T

Design of Fractionating Columns

MOTT SOUDERS, J R., AND
HE capacity of a frac-
tionating column may be
limited by the maximum
quantity of liquid that can be
passed downward or by the
maximum quantity of vapor that
can be passed upward, per unit
time, without upsetti ng the
normal f uncti oni ng of the
column. The liquid capacity is
determined by the capacity of
the weirs and downspouts; but,
if the resistance to the flow of
vapor through the plates exceeds
the available head of liquid be-
tween plates, the normal flow of
l i qui d is interrupted and the
column is said to prime. The
vapor capacity is usually below
the priming point and is limited
largely by the quantity of en-
trainment that may be tolerated.
For this reason if the column is
of good mechanical design and
has adequate liquid capacity,
I. Entrainment and Capacity
GEORGE GRANGER BROWN, University of Michigan, Ann Arbor, Mich.
Entrainment in fractionating columns affects
the color of the product and the plate eficiency,
and limits the maxi mum vapor velocity that
will give satisfactory operation. Using a theo-
retical equation and empirically derired con-
stants, an expression is obtained f or the maxi-
mum allowable vapor velocity in a column which
is dependent upon the quantity of entrainment
that can be tolerated according to operating
conditions. Values of the constants to be used
f or determining the capacity of columns f or
different types of services under ordinary operat-
ing conditions are suggested. The effects of
intermediate rejlux, the ratio of liquid overflow to
vapor rising in the column, plate spacing, and
allowable entrainment are discussed so that the
maxi mum vapor velociiy f or satisfactory opera-
tion of fractionating columns under various
conditions may be estimated in a fairly satis-
factory manner.
crease in entrainment from that
plate as compared wi th the
plate below.
An abrupt decrease in en-
trainment and corresponding in-
crease in liquid overflow from
a plate may be caused by an
increase i n spaci ng between
that plate and the plate above.
Similarly, at the point of intro-
duction of an intermediate cold
reflux, there is an increase in
the liquid overflow owing to the
decreased entrainment caused
by the decrease in velocity of
vapor rising from that plate.
As compared with the normal
operati on of introducing al l
reflux at the top of the column,
the use of intermediate cold re-
flux thus serves to decrease the
vapor vel oci ty, entrainment,
and liquid overflow of a column
above the point of introduction
of the intermediate reflux and
vapor capacity ai limited by entrainment is the controlling
factor determining column capacity.
ENTRAINMENT
I n a fractionating column, entrainment signifies the upward
displacement of liquid particles, from plate to plate, caused
by the dynamic action of the vapor. Entrainment may be
defined as the quantity of liquid carried upward from plate
to plate by the vapor per unit of time, but possesses little
quantitative significance unless expressed as a ratio, such
as the quantity of entrained liquid to the quantity of vapor
rising from a plate per unit of time ( E/ V) or as the ratio of
entrained liquid carried upward by the vapor to liquid over-
flow from the plate ( EI L) .
The effects of entrainment in fractionating equipment are
largely the impairment of color, loss of liquid overhead as in
oil absorbers, increase in the quantity of liquid flowing from
plate to plate due to abrupt changes in vapor load, and de-
crease in plate efficiency.
The effect of entrainment on color is most important in
flash jugs or chambers without fractionating plates since a
small quantity of dark residual material may have a rela-
tively large effect on the overhead material. I n fractionat-
ing columns where several plates separate bottoms and
overhead, the effect of entrainment on color is less important
since each plate acts as an entrainment separator, and the
dark material is progressively diluted by the liquid overflowing
from plate to plate.
By a material balance it is clear that the liquid overflow
from a plate must be equal the sum of the quantities-vapor,
liquid entrained from the plate below, and liquid overflow
from the plate above-less the vapor and entrainment
carried to the plate above. Therefore the liquid overflow
from a plate is increased by an amount equal to the de-
to increase the liquid load at the plate where the intermediate
reflux is introduced.
For this reason the probable entertainment in various
parts of the column should be considered in calculating the
sizes of weirs and downspouts.
The function of a plate is to change the composition of the
vapor rising through the plate. This in turn depends upon
the fact that there is a difference in composition between the
liquid and vapor leaving the plate. Entrainment of liquid
particles in the vapor stream diminishes the effective differ-
ence in composition between the vapor and liquid and de-
creases the change in composition of the total material (dry
vapor and entrainment) rising through the plate. For these
reasons entrainment is an important factor in limiting the
fractionating efficiency of a plate.
FACTORS DETERMINING ENTRAINMENT
Entrainment may be regarded as the result of two distinct
effects of the flowing vapor, the actual carrying of droplets
by the rising vapor and the throwing of liquid particles by
the dynamic action of vapor jets. The first effect is a func-
tion of the mass velocity of the vapor, the densities of the
liquid and vapor, and the diameter of the particle which in
turn is influenced by the surface tension of the liquid, density
of the vapor, and agglomeration of individual particles into
larger masses. The entrainment produced by the throwing
of liquid particles is a function of the kinetic energy of the
vapor jets, which, in turn, depends upon the density and the
linear velocity of escape of the vapor, and is closely related to
the spacing between plates.
Because of the complex nature of relationships between
these numerous variables and the limited amount of quantita-
tive data at present available, it is necessary to adopt el e
mentary simple relationships between the most important
98
J anuary, 1934 I N D U S T R I A L A N D E N G I N E E R I N G C H E M I S T R Y 99
TABLE I. OPERATING CONDITIONS OF COMMERCIAL FRACTIONATINQ COLUMNS
GENERAL DESCRIPTION
Commercial alcohol-water column
Commercial presaure distillate rerun column
Commercial vacuum column, gas oil overhead
Commercial natural gasoline stabilizer
Commercial natural ea8 absorber
COLUMN PLATE SURFACE TENSION OB
PRES SUR^ SPACINQ LIQUID ON PLAT^
Lb. / sq. in. abs. Inches Dunee/cm. Lb./ft. X 10-
20 12 60 41
55 12 13 8.9
0.96" 16 23 15.6
155 16 9 6.17
190 21 .. ..
Commercial natural Pas abaorber 465 21 .. ..
Cbmherci a natural gas absorber 465 24 ..
Commercial vacuum column, gas oil overhead 0.632d 30 24 16: 4
a 50 mm. b 10 mm. C 20 mm. d 33 mm.
variables in order to arrive at a practical solution. I n the
following treatment it; is assumed that the mass velocity of
vapor upwards through the free space of the column controls
the quantity of entrainment in the same manner as the upward
mass velocity of any fluid is able to suspend solid or liquid
particles, depending upon their density and size.
THEORETICAL SUSPENDING VELOCITY
The upward velocity of a fluid required to suspend a body
in the fluid stream may be determined from the resistance of
the body to the moving fluid and the force of gravity on the
body.
,'The resistance of a sphere in a moving fluid is given by the
expression (5) :
where F, =total force on drop
=constants which must be evaluated empirically
K, k
p =viscosity of fluid
D =diameter of particle
dz =density of fluid
v =linear velocity of fluid relative to drop
I n a fractionating column the first term on the right may be
neglected, since the viscosity of the vapor is small (0.01 to
0.001 centipoise), so that:
The force of gravity (less buoyancy) on a spherical particle,
(3)
where dl =density of particle
g =acceleration of gravity
When the force of gravity is equal to the resistance to the
moving vapor, the particle remains suspended,
and the suspending velocity,
If D and k are constants:
(4)
v =linear velocity of vapor, feet per second
Since mass velocity of the vapor W =3600 vdz in pounds
per square foot per hour,
W =C[dZ(di - dz) ]' / l
(6)
OBSD. C IN
MASS VELOCITY
EQUATION
260
235
430
385
400
400
595
630
060
550
635
440
690
I
where C =a factor depending upon conditions
d2 =density of vapor, pounds per cubic foot
dl =density of liquid, pounds per cubic foot
PRACTICAL APPLICATION TO ENTRAINMENT
Although the value of C in the above derivation is 3600
$$, Equation 6 is used to include the effects of other
variables than the theoretical suspending velocity, which are
incorporated in the factor C.
The range of sizes of the liquid particles (D) which compose
the entrained liquid is an indeterminate variable. Particle
sizes probably are related to the surface tension of the liquid
on the plate, since the dispersion of the spray produced by
a bursting bubble appears to vary inversely with the surface
tension of the film. Increase in density of the vapor also a p
pears to promote atomization (3). The tendency for in-
dividual particles to coalesce into larger drops which are less
readily entrained may be influenced by surface tension, and
spacing between plates, or time, which may have a bearing
on the probability of collisions between particles.
I n addition to the carrying of droplets by the rising vapor,
entrainment is produced by the jet action of the vapor caused
by the contraction of the path of flow through the bubble caps
and vapor-liquid mixture on the plate. The throwing of
droplets by vapor jets is related to the density and velocity
of the vapor flowing through the slots in the bubble caps and
the depth of the "liquid seal." I n general, the penetration
of droplets thrown by a jet decreases, and the dispersion
increases with increase of the density of the vapor (3). With
other conditions constant (velocities, densities, surface ten-
sion, etc.), it appears that the entrainment due to the throw-
ing of droplets should be influenced largely by the distance
between plates.
Increasing the velocity of the vapor through a column not
only tends to increase the height to which droplets may be
thrown but also decreases the free space above the vapor-
liquid mixture on the plate. This effect is due to the vapor-
lift action of the flowing vapor which raises the froth level as
the velocity is increased.
For these reasons if the limiting vapor velocity causing
entrainment is to be expressed by a simple equation such as
6, factor C will depend upon surface tension, distance between
plates, and the nature of the materials or service conditions.
The numerical values to be used for C in Equation 6 can be
best determined in an empirical manner.
Table I presents operating conditions of a number of com-
mercial fractionating columns operating at approximately
the maximum vapor load compatible with satisfactory prod-
ucts. The operating data include conditions from 0.192
pound per square inch (10 mm. of mercury) to 465 pounds per
square inch (0.013 to 32 atmospheres) total pressure, 12 to
30 inches plate spacing, and materials from lubricating oils
to natural gasoline. I n each case the value of factor C cor-
responding to the maximum capacity of the column has been
calculated by means of Equation 6.
100 I N D U S T R I A L A N D E N G I N E E R I N G C H E MI S T R Y lr0I. 26, No. 1
I n Figure 1 the values of C thus obtained from fractionating
columns are plotted against the center-to-center distance be-
tween the bubble plates. Points where the surface tension
of the liquid on the plates is above 12.7 X pounds per
foot (20 dynes per cm.) (plotted as circles) define the solid
curve; points where the surface tension is less than 20 dynes
suggest the location of the dotted curve for surface tension
of 6.85 x 10-4 pounds per foot (10 dynes per cm.).
The computed
BC.0. , values of C for the
gas absorbers
702. (Table I) are uni-
formly lower than
2 600 for the fractionat-
8 ing columns. It is
possible that part of
this difference may
be due to surf ace
tension conditions,
but these data are
not available. Any
c
j 5m-
9
g 400
9
-3x
72 d,l
iw =MA55 VEL OCl TI - L B S/ SO 'T/HIR. 1 entrainment in ab-
sorbers resul ts in
d,- DENSI TY OF YIPOR -is5 I cu F T. 1 the loss of absorbent
c = CONSTANT, DEPLI?, NG ON
DISTANCE BETWEEN PLATLS 1
d; DENS- " OF L I OVI D- _BS. / Cu FT
oil and contamina-
tion of the dry gas;
I O I5 2~ 25 30 35 for this reason, if
for no other, absorb-
with much l ower
mass velocity (1) than other fractionating columns in which
limited entrainment can be tolerated.
Since the values of factor C expressed in Figure 1 are based
on data obtained from the upper or fractionating sections of
petroleum columns, these values of C should be modified when
applied to gas absorbers and probably require modification
when applied to columns or parts of columns which are in
different services, such as stripping columns. &o, since
Figure 1 represents average maximum operating limits, a
factor of safety should be applied to these values of C when
they are used for design purposes.
Figure 2 is a chart for evaluating the allowable mass velocity
of the vapor in a column from the liquid and vapor densities
and the value of C obtained from Figure 1, a graphical solu-
tion for Equation 6. For example, in a topping column with
gasoline overhead, dz is 0.094 and dl is 39.55 pounds per cubic
foot, so that dz (dl - dz) is 3.7. With plates spaced at 24
inches, C (from Figure 1) is 640,. and (from Figure 2) W,
the allowable mass velocity of the vapor, is 1200 pounds per
square foot of column cross-sectional area, per hour.
All material comprising the vapor stream (including
products, internal reflux, fixed gases, and steam) should
be included when calculating the density of the vapor and the
vapor load of the column.
QUANTITY OF ENTRAISMENT
I n a
semiplant vacuum column with 30-inch plate spacing on a
straw oil vapor-liquid system at 10 mm. and at 20 mm. total
pressure, entrainment was estimated colorimetrically by the
use of a nonvolatile dye in the feed to the column. Although,
by plotting quantity of entrainment against mass velocity of
the vapor, there was considerable scattering of the points,
it was possible to draw a representative curve for each operat-
ing pressure.
Chillas and Weir (8) reported the quantity of entrainment
as a function of the superficial linear velocity of the vapor for
an air-water system at atmospheric temperature and pressure
with plates spaced 16 inches center to center, using a constant
' :1;.T1
D STANCE BETWEEN PLATES CENTER TO CENTER- - I NCH( ~
FIGURE 1. EFFECT OF PLATE SPACING ers are operated
Data on the quantity of entrainment are meager.
ratio of liquid to vapor. Comparable data on other systems
with various plate spacings and reflux ratios are much needed.
Comparison between data on the quantity of entrainment
obtained under conditions of different liquid and vapor
densities, and in different apparatus with different spacing
between plates, requires a general equation prope'rly evaluat-
ing the effect of these variables. The data of Chillas and
Weir on the air-water system and the vacuum column data
on the straw oil system were compared using the velocities
obtained from the constant, C, corresponding to the different
plate spacings (Figure 1) as the basis of the comparison. I n
Figure 3 quantity of entrainment, expressed as gallons at
plate temperature of entrained liquid per pound of dry
vapor, is plotted against the ratio of the observed mass
velocity to the mass velocity calculated from Equation 6 and
Figure 1.
Since the quantity of liquid flowing across the plate varied
with the mass velocity of the vapor (constant liquid-vapor
ratio), the data on the two different systems were also plotted
(Figure 4) as gallons of liquid entrained per gallon of liquid
overflow against the ratio of observed mass velocity to mass
velocity calculated by Equation 6 and Figure 1.
Both Figures 3 and 4 show reasonably satisfactory correla-
tion within the limits of accuracy of the experimental data
and variations in the mechanical details of the apparatus.
More experimental work is required to indicate the effect of
vapor-liquid ratio and quantity of liquid flowing across the
plate, in order to establish a wholly satisfactory basis of
correlation.
01 0 2 0 5 , 2 5 10 20
d2(d,-d2)
FIGURE 2. CHART FOR EVALUATING ALLOW~BLE MASS VE-
LOCITY OF THE VAPOR IN A COLUMN FROM THE LIQUID AND
VAPOR DENSITIES, AND THE VALUE OF C FROM FIGURE 1
Although this discussion deals exclusively with plate
fractionating columns, it is well to indicate that much greater
entrainment may be expected in other types of equipment
which do not contain plates or other types of entrainment
separating devices. The actual entrainment in a flash cham-
ber of a cracking plant (chamber free of any entrainment
separating device) is represented in Figure 3 by the circled
point which indicates more than twice the entrainment ob-
served in a plate fractionating tower. The vapor-liquid
mixture in this case entered the large chamber through a single
pipe at high velocity, and the large kinetic energy of this
stream was an important factor in increasing the entrainment
over that of a plate column, although the stream was directed
against the lower end of the side of the chamber.
ENTRAINMENT AND PLATE EFFICIENCY
Defining plate efficiency as the ratio of the actual change
in composition of the wet vapor passing through the plate
to the change that would take place were the vapor leaving
the plate in equilibrium with the liquid overflowing from the
plate:
J anuary, 193i I T D U S T R I A L - 4UD E N G I N E E R I K G C H E hl I S T K 1 101
RATI O- ML OCI TY OBSERVED TO VEL CITY CALCD BY Ea. 6 RATIO-VELOCITY OBSERVED TO VE CITY CALU). BY EP. 6
W= C [ dJd -4d W-C[ d, U, -4% - .
FIGURE 3 FIGURE 4
FI GURES 3 4ND 4. EFFECT OF \-4POR VELOCI TY OU QU~ \ TI TY OF EUTRI I \ J f ENT
( 7)
where yn =inole fraxtioii of a component in mixmre of vapor
anti entrained liquid rising from plate n
Y , +~= mole fraction of the same component in the mix-
ture of vapor and entrained liquid from plate
below plat'e n
I< =equilibrium constant (4)-i, e., the ratio between
mole fraction in dry vapor and mole fraction in
liquid under equilibrium condition3
2, =mole fraction of same component i n liquid over-
flowing from plate n
From a inaterial balance around plate n, assuming L and V
to be constant:
nhere f =a factor nhich iepresnts the degree of approach
towaid equilibrium betn een dry vapor and liquid
overflow leaving the plate, and is dependent
upon a large number of variables including the
mechanical design of the plate
The symbol, f, should not be confused with plate efficiency as
defined by Equation 7 .
Therefore T'y, =T"j Kx, +Exn
and from Equations S and 12,
Combining Equations 9 and 13,
=
x1,[VKj' - E(Kf - l )] (12)
VY, +1 xn[VKf - E(Kf - 1) I - L(xn - I - xu) (13)
whereL =total nioles of liquid overflowing from plate to
F~~plates above the feed plate,
plate per unit of time Lx,-i Vy, - DXD D =V - L
1- =total nioles of mixture of vapor and entrained where D =total moles of withdrawTn from column
above plate n
liquid rising from plate to plate
flowing from plate above
- = mole fraction of same component in liquid over-
ZD =mole fraction in composite overhead materials
T7 =V' + E (10)
where T" =total moles of dry vapor rising from plate to plate
E =total moles of entrained liquid carried with dry
Under actual conditions when equilibrium between liquid
and dry vapor may not be attained, the actual mole fraction of
the component in the dry vapor may be expressed by the
equation:
per unit of time
vapor from plate to plate
A similar derivation for plates below the feed plate,
where Lx, - =Vyn +B ~ B
B = L - V
B =total moles of materials u-ithdrawn from column
xg =mole fraction of same component in composite
below plate n
materials drawn below plate n
leads to the equation:
102 I N D U S T R I A L A N D E N G I N E E R I N G C H E M I S T R Y Vol. 26. No. I
Thus the general equation for the effect of entrainment on
plate efficiency becomes:
E
K(1 - f l +5 (Kf - 1)
L L
K - 3 ( 1 - 7) - p
e = l -
2"
where xp =composite mole fraction in products withdrawn
either (a) above plate n for plates above feed
plate or (b) below plate n for plates below feed
plate
Equation 17 may be simplified for special conditions, such
Equilibrium plate with total reflux:
as:
f =1, and V/ L =1
e =l - E / V (18)
Total reflux and no entrainment:
Top plate of a column where the reflux has the same composi-
tion as the overhead distillate-i. e., xn - =yn:
(20)
1 - E/ V
E E
e =
V( K - Kf) +1 - - +-
UKf - 1) V L
And for a top plate which is an equilibrium plate-i. e., f =1:
1 - v
e =
Using Equation 21 for the equilibrium top plate of a column
and the straw-oil curve of Figure 3 for the quantity of en-
trainment to compute E/ V for different relative vapor veloci-
ties, Figure 5 was constructed giving plate efficiency as a
function of relative mass vapor velocity for various values of
L/V. From Figure 5 it is clear that plate efficiency may be
maintained constant with a greater mass velocity (greater
entrainment) if the value for L/ V (ratio of liquid overflow to
vapor) is increased accordingly, and if E/V is substantially
independent of L/ V.
The ratio L/V (liquid to vapor) in gas absorbers is much
lower than in fractionating columns, usually about 0.2 in
high-pressure absorbers. Under such conditions high vapor
velocities might cause large decreases in plate efficiency.
This is possibly another reason for the lower vapor velocities
used in gas absorbers.
COLUMN CAPACITY DETERMINED BY ALLOWABLE
ENTRAINMENT
The relationship between plate efficiency, liquid-vapor
ratio ( L/ V) , and relative mass velocity of the vapor as
plotted in Figure 5 indicates that the probable maximum
vapor load or capacity for satisfactory operation as based
on plate efficiency varies more or less directly with the
liquid-vapor ratio, L/ V. Therefore it is to be expected that
fractionating equipment may be operated satisfactorily at
higher capacities when the ratio of liquid to vapor ( L/ V) is
larger. Thus, greater mass velocity might be tolerated in
the stripping sections of stabilizers or steam strippers than
in the rectifying sections of the same columns. Similarly,
the permissible mass velocity would be greater at the top
of a topping column than at plates immediately above the
feed plate, and the upper part of a topping column would be
operated with greater mass velocity of the vapor than the
upper or rectifying part of a column for stripping natural
gasoline from absorbent oil.
This is found to be the case, and the relative maximum
vapor velocities or capacities from actual plant operation
for different services are indicated on Figures 3 and 4. These
relative velocities are for average conditions and are subject
to variations in liquid-vapor ratio, mechanical design, and
other factors which may control entrainment or the allowable
quantity of entrainment.
It has been assumed in the application of Equation 21 and
Figure 5 that E/ V is independent of L/ V. Actually there is
RATIO- VCLWTY OBSERVCD TO VE CITY CUR BY EO. e
w-c[d&,-dJ~
FIGURE 5. EFFECT OF VAPOR VELOCITY ON EFFI-
CIENCY OF TOP EQUILIBRIUX PLATE
some evidence that E/ V may increase with increasing values
for L/ V. If this effect is appreciable, i t will tend to mini-
mize the differences in maximum capacity for the different
services.
The allowable quantity of entrainment may be limited
by considerations other than plate efficiency, such as color
impairment of overhead or loss of liquid in vapors, and may
vary from point to point in a single column. As has been
pointed out, the introduction of cold intermediate reflux de-
creases the entrainment and vapor load in the column above
its point of introduction as compared with the introduction
of all of the reflux at the top of the column. The allowable
entrainment at the top of the column is usually the factor
controlling the capacity of the column rather than the en-
trainment at some lower part, such as below the point of
introduction of intermediate reflux, because the entrainment
in the top part of the column controls the quality of the over-
head product. For this reason the removal of heat from
the column at an intermediate point, as by the introduction
of cold intermediate reflux, increases the capacity of the
column although at a loss in effectiveness of fractionation.
If overhead side streams are removed from the column
below the point of introduction of cold intermediate reflux,
the effect of entrainment on such side streams must be con-
sidered, as the allowable entrainment below such side stream
may be the factor limiting column capacity. If cold inter-
mediate reflux is not used, the allowable entrainment at the
top is usually the controlling factor even when side streams
are removed, because the maximum vapor load in such a
column is usually just below the top plates.
103
Phase Equilibria in Hydrocarbon Systems
I. Methods and Apparatus
BRLIX LI. SAGE AIW WILLIAM N. Lactry, California Institute of Technology, Pasadeoa, Calif.
KXVOWLEIXE OS tlii: Apparatus arid methodv for studies of phase \.ls~suitlnie?iT OF ENTERING
~~
behavior Of coindex quilibria in hydrocarbon mixtures at pressures hf.4TElilALs
drocarbon systems under
up to 200 atmospheres in the temperature range
llelatively rlonvolatile liquids
equilibrium conditions corn-
sponaingtotiiosefollnl~in iiIlder. f rom 20" to looy (2. are described. The ddn ob- ww measured by wei ghi ng
ground peiroleurn reservoirs is of taimd permit lhe prediction of the density, cam- a suitable container before and
pri mary i ni portance to t!ir position,, an,d relalire mass of each phase present after pouring tlie sample into
petroleum proditction tcclitiolw lohen ,,liz.are ofuny total is brought t,lheequilibrium vessel. The
gist. The simpler exnmples of
latter was then closed and con-
to equilibrium at any set of temperalure and pres-
netted for If the liquid
such systeiiis are also of interest
froln tilc jrlii.clg scie,,tific ljoil,t sure conditions wifhin the rurtge studied. Subse- a pure of volatile
of view.
iiuent arlicles of lhis series icill presenl dalafor i,imracter, a portion of it was
Tlie ai111 of t,liis rvork nai to 60th simple and complr:? rnidnrcs.
ilistilled into the equilibriuni
follow the behavior of gaseous
vessel from a weighed con-
and liquid p1iai;es prcaent i n taiiier. I'olatile complex liquids
equilibrium at timperatiircs rmgiiig Sboiu 20" to 100" C. were i:oaied to a sufliciently low temlxrature to be handled
(68'P. to 212' F.) and at preisiires from I to 2%) :stinosp!ieres by tibe inetliotl i i scd for notivolatile liquids.
(approximately 15 to :joO() pounds per aquzrc incii absolute). Gases entering tile system were. measured Og willidraarving
I n order to ascertain completely the state of the system, from a calilratetl reservoir constant volunle and nptjng
measurements of the density, voluine, an11 composition of encli the resulting drop i n pressure. The sample bomb cOtitalNng
of the pliases present ware required. These tncaiurements the gas supply was heated in a rlietliylene glycol bath to a
were made over a series of temperatures, pressures, and total sufficiently high temperature to insure complete vaporiza-
compositions in order that tlip effect of these variables miolil. tion. That the temperature used was high enough could he
!)I: determined.
A
The equilibriuni
method uscd in this
work resolved itself
into the following
steps: the measure-
ment of the amount
of an original liquid
phase placed in an
equilibrium vessel,
the measurement of
a series of quantities
of gaseous materinl
md their quant.ita-
tive compressi on
into the equi l i b-
ri um vessel, the
attai nment of
equi l i hri um, rand
the determination
of the state of the
system after each
addi ti on of ma-
terial.
verified by deter-
mination of the dew
point of the gas at
the pressure eltist-
ing in the sample
bomb, as described
below. The pa8
was then admitted,
through heated tub-
ing l i nes, to the
reservoir bomb in a
di ethyl ene glycol
bath carefully ther-
mostatrd at 100.0'
C. The reservoir
was so calibrated
that the quantity of
gas in it was !mown
for any given pree
sureup to a maxi-
mum of ahout 20
atmospheres. The
cal i brati on was
F K X ~E 1. Coi r~~esso~ AKD CONTHOL PANBL made for each gas