296 K)
EROSION OF KAPTON H BY HYPERTHERMAL ATOMIC OXYGEN 323
4.5
4.0
3.0
2.5
2.0
1.5
1.0
0.5
300 400 500 600
Temperature/Kelvin
Figure 5. Kapton erosion ratio as a function of temperature. The points r represent sample position
#7, and the points represent sample position #4. The dashed curve is the best fit through all the
points with a two-term Arrhenius-like function (see text)
4. Discussion
Figure 6. AFM images of unexposed (A) and exposed (B-G) Kapton H. All images are 2.5 2.5
mm and have a z scale of 500 nm
70
60
50
Roughness/nm
40
30
20 unexposed
sample
10
0
0 298 333 373 423 493 573
Temperature/Kelvin
Figure 7. Surface roughness as a function of sample exposure temperature
EROSION OF KAPTON H BY HYPERTHERMAL ATOMIC OXYGEN 325
they studied the mass-loss rate over a narrower temperature range. They concluded
that the erosion rate of the polyimide had an Arrhenius temperature dependence,
with an activation energy of 5.7 104 eV for the temperature range 253353 K.
This measurement employed a quartz crystal microbalance, and mass-loss data
were obtained during the initial erosion period of the polymer. We also observed a
negligible temperature dependence in the similar temperature range 298373 K.
In fact, our results agree with those of Tagawa and coworkers in this temperature
range, within the uncertainty of our data. The obvious temperature dependence
that we observed became apparent at temperatures above those used by Tagawa
and coworkers.
The two-term fit to the data in figure 5 suggests that the loss of material
from a Kapton H surface is the result of temperature-dependent and temperature-
independent mechanisms. The constant term, C, results from a mechanism which
is temperature-independent. This mechanism likely involves the direct reaction
of O atoms with the surface on a time scale too short for thermal equilibrium
to be achieved. Direct reactions of 5 eV O atoms with the surface should be
able to overcome even significant reaction barriers, so reaction would proceed
independently of surface temperature and yield an apparent activation energy of
zero. For example, hydrogen abstraction has a barrier of aaproximately 0.25 eV
[20], while hydrogen elimination and CC bond-breaking reactions have barri-
ers near 2 eV [21, 22]. Any of these reactions may occur, given the available
collision energy. Our results do not permit a conclusion about the nature of the
temperature-independent mechanism. If the initial reactions of hyperthermal O
atoms with a hydrocarbon polymer are similar to those that occur with gas-phase
hydrocarbons [15, 2022], then perhaps all these reactions, as well as others, are
involved.
A temperature-dependent reaction process becomes important as the sample
temperature increases. An Arrhenius functional form describes well the tempera-
ture dependence of Kapton H erosion depth at the higher temperatures used. This
dependence suggests a mechanism that may take place in thermal equilibrium with
the surface, implying that in order for this mechanism to occur, O atoms must trans-
fer their energy to the surface and become trapped (either through physisorption
or through reaction) before the rate limiting erosion process(es) proceed(s). The
trapping probability increases with surface roughness, which allows for multiple
bounces at the surface that drive incident atoms toward thermal equilibrium and
increase the likelihood that a reaction will occur. In the studies reported here, and
in any situation where macroscopic amounts of material are removed by highly
directional O atoms, the surface reaches a steady-state roughness. Therefore, the
conclusion of a mechanism that depends on thermal accommodation (or trapping)
of incident O atoms is consistent with a steady-state erosion of rough surfaces.
Trapping would be expected to be less probable on smooth surfaces, possibly
EROSION OF KAPTON H BY HYPERTHERMAL ATOMIC OXYGEN 327
reducing the importance of the thermal mechanism and thus the temperature de-
pendence of erosion.
The temperature dependence of the Kapton H erosion yield is undoubtedly
linked to the dependence of surface roughness on temperature, but the relationship
is not clear. At lower temperatures, where the temperature-independent mecha-
nism dominates, a very rough surface is produced. But as the temperature is in-
creased and the temperature-dependent mechanism starts to dominate, the eroded
surface does not become as rough. The thermal mechanism retains no memory
of the incident beam direction, so O atoms that become trapped may react with
equal probability anywhere on the surface. In addition, the reactivity of trapped
atoms increases at higher temperatures. Both these factors may contribute to the
smoothing of the surface. But why the roughness reaches a minimum and begins
to rise again with temperature is difficult to explain. The ultimate dependence of
surface roughness on temperature represents a complicated interplay between the
temperature-dependent and temperature-independent mechanisms.
5. Conclusion
determined to be 0.31 eV, which is similar to the barrier for an H-atom abstraction
reaction.
Ackowledgments
References
19. Cline, J. A., Minton, T. K., and Braunstein, M. (2004) In Proceedings of the AIAA 37th Ther-
mophysics Conference, Paper No. 2004-2685, Portland, OR, 28 June1 July 2004.
20. Andresen, P. and Luntz, A. C. (1980) Journal of Chemical Physics 72, 5842.
21. Gindulyte, A., Massa, L., Banks, B. A., and Rutledge, S. K. (2000) Journal of Physical Chemistry
A 104, 9976.
22. Troya, D., Pascual, R. Z., Garton, D. J., Minton, T. K., and Schatz, G. C. (2003) Journal of
Physical Chemistry A 107, 7161.
TRANSPARENT ARCPROOF PROTECTIVE COATINGS:
PERFORMANCE AND MANUFACTURABILITY ISSUES
Abstract. Highly transparent thin films with moderately high sheet resistivity
approximately 108 ohmssquare1 ( 1 ) are needed for protection of photo-
voltaic arrays and other spacecraft surfaces from static charging in space. They
may also be useful for dust control on Mars and the Moon. Indium tin oxide (ITO)
codeposited with the transparent insulator MgF2 is promising for these applica-
tions, but it is difficult to deposit films with reproducible sheet resistivity. We report
experiments on a small dual RF magnetron sputter coater, using plasma emission
monitoring (PEM) to improve control of film composition. Results show that PEM
is useful for composition control but must be supplemented by periodic in situ
measurements of coatings optical or electrical properties. We have successfully
coated both rigid (glass, quartz) and flexible substrates.
1. Introduction
331
J.I. Kleiman (ed.), Protection of Materials and Structures from Space Environment, 331340.
C 2006 Springer. Printed in the Netherlands.
332 JOHN GRIFFIN ET AL.
1e+11
1e+10
1e+9
Sheet Resistivity (/ )
1e+8
1e+7
1e+6
1e+5
1e+4
1e+3
1e+2
1e+1 Avg. film thickness = 65 nm
1e+0
0 5 10 15
Calculated Wt. % MgF2
0.90
Total Solar Transmittance
0.88
0.86
0.84
0.82
Avg. film thickness = 65 nm
Transmittance of SiO2 substrate = 0.93
0.80
0 2 4 6 8 10 12 14 16
Calculated Wt. % MgF2
Figure 1. Sheet resistivity and solar transmittance of ITOMgF2 vs. composition [2]
2. Experimental Techniques
PLAN VIEW
Optical
Optical
Sample Fiber
Fiber
QCM
below Pinhole
Pinhole sample shield
shield
ITO MgF2
Gun Gun
Shield
a profilometer [4]. MgF2 /ITO composition ratios were estimated from deposition
rate measurements made on each sputter gun at least once during each deposition
run. (We had previously found the deposition rate to be approximately a linear
function of RF power.)
Light from the plasma in front of each gun was collected by an optical fiber
oriented approximately as shown in figure 2. Each fiber terminated in a short tube
with a small hole at its outer end. This reduced the number of scattered particles
reaching the fiber [5]. Emissions were analyzed by a two-channel grating spec-
trometer with wavelength resolution approximately 0.4 nm. Experiments showed
that each fiber received a negligible amount of light from the other sputter gun.
Substrate temperature during deposition, estimated from thermocouple mea-
surements, was <40 C. Thus, it is likely that our samples were highly disordered
or amorphous. However, this demonstrates the feasibility of coating flexible poly-
meric substrates.
Most samples discussed in this paper were deposited on borosilicate glass.
However, we have successfully coated Mylar, Kynar, and Upilex. Each substrate
was covered by an aluminum mask to produce a sample measuring 0.3 1.9 cm2
with electrical contact arms along the edge. Electrical resistance measurements
were made at room temperature in ambient atmosphere by four-terminal methods
to eliminate the effect of contact resistance, using appropriate guarded cabling
and high-input resistance electrometers.
A typical PEM spectrum for each target is shown in figure 3. The spectra clearly
are very different.
We chose to use ITO and MgF2 lines at 453 and 384 nm, respectively, because
they showed the best correlation with sheet resistivity. Higher resolution plots of
these lines are shown in figure 4.
The possible benefits of composition control by plasma emission monitoring
are shown in figure 5, where we plot sheet resistivity for the same samples vs. the
MgF2 /ITO plasma intensity ratio and vs. the MgF2 concentration determined by
the quartz crystal monitor. Note the closer correlation of sheet resistivity to the
intensity ratio than to the estimated MgF2 concentration.
On the other hand, we sometimes found large shifts of sheet resistivity relative
to plasma intensity ratio and QCM data, usually after opening the chamber to
remove samples. This is shown in figure 6.
Based on these results, it appears that PEM can facilitate production of ITO
MgF2 but must be supplemented by periodic in-situ measurements on coated
products or witness coupons.
ITO
100
Intensity
50
0
200 300 400 500 600
Wavelength (nm)
MgF2
800
600
Intensity
400
200
0
200 300 400 500 600
Wavelength (nm)
Figure 3. Broadband optical spectrum of each targets discharge. Intensity is measured in arbitary
units
ITO
100
ITO
Intensity
Indium
50 Tin
Oxygen
Argon
0
448 450 452 454 456 458
Wavelength (nm)
MgF2
400
MgF2
Intensity
Magnesium
Fluorine
200 Argon
0
382 383 384 385 386
Wavelength (nm)
Figure 4. Higher-resolution spectra of the monitored plasma emissions, together with data on rele-
vant elemental emission lines
336 JOHN GRIFFIN ET AL.
// )
1.E+11 Run 272
Sheet Resistivity (
1.E+10 Run 273
Run 274
1.E+09
1.E+08
1.E+07
1.E+06
1.E+05
1.E+04
1.0 1.5 2.0 2.5 3.0 3.5 4.0
MgF2/ITO Optical Intensity Ratio
1.E+11
// )
Run 273
1.E+09
Run 274
1.E+08
1.E+07
1.E+06
1.E+05
1.E+04
15 20 25
Vol% MgF2 (from QCM)
Figure 5. Sheet resistivity vs. MgF2 /ITO intensity ratio (upper graph) and vs. estimated MgF2
concentration (lower graph). Sample thickness 400 A on glass
The sheet resistivity of several films measured at room temperature is plotted vs.
time since deposition in figure 7 (glass substrates) and figure 8 (Upilex substrates).
All samples were stored in ambient air.
As noted by the authors of [2], sheet resistivity generally increases over time,
probably due to absorption of atmospheric oxygen. However, the stability is sim-
ilar to that of [2] even though our films are somewhat thinner. In addition, we
note that stability of our films on Upilex substrates appears to be similar to that of
films on glass. Some films in figures 7 and 8 show unusually large resistivity in-
creases. The reason for this is unknown, and the films microstructure has not been
examined.
TRANSPARENT ARCPROOF PROTECTIVE COATINGS 337
// )
1.E+09
Sheet Resistivity (
1.E+08 Run 325
1.E+09
1.E+08 Run 326
1.E+07
1.E+06
1.E+05
1.E+04
1.E+03
1.E+02
0 10 20 30
Vol% MgF2
Figure 6. Data similar to figure 5, but showing poor correlaton of sheet resistivity with PEM and
QCM data
1.E+11
1.E+10
// )
1.E+09
Sheet Resistivity (
1.E+08
1.E+07
1.E+06
1.E+05
1.E+04
1.E+03
0 50 100 150
Age (days)
Figure 7. Sheet resistivity vs. time since deposition (glass substrates). Samples stored in ambient
air
the project, that does not affect the remainder of this paper. It appears impossible
to correct for this error in a reliable manner, so we have not attempted to do so.
The data of figure 9 for zero air flow suggest an abrupt increase in sheet
resistivity at a critical MgF2 concentration, the origin of which is unknown. It
1.E+10
Sheet Resistivity (
1.E+09
1.E+08
1.E+07
1.E+06
1.E+05
1.E+04
1.E+03
0 100 200 300 400 500 600
Age (days)
Figure 8. Sheet resistivity vs. time since deposition (Upilex substrates). Samples stored in ambient
air
TRANSPARENT ARCPROOF PROTECTIVE COATINGS 339
Air Flow = 0
Air Flow = 0.1 SCCM
1.E+12
1.E+11
1.E+10
// )
1.E+09
Sheet Resistivity (
1.E+08
1.E+07
1.E+06
1.E+05
1.E+04
1.E+03
0 5 10 15 20 25 30 35 40 45 50
Figure 9. Sheet resistivity vs. MgF2 concentration with and without addition of air. Argon flow 0.9
SCCM in both cases
4. Conclusion
Acknowledgments
The authors gratefully acknowledge the support of NASA Glenn Research Center
Cooperative Agreements NCC3-740, NCC3-1023, NCC3-1033, and NCC3-1065.
We thank Bruce A. Banks, Joyce A. Dever, Thomas W. Kerslake, Craig H.
Marshall, and Deborah L. Waters for many helpful discussions.
References
1. Kerslake, T. W., Waters, D. L., Scheiman, D. A., and Hambourger, P. D. (2003) In 1st International
Energy Conversion Conference, Paper AIAA 2003-5919, Portsmouth, VA.
2. Dever, J. A., Rutledge, S. K., Hambourger, P. D., Bruckner, E., Ferrante, R., Pal, A. M., Mayer,
K., and Pietromica, A. J. (1998) NASA Technical Memorandum 1998-208499, August 1998.
3. Cashman, T., Kaur, J., Muhieddine, L. K., Shanbhag, M., Ubaid, S. H., Welch, B., Vemulapalli, J.,
and Hambourger, P. D. (2002), In Proceedings of ICPMSE-6. Toronto, Canada, 13 May, 2002,
Kluwer Academic Publishers, Dordrecht, The Netherlands, pp. 7380.
4. Cashman, T., Demko, R., Uppala, N., Vemulapalli, J., Welch, B., and Hambourger, P. D. (2003)
Vacuum Technology and Coating, September 2003, p. 38.
5. C. H. Marshall (private communication).
THE STUDY OF THE EFFECTS OF ATOMIC OXYGEN EROSION ON
THE MICROSTRUCTURE AND PROPERTY OF VO2
THERMOCHROMIC COATING USING CSAS SPACE SIMULATION
APPARATUS
Abstract. In this study, a thermochromic VO2 coating, which has been studied
for spacecraft smart thermal radiator application, was exposed to a ground-based
atomic oxygen flux to determine the potential effects of space environment on its
performance. The coating sample was prepared using laser ablation deposition
technique. Atomic oxygen exposure experiment was conducted on the samples
for equivalent 6 months and 3 years in typical LEO environment. Mass loss of the
coating samples due to atomic oxygen exposure was measured. Characterization
of total hemispherical emittance of the coating before and after atomic oxygen
exposure indicates that the atomic oxygen erosion affects the thermal-optical
performance of the coating to an extent. X-ray photoelectron spectroscope (XPS)
analysis of the coating samples was performed and an increase in the oxygen
concentration in outermost layer of the coating due to the atomic oxygen exposure
was identified. Possible mechanism for the change in thermo-optical property of
the coating was discussed.
Key words: smart material, thin-film coating, atomic oxygen, space environment
effects, thermochromic
1. Introduction
2. Experimental Procedure
In this study, two samples of VO2 coating were deposited on aluminum alloy
substrates using laser ablation deposition technic [4]. Metallic vanadium target
EFFECTS OF ATOMIC OXYGEN EROSION 343
was used. A pulsed laser beam generated by a XeCl Excimer laser at wavelength
of 308 nm and pulse duration of 20 ns is introduced into the deposition chamber
through a quartz window and focused by quartz lens on the target. The deposition
was performed in a controlled gas phase consisting of O2 and Ar gas mixture at
a pressure of 90 mTorr so that reaction between metallic vanadium atom vapor
and oxygen produce desired VO2 instead of other forms of vanadium oxide, such
as V2 O5 . The substrate is heated and maintained at 520 C to facilitate the thin
film crystalline growth. The final thickness of the coating after deposition is about
150 nm.
2.2. CHARACTERIZATION
X-ray diffraction (XRD) analysis of the coating sample was firstly performed us-
ing Philips diffractometer to determine its crystalline structure and particularly if
other form of vanadium oxide other than VO2 was formed during the deposition
process. A 6300 F scanning electron microscope (SEM) was used to observe the
microstructure of coating surface, such as the surface smoothness and crystal-
lization of the coating. An important objective of this study is to understand the
mechanism of the potential effects of atomic oxygen exposure on the microstruc-
ture of the VO2 thermochromic coating. For this purpose, the surfaces of the
coating samples before and after atomic oxygen exposure were specifically ana-
lyzed using ESCALAB 220I-XL X-ray photoelectron spectrometer to determine
if there would be changes in surface chemistry.
Total hemispherical emittance was measured using AZ Technologys reflec-
tometer (Temp 2000) on the coating samples before and after atomic oxygen
exposure. A series of such tests was performed at different temperatures ranging
from 25 to 60 C to determine the thermochromic behavior of the coatings.
The atomic oxygen exposure experiment on the coating samples was performed
using Canadian Space Agencys space simulation apparatus, as shown in fig-
ure 1. It is a RF plasma based atomic oxygen generator and is also equipped with
barometers, thermocouples, a flashlamp, an oxygen lamp, Krypton and Xenon
continuum lamps for generation of UV/VUV radiation, and an irradiance profiler.
However, the UV/VUV lamps were turned off during the exposure experiment due
to the reason to explicitly study the interaction between atomic oxygen and the
coating.
Vacuum is generated using a Danielson Tribodyn 200/57 three-module molec-
ular drag pump, with typical operating pressure at 100 mTorr inside the test
chamber for atomic oxygen flux to be produced, although before the introduc-
tion of oxygen for producing atomic oxygen, pressure on the order of 104 Torr
was achieved. Oxygen flow rate is set to 15 ccm1 . Previous experiment on the
344 XIN XIANG JIANG ET AL.
characterization of the facility indicated that the atomic oxygen source is of en-
ergy level of 0.1 eV. Two coating samples were used in atomic oxygen exposure
experiment. The exposure times were 2 and 12 h for the first and second samples
respectively. To determine the effective atomic oxygen fluence so as to determine
the equivalent LEO (low Earth orbit) exposure time, Kapton thin film was used
as witness sample during the atomic oxygen exposure experiment in accordance
with ASTM Standard E 2089-00. The change in thickness of the Kapton wit-
ness sample was used to calculate the effective fluence. The masses of the VO2
coating samples before and after atomic oxygen exposure were also measured in
accordance to ASTM standard in anaerobic chamber using a microbalance with
an accuracy of 0.002 mg.
Figure 2 shows the XRD pattern for VO2 coating sample deposited on aluminum
substrate. Because of the penetration of X-ray through the VO2 coating, the diffrac-
tion peaks from aluminum substrate are picked up and are relatively strong. VO2
diffraction peak is easily distinguished although it is relatively weak, indicating
probably imperfect crystallization.
Concerning the degree of surface degradation, the mass losses of the coating
samples after the atomic oxygen exposures were firstly measured. Table 1 lists the
parameters of atomic oxygen exposure experiment and the mass losses occurred
on two samples. Mass loss of the coating sample after equivalent 6 months LEO
exposure was found to be too small to be accurately measured (within the error
of microbalance). However, when the equivalent LEO exposure time increases to
EFFECTS OF ATOMIC OXYGEN EROSION 345
Sample no. #1 #2
3 years, mass loss of 0.018 mg was measured. Giving the density of VO2 of 4.34
gcm3 [13], the atomic oxygen reaction efficiency (Re ) of the VO2 coating can be
calculated according to the eq. (1) [14]:
Volume of material lost (M/)
Re = = (1)
F F
where
MMass loss of the sample (g)
FFluence of atomic oxygen (atomscm2 )
Density of the material (gcm3 )
It can be seen from the results that atomic oxygen reaction efficiency of VO2
coating is very low, in the same order of magnitude as that of high quality SiO2
Al VO2/Aluminum
700
600
Al
500
Intensity
400
Al
300
200 Al
VO2
100
0
10 20 30 40 50 60 70 80
2 (degree)
0.5
Semi-hemisphere emittance
0.4
0.3
0.2
Non AO exposed
0.1 Equivalent 3 years
LEO AO Exposure
0
20 30 40 50 60 70
Temperature (C)
coating (<0.8 1027 cm3 atom1 ), which is commonly used for protecting mate-
rial in atomic oxygen environment. This means the surface erosion of VO2 coating
is rather limited from the mass loss point of view. However, degradation or change
in the thermo-optical properties of the coating cant be ruled out because several
other factors, such as the change in surface texture and stoichiometry of VO2
coating, could still affect the thermo-optical properties of the coating.
A significant feature of the thermochromic characteristic of the VO2 coating
is that its thermo-optical properties change drastically within a relative narrow
temperature range, as confirmed by previous study [4, 11]. Figure 3 shows the
thermochromic behavior (change of thermal emittance with respect to the change
of temperature) of VO2 coating sample before and after equivalent 3 years LEO
atomic oxygen exposure. It can be seen that although the atomic oxygen exposure
does not alter the basic thermochromic behavior of the coating, it elevates overall
total hemispherical emittance and also broadens the transition temperature range
slightly. In a study by Dillon et al. [15] on VO2 coating deposited on single crystal
Si substrate using sputtering process, it was also found that plasma oxygen bom-
bardment tends to increase the thermal emittance of the coating although specific
causes could not be identified. It is noted the VO2 samples used in this study, which
are deposited on Al substrate by laser ablation process, are somewhat different
from those used by Dillon et al. in terms of thermo-optical property, probably
due to the difference in coating/substrate interfacial structure and deposition con-
ditions. However, one conclusion may be drawn from this experiment and the
study done by Dillon et al. [15] is that atomic oxygen can generally affect the
thermo-optical property of VO2 coating to an extent.
Characterization of the microstructure of the coating before and after atomic
oxygen exposure was performed. Figure 4 shows the SEM observations of the
surface of the coating samples before and after atomic oxygen exposure. By
EFFECTS OF ATOMIC OXYGEN EROSION 347
(a) (b)
Figure 4. SEM images of VO2 coating surface (a) before and (b) after equivalent 3 year LEO atomic
oxygen exposure
5,0 104
1,0 106
Intensity
5,0 104
1,0 106
VO2/ Al
Before AO irradiation
5,0 104
0,0
540 535 530 525 520 515 510 505
Binding energy (eV)
Figure 5. XPS analysis of the coating before and after equivalent 6 months and 3 years LEO/AO
exposures
crystalline structure. The defects reduce the energy bandgap for photon emission,
this is probably the main reason why thermochromic behavior (thermal emittance
with respect to temperature) of the coating is affected by atomic oxygen to an
extent.
4. Conclusion
broadening the transition temperature range. SEM and XPS analysis of coating
samples before and after the atomic oxygen exposure were performed to explore
the causes for such change. The analysis suggests that an increased density of
surface and crystalline structure defects, caused by atomic oxygen bombardment,
is probably the main reason responsible for the changes in thermo-optical property
of the coating.
References
1. Gilmore, D. G (Ed), (1994) Satellite Thermal Control Handbook, The Aerospace Corporation
Press, El Segundo, California
2. Darrin, A. G., Osiander, R., Champion, J., Swanson, T., and Douglas, D. (2002) Variable emis-
sivity through MEMS technology, In Proceedings of Space Technology and Application Inter-
national Forum 2000, M. S El-Genk (ed.), AIP Conference Proceedings, American Institute of
Physics Vol. 504, January 19, pp. 803808.
3. Tachikawa, S., Shimazaki, K., Ohnishi, A., Hirosawa, H., Shimakawa, Y., Ochi, A., Okamoto,
A., and Nakamura, Y. (1620 June 2003) In Proceedings of 9th International Symposium on
Materials in a Space Environment, Noordwijk, The Netherlands, European Space Agency, pp.
3447.
4. Soltani, M., Chaker, M., Haddad, E., Kruzelecky, V., and Nikanpour, D. (May/June 2004).
Optical switching of vanadium dioxide thin films deposited by reactive pulsed laser deposition,
Journal of Vacuum Science and Technology. A22 (3), pp. 859864.
5. Takahashi, I., Hibino, M., and Kudo, T. (October 1999) Thermochromic properties of double-
doped VO2 thin films fabricated from polyvanadate-based solutions, In Proceedings of SPIE,
Vol. 3788, pp. 2632.
6. Sahana, M. B., Subbanna, G. N., and Shivashankar, S. A. (2002) Phase transfor-
mation and semiconductor-metal transition in thin films of VO2 deposited by low-
pressure metalorganic chemical vapor deposition, Journal of Applied Physics 92, 6495
6504.
7. Lee, M. H., and Kim, M. G. (1996) RTA and stoichiometry effect on the thermochromism of
VO2 thin films, Thin solid films, 286, 219222.
8. Jerominek, H., Picard, F., and Vincent, D. (1993) Vanadium oxide films for optical switching
and detection. Optical Engineering 32, 2092p.
9. Partlow, D. P., Gurkovich, S. R., Radford, K. C., and Denes, L. J. (1991) Switchable vanadium
oxide films by a sol-gel process, Journal of Applied Physics 70, 443p
10. Kim, D. H., and Kwok, H. S. (1994) Pulsed laser deposition of VO2 thin films. Applied Physics
Letters 65, 3188p
11. Kruzelecky, R. V., Hadda, Jamroz, E., Soltani, M., Chaker, M., Nikanpour, D., and Jiang, X.
(July 2003) Passive Dynamically Variable Thin-Film Smart Radiator Device, In Proceeding of
2003 SAE Conference, 2003-01-2472, Vancouver, BC, Canada.
12. Parker, S. P. (1992) Encyclopedia of Chemistry, S. P. Parker (ed), 2nd ed., McGraw-Hill,
1152p.
13. Chudnovskiy, F., Luryi, S. (2002) Switching Device Based on a First-Order Metal-Insulator
Transition Induced by an External Electric Field, in Future Trends in Microelectronics:
the Nano Millennium, S. Luryi, J. M. Xu, and A. Zaslavsky, (ed), Wiley Interscience,
pp. 148155.
350 XIN XIANG JIANG ET AL.
14. Silverman, E. M. (August 1995), Space environmental Effects on Spacecraft: LEO Materials
Selection Guide, NASA contract report 4661, E. M. Silverman (ed.), part 1
15. Dillon, R. O., Le, K., Ianno, N. (2001) Thermochromic VO2 sputtered by control of a vanadium-
oxygen emission ratio, Thin Solid Films, 398399, pp. 1016
16. Kim, H. S., Lee, C. H., Lee, C. E., Kim, K. M., Noh, S. J., Hong, C. S., Hur, N. H., Shim, S.
Y., and Ri, H. C. (2001) Oxygen-plasma effects of a La0.7 Ca0.3 MnO3 single crystal, Applied
Physics Letters 79(25), 41774179.
DAMAGE KINETICS OF QUARTZ GLASS BY PROTON RADIATION
Abstract. The change in optical transmittance of JGS3 optical quartz glass was
studied using ground simulation for the space environment conditions of high
vacuum, heat sink, and 140 keV low energy protons and the kinetic model for
the evolution of color centers in the process of radiation damage was explored.
The experimental results show that radiation damage occurs in the surface layer
of quartz glass under large flux and low-energy proton radiation. The optical den-
sity change increases rapidly and then a saturation trend appears with increasing
absorption dose. A kinetic model for the evolution of color centers in quartz glass
irradiated with protons is proposed based on experimental data, from which the
change in optical density can be given. The model fitted curve is similar to ex-
perimental ones. It is believed that the proposed kinetic model can be used in the
quantitative description for the change in optical property of quartz glass with
increasing absorption dose under proton radiation with low energy.
Key words: quartz glass, protons radiation, color center, optical density
1. Introduction
Quartz glass has a number of excellent properties, and is extensively used as optical
windows and optical lens in spacecraft [1, 2]. The properties of optical elements
or materials would be degraded [3, 4] as the spacecraft flies in orbits subjected
to the radiation by protons and electrons with a wide energy spectrum of the Van
Allen radiation belts. In the past years, the study concerning radiation effects of
space charged particles on optical materials focused on the particles with energies
over 1 MeV. However, in the Van Allen radiation belt the flux of particles increases
with their energy decreasing. The protons and electrons with energies below 200
keV have large fluxes up to 108 particlescm2 s1 . The lower the particle energy
is, the relatively shorter is the penetration depth of the particle into materials. The
absorbed energy and the particles mainly concentrate in the surface layers of the
materials. Consequently, the effects of proton radiation with such low energies on
351
J.I. Kleiman (ed.), Protection of Materials and Structures from Space Environment, 351357.
C 2006 Springer. Printed in the Netherlands.
352 Q. WEI ET AL.
optical materials are important, especially for long-term mission spacecraft. The
present study is aimed at investigation of the change of spectrum properties of a
JGS3 quartz glass induced by irradiation with protons with energy of 140 keV in
vacuum environment and heat sink, and analyze the color center evolution kinetics.
2. Experimental
The experimental material is JGS3 optical quartz glass with the impurity level of
less than 5 103 %, in which OH groups are removed. The sample dimensions
are 20 20 2 mm, and its surfaces are polished. The average transmittance of
the samples in 2003200 nm region is 93%.
The radiation experiments were conducted in a system that can simulate the
radiation of protons with 30200 keV under 104 Pa vacuum and 77 K heat sink
environment.
The protons energy of 140 keV is chosen, the beam density is 0.2 Acm2 , and
the highest protons fluence is 2 1016 cm2 . The transmittance of the samples be-
fore and after irradiation was analyzed using a spectrophotometer of UV-3101PC
type made by SHIMADZU, Japan.
The change in transmittance spectrum of the JGS3 optical quartz glass irradiated
with the protons is shown in figure 1. With increasing fluence, the transmittance of
the quartz glass decreases, mainly in the near-ultraviolet to the visible region, with
little changes in the infrared region, as shown in figure 1(a). Figure 1(b) shows
that the obvious transmittance change occurs in the region of 220250 nm.
E centers are the primary radiation defects in the quartz glass [5, 6]. The
absorption peaks of the quartz glass at 220250 nm can be related to the E
Figure 1. The spectral transmittance (a) and transmittance change for JGS3 quartz glass before and
after protons irradiation with 140 ke V
DAMAGE KINETICS OF QUARTZ GLASS BY PROTON RADIATION 353
Figure 2. The change in optical density at 230 nm with the absorption dose for quartz glass after
140 keV protons radiation
centers and their variants [7, 8]. The optical density change of quartz glass at
230 nm is discussed in this paper. Generally the degree of optical absorption of a
material is indicated by the optical density, expressed as the logarithm of the ratio
of the transmitted light intensity T to the incident light intensity T0 , i.e.,
T
A = lg (1)
T0
The absorption dose and the radiation fluence are correlated with each other
by the following formula:
= 1.6 106 E/R (2)
Here is the absorption dose in Gy, E the particle energy in MeV, the radiation
fluence in protonscm2 , R the incidence particle penetration depth in material in
cm, and the material density in gcm2 .
Figure 2 shows the increase in optical density at 230 nm with the absorption
dose for the samples irradiated with 140 keV protons. When the absorption dose
reaches a critical value, the change in optical density tends to level off. This trend
might be related to a dynamic balance between the production and annihilation of
color centers induced by the proton radiation.
4.1. HYPOTHESES
Let J be the absorption dose rate, and n i the concentration of color centers at time
t. According to the hypotheses above, the growth rate of color centers is given
by
dn i
= J [ f i (Ni (x) n i ) n i qi ] (3)
dt
If the concentration of defect traps Ni is constant, the equation can be solved
as follows:
f i Ni (x)
n i = n i0 + n i0 1 e( fi +qi ) J t
f + qi
i
f i Ni (x)
= n i0 + n i0 1 e( fi +qi ) (4)
f i + qi
Based on assumption (3) above, n i = n i0 = 0 at t = 0, the concentration of
color centers can be given as
f i Ni (x) f i Ni (x)
ni = 1 e( fi +qi ) J t = 1 e( fi +qi ) (5)
f i + qi f i + qi
The eq. (5) presents a relation between the absorption dose and the average
concentration of color centers per unit area of surface, and reflects the accumula-
tion kinetic of color centers.
DAMAGE KINETICS OF QUARTZ GLASS BY PROTON RADIATION 355
It was found that the relation between the integral areas of absorption curve and
the oscillator strength and concentration of color centers could be expressed as
[9]
9mc nZ 1017 n Z
nf = m W = const m W (6)
2e2 h (n 2Z + 2)2 (n 2Z + 2)2
where n is volume concentration of color centers; e the unit charge; f the oscillator
strength; n Z the refractive index of media; m the electronic reduced mass of
color center; (E ) the absorption coefficient, and E the photon energy. m the
absorption coefficient of absorption peak, W the full width at half height.
According to the above equation, when W is constant, the concentration of
color centers is proportional to the absorption coefficient, i.e., n = . Meanwhile,
there is a relation of = 2.303
d
A between the absorption coefficient and optical
density, and then the relation between optical density and concentration of color
centers can be given by
0.434
A = 0.434d = dn = gdn (7)
where and g are scale factors. Because the effects of proton radiation with low
energy mainly focus on the materials surface, the color centers concentrate in
a very thin layer on the surface, and d is thickness of this layer. According to
assumption (3), n io = 0. The concentration of color centers is proportional to the
optical density change induced by radiation, i.e.
Ai = Ai Ai0 = dg (n i n i0 ) = dgn i = Ai (8)
Then, according to eqs. (5), (7), and (8), the optical density change can be
given by
f i Ni (x)
Ai = dg 1 e( fi +qi ) J t
f i + qi
f i Ni (x)
= dg 1 e( fi +qi ) = Ais (1 e ) (9)
f i + qi
Parameters
Figure 3 shows the results of mathematical fit to the experimental curve. The
parameters are listed in table 1. Comparing the experimental curve with the fitted
one, it can be suggested that the functions fitted from the kinetic model of color
centers can approximately describe the optical property change for the quartz glass
under proton radiation with low energy.
5. Conclusion
Space proton irradiation with low energy is characterized by a high flux and results
in obvious radiation damage effects on the surface layer of quartz glass. Initially,
the change in optical density increases rapidly, and then a saturation trend appears
with increasing absorption dose.
Based on the analysis of the experimental results, a kinetic model for the
evolution of color centers in quartz glass irradiated with protons is proposed.
The change in average optical
density per unit area of surface can be given by
Ai = Ais 1 e . The mathematic fitted curve is similar to the exper-
imental one. We believe that the proposed kinetic model can be used in the
Figure 3. The fitted curves for the change in optical density of quartz glass at 230 nm under proton
radiation
DAMAGE KINETICS OF QUARTZ GLASS BY PROTON RADIATION 357
References
1. Introduction
When space vehicles travel in low Earth orbit (LEO) or geosynchronous Earth
orbit (GEO), their surfaces are under constant bombardment by energetic
atoms/molecules, ions, electrons, and various sources of electromagnetic radi-
ation. Among the collision processes, atomic/molecular ions are responsible for
the space vehicle charging and atmospheric drag. Under the LEO and GEO envi-
ronments, ions may have kinetic energies ranging from 1 to 106 eV [1, 2]. As a
consequence, various chemical processes that carry mass away from the surfaces
can be induced and enhanced due to ion/surface collisions that cause chemical
reaction, energy transfer, and surface ionization. Therefore, it is important to
understand the ion/surface erosion mechanisms at the microscopic level and to
evaluate the relative importance of ions compared to neutrals (such as atomic
oxygen) in producing erosion. Although there have been extensive studies on the
O(3 P) erosion mechanisms of the surface materials of spacecraft [3], surprisingly,
the microscopic reaction mechanisms of the O+ (4 S) ion, the most abundant ionic
species in LEO, are mostly unknown [4]. Recent experimental work has been
carried out for O+ reaction with small alkane molecules in the gas phase us-
ing the guided-ion beam time-of-flight (TOF) technique[5] and for O+ reaction
359
J.I. Kleiman (ed.), Protection of Materials and Structures from Space Environment, 359364.
C 2006 Springer. Printed in the Netherlands.
360 LIPENG SUN AND GEORGE C. SCHATZ
2. Results
Under the LEO environment the kinetic energy of O+ is about 5eV relative to
the traveling space vehicle. With this energy, a large number of reactions may be
energetically accessible. To evaluate the reaction energies, the ab initio electronic
structure calculations were performed by using the quantum chemistry software
package GAMESS [7]. It has been shown from previous studies that for ion-
molecule reactions the MP2/6-31G level of calculation gives reasonable ener-
getics and geometries. The calculated structures and energies are shown in figure 1.
Charge transfer can be viewed as a simple type of chemical reaction in which
one or several electrons move from one chemical species to another. For the
[O rCH4 ]+ system, charge transfer, O+ + CH4 CH4 + + O, is exothermic and
can be described by a near-resonant electron transfer mechanism. In this sce-
nario, charge transfer occurs at primarily large impact parameters with negligible
momentum transfer between the colliding species. The recently measured exper-
imental charge transfer cross section supports this picture [8]. The large charge
transfer cross section (about 80A 2 ) allows us to model the subsequent chemical
processes by studying the O + CH4 + collision system. The equilibrium structures
of the charge transfer intermediate CH4 + have been extensively studied in the past
twenty years [912]. Its symmetry, due to JohnTeller effects, is reduced from Td
to C2v , D2d and C3v with C2v as its most stable structure as depicted in figure 1.
After charge transfer, many processes can take place on the several accessi-
ble electronic states. On the lowest quartet electronic state, the primary products
that are thermodynamically favorable are found to be CH3 + OH+ , H2 CO+ +2H,
H2 O+ +CH2 , and CH3 O+ + H. All of these reactions are exothermic. In addition,
when O approaches the CH4 + ion there is a reaction intermediate formed, and
the stable structure due to the ion-molecule electrostatic interaction, as shown in
figure 1, is the configuration formed when O approaches the vertex of CH4 + . The
MICROSCOPIC MECHANISMS AND DYNAMICS SIMULATIONS 361
+ CH3+ + H + O
O+ + CH4
0.0 OH+ + CH3
+ H2CO+ + 2H
2.0 + CH3O+ + H
+
O
CH2+ + H2O
CH3+ + OH
4.0
H2CO+ + H2
6.0
OH
+
8.0
CH4 moiety of the reaction intermediate is very close to the C3v symmetry of
CH4 + with the charge located on the CH4 moiety. It is interesting to notice that
there is no reaction barrier found from [H3 CHO]+ CH3 + OH+ which indi-
cates that direct dissociation should readily occur. With such a strong electrostatic
interaction (1.57 eV) between CH4 + and O, a reaction mechanism is expected in
which O and CH4 + first form the [H3 CHO]+ complex and then this is trapped in a
deep potential energy well for a certain period of time and then this either directly
dissociates to OH+ + CH3 or undergoes further chemical reactions such as OH+ +
CH3 CH2 + H2 O+ . This reaction mechanism should be especially important
for low collision energies where all the above exothermic reactions are expected
to occur. With increase in the colliding energy, however, this complex mediated
mechanism becomes less and less important. Instead, a direct reaction mechanism
becomes more important. For example, for the [H3 CHO]+ CH3 + OH+ re-
action, the incoming O atom can directly grab one H atom from CH4 + without
exchanging much of its kinetic energy. This mechanism is usually referred as a
stripping mechanism. There are also endothermic reactions, i.e., CH3 + + H + O
and CH2 + + H2 +O possible on the quartet potential energy surface. Since the
available energy for the reaction (5eV) is much larger than the dissociation
362 LIPENG SUN AND GEORGE C. SCHATZ
threshold (0.8 eV for CH4 + CH3 + + H and 1.7 eV for CH4 + CH2 + +
H2 ), these reactions occur with a collision induced dissociation mechanism in
which the CH+ 4 cation received enough internal energy from the collision with O
to dissociate into CH3 + + H or CH2 + + H2 .
The reactions on the electronic doublet potential energy surface are highly
exothermic. Also, a reaction intermediate, formed by O atom insertion to the CH
bond, is found. The major product ions are CH2 + , CH3 + , and H2 CO+ . It is worth
noting that at low collision energies CH3 + can only be produced on the doublet
surface which implies that intersystem crossing to the doublet state is required.
Although spin-orbit coupling is usually weak, at low collision energies, when the
system is trapped as a reaction intermediate, the crossing seam can be accessed
multiple times before dissociation, so that the transition from quartet to doublet
states is enhanced.
Since a majority of the energetically accessible reaction channels are exother-
mic, at high impact energies, the primary products such as CH3 O+ and H2 CO+
can further isomerize to H2 COH+ and HCOH+ or dissociate directly to HCO+ .
These products are also shown in the figure 1.
TABLE 1. The reaction products and their branching fractions, based on PM3 calculations on the
lowest quartet potential surface
Branching
fraction 12.4% 7.0% 12.4% 29.5% 2.3% 24.8% 6.2% 5.4% <1%
MICROSCOPIC MECHANISMS AND DYNAMICS SIMULATIONS 363
3. Conclusion
Acknowledgment
References
1. Introduction
Current address: Department of Chemistry, Virginia Tech. 107 Davidson Hall, Blacksburg, VA, 24061-0212
USA
365
J.I. Kleiman (ed.), Protection of Materials and Structures from Space Environment, 365375.
C 2006 Springer. Printed in the Netherlands.
366 DIEGO TROYA AND GEORGE C. SCHATZ
the F : C ratio of the eroded sample decreases 20% with respect to that of the
pristine film, in agreement with the set of experiments by Rasoul et al. [14] and
Nakayama et al. [15]. However, this last set of experiments disagrees with the
other two in the fact that oxygen was not detected in the surface of the irradiated
polymers [16]. Irrespective of this, the small decrease of the F : C ratio in all of
the experiments suggests that the mechanisms of mass loss are not dominated by
processes in which single F atoms are abstracted or knocked-out from the polymer,
as these processes would give rise to more noticeable decreases in the F : C ratio.
The unclear origins of the superior performance in LEO of fluorinated alkane
polymer with respect to unsubstituted alkanes and the disagreement between dif-
ferent experiments indicate the need for further studies. In this paper, we present
a theoretical study of the reactions of O(3 P) with fluorinated alkanes based on
quantum-mechanical (QM) calculations and molecular dynamics simulations.
First-principles QM calculations are developed to identify all of the relevant
pathways that can lead to degradation of fluorinated alkanes in LEO environ-
ment. Classical trajectory calculations that use a high-accuracy semiempirical
QM Hamiltonian are carried out to simulate the collision of O(3 P) with fluori-
nated species in LEO and derive information about the relative importance of the
different open reaction channels and their dynamics properties. In order to gain ad-
ditional insight, the results of the simulations developed here are compared with
earlier calculations of the reactions of hyperthermal O(3 P) with unsubstituted
alkanes [13, 1721].
TABLE 1. Calculated reaction energies and barriers (in eV) for the O(3 P) + CF4 , CH4 , C2 F6 , and
C2 H6 reactionsa
Reaction Energy
O(3 P) + CX4 (X = F, H)
OF + CF3 OH + CH3 F + OCF3 H + OCH3
2.966 0.161 1.168 0.548
O(3 P) + C2 X6 (X = F, H)
OF + C2 F5 OH + C2 H5 F + OC2 F5 H + OC2 H5 OCF3 + CF3 OCH3 + CH3
2.549 0.061 0.831 0.368 0.534 0.234
Reaction Barrier
O( P) + CX4 (X = F, H)
3
Figure 1. Schematic representation of the structures and energy profile of the primary reaction
pathways involved in O(3 P) + C2 F6 collisions at hyperthermal energies
THEORETICAL STUDY OF REACTIONS OF HYPERTHERMAL O(3 P) 369
is approximately thermoneutral and proceeds through a barrier well below 0.5 eV,
more than 2.5 eV are needed to generate OF in O(3 P) collisions with fluorinated
alkanes. O and F are very electronegative species, and the formation of a bond
between them is not favored.
It can be also observed that the F elimination reaction (which yields
F + OCn F2n+1 ) is about 0.5 eV more endoergic than for the corresponding H
elimination reaction. Also, the barriers for F elimination are 1.2 eV larger than
those for H elimination. The products generated in the CC breakage reaction of
O(3 P) with C2 F6 are thermodynamically more stable than those of the analogous
reaction with C2 H6 . However, perfluorination leads to a 1 eV increase in the CC
breakage barrier height.
The fact that CF bonds (dissociation energy, De = 4.93 eV at the B3LYP/6-
31+G* level for C2 F6 ) are slightly stronger than CH bonds (De = 4.64 eV at the
same level of theory for C2 H6 ) does not fully account for the large differences in
the barrier heights between fluorinated and unsubstituted alkanes noted in table
1 for all of the primary processes. We have just discussed the reasons for the
unfavourable formation of OF. As it can be seen in the saddle point structures in
figure 1, F elimination or CC breakage are produced when the incoming O(3 P)
atom forms a bond with the carbon atoms of the alkane backbone. In the case of
unsubstituted alkanes, H atoms bonded to the hydrocarbon chain do not hinder
significantly the approach of O(3 P) to the C atoms of the chain. However, the
larger size of F atoms induces repulsive interactions with the electronic cloud of
the incoming O(3 P). This shielding effect makes it more difficult for O(3 P) to
reach the C atoms in fluorinated alkanes and leads to an increase in the reaction
barriers for F elimination and CC breakage that is larger than what would be
expected based on bond-strength arguments alone.
The CC breakage saddle points for O(3 P) + perfluorinated alkane reactions
have been studied before by Massa and coworkers [22]. However, it should be
noted that the saddle point reported in that work for the O(3 P) + C2 F6 reaction
is not the lowest energy one. Instead, it corresponds to a higher-energy reaction
pathway.
2
3
O( P) + CF4 -> F + OCF3 (F elimination)
3
O( P) + CH4 -> H + OCH3 (H elimination)0.1
cross section / a.u. 1.5
0.1
0.5
0
4.5 5 5.5 6
Ecoll / eV
Figure 2. Excitation functions (cross sections vs Ecoll ) for the O(3 P) + CX4 X + OCX3 (X = F,
H) reactions. The cross sections of the H elimination reaction are divided by 10
elimination are very small. We carried out additional calculations to locate the
threshold of reactivity for this channel, and 1 reactive trajectory was found out
of 10,000 trajectories calculated at E coll = 4.0 eV. This suggests that the lowest
collision energy for F elimination is about 1 eV larger than the energy of the saddle
1
3
O( P) + C2F6 -> OCF3 + CF3 (C-C breakage)
3
0.8 O( P) + C2F6 -> F + OC2F5 (F elimination)
3
O( P) + C2H6 -> OCH3 + CH3 (C-C breakage)0.1
cross section / a.u.
3
O( P) + C2H6 -> H + OC2H5 (H elimination)0.01
0.6
0.4 0.1
0.2 0.01
0
4.5 5 5.5 6
Ecoll / eV
Figure 3. Excitation functions (cross sections vs Ecoll ) for the O(3 P) + C2 X6 X + OC2 X5 (solid
lines) and the O(3 P) + C2 X6 OCX3 + CX3 (dotted lines) (X = F,H) reactions. The cross sections
of the X = H reactions have been divided by 100 and 10, respectively
372 DIEGO TROYA AND GEORGE C. SCHATZ
point for this process (3 eV, see table 1). Although the ROHF/MSINDO barrier
for F elimination in the O(3 P) + C2 F6 reaction is about 0.5 eV smaller than that in
O(3 P) + CF4 , the size of the F elimination cross section in O(3 P) + C2 F6 remains
quite small. In fact, the F elimination cross section in O(3 P) + C2 F6 is much
smaller than that in O(3 P) + CF4 (a factor of 6 at E coll = 6 eV). The threshold
for reactivity is also well above the minimum energy saddle point located for this
reaction pathway. Thus, much as it happens for the F abstraction reaction, the fea-
tures of the potential energy surface are such that there is a dynamical barrier to
reaction induced by the inefficacy of collision energy to explore geometries close
to those of the minimum energy saddle point. F abstraction or elimination reac-
tions require a large elongation of the CF bond that breaks. Whereas vibrational
excitation of the CF bond would promote reactivity, a fast O(3 P) atom is not
able to efficiently promote the CF stretching needed for reaction. In addition, in
collisions in which a substantial fraction of the momentum of the incoming O(3 P)
is transferred to CF stretching, motions other than those leading to reaction are
strongly coupled and can inhibit reaction, relaxing the instantaneous excitation
of the CF stretch to other modes. This is known as frustrated collisions and
has been described before in studies of H with vibrationally excited HF and H2 O
[2426].
The cross sections for the CC breakage reaction, O(3 P) + C2 F6 OCF3 +
CF3 , are displayed in figure 3 in comparison with the results for O(3 P) + C2 H6 .
No attempts to accurately locate the threshold for reactivity were made in this
case due to the large computational expense involved. However, it can be inferred
from the small cross section found at E coll = 4.5 eV, that the minimum collision
energy leading to CC breakage is well above the energy of the minimum-energy
saddle point located for this reaction. Therefore, dynamical barriers seem to be
present in all of the reactions of O(3 P) with perfluorinated alkanes investigated in
this work.
It is remarkable that the CC breakage cross sections in the O(3 P) + C2 F6
reaction is larger than the F elimination cross section. This behavior is opposite to
that of unsubstituted alkanes [13, 18, 20, 21]. For instance, in the O(3 P) + C2 H6
reaction at E coll = 5.75 eV, the H elimination: CC breakage cross-section ratio is
4.00 (16.64 : 4.16 a.u.). However, in the O(3 P) + C2 F6 reaction the F elimination:
CC breakage cross section ratio is 0.12 at E coll = 5.5 eV and 0.24 at E coll =
6.0 eV. The H elimination and CC breakage reaction barrier heights in O(3 P) +
C2 H6 are very similar. Thus the preference for H elimination over CC breakage
comes from the fact that when O(3 P) strikes one of the carbon atoms of the ethane
molecule, 3 CH bonds can break leading to H elimination, but only 1 CC bond
can break leading to CC breakage. In O(3 P) + C2 F6 the barrier height for CC
breakage is slightly smaller than that of F elimination. Also, the CC bond is in
this case the most labile bond in C2 F6 (De = 3.66 eV, compared to De = 4.93 eV
for a CF bond at the B3LYP/6-31+G* level) and this is why CC breakage is
THEORETICAL STUDY OF REACTIONS OF HYPERTHERMAL O(3 P) 373
4. Concluding Remarks
Our calculations indicate that fluorinated alkanes are much less susceptible to
reaction than hydrogenated alkanes upon collisions with hyperthermal O(3 P),
with the O(3 P) + C2 H6 / O(3 P) + C2 F6 ratio of total reactive cross sections at
5.0 eV being 39.3/0.14 = 270. This is to be compared with the EOIM-III on-orbit
experimental measurements that reported a reaction efficiency for polyethylene
88 times larger than that of TEFLONR FEP [6]. Our calculations are also in
agreement with the gas-phase crossed-beams experiments of the Minton group
that indicate no oxygen-containing products in the O(3 P) + C3 F8 reaction at
E coll = 4.6 eV [12], but substantial reactivity for hyperthermal collisions of O(3 P)
with C3 H8 [13]. The reasons for a decrease in the reactivity of alkanes with respect
to hyperthermal O(3 P) upon fluorination are manifold
CF bonds in fluorinated saturated alkanes are 10% stronger than CH bonds
in unsubstituted species.
The barriers for CC breakage or F elimination in fluorinated alkanes are about
1 eV larger than the corresponding barriers in unsubstituted alkanes as a con-
sequence of the larger size of F atoms, which inhibits the approach of O(3 P)
to the carbon backbone,
H atom abstraction by OH is dominant in O(3 P) + alkane reactions, but is
negligible in O(3 P) + fluorinated alkane reactions.
Additional dynamical barriers arise due to the inefficacy of collision energy to
provide geometries close to those of the first-order saddle points for reaction.
The results of our calculations can also contribute to the understanding of
XPS experiments of surface composition after hyperthermal atomic oxygen ex-
posure [1416]. In all of these experiments it was found that the F : C ratio of the
374 DIEGO TROYA AND GEORGE C. SCHATZ
irradiated polymer was only slightly smaller than that of the pristine polymer. This
suggests that mass removal could involve ejection of CF2 units into the gas phase.
This is in contrast with the microscopic mechanisms characterized in erosion ex-
periments of unsubstituted hydrocarbon polymers by O(3 P), in which CO, CO2
and H2 O are the volatile species [27]. Our calculations show that O addition to
fluoropolymers is only minor and the dominant erosion process is CC breakage.
Therefore we can tentatively infer that erosion of fluorinated polymers mainly
involves CC breakage with formation of volatile OCF2 products that scatter into
the gas phase. Although OCF species have not been found in O(3 P) + C3 F8 colli-
sions at E coll = 4.6 eV [12], larger collision energies in the beams that have been
used to study TEFLONR erosion, together with the presence of UV light or ions
and defects in the exposed fluoropolymers would facilitate OFC species forma-
tion. Future calculations involving condensed-phase fluorinated alkanes, such as
self-assembled monolayers, will help to verify this. Time-of-flight detection of
the scattered species in a hyperthermal O(3 P) steady-state bombardment exper-
iment of TEFLONR would directly provide information about the microscopic
mechanisms of fluoropolymer erosion in LEO.
Acknowledgments
References
1. de Groh, K. K., Gaier, J. R., Hall, R. L., Espe, M. P., Cato, D. R., Sutter, J. K., and Scheiman,
D. A. (2000) Journal of High Performance Polymers 12 (1), 83.
2. Dever, J. A., de Groh, K. K., Banks, B. A., Townsend, J. A., Barth, J. L., Thomson, S., Gregory,
T., and Savage, W. (2000) Journal of High Performance Polymers 12 (1), 125.
3. de Groh, K. K., Dever, J. A., Sutter, J. K., Gaier, J. R., Gummow, J. D., Scheiman, D. A., and
He, C. (2001) Journal of High Performance Polymers 13, S401.
4. Dever, J. A., de Groh, K. K., Messer, R. K., McClendon, M. W., Viens, M., Wang, L. L., and
Gummow, J. D. (2001) Journal of High Performance Polymers 13, S373.
5. Jursa, A. (1976) U. S. Standard Atmosphere. (U.S. Government Printing Office, Washington D.
C., 1976).
6. Koontz, S. L., Leger, L. J., Visentine, J. T., Hunton, D. E., Cross, J. B., and Hakes, C. L. (1992)
Journal of Spacecraft and Rockets 32(3), 483.
7. Leger, L. J. and Visentine, J. T. (1986) Journal of Spacecraft and Rockets 23(5), 505.
8. Koontz, S. L., Albyn, K., and Leger, L. J. (1991) Journal of Spacecraft and Rockets 28(3), 315.
9. Cazaubon, B., Paillous, A., and Siffre, J. (1998) Journal of Spacecraft and Rockets 35(6), 797.
THEORETICAL STUDY OF REACTIONS OF HYPERTHERMAL O(3 P) 375
10. Wagner, A. J., Wolfe, G. M., and Fairbrother, D. H. (2004) Journal of Chemical Physics 120(8),
3799.
11. Garton, D. J., Minton, T. K., Maiti, B., Troya, D., and Schatz, G. C. (2003) Journal of Chemical
Physics 118, 1585.
12. Minton, T. K., private communication (2004).
13. Garton, D. J., Minton, T. K., Troya, D., Pascual, R., and Schatz, G. C. (2003) Journal of Physical
Chemistry A 107(23), 4583.
14. Rasoul, F. A., Hill, D. J. T., George, G. A., and ODonnell, J. H. (1998) Polymers for Advanced
Technologies 9, 24.
15. Nakayama, Y., Imagawa, K., Tagashira, M., Nakai, M., Kudoh, H., Sugimoto, M., Kasai, N.,
and Seguchi, T. (2001) Journal of High Performance Polymers 13, 2001.
16. Gonzalez, R. I., Phillips, S. H., and Hoflund, G. B. (2003) Journal of Applied Polymer Science
92 (3), 1977.
17. Troya, D., Pascual, R. Z., and Schatz, G. C. (2003) Journal of Physical Chemistry A 107 (49),
10497.
18. Troya, D., Pascual, R. Z., Garton, D. J., Minton, T. K., and Schatz, G. C., Journal of Physical
Chemistry A 107(37), 7161.
19. Troya, D., Schatz, G. C., Garton, D. J., Brunsvold, A. L., and Minton, T. K. (2004) Journal of
Chemical Physics 120(2), 731.
20. Troya, D. and Schatz, G. C. (2003) European Space Agency Special Publication SP-540, 121.
21. Troya, D. and Schatz, G. C. (2004) Journal of Chemical Physics 120, 7696 .
22. Gindulyte, A., Massa, L., Banks, B. A., and Miller, S. K. R. (2002) Journal of Physical Chemistry
A 106, 5463.
23. Ahlswede, B. and Jug, K. (1999) Journal of Computational Chemistry 20(6), 563.
24. Schatz, G. C. and Kuppermann, A. (1980) Journal of Chemical Physics 72(4), 2737.
25. Lendvay, G., Bradley, K. S., and Schatz, G. C. (1999) Journal of Chemical Physics 110, 2963.
26. Barnes, P. W., Sims, I. R., Smith, I. W. M., Lendvay, G., and Schatz, G. C. (2001) Journal of
Chemical Physics 115, 4586.
27. Minton, T. K. and Garton, D. J. (2001) In Advances Series in Physical Chemistry Vol 11: Chemical
Dynamics in Extreme Environments, R. A. Dressler (ed.) (World Scientific, Singapore, 2001),
pp. 420.
SIMULATION OF UV INFLUENCE ON OUTGASSING
OF POLYMER COMPOSITES
R. H. KHASSANCHINE, A. N. GALYGIN,
A. V. GRIGOREVSKIY, AND A. N. TIMOFEEV
Joint-stock company Kompozit, 4, Pionerskaya str., Korolev,
Moscow region, Russia
1. Introduction
377
J.I. Kleiman (ed.), Protection of Materials and Structures from Space Environment, 377383.
C 2006 Springer. Printed in the Netherlands.
378 R. H. KHASSANCHINE ET AL.
2. Mathematical Model
The outgassing rate of VP source and deposition and re-emission rates on/from
the condensation surface respectively are defined mainly by the constants in eqs.
(7) and (8).
The total mass of VP deposited on the condensation surface can be determined
by summing masses of individual components:
N
Mc total (t) = [Mcn (t) +
n (t, cn )], (9)
n=1
where
n (t, cn )n-type part of VP mass that has chemically reacted on the
condensation surface by time t; N number of VP components.
4. Experimental Set-Up
Figure 2 shows a schematic diagram of the Vesy equipment that was used
to study influence of UV radiation on outgassing kinetics at different sample
temperatures.
7
70 mm 4
170 mm
5
6
Figure 2. Schematic diagram of the Vesy equipment: (1)reservoir that is filled with liquid nitro-
gen to control thermostatically the mass flow sensor; (2)vacuum chamber body; (3)cryogenic
shield; (4)mass flow sensor; (5)heating table with sample; (6)UV sources; (7)quartz
window
382 R. H. KHASSANCHINE ET AL.
Figure 4. Experimental curves of outgassing kinetics: (a; c)ECOM-1; (b)model material with
equal binder/filling agent volumes; (d)influence of actuation of UV source at t = 24 h on outgassing
rate
the basic parameters characterizing the material itself and on the emission source
spectrum.
Experimental data show that the outgassing rate depends on the temperature
of the material and the volume ratio of the polymeric binder/filling agent in the
case when polymeric materials are subjected to UV radiation.
References
1. Khassanchine, R. H., Grigorevskiy, A. V., and Galygin, A. N. (2002) Some aspects of simulation
of outgassing processes under thermal vacuum exposure to coatings applied to space vehicles,
Protection of Materials and Structures from Space Environment, ICPMSE-6, Toronto-Canada,
May 2002.
2. Kostiuk, V. I. and Khassanchine, R. H. (2002) Journal of Cosmonautics and Rocket Engineering,
28, 155163.
3. Grigorievski, A., Gordeev, J., Gurov, A., Kiseleva, L., and Shuiski, M. (2000) In Proceedings
8th International Symposium, Material in a Space Enviroment and 5th International Conference,
Protection of Materials and Structures from Space Environment, Arcachon, France, 59 June,
2000 pp. 8187.
THE IMPACT OF HIGH-VELOCITY PARTICLES ON THERMAL
PIPELINES IN SPACECRAFT
1. Introduction
During long-duration space flights spacecraft components and systems are ex-
posed to the deep space factors. High-velocity particles (space debris, microme-
teorites) present the greatest danger to high-temperature thermal pipelines (TP),
which are important elements of the on-board installation. Small particles lead to
changes in optical characteristics, while large ones cause depressurization of TP
with a resulting release of the heat-carrier into open space, with the contamination
of the space environment becoming one of dangerous consequences.
The performance of Thermal Pipelines and their diagnostics (control of func-
tional integrity and TP emission power) are discussed in this paper together with
experiments simulating high-velocity particle interaction with Metal-Insulator-
Metal-Semiconductor (MIMS) structures being a part of the TP wall design. Ex-
periments simulating the high-velocity particle interaction with TP using various
types of accelerators were conducted and are described in this paper.
where M is the mass of the heat-carrier, j is the density of the vapour stream.
Assuming the section of the hole is S0 = 102 cm2 , then te 2 hours.
Let us estimate the time of TP cooling after the escape of the heat-carrier,
assuming the stream from the heat source q = 0. Then the equation describing TP
cooling will look like:
d CmT
= T 4 S, (2)
dt
where Cm is an average thermal capacity of a niobium TP wall, m is the mass
of Niobium, S is the area of the lateral surface. The initial condition for eq. (2)
is: at t = 0, T = 1150 0 K. Integrating eq. (2), we shall obtain the dependence of
THE IMPACT OF HIGH-VELOCITY PARTICLES 389
Figure 4. Current vs. Temperature dependence for TP with various black covers
T0
T (t) = , (3)
3t
3
1+
t0
where t0 = CmT 0
T0 4 S
is a typical time of cooling.
During TP cooling, the resistance Rbc of the black cover and Rd of the di-
electric (figure 4) are increased. Figure 4 shows experimental dependence of
the dielectric and black cover currents Ibc and IL of a thermal pipe model on
temperature.
Since the temperature dynamics during the vapour escape (T = T0 = const)
and after the escape are different and te t0 , it is possible to determine the time of
depletion of the heat-carrier mass based on the change of the conductivity current
of MIM-structure in time, ddtId .
The beginning of the escape process is determined by the fact of closure-
restoring MIM-structure, and the value Sh is evaluated from formula (1).
Given the size of the hole, and using models of open penetration of a MIM-
structure, the parameters of a high-velocity particle are determined. The process
of vapor escape with its condensation on the walls of the hole channel, can be
controlled by burning of a conducting jumper. The moment of full depletion of
the heat-carrier is determined by the moment when the jumper in the created
penetration hole disappears.
In this model a probability remains of an incomplete depletion of the heat-
carrier mass.
390 SEMKIN N. D. ET AL.
The third model assumes that liquid sodium flows from a TP wick into the
hole. In this case it is likely that higher breakthrough voltage will be necessary for
restoring a MIM-structure. Knowing the value of the breakthrough voltage and
taking into account the model of formation of the conducting jumper, it is possible
to identify the loss of hermeticity in the TP, as a result of its perforation, as well
as to restore the information about the penetration hole size. Let us calculate the
resistance of a MIM-structure dielectric in an operating mode (T = 900 C) from
knowing the dependencies between TP temperature along the length Z of a pipe
T = T (Z), and the specific conductivity on temperature = (T ). The combined
conductivity of the whole TP is defined by the following expression:
h
2 r
G= (Z ) d Z , (4)
h
0
where r is the radius of the pipe, h is the thickness of dielectric (h T). The eq.
(4) with allowance for the above dependencies will be transformed to:
T
2 r (Z ) d Z
G= dT .
h dT
To
2r l0 1
If = const = 0 , then G = , or R = .
h G
Let us evaluate the electrical energy power necessary to burn the short-circuit
jumper:
2
E
P= Rcc .
Rcc + Rl
where Rl is load resistance, Rcc is short-circuited jumper resistance. The energy
necessary for evaporation of the heat-carrier mass can be expressed as:
Ee = Cm(T e T ) + qm
where Cm is mass thermal capacity, qm is the heat of vaporization, and Te is the
temperature of evaporation.
4. Experimental Results
mm and a black-cover (spinel h = 0.1 mm), the perforating velocity for a particle
of this size is estimated as 5.3 kms1 . Figure 5 (the left view) shows a crater
formed by a particle with velocity 2.5 kms1 and a full penetration ( the right
view). Perforation of the black cover is observed on a large area as a result of
an unloading wave on the TP surface. For a TP wall, heated up to T = 900 C,
penetration velocity of a particle decreases to 4.8 kms1 .
The influence of fine particle fluxes on optical properties of TPs was investi-
gated experimentally using an explosion type accelerator [1], and an electrostatic
accelerator over the range of masses 5 1012 1015 kg [2] with particle mate-
rials chosen from: C (graphite, black soot), Al2 O3 , Al, TiC, Fe, W. The velocity
range of particles was 1.58 kms1 . The particle density in all experiments was
within 0.10.65 gm2 . Experiments demonstrated that the absorption coefficient
As is considerably increased from 0.15 up to 0.430.6, and blackness factor E is
changed slightly from 0.9 up to 0.92. Lanthanum manganate, lanthanum chromide
and spinel were used as black covers for TP.
5. Conclusions
1. The absorption coefficient of the black surfaces of the investigated TPs was
found to increase substantially from 0.15 to 0.60 upon interaction with a flux of
particles 0.10.65 gm2 that is equivalent to interaction with particles having
a mass of 1015 1012 kg and velocity 130 kms1 in space over a period of
22.5 years [3].
2. The conducted experiments demonstrated that hermeticity and emissivity
control of TPs is possible by presenting their structure as a distributed
392 SEMKIN N. D. ET AL.
References
1. Novikov, L. S., Akishin, A. I., Semkin, N. D., and Voronov, K. E. (1997) In Proceedings of
the 7th International Symposium on Materials in Space Environmen,t Toulouse, France, 1620
June, 1997 (SP-399, August 1997), pp. 493496.
2. Emma, A. T., Mark, K. H., Kay, L. (1977) In Proceedings of the Second European Conference
on Space Debris, ESOC, Darmstadt, Germany, 1719 March, 1997, (ESA SP-393, May 1977),
pp. 429434.
3. Schafer, F. and Schneider, E. (1997) In Proceedings of the Second European Conference on
Space Debris, ESOC, Darmstadt, Germany, 1719 March, 1997, (ESA SP-393, May 1977),
pp. 435443.
PHYSICAL MECHANISM OF SOLAR CELL SHUNTING UNDER
HIGH VELOCITY IMPACT OF SOLID PARTICLES
1. Introduction
393
J.I. Kleiman (ed.), Protection of Materials and Structures from Space Environment, 393400.
C 2006 Springer. Printed in the Netherlands.
394 V. A. LETIN, A. B. NADIRADZE, AND L. S. NOVIKOV
2. Shunting Model
From the theory of high-velocity collisions [6] it is known that immediately after
collision of a particle with a solid body surface the impacted material is com-
pressed, at first, in the shock wave and then expands in relief waves coming from
the target free edges. The thermodynamic state of the substance changes within
very wide limits. At collision velocities of 515 kms1 the initial pressure behind
the shock-wave front reaches several million atmospheres, whereas the energy
density exceeds several times the sublimation energy.
Under certain conditions of shock-wave loading, the semiconductor starts
melting. During the melting-crystallization process, precipitates (impurity clus-
ters) having a high electric conductivity may originate. The disturbance of the
structure with impurities may in turn bring partial or complete loss of the pn junc-
tion rectifying properties, increase in leakage current, and substantial deterioration
of the solar cell characteristics even when the impact zone area is relatively small.
a) Dsh Dk b) Iph Rs IL
Id Ish Ish
Ish
n-region Rsh Rsh UL RL
p-region. Strike zone.
Figure 1. (a) Impact zone and (b) equivalent circuit diagram of a SC under high-velocity impact by
a solid particle: Rsh initial shunting resistance; Rsh shunting resistance of region, originated at
impact
PHYSICAL MECHANISM OF SOLAR CELL SHUNTING 395
accumulated internal energy turns to be higher than the target substance melting
point then the target substance transforms into the liquid state.
Entropy increment, due to the substance impact compression, was found ap-
plying the RankineHugoniots equations [6].
The pressure value at the substance melting initiation was determined by the
intensity of shock wave within which the entropy increase will be equal to the
entropy difference of the substance in its initial state and at its melting point.
For silicon, this value is estimated at S = 1.08 kJ.(kg.K)1 . In this case, the
calculated value of shock-wave pressure sh obtained according to reference [6],
is equal to 68 GPa.
At a known critical amplitude of the shock wave sh , the critical velocity of
particles would be equal to:
ck 2 sh c1 2 sh ck c1
Vsh = + + + + (3)
2sk k sk 2s1 t st 2sk 2s1
where c1 , s1 constants of Hugoniots adiabat, mass density of the material,
k, tall terms with subscripts k, t refer to the materials of the particle and target
accordingly.
Applying the expression describing the crater diameter as a function of impact
parameters (1), (2), it is easy to find the critical radius of particles ash assuming
that the crater diameter Dk = Dsh :
ash = (/K )(m /k )1/3 (sh,21 /m Vk2 )1/ (4)
The obtained relations define the collision critical parameters (critical veloc-
ity and particle radius) corresponding to the semiconductor melting initiation,
depending on the striker and target properties as well as on the thickness of the
solar cell protective shielding. Outcome of calculating the particle critical radius
depending on the collision velocity and the particle material density is shown in
figure 2. In the calculations it was assumed that sh,21 = 54.8 GPa. The curves
shown in figure 2 are restricted from the left side (marked by black circles) and
correspond to critical velocity values calculated by formula (3).
Al
0.5
0
1 10 100
Impact Velocity, km/s
In this experiment, changes of the solar cell parameters at each particle impact
were recorded. The solar cell reverse current at the fixed bias voltage value (around
2.5 V) was recorded. This enabled to control reverse voltage changes of the cells
at each particle impact.
The solar cell typical IV characteristics are given in figure 3 and inner resis-
tance and back bias current of SC during irradiationin figure 4.
The critical velocity of collisions was determined as Vsh = 7.512.5 kms1 ,
specific resistance of the shunting region sh = 0.068 cm, and critical
pressure value corresponding to the obtained Vsh value is equal to sh =
74.5 GPa.
Figure 4. Inner resistance (marked open squares) and back bias current of SC (marked black circles)
during bombardment
Figure 5. IV characteristic of the solar element from returned solar panel: 1, 2 standard IV charac-
teristics; 3, 4anomalous IV characteristics of the solar element with increased series resistance;
5, 6, 7anomalous IV characteristics of the solar element with shunting p-n junction
Rsh = 0.271 , Dsh = 0.075 cm, L sh = 0.235 cm, sh = 0.064 cm was in good
agreement with the results obtained for the spacecraft VEGA and using the
electrostatic accelerator.
5. Conclusions
Acknowledgments
References
1. Letin, V. A., Zayavlin, V. R., and Gubanova, I. A. (1988) Journal of Electrotechnical Industry
V. 22, N. 13, pp. 144.
2. Rauschenbach, H. S. (1980) Solar Array. Handbook. Litton Educational Publishing Inc., New
York, pp. 1549.
3. Letin, V. A., Holeva, M. N., and Zayavlin, V. R. (2002) Journal of Uzbekistan Academy of
Science, 1, pp. 716.
4. Rijov., Y. A., Burgasov, M. P., Nadiradze, A. B., and Svirschevsky S. B. (1990), In 17th Inter-
national Symposium on Rarefied Gas Dynamic, Aachen, FRG.
5. Burgasov, M. P., and Nadiradze, A. B. (1993) In Proceedings of ESPC-93, Austria, 2327 August
1993, p. 767.
6. Kinslow, R. (1970) High Velocity Impact Phenomena, (ed.) Academic Press, New York and
London, pp. 1511.
7. Novikov, L.S. (2003) High Velocity Collisions in Space, Scobeltsyn Institute of Nuclear Physics,
Moscow, 2003, pp. 111.
8. Letin, V.A. (2002) In Proceedings of 6th European Space Conference, Porto, Portugal, 610
May, 2002, pp. 713718.
9. Letin, V. A., and Babayevsky, P. G. (2001) Journal of High Performance Polymers 13 S453S460.
10. Letin, V. A., Deev, I. S., Scurat, V. E., and Tsetlin, V. V. (2002) II Stage Report on after Flight
Studies of a Retrieved Fragment of Mountable Solar Array 17KS5810-0, RSC Energia-SPRE
Kvant, Korolev, Moscow.
DETERMINATION OF GROUND-LABORATORY TO IN-SPACE
EFFECTIVE ATOMIC OXYGEN FLUENCE FOR DC 93-500 SILICONE
Key words: space environment, environmental durability, low Earth orbit, atomic
oxygen, oxidation, silicones, nanomechanical hardness, ground testing, space
flight experiment, calibration
1. Introduction
2. Experimental Procedures
2.1. MATERIALS
Figure 1. Preflight and postflight images of the EOIM III DC 93-500 sample: (a) preflight, and (b)
postflight showing surface crazing due to AO oxidation
0 0
18 4.2
20 4.7
24 5.6
30 7.0
40 9.4
DETERMINATION OF GROUND-LABORATORY 405
During indent testing of the pristine DC 93-500, negative unloading force values
were observed in the real time force versus displacement data during the second
set of indents, possibly indicating tip contamination. Tests were conducted that
verified that the sample tip was contaminated from the silicone (i.e., repeated
indents were consistently softer than the original data taken at the same location
prior to tip contamination). After tip cleaning, additional data was taken to verify
that the tip was not changed. The results of the noncontaminated pristine data
obtained during two separate indent sessions at two different locations are provided
in figure 2 along with a curve fit for the data. An image of the sample surface is also
DETERMINATION OF GROUND-LABORATORY 407
25
20
Hardness (MPa)
15
0.2 mm
10
5
Y = 5175/(0.01+1.0X1.5)+0.45
0
0 500 1000 1500 2000 2500 3000 3500
Contact Depth (nm)
Figure 2. Hardness versus contact depth for the pristine DC 93-500 ECR control sample
provided in figure 2. As can be seen, the hardness of the pristine silicone is greater
at the near surface than deeper in the bulk of the material. This is commonly
found for both metal and polymer materials using nanoindentation techniques.
Polymers typically display greater hardness values at the surface because of air or
light induced cross-linking and ambient oxidation of the sample.
The 18 h ECR exposed sample received a Kapton effective fluence of 4.2 1020
atomscm2 . Hardness data was acquired on the 18 h AO exposed sample during
two indentation sessions at different locations. The hardness versus contact depth
data is provided in figure 3 along with a curve fit for the data. An image of the
sample surface is also provided in figure 3 and shows that very fine microcracks
have developed.
An interesting effect was observed in the 18 h sample data at higher applied
indent forces (400, 600, 800, and 900 N). An apparent film breakthrough
was displayed in the loading curve. It appears that the tip was broken through a
hardened surface layer and down into a softer underlying layer. An example is
provided in figure 4 for the 900 N applied force indent in the 18 h AO exposed
sample. This type of breakthrough was observed for most of the oxidized films
after some loading force is applied. It should be noted that the screen image shown
in figure 4, provides the displacement at breakthrough, not the actual contact
depth.
408 KIM K. DE GROH ET AL.
25
20
Hardness (MPa)
15
0.2 mm
10
5
1.5
Y = 0.86/(0.01+0.00005X )+0.3
0
0 500 1000 1500 2000 2500 3000 3500
Contact Depth (nm)
Figure 3. Hardness versus contact depth for the 18 h ECR AO exposed DC 93-500
Figure 4. TriboScope software sample force versus displacement curve for 900 N applied force
indent in the 18 h AO exposed DC 93-500 sample
DETERMINATION OF GROUND-LABORATORY 409
25
20
Hardness (MPa)
15
0.2 mm
10
5
Y = 6950/(0.01+0.5X1.5)+0.5
0
0 500 1000 1500 2000 2500 3000
Contact Depth (nm)
Figure 5. Hardness versus contact depth for the 20 h ECR AO exposed DC 93-500
versus contact depth data is provided in figure 5 along with a curve fit for the data.
There was a fair amount of scatter in the data for contact depths between approx-
imately 200500 nm. Film breakthrough was observed when using the 500, 700,
and 800 N applied indent forces. An image of the sample surface is also provided
in figure 5 and shows that the surface has finer craze line than the 18 h exposed
sample.
25
20
Hardness (MPa)
15
10 0.2 mm
5
Y = 4350/(0.01+0.25X1.5)+0.5
0
0 500 1000 1500 2000 2500 3000 3500
Contact Depth (nm)
Figure 6. Hardness versus contact depth for the 24 h ECR AO exposed DC 93-500
25
20
Hardness (MPa)
15
10 0.2 mm
5
Y = 4825/(0.01+0.23X1.5)+0.5
0
0 500 1000 1500 2000 2500
Contact Depth (nm)
Figure 7. Hardness versus contact depth for the 30 h ECR AO exposed DC 93-500
DETERMINATION OF GROUND-LABORATORY 411
25
20
Hardness (MPa)
15
10 0.2 mm
5
Y = 0.46/(0.01+0.000015X1.5)+0.5
0
0 500 1000 1500 2000 2500 3000
Contact Depth (nm)
Figure 8. Hardness versus contact depth for the 40 h ECR AO exposed DC 93-500
is also provided in figure 8 and like the 24 and 30 h exposed samples, shows a
well-developed crazed texture.
The hardness versus contact depth data for the EOIM III control sample is provided
in figure 9 along with a curve fit for the data.
25
20
Hardness (MPa)
15
10
5
Y = 630/(0.01+0.065X1.57)+0.45
0
0 500 1000 1500 2000 2500 3000 3500
Contact Depth (nm)
Figure 9. Hardness versus contact depth for the EOIM III control sample fabricated at the same
time as the flight sample
412 KIM K. DE GROH ET AL.
25
20
Hardness (MPa)
15
0.2 mm
10
5 Y = 169.5/(0.01+0.01X1.5)+0.9
0
0 500 1000 1500 2000 2500 3000
Contact Depth (nm)
Figure 10. Hardness versus contact depth for the EOIM III flight sample
The EOIM III flight sample was exposed to a LEO AO fluence of 2.3
1020 atomscm2 . The hardness versus contact depth data is provided in figure 10
along with a curve fit for the data. Film breakthrough was observed for indents at
600 and 700 N applied indent forces. An image of the sample surface is also pro-
vided in figure 10 and shows that the surface has a well-developed crazed texture.
16
Ground Data (ECR Source) y=1.054E-20x + 3.693E+00
14
Space Data (EOIM-III)
12
Hardness (MPa)
10
8 6.10 1020 atoms/cm2 in
ECR facility equivalent to
6
2.3 1020 atoms/cm2 in LEO
4
2
0
0 2E+20 4E+20 6E+20 8E+20 1E+21
Kapton Effective AO Fluence (atoms/cm2)
a.
16
8
y = 4.521E-21x + 1.376E+00
6
0
0 2E+20 4E+20 6E+20 8E+20 1E+21
Kapton Effective AO Fluence (atoms/cm2)
b.
Figure 11. Hardness versus Kapton effective AO effective fluence: (a) at a contact depth of 150 nm,
and (b) at a contact depth of 300 nm
This would require additional analyses of the force versus distance curves to de-
termine the specific depth of contact and applied load for which breakthrough
occurred. This information was not readily available. As previously mentioned,
the observation of film breakthrough confirms that an embrittled oxidized layer
has developed over the soft underlying silicone layer.
Using values obtained from the curve fit equations for each data set (the curve
fit equations are given in the hardness vs. depth charts); plots of hardness versus
Kapton effective fluence at contact depths of 150, 200, 250, and 300 nm were
produced. These depths represent the near-surface data. The 150 nm contact
depth and 300 nm contact depth plots are provided in figures 11(a) and 11(b),
414 KIM K. DE GROH ET AL.
respectively. The hardness for the ECR pristine sample and the EOIM III control
samples were very similar. The EOIM III control was slightly harder at a contact
depth of 150 nm, which is not surprising. Even if the mix ratios and degrees of
cure were the same for both controls, because the EOIM III control was made
years earlier there would be a difference in the quantity of residual outgassing
products and oxidation that could increase the hardness slightly over time. These
aging differences may also affect the flight sample hardness, but are considered
negligible, as the hardness difference of the two controls was determined to be
insignificant at a contact depth of 150 nm (and they were essentially the same at
300 nm).
Based on the linear fits for the hardness versus exposure time data, the
ECR Kapton effective fluence that provided the same hardness as the EOIM
III sample for 150, 200, 250, and 350 nm depth was 6.10, 6.00, 6.04, and
6.16 1020 atomscm2 , respectively. Averaging these values provides 6.08
1020 atomscm2 , therefore the Kapton effective atomic oxygen fluence in the ECR
facility needs to be 2.64 times higher than in LEO to replicate equivalent exposure
damage in the ground-test silicone as occurred in the space exposed silicone.
When comparing the surface crazing, the extent of crazing in the 18 and 20 h
ECR exposed samples appears to be less developed than the flight sample, while the
24 h ECR exposed sample seems more substantially crazed like the flight sample.
This is a qualitative assessment only, and may be contributed to through sample
handling procedures. For comparison purposes, the hardness versus contact depth
curves for the EOIM III flight sample and the 24 h ECR sample, which was closest
for the necessary ECR fluence for equivalent damage, were overlaid along with
the pristine ECR sample data and are provided in figure 12.
It should be noted that consistency in the sample thickness, sample holder,
and nanoindentation technique for obtaining nanomechanical hardness data is
very important. This is particularly true when the hardness changes substantially
with depth such as with these samples. It is believed that these results will apply
to chemically similar silicones in thermal energy AO systems. Further testing
should be conducted to verify this. Similar tests should be conducted to get a
ground-laboratory to in-space effective fluence correlation for silicones in higher
energy AO facilities.
4. Relevance to ISS
20
Pristine
15
24 hrs ECR AO (5.6 E 20 at/cm2)
Hardness (MPa)
0
0 500 1000 1500 2000 2500 3000
Contact Depth (nm)
Figure 12. Hardness versus contact depth for pristine, 24 h ECR exposed and LEO space exposed
DC 93-500 silicone
controls for the ECR tests and for the EOIM III flight sample were also measured.
The ground-to-space correlation value was determined based on correlation val-
ues for four contact depths (150, 200, 250, and 300 nm), which represent the
near-surface depth data. The results indicate that the Kapton effective AO flu-
ence in the ECR facility needs to be 2.64 times higher than in LEO to replicate
equivalent exposure damage in the ground-test silicone as occurred in the space
exposed silicone.
Acknowledgments
The authors would like to thank Phil Abel of NASA Glenn Research Center for his
consistent help with hardware setup and introductions to the Park Scientific AFM
and Hysitron systems. They also greatly appreciate all the many helpful insights
and advice on calibration and testing procedures received from Lance Kuhn and
Richard Nay of Hysitron.
References
1. Hung, C. and Cantrell, G. (1994) Reaction and Protective of Electrical Wire Insulators in Atomic-
Oxygen Environment, NASA TM 106767.
2. ASTM E 2089-00. (2000) Standard Practices for Ground Laboratory Atomic Oxygen Interaction
Evaluation of Materials for Space Applications, American Society for Testing and Materials,
Philadelphia, USA, June 2000.
3. Knootz, S. L., Leger, L. J., Rickman, S. L., Hakes, C. L., Bui, D. T., Hunton, D. E., and Cross,
J. B. (1995) Journal of Spacecraft and Rockets 32(3), 475482.
4. Smallman, R. E. (1985) Modern Physical Metallurgy, Butterworths and Co. (Publishers) Ltd,
England.
5. Avner, S. H. (1974) Introduction to Physical Metallurgy, 2nd ed., McGraw-Hill Inc., New York,
1974.
6. Oliver, W. C. and Pharr, G. M. (1992) Journal of Material Research 7(6), 15641583.
ATOMIC OXYGEN CONCENTRATION USING
REFLECTING MIRRORS
Key words: atomic oxygen, low Earth orbit, concentrator, space environmental
utilization
1. Introduction
Corresponding authors.
417
J.I. Kleiman (ed.), Protection of Materials and Structures from Space Environment, 417429.
C 2006 Springer. Printed in the Netherlands.
418 MASAHITO TAGAWA ET AL.
is because atomic oxygen resistant materials require greater atomic oxygen flu-
ences to see the effect of atomic oxygen exposure. Moreover, high-atomic oxygen
flux reduces a required exposure time in LEO and may eliminate contamination
problems in material tests. Therefore, use of atomic oxygen concentrator will
drastically increase the degree of freedom in material tests conducted in LEO.
One of the ways to achieve higher atomic oxygen flux in LEO is lowering
the altitude of space vehicle. However, it is practically impossible to change the
altitude of International Space Station (ISS) only for a material exposure test. The
other way to achieve higher atomic oxygen flux in a given altitude is to concentrate
(or to focus) atomic oxygen in a small targeting area. For this purpose, a high-
efficiency atomic oxygen concentrator needs to be developed. An attempt has
already been tried by STS-85 in 1997 (see section 2 for detail). Three parabolic
horns with different size have been tested to collect atomic oxygen. However, it
has been reported that the concentration factor was not proportional to the size of
the parabolic horn due to gas build-up [1].
Not only a material testing in LEO, but also a new thruster system requires
to concentrate atomic oxygen. The new ion engine system uses atomic oxygen in
LEO as a propellant (a conventional ion engine carries Xenon as a propellant) [2].
If the air breathing ion engine (ABIE) were successfully developed, it would be
applied to compensate a drag of satellite at relatively low altitude (100200 km)
[3]. However, ABIE requires a high-efficiency atomic oxygen concentration device
to maintain plasma in its reactor. Since the efficiency of the concentrator limits
the operating altitude of ABIE, the development of high-efficiency atomic oxygen
concentration device is a key for realizing this new ion engine system.
In this paper, an attempt to develop a high-efficiency atomic oxygen con-
centration device is reported. Basic properties of atomic oxygen scattering at
solid surfaces were investigated to develop atomic oxygen concentrator. Some
key points to achieve high-efficiency in atomic oxygen concentration will be ad-
dressed based on the computational and experimental results obtained in atomic
oxygen beam facilities at Kobe University [4].
Figure 1. Atomic oxygen concentrators used in ESEM experiment flown on STS-85 [1]
Acceleration factor
Material 4 9 16
Kapton 9 m 18 m 15 m
3. Experimental Details
The atomic oxygen source used in this study was a laser detonation atomic oxygen
source. This type of source was originally developed by Physical Sciences Incor-
poration (PSI) [5]. It uses laser-induced breakdown phenomenon of the oxygen
gas. Principle of operation in the atomic oxygen source can be described as fol-
lows; pure oxygen gas is introduced into the nozzle throat through a PZT pulsed
supersonic valve. A powerful laser pulse of TEA CO2 laser is focused on to the
oxygen gas cloud at the nozzle throat. By absorbing the laser energy, high-density
and high-temperature oxygen plasma is formed at the nozzle throat. Once plasma
is formed, it propagates forward absorbing the laser energy that is included in the
tail of the laser pulse. The plasma propagation is along the incident laser axis, and
oxygen molecules are decomposed and accelerated in the front of the propagating
plasma wave. Thus, an intense hyperthermal atomic oxygen beam pulse is formed.
The PSI-type atomic oxygen source is attached to the space environment sim-
ulation facility at Kobe University (figure 2) [4]. The atomic oxygen beam is
Figure 2. Laser detonation (PSI-type) atomic oxygen beam source at Kobe University used in this
study
ATOMIC OXYGEN CONCENTRATION USING REFLECTING MIRRORS 421
1.2
1.0
0.6
0.4
0.2
0.0
0 2 4 6 8 10 12 14 16 18 20
Translational Energy (eV)
Figure 3. Typical translational energy distribution of the atomic oxygen beam used in this experi-
ment. Mean translational energy is 5 eV
The flux of concentrated atomic oxygen was also measured using QCM with
silver electrodes [8]. The resonant frequency of a QCM decreases with time when
atomic oxygen oxidizes the silver electrode of QCM. Since silver is not oxidized
by molecular oxygen but by atomic oxygen at room temperatures, formation of
silver oxide signifies atomic oxygen reaction at the silver surface. A change in
mass of the silver film was recorded by measuring the resonance frequency of
QCM with a resolution of 0.1 Hz during atomic oxygen exposure. Resonance
frequency of QCM is expressed by the following formula;
Metals Oxides
Recombination Recombination
Material cross-section Material cross-section
Figure 4. Experimental configuration of the reflectivity measurements of the atomic oxygen reflect-
ing mirror with QCMs
2.5
2.0
Pyrex
Mass gain (g)
1.5
Gold
1.0
0.5
0.0
0 500 1000 1500 2000 2500 3000
Elapsed time (s)
Figure 5. Mass gain of the Ag-QCM due to oxidation by the atomic oxygen reflected by flat mirrors
424 MASAHITO TAGAWA ET AL.
2.5
2.0
Concave
mirror
1.0
0.0
Figure 6. Mass gain of the Ag-QCM due to oxidation by the atomic oxygen reflected by Pyrex
mirrors
1.7 1014 atoms.cm2 .s1 . By comparing these values, the efficiency of reflection
at Au and Pyrex surfaces were evaluated to be 66.5 and 87.1%, respectively. It was
concluded that the recombination cross-section is important as a material property
for a reflecting surface. Pyrex was the material to be selected as a reflecting surface
in the following experiments.
The effect of the shape of the reflecting surface was examined. Two Pyrex reflect-
ing surfaces were prepared: flat and concave mirrors. The concave mirror had a
15.24 cm radius of curvature of the reflecting surface (7.62 cm in focal length).
Configuration of the experiment was the same as that described in previous sec-
tion. Figure 6 indicates the frequency shift of QCM during the test. Atomic oxygen
fluxes concentrated by flat and concave reflecting surfaces were 1.60 1014 , and
3.18 1014 atoms.cm2 s1 . By comparing these values, it was obvious that the
concave Pyrex mirror provided an atomic oxygen flux twice as high as the flat
mirror. It was, thus, concluded that the use of Pyrex concave mirror is effective to
concentrate atomic oxygen.
mg
ui
ut
u in i r
u rn
u it u rt
ms (urt=u it)
vs
Figure 7. Principle of the Hard-Cube model
5 deg. off
6
4
qr - qi (deg.)
0 Specular reflection
0 20 40 60 80
Incident angle (deg.)
Figure 8. Computational results on difference between incident and exit angles of 5 eV atomic
oxygen reflected at the solid surface
426 MASAHITO TAGAWA ET AL.
Three concave mirrors were used to increase the atomic oxygen flux concentration.
The experimental setup is shown in figure 11. Incident angles of atomic oxygen
ATOMIC OXYGEN CONCENTRATION USING REFLECTING MIRRORS 427
6
(a) (b)
atoms/cm2/s))
5
Flux (1014 a
2
0
0 20 40 60 80 100
Exit angle (deg.)
Figure 10. (a): Experimental setup of the angular distribution of reflected atomic oxygen, and (b):
its results with incident angle of 45 . The atomic oxygen flux peaks at exit angles between 4560
for upper and middle mirrors were 45 , whereas that for the lower mirror was
55 . These mirrors were installed to point their exit directions toward the QCM
detector. Atomic oxygen flux detected by Ag-QCM is shown in figure 12. Atomic
oxygen fluxes concentrated with a single mirror are distributed between 0.6 and
3.8 1014 atoms.cm2 /.s1 depending on the mirror-QCM distances (Average
flux: 1.92 1014 atoms.cm2 .s1 ). With three mirrors, atomic oxygen flux reaches
as high as 1.31 1015 atoms.cm2 .s1 . Note that the atomic oxygen flux detected
without reflecting mirror was subtracted as background, the fluxes mentioned
above are the net atomic oxygen fluxes reflected by the reflecting mirrors. From
the experiment described herein, the possibility for concentration of atomic oxygen
by reflecting surfaces was clearly demonstrated.
upper mirror
center mirror
Ag-QCM detector
lower mirror
Figure 11. Experimental configuration of the atomic oxygen concentration with multiple concave
mirrors
428 MASAHITO TAGAWA ET AL.
14
10
0
1 2 3
Number of mirrors
Figure 12. Average atomic oxygen fluxes amplified by the multiple concave (13 mirrors)
5. Conclusion
Use of Pyrex reflecting mirrors to concentrate atomic oxygen in LEO was pro-
posed in this study. The reflection efficiency of 5 eV atomic oxygen at Pyrex
surface was measured to be as high as 87.1%. The focal point of reflected atomic
oxygen was predicted by the simple computer simulation based on the Hard-Cube
model. It was calculated that the difference between incident and exit angles of
atomic oxygen due to energy loss during inelastic scattering event was at most
5 depending on the incident angle. The computational results on trajectories of
atoms scattered at the Pyrex concave mirror were compared to experimental re-
sults, and good correlation was observed. As a result, the atomic oxygen flux at
the focal point increased one order of magnitude higher than the initial exposure
condition by using three Pyrex mirrors. Even though the directionality and en-
ergy distribution of the incoming atomic oxygen beam may be partially lost by
the reflection at the concentrator wall, this new technology allows us to see early
previews of atomic oxygen effects that wont be seen until later on ISS materi-
als and components, especially for atomic oxygen resistant materials. It is also a
key technology for the development of ABIE and for the other applications using
atomic oxygen as a natural resource for space environmental utilization. Use of
the concentrator opens the new possibility in atomic oxygen research in LEO.
Acknowledgments
A part of this work was supported by the grant-in-aid of scientific research contract
No. 14350511 and 15560686 from the Ministry of Education, Culture, Sports, Sci-
ence and Technology, Japan, and the Kurata Memorial Hitachi Science Foundation.
ATOMIC OXYGEN CONCENTRATION USING REFLECTING MIRRORS 429
The author would like to thank H. Kinoshita of Kobe University for stimulating
discussion.
References
1. United States Developed ESEM Experiments. Evaluation of Space Environment and Effects on
Materials, Final Report, available at the website http://setas-www.larc.nasa.gov/esem/AOE.html
2. Stuhlinger, E. (1967) Ion Propulsion for Space Flight, McGraw-Hill, New York.
3. Nishiyama, K. (2003) In Proceedings of 54th International Astronautical Congress, Bremen,
Germany, 29 September3 October, 2003.
4. Yokota, K., Seikyu, S., Tagawa, M., and Ohmae, N. (2003) In Proceedings of the 9th International
Symposium on Materials in a Space, Noordwijk, The Netherlands, ESA SP-540, 1620 June,
2003, pp. 265272.
5. Caledonia, G. E., Krech, R. H., Upschulte, B. L., Sonnenfroh, D. M., Oakes, D., and Holtzclaw,
K. W. (1992) Fast Oxygen Atom Facility for Studies Related to Low Earth Orbit Activity, AIAA
Paper 92-3974.
6. Tagawa M., Yokota K., Kinoshita H., and Ohmae, N. (2003) In Proceedings of the 9th Inter-
national Symposium on Materials in a Space, Noordwijk, The Netherlands, 1620 June, 2003,
pp. 247252.
7. Kinoshita, H., Tagawa, M., Yokota, K., and Ohmae, N. (2001) High Performance Polymers
13(4), 225234.
8. Matijasevic, V., Garwin, E. L., and Hammond, R. H. (1990) Review of Scientific Instruments
61, 17471749.
9. Yokota, K., Tagawa, M., and Ohmae, N. (2003) Journal of Spacecraft and Rockets 40(1),
143144.
10. Greaves, J. C. and Linnet, J. W. (1959) Transactions of the Faraday Society 55, 13231330.
11. Logan R. M. and Stickney R. E. (1966) The Journal of Chemical Physics 44, 195201.
12. Minton T. K., Giapis K. P., and Moore T. A. (1997) Journal of Physics and Chemistry A 101,
65496555.
ATOMIC OXYGEN SOURCE CALIBRATION ISSUES:
A UNIVERSAL APPROACH
1. Introduction
It is known that many institutions in the academia, military, and government sectors
have facilities that simulate one or more aspects of the environment present in Low
Earth Orbit (LEO), and that the facilities available to reproduce the atomic oxygen
(AO) in LEO do so with different degrees of success.
The objective of this paper is to highlight the need for a round robin charac-
terization of the major AO facilities in order to provide a better understanding of
the factors affecting the erosion mechanism and rate. Due to the synergistic nature
of the effects of the space environment on spacecraft materials, measurements of
the UV dose and contamination must be included in a round robin experimental
package.
This paper includes an overview of some methods to produce LEO-like AO
as well as a brief discussion on the major methods to evaluate AO sources.
431
J.I. Kleiman (ed.), Protection of Materials and Structures from Space Environment, 431441.
C 2006 Springer. Printed in the Netherlands.
432 CARL WHITE ET AL.
It is widely accepted, that the atomic oxygen in LEO has the following character-
istics:
r Kinetic Energy of approximately 5 eV
r Flux of approximately 5 1015 atoms cm2 s1
r Ground Quantum Level
r Very low level of ionized species
The AO flux in LEO varies with altitude and time (the latter related to the solar
cycle), and this variation can be of the orders of magnitude level.
Experience with ground-based simulation facilities has shown that different
sources produce AO with characteristics different from those mentioned above.
Hence it is difficult to compare the results produced in different facilities. The sit-
uation is further complicated by the lack of a reliable and exact standard technique
to measure AO flux.
The following discussion will only address those methods that are successful
in producing an AO flux and kinetic energy similar to those found in orbit.
This type of source has enjoyed great interest during the last fifteen years. A beam
of positive or negative ions that is later electrostatically accelerated and focused
is created by either RF excitation or electron bombardment; an energy of around
5 eV can be achieved. In a later process, the ions are neutralized by either charge
exchange or surface neutralization.
Charge exchange can be achieved by the injection of fast ions into the beam
that would also result in an overall increase of the kinetic energy of the beam.
However, this may not always be possible due to divergence of the beam induced
by the space charge that in turn results in a reduction of the kinetic energy of the
species present in the beam. Gaseous charge exchange is inefficient due to low
cross sections.
Surface neutralization presents the complication that the reflected beam would
not have a uniform kinetic energy. Unfortunately, and in addition to the charge
exchange limitations of this method, the Columbian repulsion of the ions in the
beam before neutralization, prevents this method from producing very high fluxes,
although in theory LEO-like densities can be achieved.
plasma expands into the conical nozzle, creating a detonation/blast wave and
generating a beam comprising a mixture of mainly neutral AO but with some
ions. The flux that can be achieved in this kind of source is similar to LEO, with
energies (that can be controlled) in the range of 38 eV [1].
A difficulty of this method is that it generates UV emission together with the
AO atoms. It has been reported [2] that in the spectral range of 115180 nm, a
laser pulsed detonation source produced an intensity of 104 equivalent suns.
It should be noted that the production of plasma can also be achieved with
a continuous source of laser light, but the achieved beam intensities tend to be
lower than with a pulsed laser. This approach also produces ionized species and
UV radiation.
This method is not very different from the laser discharge methods described
above. This technique creates a plasma by means of a microwave or RF discharge,
which is later accelerated in a nozzle. A kinetic energy of about 5 eV can be
accomplished with this technique.
As with the laser, microwave and RF discharges ionized species and UV radi-
ation in addition to AO.
One side of a specially designed ceramic is placed under a high pressure of molec-
ular oxygen, while the other side is exposed to vacuum; the latter side is usually
coated with a thin film of silver. The molecular oxygen on the high-pressure side
dissociates into AO and migrates through the ceramic towards the low pressure
side. Upon reaching the other surface, and before it recombines into molecular
oxygen, a flux of electrons impinges it causing the atomic oxygen atoms to leave
with a kinetic energy similar to LEO conditions.
This method has provided [3] fluxes similar to LEO for a short amount of time,
without a significant production of ions or any UV emission. Further research is
needed to increase the flux of oxygen through the ceramic membrane so as to
allow the production of higher AO beam densities. The energy distribution in the
beam under high flux conditions still needs to be determined.
3.1. KAPTON
thus preventing it from being chemisorbed into the silver film; these recombined
atoms would not be detected as part of the flux by the QCM.
Further research on silver-coated QCMs needs to be performed to improve
their accuracy, since the sticking coefficient between the AO and the silver is
not accurately known, and there are some uncertainties related to the chemistry
between those two species and the diffusion rate of AO on silver [3].
QCMs can also be used to monitor the deposition of contaminants in a vacuum
chamber. For this application, it is enough to use an uncoated crystal. Provided that
this crystal is kept at about the same temperature as the chamber, contaminants
should deposit on the crystal as the same pace as on the rest of the chamber.
The corresponding change in mass would in turn change the oscillation frequency
of the crystal, what can be used to accurately determine the deposition rate of
contaminants.
This method is simple, reproducible, and very sensitive. The probe consists of a
wire, coil, or foil and it may be movable in the test area. Its highly catalytic surface
is heated by the recombination of reactive species.
A method very appropriate to absolute measurements of atom concentrations
was applied to the O-atom reactions by Elias, Ogryzlo, and Schiff [9]. The probe
is a silver-coated platinum wire coil, large enough so that the temperature rise
due to recombination is not large (T < 100 C). It is operated isothermally at an
elevated temperature and the difference in electrical power to produce the same
wire temperature in the presence and absence of O-atoms is measured.
The obvious disadvantage of this method is the lack of specificity to detect AO.
It has been reported that oxygen molecules in the metastable state 92.2 kJ.mole1
above the ground state will produce an erroneously large heat release in the de-
tector. This error is avoided with the complete absence of O2 in the system. In this
scenario, measurements using this method are in excellent agreement with values
obtained by NO2 titration.
This method is highly specific for the detection of atomic oxygen recombination.
The resonance lines are easily identified and other information such as temperature
can be obtained for any parametric excited state.
Among its principal disadvantages are cost and complexity. Also, the fact that
it gives only space averages of concentration over a considerable length of the flow
tube. Absolute or even relative concentrations are calculated with some difficulty
from experimental data. The signal depends on the average collision frequency of
the atom and this complicates an experiment with added gases.
436 CARL WHITE ET AL.
The only rate data so far reported by this method are neither qualitatively nor
quantitatively reliable.
NO2 + O NO + O2
NO + O NO2 + h
The titration involves the careful measure of the flux of NO2 in a fast flow tube. As
the first reaction is about five orders of magnitude faster than the second, no energy
emission would take place from the second reaction when the number density of
O is less or equal than the number density of NO2 .
When the flux of O is higher than that of NO2 , the second reaction begins
to take place and light emission would be present. This emission is a greenish-
yellow afterglow that has been well determined by spectral analysis. When the
number density of O doubles that of NO2 , the maximum light emission occurs;
the concentration of NO remains constant during the peak emission.
This glow can be detected by means of a photomultiplier, so that the end-
point of the titration can de determined with precision. This method is highly
sensitive and specific. It has been successfully applied to pressures below one
torr, and the presence of neutral gases such as Argon does not interfere with the
measurements. Unfortunately, the implementation of this method is not easy, since
the equipment required is not only bulky, but it is not commercially available and
has high demands in terms of vacuum requirements.
The main principle behind this kind of instrumentation is the ionization of all the
species in the beam so as to later divert and detect them. This technique detects
mass-to-charge ratios, and as such, would give the same reading for both O+ and
O+2
2 .
Mass spectrometers have a virtually indefinite useful detection life and can
also be used to detect other species present in the flux. The most common type is
the quadruple mass spectrometer, which has been successfully used on numerous
occasions for both flight and ground AO measurements.
A disadvantage of this technique is that the equipment required may be expen-
sive, consume a large amount of power, and it is generally bulky. Other difficulties
may include the recombination of AO before detection and that its inability to
detect kinetic energies (unless supplemented by additional equipment for a time-
of-flight measurement or other technique).
ATOMIC OXYGEN SOURCE CALIBRATION ISSUES 437
A good and reliable way to measure the energy of a beam of AO is the Time-of-
Flight technique. Its basic principle is to measure the time oxygen atoms in the
beam require to travel a known distance. By dividing the distance by the time we
would know the velocity.
The kinetic energy can be calculated by means of the following equation:
E = mv2/2
where E is the kinetic energy, v is the speed, and m is the mass of an atom of
oxygen.To provide an accurate speed measurement, beam modulation methods are
typically used. In many instances, mechanical modulation is achieved by means
of a chopper.
A good overview of this method and of some of the instruments required to
implement it can be found in [14].
We can conclude from the discussion presented above, that different AO sources
produce fluxes with different excited states and energies and that sometimes these
differences can be significant. There may also be differences in the radiation level
and contaminant levels which could either accelerate or retard an AO reaction.
With such variations a full source calibration is required if the behaviour of exposed
materials is to be accurately quantified and understood.
The rationale behind a round robin is to define a standard set of experiments
and measurements which if conducted in a source would produce a universally
accepted calibration of that source.
Thus, an ideal method or equipment to compare and evaluate AO sources
should have the following qualities:
r Ease of manufacture and operation.
r Low cost.
r Relatively small size and ease of integration with major AO simulator
facilities.
r Able to measure the flux density.
r Ability to measure the beam energy.
r Capable of providing a qualitative and quantitative chemical analysis of the
species present in the beam.
440 CARL WHITE ET AL.
For absolute confidence in the outcome, the calibration should as far as possi-
ble use a standard experimental package and be carried out under the supervi-
sion of the same team. It would also be beneficial if this package is suitable of
AO exposure in LEO, so as to allow the comparison of orbital and laboratory
fluxes.
Based on the review on AO flux measurement techniques presented in this
publication, it seems reasonable to propose the design of a small integrated cir-
cuit with a carbon-based actinometer and a silver-coated or carbon-coated QCM
crystal. If problems of reproducibility can be overcome, a ZnO actinometer could
also be included since it is potentially very sensitive and reusable. To measure
the deposition of contaminants on the internal surfaces of the space simulator, a
QCM crystal without coating can be used. It can be easily added to the circuit and
share some of the electronics from the other QCMs on it.
Kapton witness samples must also form part of the experimental package
since it is currently regarded as being the industry standard and the majority
of quoted erosion yields are based on Kapton reference samples. The proposed
experiment should also include two solar-visible blind aluminium gallium nitride
photodetectors; one of them should be covered with a sapphire window. This
sensor arrangement would allow the UV radiation to be monitored in the chamber,
with particular attention to the 120130 nm spectrum range. A small temperature
monitor should be included.
To overcome the practical difficulties associated with working in different
facilities, the development of a miniaturized integrated package is seen as being
an essential development.
7. Challenges
The discussion above presents the need for, and the method of, conducting a
series of round-robin tests of laboratory-based AO simulation facilities world-
wide in order to obtain a benchmark comparison. Although feasible in the near-
term, such a program nevertheless presents some additional challenges. These
include:
1. The chemical nature of the oxygen atoms (e.g., ionic, neutral) is harder to
determine using the proposed techniques. Mass spectrometers should be used
where available, but this may be hard to implement in all facilities.
2. The proposed round robin experiment does not measure AO kinetic energies.
Suitable techniques like time-of-flight mass spectroscopy should be used when
possible.
3. The proposed techniques should first be integrated and tested in a standard
facility. This process would ease the calibration of the instrumentation and
allow a more reliable comparison among facilities.
ATOMIC OXYGEN SOURCE CALIBRATION ISSUES 441
References
1. Kleiman, J., Iskanderova, Z., Gudimenko, Y., and Horodetsky, S. (2003) In Proceedings of the
9th International Symposium on Materials in a Space Environment, The Netherlands, ESA
Publication SP-540, 2003.
2. Grossman, E., Guzman, I., Viel-Inguimbert, V., and Dinguirard, M. (2003) Journal of Spacecraft
and Rockets 40(1), 110113.
3. Valer, J. C. (2000) Thesis for the Masters in Aerospace Engineering Degree presented at the
University of Tennessee Space Institute, August 2000.
4. Osborne, J. J., Harris, I. L., Roberts, G. T., and Chambers, A. R. (2001) Review of Scientific
Instruments 72(11), pp. 40254041.
5. Reddy, M. R. and Mater, J. (1995) Science 30, 281.
6. Leger, L. J. and Visentine, J. T. (1986) Aerospace America 24, 32.
7. Wolan, J. T. and Hoflund, G. B. (1999) The Journal of Vacuum Science And Technology A 17(2),
662664.
8. Tim Minton, personal communication.
9. Elias, L., Ogryzlo, F. A., and Schiff, H. I. (1959) Canadian Journal of Chemistry 37, 1690.
10. Thrush, B. A. (1967) Science 156, 470473.
11. Harris, I. L., Chambers, A. R., and Roberts, G. T. (1997) Review of Scientific Instruments 68(8),
32203238.
12. White, C. B., Rao, J., Roberts, G. T., Chambers, A. R., Lawson, K. J., and Nichols, J. R. (2003)
In Proceedings of the 9th International Symposium on Materials in a Space Environment, The
Netherlands, ESA Publication SP-540, 2003.
13. Osborne, J. J., Roberts, G. T., Chambers A. R., and Gabriel, S.B. (1999) Review of Scientific
Instruments 70(5), 25002506.
14. Auerback, D. J. (1988) Atomic and Molecular Beam Methods, Vol. 1, Oxford University Press,
pp. 362379.
15. Dinguirard, M., Mandeville, J. C., Van Eesbeek, M., Tighe, A. P., Durin, C., Chambers, A.,
Gabriel, S., Goulty, D., and Roberts, G. (2001) American Institute of Aeronautics and Astro-
nautics Publication, # 2001-5070.
LOW-COST SPACE MISSIONS FOR SCIENTIFIC AND
TECHNOLOGICAL INVESTIGATIONS
Abstract. The Space Flight Laboratory at the University of Toronto Institute for
Aerospace Studies (UTIAS/SFL) is pioneering the use of commercial off-the-shelf
technologies in space to support education and rapid access to space for Canadian
researchers. The Canadian Advanced Nanospace eXperiment (CanX) Program
empowers researchers across Canada by providing spacecraft under 10 kg that
facilitate inexpensive research in low Earth orbit. The limited resources available
to these nanosatellites force innovation in the way systems are built, how exper-
iments are defined, and in the instrumentation that is used to support international
class science and technology missions. Graduate students at UTIAS/SFL build
nanosatellites under the supervision and mentoring of UTIAS/SFL staff while
payloads are defined and developed by other universities and businesses across
Canada. CanX-1, the programs first nanosatellite, was launched on 30 June 2003.
The objectives of CanX-1 were to qualify critical system technologies in support
of future nanosatellite missions, such as CanX-2. Currently under development,
CanX-2 is planned to be the first nanosatellite in the CanX program to support a
variety of experiments for researchers across Canada. CanX nanosatellites are an
excellent means to acquire fast heritage for new technologiestechnologies that
will be critical in supporting the demanding satellite missions of tomorrow. High-
performance radiation-robust computer systems, high-speed radio transceivers,
and electric propulsion systems, are a few examples of microspace technologies
under development at UTIAS/SFL that can be tested and proven through the CanX
Program.
1. Introduction
Studies Space Flight Laboratory (UTIAS/SFL). The CanX program began with
the intention of providing an opportunity for Canadian graduate engineering stu-
dents to learn about the field of microsatellite engineering, while at the same time
providing a low-cost orbital platform for Canadian scientists. CanX missions use
the CubeSat standard developed by Stanford and CalPoly universities, with the
aim of lowering satellite launch costs through standardization that allows small
cube shaped satellites to be launched inside a compatible deployment system. A
single CubeSat has a mass of less than 1 kg and a side length of 10 cm, and is
cube shaped as the name suggests. The CanX program also takes advantage of the
latest advances in technologies that are applicable to space, by using a relatively
short design cycle.
As an integral part of the CanX program, student participation is essential for
the success of the program. The expected design cycle of a CanX satellite, lasting
approximately 1824 months, nicely coincides with the length of time it typically
takes for students to complete a Masters degree. In this way, the students are able
to experience a complete satellite development cycle, and leave the CanX program
with training in all phases of satellite design, construction, testing, and operations
(see figure 1).
In addition to the student team responsible for much of the work on a CanX
satellite, there is also a team of UTIAS/SFL staff members who may design some
subsystems of the satellite.
The range of fields covered by the SFL staff members includes computer en-
gineering, power systems engineering, radio frequency communications, systems
engineering, propulsion design, and satellite testing. The staff members have pre-
vious experience in microsatellite design, and are able to mentor the students and
share the lessons learned through previous SFL satellite experiences to help ensure
the success of the current CanX mission.
LOW-COST SPACE MISSIONS 445
2. Facilities
The Space Flight Laboratory (SFL) at the University of Toronto Institute for
Aerospace Studies (UTIAS) is a modern satellite engineering facility built within
the confines of a world-recognized centre for space research. The laboratory has
incorporated significant facilities to allow most of the design, assembly, and test-
ing of SFL satellites to be accomplished in-house. For satellite design, there are
significant computing facilities along with many of the latest software packages
for aiding the development mechanical, electrical, and software designs. There are
also facilities for constructing basic mechanical and electrical prototypes of flight
systems. After prototypes are constructed, SFL has facilities to conduct tests on the
prototype systems. SFL has two thermal cycling chambers, which can test items
within a temperature range of 70 C to +180 C. SFL also possesses equipment
to operate a small vacuum chamber which can be used within a thermal chamber,
thus allowing in-house thermal-vacuum testing of spacecraft components, or in the
case of CanX series satellites the entire spacecraft. For radio testing, SFL possesses
a small anechoic chamber. There are also facilities available for spacecraft testing,
as well as instruments such as oscilloscopes, spectrum analyzers, and signal gen-
erators. For vibration testing and EMI/EMC testing, UTIAS/SFL has relationships
with other departments at the University of Toronto, some Canadian government
agencies, and industry partners to provide support for testing that cannot be con-
ducted in-house. For final spacecraft integration and assembly, there is a Class
10,000 clean room located at UTIAS/SFL. The clean room facility at SFL allows
for integrated functional testing in-house, saving the cost of having to move the
established test support equipment to an off site clean room facility (see figure 2).
UTIAS/SFL also contains ground station facilities for communicating with
and tracking SFL satellites. There are two separate ground stations at the present
time: the MOST ground station, and the CanX ground station. The MOST ground
station has facilities for VHF/UHF/S-band satellite communications using Yagi
antennas, as well as S-Band downlink capability with a 2.3 m parabolic dish
antenna. At present, the CanX ground station works in half-duplex with UHF
uplink and downlink using a dual Yagi antenna array. The ground station control
area at SFL also has a large screen projection system, so that major spacecraft
events can be watched easily by larger groups (see figure 3).
3. CANX-1
The CanX-1 OBC is responsible for managing all of the satellite systems
including power, payloads, attitude control, and the UHF radio. The OBC operates
at 40 MHz, with application programs running in 512 kB of error detection and
correction (EDAC) controlled SRAM memory. For bulk data storage, there is also
32 MB of flash memory on board. The satellite bootloader is written on a 128 kB
EPROM, in order to increase the reliability of this critical system.
Power is supplied to CanX-1 using an array of Emcore triple junction Gallium
Arsenide solar cells. These cells provide approximately 2 W of power to the system
when the satellite is in sunlight, with the exception of the +Z face of the satellite
that only has one cell instead of the nominal two.
The UHF radio system is the only means of communication between the ground
and CanX-1. It operates in the amateur band of frequencies with a downlink
frequency of 437.88 MHz, and was custom-designed for CanX-1 by students at
SFL. The radio system is half-duplex, working at a data rate of 1200 bps.
The magnetic Attitude Control System (ACS) for CanX-1 utilizes three air-
wound magnetic torquer coils. These torquers, orthogonally mounted underneath
three of the CanX-1 solar panels, provide the torques needed to implement a
B-dot control law. This controller is used to reduce the angular momentum of
the satellite, i.e., detumble the satellite. Through the use of a 3-axis magnetometer
supplied by Honeywell, the rate of change of the magnetic field is sensed and
current is applied to the magnetic torquer coils in order to slow the rotation of the
spacecraft. The software needed to accomplish these tasks is run by the OBC on
the EDAC protected SRAM.
The CMOS imagers are included on the CanX-1 mission for two purposes.
The monochrome imager, with a field of view of approximately 5 , is intended
mainly to conduct ground based star-tracking experiments on the returned images.
The colour imager, with a field of view of approximately 40 , is to be used to
448 DANIEL RANKIN ET AL.
take pictures of Earth and its moon. Both imagers are manufactured by Agilent
Technologies and have a resolution of 640 480 pixels.
CanX-1 also contains two other payloads. The Xiphos Q4 board was provided
by Xiphos Technologies of Montreal, Quebec. The Q4 board contains an array
of sensors for on-orbit evaluation, including angular rate sensors, temperature
sensors, and magnetometers. The other payload is a CMC Electronics GPS re-
ceiver. This GPS receiver acquires signals through two small patch antennas on
the Y panels of the spacecraft. The GPS receiver is able to determine spacecraft
position and velocity when multiple GPS satellites are in view.
The CanX-1 design was completed in October 2002, with acceptance testing
taking place through JanuaryApril of 2003. CanX-1 was launched from Plesetsk
Cosmodrome in northern Russia on June 30th, 2003, along with the MOST mi-
crosatellite also assembled at UTIAS/SFL. CanX-1 was launched in a deployment
tube along with two other CubeSats (DTUSat and AAU CubeSat) developed by
Danish university student teams (see figure 5 and 6).
Unfortunately, no contact was ever made with CanX-1 after launch. In the first
couple of months following launch, none of the three satellites, CanX-1, DTUSat,
or AAU CubeSat was contacted. Various large dish antennas were commissioned
in an attempt to make contact. AAU CubeSat was finally contacted after securing
the use of a 8 m dish antenna in Denmark. However, they were only able to down-
load minimal telemetry and were not able to complete their mission. To this day
neither CanX-1 nor DTUSat have been contacted. After going through numerous
contingency scenarios, and after obtaining the help of the Algonquin Radio Ob-
servatory and a 9 m ground station at Defence Research and Development Canada
(Ottawa) to try and make contact with CanX-1, the UTIAS/SFL student team was
unable to do so.
The absence of data makes it difficult to draw definitive conclusions. A com-
plete failure report has been written based on the format used for recent NASA
failures, and the most plausible causes of failure have been identified. All that
LOW-COST SPACE MISSIONS 449
Figure 6. NLS-1 launch tube containing CanX-1, DTUSat, AAU CubeSat mounted on rocket upper
stage
4. CANX-2
Initiated in September 2003, CanX-2 is the second in the series of CanX satellites.
At present, CanX-2 is in the detailed design phase. CanX-2 is planned to be one
450 DANIEL RANKIN ET AL.
of the first operational science nanosatellites for Canadian researchers. The size
of the satellite is roughly double that of CanX-1, allowing for greater available
volume and surface area for power generation (figure 7).
CanX-2 uses a UTIAS/SFL developed OBC, this time based on an ARM9
processor at 18 MHz. The OBC will use 512 kB of EDAC protected SRAM and
32 MB of flash memory for mass data storage (see figure 8). Power is provided by
triple junction GaAs solar cells, and with more surface area compared to CanX-1,
the maximum power generation is increased to 4 W. Primary radio communications
is accomplished using half-duplex UHF, and the bit rate for CanX-2 is planned to
be 4800 bps.
CanX-2 also includes an extensive suite of science and engineering payloads.
These payloads consist of:
r A three-axis momentum-bias coarse pointing attitude control system.
r A high-data-rate S-Band transmitter.
r Two CMOS imagers (colour and monochrome).
r A Nanocalorimeter.
r An Atmospheric Spectrometer.
r A GPS receiver for GPS radio occultation studies.
r A Surface Materials experiment.
Like CanX-1, the CanX-2 Attitude Control System (ACS) utilizes three mag-
netorquer coils. In addition, CanX-2 will include a momentum wheel provided
by Dynacon, Inc. The goal for the system is to have both attitude determination
and pointing within one degree of accuracy. The ACS is included to both test the
LOW-COST SPACE MISSIONS 451
momentum wheel technology, and to support the other science and engineering
payloads.
The S-Band transmitter is an SFL designed product to test the performance
of miniaturized S-Band technology for nanosatellites, while also increasing the
amount of science data that can be received by the ground. The maximum designed
information rate for the S-Band transmitter is 256 kbps, and this can be scaled on
the fly; this allows higher data rates when the satellite is at higher elevations in
the sky.
The two CMOS imagers that are being planned for CanX-2 are very similar to
the CanX-1 imagers. These imagers, manufactured by National Semiconductor,
are slightly larger than the CanX-1 imagers, and both have planned fields of view
of about 30 with a resolution of 1200 1024. Again, the monochrome imager
provides the option of doing ground-based star tracking experiments, while the
colour imager will be used mainly to take pictures of interesting targets like Earth
and its moon. Both cameras may be used for on orbit calibration of the ACS
performance.
The Nanocalorimeter experiment is provided by Dr. Harry Ruda of the Uni-
versity of Toronto. The proposed experiment is aimed at studying the influence of
space-based radiation on the properties of semiconductor nanowhiskers with and
452 DANIEL RANKIN ET AL.
without double barrier structures. The end goal of this experiment is to develop a
semiconductor-based heat-sensing array with sensors as small as 50100 nm.
The Atmospheric Spectrometer, provided by Dr. Brendan Quine of York Uni-
versity, is an Earth imaging spectrometer. It provides measurements of airborne
greenhouse gases to support the goals of the Kyoto protocol. The payload operates
in the near infrared band, and features a surface resolution of 1 km, which will
enable the identification of local variation and sources of pollution emission.
The GPS experiment, designed by Dr. Susan Skone of the University of
Calgary, uses a directional GPS antenna mounted on the outer surface of the
satellite to take measurements as GPS satellites are occulted by the Earths at-
mosphere. From this data, a detailed profile of tropospheric water vapor can be
generated, along with atmospheric electron densities.
The surface materials experiment for CanX-2 is provided by Dr. Jacob Kleiman
of the University of Toronto in collaboration with Prof. Alan Chambers from the
University of Southampton, UK. This experiment uses a specifically developed
array of 4 atomic oxygen sensors. One or more sensors will be treated by Dr.
Kleimans group to enhance their oxidative stability. The behavior of all sensors
will be evaluated during the flight. The plan is to monitor the changes in sample
thickness as a result of atomic oxygen erosion to evaluate the effectiveness of the
special surface treatment.
In addition, a network communications experiment involving an innovative
satellite communication protocol developed by Dr. Michel Barbeau of Carleton
University is included in CanX-2. This experiment will test a satellite networking
protocol under the open-source operating system of eCos.
With the design of CanX-2 well underway, the flight hardware is planned for
completion by the end of 2004. With a projected launch in late-2005, CanX-2 will
be the second Canadian nanosatellite launched into space.
5. Conclusion
The CanX program serves Canada and international partners by providing ex-
ceptional hands-on training to graduate students in the area of space systems en-
gineering. At present, this program is unique in Canada. University researchers,
industry and government have opportunities to fly science and technology experi-
ments in space cheaply and rapidly. The CanX program is intended for aggressive
experimentation in space. In combination with the training aspect of the program,
missions are completed in less than two years for a few hundred thousand dol-
lars. In keeping with the nano/microspace philosophy, redundancy in the satellite is
traded for simple, good design, with mission risk distributed over multiple low-cost
missions rather than over multiple components in a single mission. The CanX pro-
gram is pioneering the low-cost exploitation of space and intends to revolutionize
LOW-COST SPACE MISSIONS 453
Canadian space activity in the 21st century. The UTIAS Space Flight Labora-
tory encourages companies, university groups and government agencies across
the world to collaborate on these low-cost missions. For more information please
visit www.utias-sfl.net.
Acknowledgements
The UTIAS Space Flight Laboratory gratefully acknowledges the following spon-
sors of the CanX program:
r Defense Research and Development Canada (Ottawa)
r Natural Sciences and Engineering Research Council (NSERC)
r MacDonald Dettwiler Robotics
r Dynacon Incorporated
r Ontario Centers of Excellence, CRESTech Division
r Canadian Space Agency
r Xiphos Technologies Inc.
r Radio Amateur Satellite Corporation (AMSAT)
ARC International
ATI
Autodesk
@lliance Technologies
Cadence
CMC Electronics
EDC
E Jordan Brookes
Emcore
Encad
Honeywell
Micrografx
National Instruments
Natural Resources Canada
Raymond EMC
Rogers Corporation
Stanford University
Texas Instruments
The Mathworks
Wind River
SUBJECT INDEX
455
456 SUBJECT INDEX
173, 175, 176, 179190, 217219, 221223, methyl silicone rubber, 35, 40
233, 235, 237, 240, 241, 243245, 258, 266, microcrack(s), 171, 219, 221, 222, 257, 258,
360, 362, 363, 370, 385, 393397, 421 262, 309, 315, 407
inclination, 115, 116, 186 microhardness, 248, 313, 314, 401, 402
inelastic scattering, 422, 428 mitigation, 9, 16, 18, 21, 71, 7678, 8085, 93,
inorganic, 173, 247, 248 101, 102
insulator, 277280, 286, 288, 289, 292, 331, model(s), 4449, 62, 76, 8490, 99101, 115,
339, 342, 385 116, 120, 125127, 153, 158, 161165, 176,
International Space Station (ISS), 18, 186, 204, 205, 252, 272274, 351, 354, 356,
7185, 93104, 168, 175183, 186, 187, 360, 378, 379, 383, 386390, 393, 394, 417,
190198, 205208, 295299, 414, 415, 419, 424428, 450
418, 428 modeling, 71, 76, 84, 101, 239, 267, 271, 399
ionizing radiation, 43, 44, 153, 296, 300, 302, molecular dynamics, 365, 367
303, 307 molecular oxygen, 421, 433
Mylar, 116, 155, 233, 238245, 334
Kapton, 17, 18, 51, 54, 58, 75, 76, 102, 142,
153, 158164, 247, 252258, 262, 278, nanomechanical hardness, 402, 405, 412, 414
281299, 302, 304, 306, 317328, 344, 366, N -nitrosodimethylamine (NDMA), 193, 194,
401404, 407410, 413416, 419, 420, 434, 198208
437, 440
Kapton H, 247, 252, 253, 255, 256, 262, 286, O-atom, 317328, 435
287, 317328 on-board computing platform, 2132
Kapton HN, 286, 287 on-orbit, 17, 18, 80, 85, 176, 200, 295299,
Kapton polyimide, 278 365, 373, 414, 415, 448
optical characteristics, 93, 94, 167, 171, 173,
LEO environment, 155, 341, 360, 367 281, 284, 385
LEO erosion, 365 optical coatings, 71, 85, 93, 94, 104, 266
Long Duration Exposure Facility (LDEF), 154, optical constants, 126, 265, 269, 271274
155, 176, 278 optical degradation, 8790, 133
loss tangent (tan ), 39, 209, 213 optical density, 351, 353, 355, 356
low Earth orbit (LEO), 9, 51, 85, 141, 153, 167, optical materials, 107, 351, 352
175, 176, 247, 248, 277, 278, 306, 317, 318, optical microscopy, 36
331, 344, 359, 365, 401403, 417, 443 optical properties, 6163, 6669, 75, 77, 107,
low temperature, 1, 108, 292, 309315, 323 125, 138, 159, 168, 225, 226, 230, 231, 233,
239, 241, 245, 252, 259, 260, 266, 289, 339,
magnetic field, 447 346, 391
mass loss, 35, 37, 41, 4752, 5558, 142144, optical quartz glass, 107, 108, 351, 352
151, 154, 169172, 209, 210, 212, 214, 215, orbital debris, 73, 155, 175, 176, 181186, 189,
252, 257, 258, 262, 290, 299, 306, 317, 318, 190, 235
321, 341, 344348, 367, 379, 382, 401404, organic, 43, 124, 128, 129, 247, 248, 317, 318,
415, 419, 434 401, 402
mass spectrometer(s), 52, 56, 319, 320, 421, oscillator, 271, 355, 379
436, 440 outgassing, 4349, 124, 212, 377383, 414
material-outgassing, 4349 oxidation, 141, 149, 150, 248, 253, 257, 317,
mechanical properties, 123125, 129, 133, 366, 402, 404, 407, 414, 422424, 431, 437,
136139, 155, 167, 170, 209, 241, 405 438
meteoroids, 154, 176, 183 oxide(s), 74, 154, 171, 249, 258, 265, 280, 282,
methane, 210, 359, 363, 365 309, 331, 332, 343, 347, 402, 412, 421, 422,
methyl, 35, 40, 41, 247, 249, 254, 255, 261, 437
402 oxide layer, 249, 258, 437
458 SUBJECT INDEX
oxygen atom(s), 150, 258, 299, 300, 318, 366, 277, 278, 285, 289292, 296, 298303, 307,
426, 432, 433, 437, 438, 440 343, 351359, 366, 377, 378, 381383, 386,
oxygen plasma(s), 167, 168, 172, 173, 247, 393, 433, 438440, 443, 451
289, 401, 420 radiation protection, 21
radio-frequency plasma, 253
paint, 44, 61, 74, 75, 87, 154, 172, 181, 182, reaction mechanism, 359363
184 reaction pathways, 367, 370
paint-thermal control, 74 reaction products, 51, 58, 193, 194, 359, 362,
paint-white, 44, 61 434
perfluorinated alkanes, 365, 366, 372 re-configurable computing, 21
pitting, 7173, 7681, 84, 85, 93, 94, 98, 104 reflectance, 6169, 125, 171173, 236,
planetary, 154 239241, 243, 252, 257, 259, 260,
plasma interaction(s), 9, 10 265267, 270275, 281, 284, 285, 290, 302,
plasma-asher, 253 339, 342
polyethylene, 141151, 162, 218, 219, 225, reflectance-spectral, 239
366, 373 reflecting mirrors, 417, 427, 428
polyimide, 5158, 141151, 153, 159, reliability, 25, 35, 107, 108, 225, 447, 449
168173, 233, 252, 278, 318, 321, 326, 401, remission, 43, 46, 48, 377
420, 422 roughness, 126, 145, 146, 170, 310, 319324,
polymer materials, 363, 407 326, 327, 347, 412
polymer(s), 44, 52, 53, 123, 124, 139, 142144,
146, 153161, 169, 242, 247, 248, 250, 255, satellite, 18, 115120, 153, 154, 289, 418,
258, 280, 318, 319, 325, 326, 359, 360, 363, 443453
366, 367, 374, 377, 407 scanning electron microscopy (SEM), 158,
polymeric composite, 47, 377, 379 181185, 190, 217, 219, 221, 222, 252, 281,
power, 14, 16, 17, 22, 24, 25, 30, 71, 84, 87, 310, 312, 343, 346, 347, 349
123, 124, 143, 159, 176, 177, 268, 270, 275, sensor, 53, 54, 142, 342, 379, 381, 386, 422,
278, 279, 291, 316, 331, 333, 334, 342, 437, 438440
385387, 390, 393, 398, 435439, 444, 447, SEU-mitigation, 21
449, 450 shielding, 154, 168, 175, 176, 369, 396
product-volatile, 43, 44, 254, 377 shunting, 393399
protective coating, 102, 167, 173, 247, 249, silica, 168, 171, 248, 252262, 395
257, 258, 262, 290, 331 silicone, 3541, 123138, 168, 171, 173,
proton (ir)radiation, 4, 6, 7, 87, 88, 110, 113, 247249, 252258, 262, 295307,
356 401407, 412416
proton flux(es), 68 simulation test, 107, 167
purge, 93 single event upset (SEU), 2132
Pyrex, 417, 422424, 426, 428 smart material, 277, 278, 279, 341
software, 30, 176, 185, 271275, 285, 323,
quantum chemistry, 359, 360, 363, 365 360, 408, 445, 447
quartz crystal microbalance (QCM), 5158, solar sail, 233, 234, 243, 245
142151, 333339, 421427, 434, 435, solar-absorptance, 124, 171, 236, 239, 241,
440 242, 277, 284, 285, 290, 293
quartz glass, 107113, 351357 solar-array, 918, 71, 7485, 93, 94, 96, 98,
101104, 154, 159, 262, 295299, 306, 310,
radiation, 18, 2125, 32, 3537, 4049, 393, 398, 399, 414
6169, 87 91, 94, 107116, 119125, 129, solar-cell, 1, 2, 4, 7, 8, 16, 17, 35, 75, 84, 94,
132, 133, 137, 141, 146, 150, 153155, 159, 98, 101, 102, 107, 123, 124, 209, 257, 258,
225, 229, 231, 233249, 252, 265, 266, 270, 393399, 447, 450
SUBJECT INDEX 459
461
462 AUTHOR INDEX