Documenti di Didattica
Documenti di Professioni
Documenti di Cultura
UNIVERSITE
U.F.R SCIENTIFIQUE DORSAY
Licence et Magist`
ere de Physique
2014-2015
Classical Electromagnetism
-
Lecture Notes
26
1.8
Additional reading
Introduction
Several mission studies for high-energy interplanetary solar sailcraft missions have been carried out at
DLR [4][5][8][9] and elsewhere [13][15]. Most of them
require a rather demanding sailcraft performance to
keep mission durations short (see Table 1). However,
taking the current state-of-the-art in engineering of
ultra-lightweight structures into account, solar sailcraft of the first generation will be of relatively moderate performance. For such near-term solar sailcraft
few mission examples can be found in the literature.
The aim of this paper is to narrow down this gap
and to get a lower bound on solar sailcraft performance for interplanetary missions that are under consideration. It will be shown, that challenging scientific missions are feasible at relatively low cost, even
with moderate performance sailcraft of the first generation. This will be demonstrated below by the trajectory analysis of a proposed sample return mission
to Near Earth Asteroid 1996FG3 (mission duration
approx. 9.4 years).
1
German Aerospace Center (DLR), Cologne
Institute of Space Sensor Technology and Planetary Exploration
Phone: +49-2203-601 {3001|3028} Fax: +49-2203-601 4655
E-Mail: {bernd.dachwald|wolfgang.seboldt}@dlr.de
ausler2
Bernd Dachwald1 , Wolfgang Seboldt1 and Bernd H
6th International AAAF Symposium on Space Propulsion: Propulsion for Space Transportation of the XXIst Century, Versailles, France, 14-16 May 2002
Copyright (c) 2002 by DLR
Sailcraft performance
ac [ mm/s2 ] [ g/m2 ]
0.5
16.0
0.7
11.4
0.75
10.7
0.85
9.4
1.0
8.0
1.0
8.0
1.0
8.0
1.0
8.0
1.25
6.4
Transfer time
[ yr]
1.4
10.4
3.3
3.0
6.8
0.6
1.0
1.2
1.2
[13]
[4][5]
[4][8]
[4]
[15]
[13]
[13]
[13]
[15]
References
P A( er n) er
= 2P A cos2 n
= 2P A( er n)2 n =
FSRP = Fr + Fr =
Fr = P A( er n) er
Fr =
.
(Peff )1AU = 2(P0 )1AU = 7.757 106 N/m2
Table 1: Fast solar sailcraft missions using advanced sailcraft (rendezvous, if not stated otherwise)
Mercury
Pluto (fly-by)
(4) Vesta
2P/Encke
21P/Giacobini-Zinner
Venus
Mars
(433) Eros
(1566) Icarus
Target body
6th International AAAF Symposium on Space Propulsion: Propulsion for Space Transportation of the XXIst Century, Versailles, France, 14-16 May 2002
Copyright (c) 2002 by DLR
ms
A
mp
ms + mp
m
=
= s +
A
A
A
is defined accordingly as the specific mass of sailcraft including the payload (index p). It should
be noted, that the term payload stands for the
total sailcraft except the solar sail assembly (i.e.
except the propulsion system).
s =
ac
5.93 mm/s2
The square solar sail consisted of four CFRP (Carbon Fiber Reinforced Plastics) booms with a specific
mass of 101 g/m and of four triangular sail segments
made of aluminum-coated (0.1 m) plastic films with
where = GMsun .
6th International AAAF Symposium on Space Propulsion: Propulsion for Space Transportation of the XXIst Century, Versailles, France, 14-16 May 2002
Copyright (c) 2002 by DLR
Near Earth Asteroids (NEAs) are a promising category of target bodies for a first solar sailcraft mission,
since they can be accessed relatively easily and since
they are of great scientific interest. Therefore, in August 2000, a dedicated mission for the exploration of
NEAs with solar sailcraft (ENEAS) was proposed by
DLR in cooperation with the Westf
alische WilhelmsUniversit
at at M
unster (Germany) as a candidate
a thickness between 4 and 12 m. The booms consisted of two CFRP shells that were bonded at the
edges to form a tubular shape, so that they can be
pressed flat and rolled up (Figure 5).
FSRP,c
A
s
ms
mp
m
ac
(50 m)2
29.2 g/m2
73 kg
65.5 kg
138.5 kg
55.4 g/m2
1/42.4
0.14 mm/s2
19.5 mN
Sail area
Sail assembly loading
Sail assembly mass
Payload mass
Total sailcraft mass
Sailcraft loading
Lightness number
Characteristic acceleration
Characteristic SRP force
within the German small satellite program for extraterrestric sciences [3][14]. Based on the successful deployment experiment described above, ENEAS
(Figure 6) was intended to feature a deployable 50 m
50 m solar sail that would be capable to transport a
micro-satellite with a mass of 65.5 kg to a NEA within
less than five years. Table 2 summarizes the ENEAS
parameters.
6th International AAAF Symposium on Space Propulsion: Propulsion for Space Transportation of the XXIst Century, Versailles, France, 14-16 May 2002
Copyright (c) 2002 by DLR
(Peff )1AU
m ,
s + s2p
ac =
Looking at the equation for the characteristic acceleration of solar sailcraft with a square sail,
It should be noted that mp and s can be chosen independently, whereas s (s) is a function of s with s /s < 0, since
the mass of the booms and the deployment module scale less
than linearly with the sail area. However, we are on the safe
side, when we assume s /s = 0 to keep calculations simple.
By different combinations of the three design parameters any desired characteristic acceleration can
be achieved2 . An increase in payload mass can, for
example, be offset with a proportional increase of s2
or with a (not inversely proportional) decrease of s .
6th International AAAF Symposium on Space Propulsion: Propulsion for Space Transportation of the XXIst Century, Versailles, France, 14-16 May 2002
Copyright (c) 2002 by DLR
1
1 + mp /s s2
1
m p =
1 + s s2 /mp
2
s =
1 + s s2 /mp
s =
3
similar to the Rosetta mission to comet 46P/Wirtanen,
which will have three intermediate gravity assist maneuvers
(Mars-Earth-Earth) and a trip time of approximately nine years
Answer to Q1: At present, the maximum acceptable mission duration seems to be determined by the
trip time required with chemical propulsion, including (eventually multiple) gravity assist maneuvers.
Due to the relatively large v-requirement of about
6 10 km/s for a mission comparable to ENEAS-SR,
but with chemical propulsion, such a mission would
require either an expensive launch vehicle and heavy
spacecraft, resulting in a short trip time of a few years,
or several gravity assists, resulting in a long trip time3 .
Since our approach aims at low-cost missions, only the
gravity assist option seems to be a reasonable conventional alternative. Thus, for the ENEAS-SR mission
we assume a total mission duration of more than ten
years as not acceptable.
Answer to Q2: Trajectory calculations show, that
an ENEAS-SR mission to 1996FG3 can be achieved
even with a characteristic acceleration of 0.10 mm/s2
in 9.40 years, including a rendezvous trajectory of 6.27
years (2290 days, Figure 10), 340 days of operations
at the asteroid and an Earth return trajectory of 2.20
years (805 days, Figure 11).
with
ac
=
ac
Those design sensitivities can be determined quantitatively using sensitivity functions, which provide an
indication of the relative importance of each design
parameter for a given point in the solar sail design
space [10]. The sensitivity function for any design parameter {s , mp , s} may be written as
6th International AAAF Symposium on Space Propulsion: Propulsion for Space Transportation of the XXIst Century, Versailles, France, 14-16 May 2002
Copyright (c) 2002 by DLR
Since for solar sailcraft of moderate performance gaining orbital energy in the Earths gravitational field is
difficult and time consuming, the launcher will insert
the ENEAS-SR solar sailcraft directly into an interplanetary trajectory with a hyperbolic excess energy
of C3 = 0 km2 /s2 . After the injection, the sail and
FSRP,c
s
ms
mp
m
ac
29.2 g/m2
143 kg
237 kg
380 kg
77.6 g/m2
1/59.3
0.10 mm/s2
38.0 mN
Those quasi-stationary hovering positions are unstable but can be stabilized using a feedback control
loop to sail attitude alone [11]. Hovering near the asteroid, the (likely complex) gravitational field of the
target body is studied, so that a coarse gravitational
field model can be determined. Thereafter, the lander with the Earth return capsule is separated from
the attitude control mast are deployed in a 3-axis stabilized mode. Then the sail is oriented to follow a
pre-calculated attitude profile, leading to an optimal
interplanetary transfer trajectory. During the transfer, the ENEAS-SR solar sailcraft will run almost autonomously, so that ground monitoring will be carried
out on a weekly basis only. At the end of the transfer
trajectory the solar sailcraft will be making a rendezvous with 1996FG3 within its gravitational sphere
of influence (Hill-sphere) of between 70 km radius (at
perihelion) and 150 km radius (at aphelion). Even in
the near-field of the asteroid, the SRP acceleration
of between 0.05 mm/s2 (at aphelion) and 0.21 mm/s2
(at perihelion) is larger than the asteroids gravitational acceleration (0.01 to 0.00005 mm/s2 in a distance ranging from 5 to 50 km), so that the sailcraft
is able to hover on an artificial equilibrium surface in
the hemisphere that is opposite to the sun (Figure 13).
6th International AAAF Symposium on Space Propulsion: Propulsion for Space Transportation of the XXIst Century, Versailles, France, 14-16 May 2002
Copyright (c) 2002 by DLR
Sail area
A
(70 m)2
We have investigated the minimum solar sailcraft performance requirements for various interplanetary missions. We were able to show, that the characteristic
Summary
We have investigated the performance of near-term solar sailcraft also for rendezvous missions with celestial
bodies other than Near Earth Objects. Table 4 gives
the minimum rendezvous times for solar sailcraft with
a characteristic acceleration of 0.10 to 0.20 mm/s2 for
several target bodies. It shows that even with nearterm solar sailcraft planetary rendezvous mission are
feasible within the inner solar system, if relatively long
trip times can be tolerated. However, a characteristic
acceleration of 0.10 mm/s2 seems to be a lower bound
for rendezvous missions within the inner solar system.
the solar sail to go into closer orbit about the asteroid. While measuring the asteroids gravitational field
with increasing accuracy, the orbit of the lander is continuously lowered until a safe landing trajectory can
be computed (some or all of those extensive computations may be performed on Earth). Once landed,
the sample is fed directly into the Earth return capsule and brought back to the hovering sailcraft. In
this mission phase, the sailcraft is waiting edge-on
(so that no SRP force is acting on the sail) at the
L2 Lagrange point for the lander in order to assist
the rendezvous. The lander design, the sample extraction mechanisms and the subsystems required to
rendezvous the waiting sailcraft require further studies and are beyond the scope of this paper. Since
1996FG3 is a binary system, it would be interesting
to land and extract samples from both bodies to investigate the origin and the collisional evolution of the
1996FG3 system. Since the gravitational acceleration
is very low near the asteroid and the required v for
the lander less than 10 m/s, a cold gas system with a
propellant mass of less than 4 kg will suffice to perform all operations. After rendezvous with the hovering sailcraft, the re-docked ENEAS-SR solar sailcraft
returns the sample to Earth. The return trajectory is
much faster than the transfer trajectory to 1996FG3
since no rendezvous is required at Earth. Thus, the
sailcraft may arrive with a relatively large hyperbolic excess velocity of about 8.4 km/s. The gravitational acceleration of Earth adds another 11.2 km/s,
so
that the Earth reentry velocity may reach about
8.42 + 11.22 km/s = 14.0 km/s. Finally, just before
the arrival of the ENEAS-SR solar sailcraft at Earth,
the return capsule is separated from the lander and
injected into an Earth reentry trajectory, where it
is decelerated by atmospheric friction and breaking
parachutes.
20.1
6.8
6.2
Mercury
Venus
Mars
References
vmin
[ km/s]
Target
body
6th International AAAF Symposium on Space Propulsion: Propulsion for Space Transportation of the XXIst Century, Versailles, France, 14-16 May 2002
Copyright (c) 2002 by DLR
[11] E. Morrow, D. J. Scheeres, and D. Lubin. Solar sail orbit operations at asteroids. Journal of
Spacecraft and Rockets, 38:279286, 2001.
[10] C. R. McInnes. Solar Sailing. Technology, Dynamics and Mission Applications. Springer
Praxis Series in Space Science and Technology.
SpringerPraxis, Berlin, Heidelberg, New York,
Chicester, 1999.
6th International AAAF Symposium on Space Propulsion: Propulsion for Space Transportation of the XXIst Century, Versailles, France, 14-16 May 2002
Copyright (c) 2002 by DLR
Chapter 2
Electrostatics
Introduction
Chapter 1 contains in principle all laws and properties necessary to solve any problem in electromagnetism. The rest of the course will therefore be applications of these laws and properties in
specific cases.
As stated in section 1.1.4.a, in the static regime (or steady state), where the sources {, J } do
not vary in time, the coupling between the electric and magnetic field vanishes. The electric field
is only due to static charges, whereas the magnetic field is only created by constant currents. One
can therefore study the electric field - Electrostatics - and the magnetic field - Magnetostatics
- separately.
In this chapter, we will concentrate on Electrostatics. In section 2.1, we will briefly go over the
relations valid in electrostatics. These are only the retranscription of the relations seen in chapter
1 in the case of time-independent sources.
2.1
2.1.1
F =
u qq
40 r2
(2.1)
where
u qq is the unit vector going from the source charge q to the charge q and 0 the vacuum
permittivity or permittivity of free space :
0 8, 85 1012 F.m1 = 8, 85 1012 kg1 .m3 .A2 .s4
In the case of multiple charges {qi }i acting on a charge q , Coulombs law can be written as:
F =
q X qi
ui
40 i ri2
(2.2)
where ri is the distance between the source charge qi and the charge q , and
ui is the unit vector
q
(P )
F (M ) =
P M dV
(2.3)
40
P
M3
(V )
Note that these three expressions of Coulombs law are strictly identical, albeit the form they
are written in has been chosen for convenience.
2.1.2
The decoupling between the electric and magnetic fields reduces the number of equations to be
taken into account. The Maxwells equations of electrostatics are :
Maxwell-Gauss
Maxwell-Faraday
2.1.3
Electrostatic field
2.1.3.a
Electrostatic field
div E = . E =
0
rot E = E = 0
(2.4)
(2.5)
E (M ) =
ur
40 r2
(2.6)
where
u r is the unit vector going from the source charge q to point M.
In the case of multiple charges {qi }, the field created at point M can be written as:
E (M ) =
1 X qi
ui
40 i ri2
(2.7)
where ri is the distance between the source charge qi and point M , and
ui is the unit vector going
from the source charge qi to point M .
as:
In the case of a distribution of charges {(P )}, the field created at point M can be written
(P )
1
P M dV
(2.8)
E (M ) =
40
P
M3
(V )
2.1.3.b
Typical values
The table below gives the orders of magnitude for a few electric fields :
0.5 mV.m1
50 V.m1
1 - 100V.m1
102 V.m1
106 V.m1
108 - 1012 V.m1
34
CHAPTER 2. Electrostatics
2.1.3.c
Conservation of charge
div J = 0
(2.9)
non-hatched surface of (), J .dS = 0. In other words, no current density goes in or out of the
tube except at the two ends S1 and S2 . Let us moreover suppose that the flux tube is sufficiently
0 =
div J
(V )
=
J .dS
(S )+()+(S2 )
1
=
J .dS +
J .dS +
J .dS
(S1 )
()
(S2 )
J 1 S1 J 2 S2
Hence J1 S1 = J2 S2 . The flux of current density going through S1 is conserved when going out of
S2 .
Gauss law
Let us consider a closed surface (S) delimiting a volume (V ) which contains a charge volume
density and a total charge Qint . Maxwell-Gausss equation gives Gauss law:
Qint
E .dS =
dV =
(2.10)
0
(S)
(V ) 0
This relation is one of the most useful relations in electrostatics. As we will see in section 2.2, along
with the use of symmetries and invariances, it allows to determine the electric field in numerous
simple situations.
E . dl = 0
(2.11)
(C)
n 12
E2 E1 =
0
(2.12)
where E i is the electric field in the medium i, is the surface charge density at the interface
2.1.4
Electrostatic potential
2.1.4.a
Electrostatic potential
E = grad
(2.13)
or equivalently :
(A) (B) =
E . dl
(2.14)
q
40 r
(2.15)
In the case of multiple charges {qi }i , the potential created at point M can be written as:
(M ) =
1 X qi
40 i ri
(2.16)
36
CHAPTER 2. Electrostatics
In the case of a distribution of charges {(P )}, the potential created at point M can be
written as:
1
(P )
(M ) =
dV
(2.17)
40
(V ) P M
Note : The electrostatic potential (r) of a group of charges vanishes at infinity. This is the
origin of the convention we will often use :
(r) 0 when r +
2.1.4.b
(2.18)
Potential propagation
In electrostatics, the potential propagation (see section 1.3.4) is written under the form of
Poissons equation :
=
(2.19)
0
In the specific case where there is no charge ( = 0), the propagation equation reduces to
Laplaces equation :
= 0
(2.20)
The Poisson (or Laplace) equation will have an unique solution if boundary conditions are
specified on a closed surface :
Either the potential is defined on a closed surface. This is called the Dirichlet boundary
condition.
Or the electric field is defined on a closed surface. This is called the Neumann boundary
condition.
Proof : Let us suppose that 1 and 2 obey to the same Poisson equation = 0 inside a
volume (V ). Let :
U = 2 1
Let us impose Dirichlet boundary condition on the closed bounding surface (S) of the volume (V ).
Then, inside (V ) :
U = 2 1 = +
=0
0
0
and on (S), U = 0.
We now need to prove Greens first identity :
+ grad . grad dV =
(V )
(S)
dS
n
grad :
grad .
n dS
div grad dV =
(S)
(V )
+ grad . grad dV =
dS
(V )
(S) n
2
U U + grad U
dV =
(V )
(S)
U
dS
n
which means that grad U = 0 inside (V ). Hence U is constant in (V ) and since U = 0 on (S),
U = 0 in all (V ). Therefore 1 = 2 and the solution is unique.
2.1.5
2.1.5.a
Field lines
Definition : Electric field lines are the lines that are, at all points in space, tangent to
E (M ) dl = 0
(2.21)
where dl is an infinitesimal vector along the field line, centered around point M .
2.1.5.b
Equipotential surfaces
Properties
Property #1 : The electrostatic potential and the electric field reproduce all symmetries
and invariances of the charge distribution (P ) that create this potential and this field.
Property #2 : Equipotential surfaces are normal to the electric field E at all points and
therefore are normal to the field lines.
Proof : Equation 2.14 shows that if points A and B are points sufficiently close to one another
on the same equipotential surface :
(A) (B) = 0 =
E . dl
with dl defining the equipotential surface. Therefore E dl .
This relation is only valid in electrostatics. In the general case, the term in
taken into account.
A
t
has to be
Property #3 : Field lines cannot be closed. They begin and end either on a charge or at
infinity.
Proof : Let us suppose that (C) is a closed field line. In that case, by definition of a field line
( E parallel to dl ) :
E dl 0
E . dl =
(C)
(C)
38
CHAPTER 2. Electrostatics
E . dl = 0
(C)
(C)
E . dl = 0 and therefore E = 0 .
In other words, when the field lines move apart, the norm of E diminishes.
E .dS > 0
()
Qint
E .dS =
0
()
2.1.6
Energetics in electrostatic
2.1.6.a
Electrostatic energy
0 E 2
2
(2.22)
U=
u dV
(V )
U=
1
2
0 E 2 dV 0
(2.23)
(V )
This energy is always positive. It implies that (+) = 0 and that the presence of a field adds
energy to the system.
39
In section 1.3.2, we have seen that q corresponds to the potential energy of a charge q placed
in a potential . In other words, if the charge q is brought from infinity (where = 0) to a point
x where the electrostatic potential is , the work done on the charge is
W = q
N
qj
qi X
40 j=1 |
rj
ri |
N
X
i=1
Wi =
N
1 X X qi qj
1 X qi qj
=
40 i=1 j<i | rj ri |
80
|
rj
ri |
(i,j)
where the terms i = j are not included in the sum. This can be rewritten as :
W =
where i =
2.2
2.2.1
1
40
qj
j |
rj
ri |
1X
qi i
2 i
Let us consider an uniformly charged infinite wire as schematized figure 2.2.1. The uniform
linear charge density is . Let us determine the electrostatic field and potential created by such a
system.
40
CHAPTER 2. Electrostatics
The problem is of cylindrical symmetry. We will therefore use the cylindrical coordinates. Since
the charge distribution (r, , z),
and the electrostatic potential conserves the symmetries of the
charge distribution, (r, , z).
Hence :
d
E = grad =
ur
dr
Another way of obtaining this result is to say that all point M of space belongs to two symmetry
,
planes : the one defined by {
u
r uz } and the one defined by {ur , u }. Since at these points M the
electric field must belong to all symmetry planes, E k
ur .
We will apply Gauss law on the following Gauss surface : a cylinder of height h and radius r :
Qint
E .dS =
0
(SGauss )
h
E .dS +
E .dS +
E .dS =
0
(S1 )
(S2 )
(Slat )
h
0 + 0 + 2rhE(r) =
0
E(r) =
20 r
E =
u
r
20 r
2.2.2
Let us consider an uniformly charged infinite plane as schematized figure 2.2.2. The uniform
surface charge density is . Let us determine the electrostatic field and potential created by such
a system.
41
The problem is of cartesian symmetry. We will therefore use the cartesian coordinates. Since
(x,
y,
z), and the electrostatic potential conserves the symmetries of the charge distribution,
(x,
y,
z). Hence :
d
uz
E = grad =
dz
Moreover, the {Oxy} plane is a symmetry plane, so that - using E k
uz :
E (x, y, z) = SP E (x, y, z)
E (x, y, z) = E (x, y, z)
We will apply Gauss law on the following Gauss surface : a cylinder of height h and radius R,
centered around the {Oxy} plane :
Qint
E .dS =
0
(SGauss )
R2
E .dS +
E .dS +
E .dS =
0
(S1 )
(S2 )
(Slat )
E(z)R2 E(z)R2 + 0 =
E(z)
R2
0
20
u
z
20
From the relation between and E , one derives : = 20 z + constant. If one imposes = 0 on
the conductor, one obtains : = 20 z.
2.2.3
Let us consider an uniformly charged sphere as schematized figure 2.2.3. The uniform volume
charge density is . Let us determine the electrostatic field and potential created by such a system.
42
CHAPTER 2. Electrostatics
Figure 2.4: Uniformly charged sphere.
The problem is of spherical symmetry. We will therefore use the spherical coordinates. Since
the charge distribution (r, ,
), and the electrostatic potential conserves the symmetries of the
charge distribution, (r, ,
). Hence :
d
E = grad =
ur
dr
are symmetry planes.
Another way of obtaining this result is to say that all planes containing
u
r
Qint
E .dS =
0
(SGauss )
3
.dS
= 4R
E
u
u
r
r
30
(SGauss )
2
4R3
E(r)r2 sind d =
30
=0 =0
E(r)4r2
E(r)
4R3
30
R3
30 r2
R3
u
r
30 r2
R3
From the relation between and E , one derives : = 3
+ constant. If one imposes = 0 on
0r
3
R
the conductor, one obtains : = 3
+
0r
2.3
2.3.1
R2
30 .
Conductors
Electrostatic field and potential inside a conductor
An electrical conductor is a material that contains numerous free electrons. This means that
these electrons - usually from the outer shells of the atoms constituting the conductor - are able to
freely move inside the material. In particular, if a conductor is submitted to an external electric
field E ext , the electrons will be set in motion by this field and will only stop once equilibrium is
E conductor = 0
(2.24)
(2.25)
43
2.3. CONDUCTORS
2.3.2
Let us consider an electrical conductor. As we have seen, the electrostatic field inside the
conductor E int is zero : E int = 0 . Then Maxwell-Gauss equation gives that the volume charge
(2.26)
In practice, all charges accumulate on a few atomic layers at the surface of the conductor. These
are therefore surface charges.
2.3.3
Thus, under the action of an external magnetic field, charges accumulate at the surface of the
conductor. These surface charge, in turn, create an electric field such that, in the conductor the
total field is zero.
Property #4 : Outside, but at the immediate vicinity of an electrical conductor, the external
electric field is :
n
E ext =
0
(2.27)
where
n is the vector normal to the conductor surface.
2.3.4
.
s
: E1 = 2
u
z
0
ii. The field E2 created by all other surface charges on the conductor.
We will call the surface of the conductor from which the surface
dS is extracted.
Since the field in Me is :
u
E = E 1 + E2 =
z
0
44
CHAPTER 2. Electrostatics
E2 =
u
z
20
Now, let us calculate the force exerted by the surface charges of on those on dS :
u
dF = dq E2 (Me ) = s dS
z
20
2
This force is equivalent to a pressure since proportional to dS. The pressure therefore is P = 2s0 .
This pressure is called the radiation pressure.
A few comments :
To say that the field E1 is the same than the one created by an infinite charged plane might
look like an approximation. However, it can be shown that it is the exact result. Indeed, the
electric field outside the conductor is :
u
E ext = E1 (Me ) + E2 (Me ) =
z
0
The field inside the conductor (at Mi ) can also be decomposed into a component due to the
field created by dS : E1 (Mi ) = E1 (Me ), and a component due to the field created by
E1 (Me )
E2 (Me )
u
z
20
s
u
z
20
It is interesting to note that the field in the vicinity of a conductor, at point Me only depends
on the local surface charge density s (Me ), i.e. the one carried by the small conductor surface
dS. The charges of the rest of the conductor (on ), organize themselves so as to cancel,
inside the conductor, the field created by those carried by dS.
2.3.5
In the presence of an constant external electric field E 0 , let us proceed with an analogous reasoning. The total external electric field can
be written :
The field inside the conductor, at point Mi can also be decomposed into
a component due to the field created by the charges on dS, which value is E1 (Mi ) = E1 (Me ),
and a component due to the field created by the charges on , which value is E2 (Mi ) = E2 (Me ) :
E1 (Me )
E0 + E2 (Me )
u
z
20
s
u
z
20
45
2.3. CONDUCTORS
This time, the charges on organize themselves so as to exactly compensate the field created
by charges on dS, taking into account the external electric field. The radiation pressure exerted on
dS corresponds to the force exerted on the surface charges on dS due to the external field E0 and
2
N.B. 2 : A conducting object placed within an uniform electrostatic field E0 charges itself at
its surface, so that the electric field outside but at the immediate vicinity of this object is not
necessarily uniform !
N.B. 2 : In this particular case, one cannot apply the principle of superposition in a simple manner
(i.e. a charged conductor in E0 = a charged conductor in zero field + a static field). This is due
to the fact that the boundary conditions, at infinity in particular, are not trivial.
2.3.6
Let us now mention the dependence of the field with respect to shape. More specifically, we
will qualitatively discuss the field near a sharp conductor.
First, let us remark that if one takes two spheres of radius R1 and R2 R1 , both charged with
Q. According to the calculation made section 2.2.3, the electric fields respectively verify :
4R12 E1
4R22 E2
Q
0
Q
0
Hence :
E1
E2
Q
4R12 0
Q
4R22 0
As can be seen, the field near the smaller sphere is larger than the one in the vicinity of the
larger sphere. In other words, any sharp point in a material has a large electric field near its surface.
This result is important for practical realizations. Indeed, above a certain electric field, the air
will break down. In this case, a loose charge - whether electron or ion - will be accelerated by the
field, will then collide with an atom, knocking out an electron of this atom. The additional ion and
electron will, in turn, be accelerated and a cascade process will begin, resulting in a great number
of charged particules in the air. Their motion creates a discharge or a spark.
Application #1 : If one wants to put a conductor to a high potential, one must avoid the
conductor surface to have any sharp surface features. Otherwise the air will discharge towards this
tip.
Application #2 : This field effect is used by field-emission microscopy to image the material of
a thin needle 1 . This technique has enabled, since 1955, to image materials with a resolution of a
few Angstr
oms, well before the advent of local probes such as the Scanning Tunneling Microscope
(STM) or the Atomic Force Microscopie (AFM).
1. For explanations on the working principle, see for example : Field Ion Microscopy, B. Gault et al., Atom
Probe Microscopy, accessible on the Internet.
46
2.4
CHAPTER 2. Electrostatics
Electrostatic dipole
The notion of electrostatic dipole is important both in physics and chemistry. Indeed, in
chemistry, it can model the behavior of polar molecules - such as HCl for instance - under a
magnetic field. In physics, we will see in the next chapter that it is at the basis of the microscopic
description of dielectric materials.
2.4.1
Definition
p =
qi APi
(2.28)
i
Example : The simplest electrostatic dipole is a charge doublet {+q, q} (q > 0). If +q is placed
p = +q AP q AN = q N P = qd
u
Note : A dipole which moment is not affected by the application of an external electric field is
called rigid.
2.4.2
2.4.2.a
dipole
p placed at point A such that AM = r
u , with
u an unit vector, is :
(M ) =
p .
u
40 r2
(2.29)
47
X
i
1
Pi M
qi
40 Pi M
Pi M
2 1/2
2
OM OPi
1/2
OPi
r
(M )
:
=
!1/2
.
2r
u
OPi2
r OPi
1+ 2
r
r2
i
!
.
X qi
r
u
r OPi
+ ...
1+
40 r
r2
i
.
X qi
X qi
u
r OPi
+
40 r
40 r2
i
i
X
u
r
.
qi OPi
0+
2
40 r
i
qi
40 r
u
r p
(M )
40 r2
p cos
40 r2
48
2.4.2.b
CHAPTER 2. Electrostatics
Electrostatic field created by an electrostatic dipole
In the dipole approximation, the electrostatic field created at point M by an electrostatic dipole
3 (
p .
u ) .
u
p
E (M ) =
(2.30)
3
40 r
Proof : Taking the same definitions as above :
V (M )
Moreover :
3 (
p .
u ) .
u
p
40 r3
=
=
2.4.2.c
p cos
40 r2
grad
2p cos
3
4
0r
p sin 3
40 r
0
p cos
+ p sin
3p cos
u
u
u
r
r
40 r3
2p cos
4 r3
p sin0
40 r3
0
From equation 2.29, the equipotential surfaces are such that r2 = k cos with k a contant.
Field lines are obtained by taking, at all points, the normal to the equipotential surfaces. Field
lines and equipotential surfaces are schematized figure 2.6.
49
Figure 2.6: Schematic representation of the equipotential surfaces and field lines of a
dipole. Taken from http://www.pstcc.edu/departments/natural_behavioral_sciences/Web%
20Physics/Experim%2001.htm. The dipole is here modeled by two opposite charges.
Note : Bear in mind that the above expressions are obtained in the dipole approximation, which
means that they are not valid within the charge distribution.
2.4.3
2.4.3.a
Proof : Let us first calculate the resultant force (first order approximation) :
X
qi E (Pi )
F (A) =
i
P
q
E
(A)
+
AP
.
grad
E
(A)
i
x
i
x
P i
qi Ey (A) + APi . grad Ey (A)
i
P
i qi Ez (A) + APi . grad Ez (A)
P
qi APi . grad Ex (A)
Pi
q AP . grad E (A)
Pi i i y
qi APi . grad Ez (A)
i
p . grad A E (A)
M(A)
APi qi E (Pi )
APi qi E (A)
p E (A)
2.4.3.b
If the considered dipole is rigid, the charges composing the dipole do not move with respect to
one another.
p is then constant in norm. Then the resulting force is :
F (A)
Ue
grad A (Ue )
p . E (A)
50
CHAPTER 2. Electrostatics
Proof :
Ue
qi (Pi )
X
i
X
i
qi ext (Pi ) +
qi int (Pi )
qi ext (A) + APi . grad ext (A) + constant taken to be 0
0+
p . E (A)
The force exerted on the dipole therefore derives from the potential energy Ue . Then, the dipole
Figure 2.7: Schematization of electrostatic dipole moment of a molecule of water. Taken from http:
//guweb2.gonzaga.edu/faculty/cronk/CHEM101pub/L18-index.cfm?L18resource=water.
Figure 2.8: Schematization of the hydration of Na+ by water molecules. Taken from http://www.
science.uwaterloo.ca/~cchieh/cact/applychem/hydration.html.
Example #2: Van der Waals interactions - Van der Waals forces 2 group different forces, all
proportional to 1/r7 . You will derive some of these in the Electrostatics exercise sheet. These
2. These intermolecular forces have first been predicted by the Dutch physicist J.D. van der Waals in his PhD
thesis, defended in 1873.
51
forces originate from dipole-dipole interactions 3 and explain most of intermolecular interactions.
They have numerous applications, from scotch tape to biology 4 and including chemistry or polymer
science.
2.5
Methods in electrostatics
There are numerous solving methods in electrostatics which we will not review here 5 . For the
purpose of this course, you will mainly be using one of the following methods :
Directly compute the electric field using symmetries, invariances and :
E =
u dV
2
40
r
(V )
Directly compute the electrostatic potential using symmetries, invariances and :
1
dV
=
40
(V ) r
Determine the electric field using symmetries, invariances and Gauss theorem :
Qint
E .dS =
0
Also note that, for a problem with spherical symmetry, the method of separation of variables as seen in Quantum Mechanics for example - can be used.
2.5.1
Method of images
We will however detail one important method : the Method of images. This method can be
useful in the case of one or a few point charges in the presence of boundary surfaces that impose
discontinuities of the electric field. For instance, point charges in the presence of conductors.
The method consists in replacing the real problem with discontinuities - for eg. the conducting
planes - by an equivalent problem in an enlarged region, with no boundaries, where one or a few
virtual charges - the image charges -, are placed outside the region of interest so that :
1. these image charge simulate the boundary conditions ;
2. the electrostatics equation - the Poisson equation - remain unchanged in the region of interest.
The virtual problem with image charges is usually easier to solve. Then, thanks to the unicity
theorem (see section 1.6.2), the electric field that is found for the equivalent-charge problem is the
same, in the region of interest, than the one in the real problem.
Example : single charge in the presence of an infinite conducting plane
52
CHAPTER 2. Electrostatics
Figure 2.9: Schematization of the equivalence between a real problem and an equivalent-image
problem.
The method of images can be understood in the simple example schematized figure 2.9.
Let us consider a charge q sitting in vacuum in the presence of an infinite conducting plane,
at a distance d, occupying a semi-infinite space. As stated earlier, the potential in a conductor is
constant and since it extends to infinity, the corresponding potential is = 0. In particular, = 0
is imposed at the boundary between the vacuum and the conductor.
Moreover, in the region of interest - outside the conductor - the electric field is determined by
the Poisson equation : = 0 .
Let us now consider another problem where there is no conductor, but two point charges q and
q, separated by 2d. Then, in the middle of these two charges = 0 and in the region where
there previously was no conductor, the electric field is determined by the same Poisson equation
as before : = 0 .
The two problems therefore have the same Poisson equation in the volume of interest - the
vacuum in the initial problem -, with the same boundary conditions - = 0 where the boundary
between vacuum and conductor was before -, so that the solution, in the region of interest, is the
same for the two problems.
The corresponding field lines are schematized figure 2.10. The plain lines are the field lines for
the real problem, whereas the dashed lines correspond to the additional field lines corresponding
to the virtual problem, outside the region of interest.
53
Figure 2.10: Field of a charge near a plane conducting surface, found by the method of images.
Taken from R.P. Feynman, The Feynman Lectures, Chapter 6.
2.5.2
Numerical methods
For more complex problem - with no obvious symmetries for example - different solving softwares
are available. They generally require the user to enter the geometry of the boundary conditions
and the position of charges. They compute the electric and/or magnetic fields. Some applications
directly solve Maxwells equations, but these are practical for 2D systems, or systems with high
symmetries.
For problems where the exact solution would be too costly in computer time, softwares using
finite element analysis (FEA) can be used. These sets of numerical methods are used to find
approximate solutions to partial differential equations - in electrostatics or any other field -.
2.6
Additional reading
The following article 6 , describes the mechanism by which geckos adhere to the surface they
walk on. The ultimate mechanism at play implies van der Waals forces.
KELLAR AUTUMN2
AND
ANNE M. PEATTIE3
INTRODUCTION
Geckos seem to defy gravity as they run along
smooth vertical surfaces at up to 20 body lengths per
second (Autumn et al., 1999a), and even upside down
on the ceiling. Over two millennia ago, Aristotle commented on the ability of the gecko to run up and
down a tree in any way, even with the head downwards (Aristotle/Thompson, 1918, Book IX, Part 9).
How geckos adhere has been a gripping topic of scientific research for well over a century (Cartier, 1872;
Haase, 1900; Gadow, 1901; Weitlaner, 1902; Schmidt,
1904; Hora, 1923; Dellit, 1934; Mahendra, 1941;
Maderson, 1964; Ruibal and Ernst, 1965; Hiller, 1968,
1969, 1975; Gennaro, 1969; Russell, 1975, 1986; Williams and Peterson, 1982; Stork, 1983; Schleich and
Kastle, 1986; Irschick et al., 1996; Autumn et al.,
2000; Liang et al., 2000; Autumn et al., 2002). Rapid
locomotion on a vertical surface requires the ability to
generate parallel (frictional) forces equal to or greater
than body weight (Fig. 1A). Detachment is perhaps
even more important than attachment. After all, even
the most common household adhesives are more than
sufficient to hold statically the weight of a large gecko
(50 g), but repeated and rapid detachment without significant detachment forces is beyond the capability of
any current synthetic adhesive. The secret of geckos
adhesive capabilities lies in the structure and function
1081
SYNOPSIS.
The extraordinary adhesive capabilities of geckos have challenged explanation for millennia,
since Aristotle first recorded his observations. We have discovered many of the secrets of gecko adhesion,
yet the millions of dry, adhesive setae on the toes of geckos continue to generate puzzling new questions and
valuable answers. Each epidermally-derived, keratinous seta ends in hundreds of 200 nm spatular tips,
permitting intimate contact with rough and smooth surfaces alike. Prior studies suggested that adhesive
force in gecko setae was directly proportional to the water droplet contact angle (u) , an indicator of the
free surface energy of a substrate. In contrast, new theory suggests that adhesion energy between a gecko
seta and a surface (WGS) is in fact proportional to (1 1 cosu) , and only for u . 608. A reanalysis of prior
data, in combination with our recent study, support the van der Waals hypothesis of gecko adhesion, and
contradict surface hydrophobicity as a predictor of adhesion force. Previously, we and our collaborators
measured the force production of a single seta. Initial efforts to attach a seta failed because of improper 3D
orientation. However, by simulating the dynamics of gecko limbs during climbing (based on force plate data)
we discovered that, in single setae, a small normal preload, combined with a 5 mm displacement yielded a
very large adhesive force of 200 microNewton (mN), 10 times that predicted by whole-animal measurements.
6.5 million setae of a single tokay gecko attached maximally could generate 130 kg force. This raises the
question of how geckos manage to detach their feet in just 15 ms. We discovered that simply increasing the
angle that the setal shaft makes with the substrate to 308 causes detachment. Understanding how simultaneous attachment and release of millions of setae are controlled will require an approach that integrates
levels ranging from molecules to lizards.
1082
K. AUTUMN
AND
A. M. PEATTIE
1083
FIG. 1. Gecko adhesive structures and methods used to measure setal adhesive function. Images (A) and (B) by Mark Moffett. Figure modified
from Autumn et al. (2000), Nature.
A) Ventral view of a tokay gecko (Gekko gecko) climbing a vertical glass surface. Arrows represent the forces acting on the geckos feet
as it climbs.
B) Ventral view of the foot of a tokay gecko, showing seta-bearing scansors.
C) Setae are arranged in a nearly grid-like pattern on the ventral surface of each scansor. In this scanning electron micrograph, each
diamond-shaped structure is the branched end of a group of four setae clustered together in a tetrad. Box shows seta enlarged in (D).
D) Single isolated gecko seta used in measurements in Autumn et al. (2000). Box shows spatulae enlarged in (E).
E) Spatular tips of a single gecko seta.
F) Isolated seta adhering to a micro-electromechanical system (MEMS) cantilever capable of measuring forces parallel and perpendicular
to the surface. Arrow shows direction of manipulation during the experiment, simulating parallel forces generated during vertical locomotion
of the gecko.
G) Single seta adhering to a 25 mm aluminum bonding wire capable of measuring detachment force perpendicular to the surface. Arrow
shows direction of manipulation during the experiment, simulating perpendicular movement during detachment of the foot. (a is the angle
between the setal stalk and the wire.
K. AUTUMN
AND
A. M. PEATTIE
FIG. 2. Single seta measurements. Figure modified from Autumn et al. (2000), Nature.
A) Submaximal force of single seta parallel to the surface with a known perpendicular preload, as a function of time. Perpendicular preload
is designated by the dashed line. ts represents the time when the seta began to slide off the sensor. The initial perpendicular force need not be
maintained during the subsequent pull. Diagrams show the stages of setal movement corresponding to the force record from the MEMS
cantilever (Fig. 1F). Arrows indicate the direction of applied force to the seta. Vertical arrow indicates a parallel force, and a horizontal arrow
indicates a perpendicular force.
B) Setal force parallel to the surface during attachment as a function of perpendicular preload force. Setal force was taken to be the adhesive
force at the time just prior to sliding (ts; Fig. 2A). The solid line represents a seta with spatulae projecting toward the surface. The dashed
line represents the setal force with spatulae projecting away from the surface (parallel force 5 0.25 perpendicular preload 20.09; r2 5 0.64;
F 5 13; df 5 1,9; P 5 0.007). The force produced by the inactive, non-spatular region increased with normal or perpendicular force, typical
of materials with a coefficient of friction equal to 0.25. The perpendicular preloading force that could be applied attained a maximum (near
15 mN), because greater forces resulted in the setal buckling.
C) Maximal force after a maximum preload (15 mN) of a single seta parallel to the surface as a function of time. Diagrams show the
stages of setal movement corresponding to the force record from the MEMS cantilever (Fig. 1F). Arrows indicate the direction of force applied
to the seta. Vertical arrow indicates a parallel force; horizontal arrow indicates a perpendicular force. The maximum force (;200 mN) following
the small rearward displacement (5 mm) was 32 times that predicted from maximal whole animal estimates (see text). The large increase in
force during the rearward displacement may be caused by an increase the number of spatulae contacting the surface.
D) Change in the orientation of the setae may facilitate detachment. Setal angle (a) with the surface at detachment as a function of
perpendicular force. Filled symbols represent seta pulled away from the surface until release. Open symbols represent seta held at a constant
force as angle is increased. Each symbol shape represents a different seta. Data collected with wire gauge (Fig. 1G). Setal angle at detachment
changed by only 15% over the entire range of perpendicular forces. This observation is consistent with an adhesive model where sliding stops
when pulling at greater than the critical setal angle and hence stress can increase at a boundary, causing fracture of the contact.
1084
that the setae are bent out of this resting state, flattening the stalks between the toe and the substrate such
that their tips point distally. This small preload and a
micron-scale displacement of the toe or scansor proximally may serve to bring the spatulae (previously in
a variety of orientations) uniformly flush with the substrate, maximizing their surface area of contact. Adhesion results and the setae are ready to bear the load
of the animals body weight.
All 6.5 million (Schleich and Kastle, 1986; Irschick
et al., 1996) setae of a 50 g tokay gecko attached maximally could theoretically generate 1,300 N (133 kg
force) of adhesive forceenough to support the
weight of two humans. This suggests that a gecko need
only attach 3% of its setae to generate the greatest
forces measured in the whole animal (20N; Irschick et
al., 1996). Less than 0.04% of a geckos setae attached
maximally are needed to support its weight of 50 g on
a wall. At first glance, gecko feet seem to be enormously overbuilt. On further consideration, however,
there are some clear advantages to possessing as many
setae as possible (at which point we might ask what
factors actually limit a geckos adhesive capacity). It
is unlikely that all setae are able to achieve the same
orientation simultaneously. The proportion of spatulae
attached may be greatly reduced on rough surfaces
(particularly those with roughness on the same scale
as spatulae or setae; Autumn and Gorb in preparation).
On dusty or exfoliating surfaces, attachment to a wellanchored substrate will not be possible for every seta.
Large forces generated by perturbations during locomotion (e.g., recovering from a fall, predator avoidance, or station-keeping in high winds) may also utilize a greater proportion of geckos adhesive capacity.
1085
to answering biological questions (Lauder, 1991; Savageau, 1991; Ryan et al., 1998; Dickinson et al., 2000;
Autumn et al., 2002). How attachment and detachment
of millions of setae during locomotion is integrated
with the function of the scansor, toe, foot, leg, and
body remains a topic of great interest and ongoing
research (Sponberg et al., 2001; Russell, 2002). Russell has suggested that in the tokay (Gekko gecko), the
perpendicular preload and 5 mm drag requirements
(Autumn et al., 2000) are controlled by hydrostatic
pressure in the highly derived blood sinuses, and lateral digital tendon system, respectively (Russell,
2002).
Since gecko setae require a preload in the normal
axis for adhesion, large forces could potentially be associated with attachment of the foot. The tremendous
adhesive capacity of gecko setae suggests that large
forces could also occur during detachment. In fact, no
measurable ground reaction forces were associated
with either attachment or detachment during vertical
climbing on a force plate of the house gecko Hemidactylus garnoti (Autumn et al., 1999a, b), indicating
that these actions are either mechanically decoupled
from the center of mass in this species, or so small as
to be undetectable.
The absence of detachment forces is consistent with
(1) the mechanism of detachment and (2) the anatomy
of the foot. Geckos peel their toes up and away from
the substrate (digital hyperextension) rather than attempting to detach an entire foot at once, much like
removing a piece of tape. Peeling minimizes peak forces by spreading detachment out over time. Since the
muscles responsible for digital hyperextension (interossei dorsales; Russell, 1975) are located in the foot,
detachment does not have to be coupled mechanically
to the center of mass, as would be the case if the gecko
used its leg musculature to break the adhesive bonds
in the foot.
The absence of attachment forces is a more complicated issue, with at least three possible explanations.
As Russell (2002) has suggested, inflation of the digital blood sinuses (such as those present in individuals
of Hemidactylus and Gekko) may satisfy the preload
requirement of the setae during attachment without
generating measurable forces acting on the center of
mass. However, control of inflation and deflation of
the sinuses remains to be demonstrated. This mechanism would not be available to those species that lack
blood sinuses.
A second potential explanation is that setal preload
and drag are a consequence of force development during the stride. Climbing geckos use all four feet similarly to produce positive fore-aft forces parallel to the
surface that propel the gecko upwards (Autumn et al.,
1999b). Left legs apply a force to the right while right
legs apply a force to the left. Therefore, all four feet
pull medially, probably dragging the setae to engage
them fully, increasing the force of attachment (Autumn
et al., 1999b). However, geckos front legs pull the
center of mass into, while hind legs push the center of
K. AUTUMN
AND
1086
lit (1934) suggest that suction is not involved. Furthermore, our measurements of 10 atm of adhesion
pressure (Autumn et al., 2000) strongly contradict the
suction hypothesis.
Electrostatic attraction (Schmidt, 1904) is another
possible mechanism for adhesion in gecko setae. Experiments using X-ray bombardment (Dellit, 1934)
eliminated electrostatic attraction as a necessary mechanism for setal adhesion since the geckos were still
able to adhere in ionized air. However, electrostatic
effects could possibly enhance adhesion even if another mechanism is operating (Maderson, 1964).
Friction (Hora, 1923) and microinterlocking (Dellit,
1934; the climbers boot model of Mahendra, 1941)
may also play a secondary role, but the ability of geckos to adhere while inverted on polished glass, and the
presence of large adhesive forces on a molecularly
smooth SiO2 MEMS semiconductor (Autumn et al.,
2000) show that surface irregularities are not necessary
for adhesion, and may in fact be an impediment (Autumn and Gorb, in preparation)
A. M. PEATTIE
A
6pD 3
(1)
FIG. 3. Normalized force versus surface polarity (measured by water droplet contact angle, u) from Hiller (1968, 1969). Hillers (1968)
measurements of whole geckos on surfaces of varying polarity (circles) showed a correlation between polarity and force (y 5 20.012x
1 1.26; R2 5 0.85). In a separate study, Hiller (1969) found a weakly
significant correlation between force and surface polarity on polyethylene films modified with corona discharge (squares; y 5
20.012x 1 1.79; R2 5 0.69).
1087
K. AUTUMN
AND
(2)
1088
force on the strongly hydrophilic (u 5 08) silicon dioxide (SiO2) semiconductor surface. We also compared the perpendicular force of single isolated gecko
setae on hydrophilic (SiO2, u 5 08) and hydrophobic
(Si, u 5 81.98) micro-electro-mechanical-systems
(MEMS) force sensors. If wet, capillary adhesive forces dominate, we expected a lack of adhesion on the
strongly hydrophobic GaAs and Si MEMS surfaces. In
contrast, if van der Waals forces dominate, we predicted large adhesive forces on the hydrophobic, but
polarizable GaAs and Si MEMS surfaces. In either
case we expected strong adhesion to the hydrophilic
SiO2 semiconductor and MEMS control surfaces.
We showed that tokay gecko setae are strongly hydrophobic (160.98; Autumn et al., 2002), probably a
consequence of the hydrophobic side groups of -keratin (Gregg and Rogers, 1984). Parallel stress of live
gecko toes on GaAs and SiO2 semiconductors was not
significantly different, and adhesion of a single gecko
seta on the hydrophobic and hydrophobic MEMS cantilevers differed by only 2%. These results reject the
hypothesis that water contact angle (u) of a surface
predicts attachment forces in gecko setae, as suggested
by Hiller (1968, 1969), and are consistent with our
reanalysis (above). Since van der Waals force is the
only mechanism that can cause hydrophobic surfaces
to adhere in air (Israelachvili, 1992), the GaAs and
hydrophobic MEMS semiconductor experiments provide direct evidence that van der Waals force is the
mechanism of adhesion in gecko setae, and that waterbased capillary forces are not significant.
A. M. PEATTIE
Alexander, R. M. N. 1992. Exploring biomechanics: Animals in motion. Scientific American, New York.
Aristotle, Historia Animalium, trans. D. A. W. Thompson. 1918. Clarendon, Oxford, http://classics.mit.edu/Aristotle/historypanim.html.
Attygalle, A. B., D. J. Aneshansley, J. Meinwald, and T. Eisner.
2000. Defense by foot adhesion in a chrysomelid beetle (Hemisphaerota cyanea): Characterization of the adhesive oil. Zoology (Jena). 103:16.
Autumn, K., S. T. Hsieh, D. M. Dudek, J. Chen, C. Chitaphan, and
R. J. Full. 1999a. Dynamics of geckos running vertically. Amer.
Zool. 38:84A.
Autumn, K., S. T. Hsieh, D. M. Dudek, J. Chen, C. Chitaphan, and
R. J. Full. 1999b. Function of feet in ascending and descending
geckos. Amer. Zool. 38:84A.
Autumn, K., Y. A. Liang, S. T. Hsieh, W. Zesch, W.-P. Chan, W. T.
Kenny, R. Fearing, and R. J. Full. 2000b. Adhesive force of a
single gecko foot-hair. Nature 405:681685.
REFERENCES
ACKNOWLEDGMENTS
The research described in this paper was the result
of collaboration between the Autumn, Full, Fearing,
and Kenny labs, and was supported under the auspices
of Alan Rudolph, director of DARPAs Controlled Biological and Biomimetic Systems program, N6600100-C-8047 and N66001-01-C-8072. Thanks to Michael Broide, Walter Federle, Valeurie Friedman,
Amanda Gassett, Wendy Hansen, Tony Russell, Simon
Sponberg, and two anonymous reviewers. We are especially grateful to Jacob Israelachvili for Equation 2.
1089
Autumn, K., M. J. Ryan, and D. B. Wake. 2002. Integrating historical and mechanistic biology enhances the study of adaptation.
Q. Rev. Biol. (In press).
Baier, R. E., E. G. Shafrin, and W. A. Zisman. 1968. Adhesion:
Mechanisms that assist or impede it. Science 162:13601368.
Bauchhenss, E. and M. Renner. 1977. Pulvillus of Calliphora erythrocephala Meig (Diptera: Calliphoridae). Int. J. Insect Morphol. 6:225227.
Bauer, A. M. and A. P. Russell. 1990. Alternative digital scansor
design in the New Caledonian Gekkonid genera Bavayia and
Eurydactylodes. Mem. Queensland Mus. 29:299310.
Betz, O. 1996. Function and evolution of the adhesion-capture apparatus of Stenus species (Coleoptera, Staphylinidae). Zoomorphology 116:1534.
Blackwall, J. 1845. On the means by which walk various animals
on the vertical surface of polished bodies. Ann. Nat. Hist. XV:
115.
Brainerd, E. L. 1994. Adhesion force of ants on smooth surfaces.
Amer. Zool. 34:128A.
Cartier, O. 1872. Studien uber den feineren Bau der Haut bei den
Reptilien. I. Abt. Epidermis der Geckotiden. Verhandl. Wurz.
Phys.-med. Gesell. 1:8396.
Cartier, O. 1874. Studien uber den feineren Bau der Epidermis bei
ber die Wachsthumserscheinungen der
den Reptilien. II. Abt. U
Oberhaut von Schlangen und Eidechsen bei der Hautung. Arb.
zool.-zoot. Inst. Wurz. 1:239258.
Christenson, H. K. 1993. Adhesion and surface energy of mica in
air and water. J. Phys. Chem. 97:1203441.
Chui, B. W., T. W. Kenny, H. J. Mamin, B. D. Terris, and D. Rugar.
1998. Independent detection of vertical and lateral forces with
a sidewall-implanted dual-axis piezoresistive cantilever. Appl.
Phys. Lett. 72:13881390.
Dellit, W.-D. 1934. Zur anatomie und physiologie der Geckozehe.
Jena. Z. Naturw. 68:613656.
Dickinson, M. H., C. T. Farley, R. J. Full, M. A. Koehl, R. Kram,
and S. Lehman. 2000. How animals move: An integrative view.
Science 288:100106.
Dixon, A. F. G., P. C. Croghan, and R. P. Gowing. 1990. The mechanism by which aphids adhere to smoth surfaces. J. Exp. Biol.
152:243253.
Edwards, J. S. and M. Tarkanian. 1970. The adhesive pads of Heteroptera: A re-examination. Proc. R. Entomol. Soc. London 45:
15.
Eisner, T. and D. J. Aneshansley. 2000. Defense by foot adhesion in
a beetle (Hemisphaerota cyanea). Proc. Nat. Acad. Sci. U.S.A.
97:65686573.
Emerson, S. B. and D. Diehl. 1980. Toe pad morphology and mechanisms of sticking in frogs. Biol. J. Linn. Soc. 13:199216.
Gadow, H. 1901. The Cambridge natural history. Vol. 8. Amphibia
and reptiles. McMillan and Co., London.
Gay, C. 2002. StickinessSome fundamentals of adhesion. Integr.
Comp. Biol. 42:11231126.
Gay, C. and L. Leibler. 1999. Theory of tackiness. Phys. Rev. Lett.
82:936939.
Gennaro, J. G. J. 1969. The gecko grip. Nat. Hist. 78:3643.
Gillett, J. D. and V. B. Wigglesworth. 1932. The climbing organ of
an insect, Rhodnius prolixus (Hemiptera, Reduviidae). Proc. R.
Soc. London B 111:364376.
Gorb, S. N. 1998. The design of the fly adhesive pad: Distal tenent
setae are adapted to the delivery of an adhesive secretion. Proc.
R. Soc. London B 265:747752.
Gorb, S. N. 2001. Attachment devices of insect cuticle. Kluwer Academic Publishers, Dordrecht.
Gorb, S. N. and R. G. Beutel. 2001. Evolution of locomotory attachment pads of hexapods. Naturwissenschaften. 88:530534.
Gorb, S. N., E. Gorb, and V. Kastner. 2001. Scale effects on the
attachment pads and friction forces in syrphid flies (Diptera,
Syrphidae). J. Exp. Biol. 204:14211431.
Green, D. M. 1981. Adhesion and the toe-pads of treefrogs. Copeia
4:790796.
Gregg, K. and G. E. Rogers. 1984. Feather keratin: Composition,
structure, and biogenesis. In J. Bereiter-Hahn, A. G. Matoltsy,
K. AUTUMN
AND
and K. S. Richards (eds.), Biology of the integument 2: Vertebrates, pp. 666694. Springer-Verlag, New York, New York.
Haase, A. 1900. Untersuchungen uber den Bau und die Entwicklung
der Haftlappen bei den Geckotiden. Archiv. f. Naturgesch. 66:
321345.
Hanna, G. and W. J. P. Barnes. 1991. Adhesion and detatchment of
the toe pads of tree frogs. J. Exp. Biol. 155:103125.
Hepworth, J. 1854. On the structure of the foot of the fly. Q. J.
Micros. Sci. 2:158163.
Hill, D. E. 1977. The pretarsus of salticid spiders. Zool. J. Linn.
Soc. 60:319338.
Hiller, U. 1968. Untersuchungen zum Feinbau und zur Funktion der
Haftborsten von Reptilien. Z. Morph. Tiere. 62:307362.
Hiller, U. 1969. Correlation between corona-discharge of polyethylene-films and the adhering power of Tarentola m. mauritanica
(Rept.). Forma et functio. 1:350352.
Hiller, U. 1975. Comparative studies on the functional morphology
of two gekkonid lizards. J. Bombay Nat. Hist. Soc. 73:278282.
Hora, S. L. 1923. The adhesive apparatus on the toes of certain
geckos and tree frogs. J. Proc. Asiat. Soc. Beng. 9:137145.
Irschick, D. J., C. C. Austin, K. Petren, R. Fisher, J. B. Losos, and
O. Ellers. 1996. A comparative analysis of clinging ability
among pad-bearing lizards. Biol. J. Linn. Soc. 59:2135.
Israelachvili, J. 1992. Intermolecular and surface forces. Academic
Press, New York.
Israelachvili, J. 2001. Tribology of ideal and non-ideal surfaces and
fluids. In B. Bhushan, Fundamentals of tribologybridging the
gap between the macro-, micro-, and nano-scales, pp. 631650.
Kluwer Academic Publishers, Dordrecht.
Kinloch, A. J. 1987. Adhesion and adhesives: Science and technology. Chapman and Hall, New York.
Koelsch, G. 2000. The ultrastructure of glands and the production
and function of the secretion in the adhesive capture apparatus
of Stenus species (Coleoptera: Staphylinidae). Can. J. Zool. 78:
465475.
Kunzig, R. 1999. Why does it stick? Bubbles, bubbles everywhere,
each one a small suction cup. Discover. 2729.
Lauder, G. V. 1991. Biomechanics and evolution: Integrating physical and historical biology in the study of complex systems. In
J. M. V. Rayner and R. J. W. (eds.), Biomechanics in evolution,
pp. 119. Cambridge University Press, Cambridge, Massachusetts.
Lee, Y. I., M. Kogan, and J. R. J. Larsen. 1986. Attachment of the
potato leafhopper to soybean plant surfaces as affected by morphology of the pretarsus. Entomol. Exp. Appl. 42:101107.
Lees, A. D. and J. Hardie. 1988. The organs of adhesion in the aphid
Megoura viciae. J. Exp. Biol. 136:209228.
Liang, Y. A., K. Autumn, S. T. Hsieh, W. Zesch, W.-P. Chan, R.
Fearing, R. J. Full, and T. W. Kenny. 2000. Adhesion force
measurements on single gecko setae. Technical Digest of the
2000 Solid-State Sensor and Actuator Workshop. 2000:3338.
Maderson, P. F. A. 1964. Keratinized epidermal derivatives as an aid
to climbing in gekkonid lizards. Nature 203:780781.
Mahendra, B. C. 1941. Contributions to the bionomics, anatomy,
reproduction and development of the Indian house gecko Hemidactylus flaviviridis Ruppell. Part II. The problem of locomotion. Proc. Indian Acad. Sci., Sec. B 13:288306.
Nachtigall, W. 1974. Biological mechanisms of attachment: The
comparative morphology and bioengineering of organs for linkage, suction, and adhesion. Springer-Verlag, New York.
Noy, A., C. D. Frisbie, L. F. Rozsnyai, M. S. Wrighton, and C. M.
Lieber. 1995. Chemical force microscopy: Exploiting chemically-modified tips to quantify adhesion, friction, and functional
group distributions in molecular assemblies. J. Am. Chem. Soc.
117:79437951.
Persson, B. N. J. 1999. Sliding friction. Surf. Sci. Rep. 33:83119.
Peterson, J. A. and E. E. Williams. 1981. A case study in retrograde
evolution: The onca lineage in anoline lizards. II. Subdigital
fine structure. Bull. Mus. Comp. Zool. 149:215268.
Roscoe, D. T. and G. Walker. 1991. The adhesion of spiders to
smooth surfaces. Bull. Br. Arachnol. Soc. 8:224226.
Rosenberg, H. I. and R. Rose. 1999. Volar adhesive pads of the
1090
A. M. PEATTIE