SOLUBILITY OF NATURAL GAS I N VARIOUS CRUDE OILS

and
PREPARAT I ON OF PURE lvlETHANE FROM NATURAL GAS

Thesis by
Char les Ki r cher Jr.

r
(
(

In Parti a l Fulfillment of the Requirement s

for the Degree of Master of Science

California Institute of Te chnology

Pasadena, California
1933
 TABLE OF CONTENT S

Solubility of Natural Ga s in Various Cr ude Oils Page

Introduction 1

Determination of solubility 2

Method of procedure and descri ption of a ppar a tus 2

Dem ~ i mination of the compressibility of the ga s used 11
Sample c a lcul a t~on 12
Discussion of results and plots of dat a 13

Preparation of Pure Methane from Natural Gas

Int r oduction 22

Method of Procedure 22
Analysis of methane produced 27

Design of cha rcoal cells 28

.App endix 32
 ~ .L)

SOLUBILITY OF N.ATUR.AL GAS IN VARIOUS CRUDE OILS

I NTRODUCTI ON
The work re portea in this paper is a continua tion of
experiments begun and carried out by B.H.Sage, and described
in his thesis for the Masters degree C.I. T. 1931. The object
of the research was to determine as quantit a tively a s possible
the solubility of natural gas in various crude oils, as it
was known that by dissolving natural gas in crude oil, the
physical pr operties of the oil could be changed in such a way
as to make possible gr e ate r recove ry of the crude oil from
the underground formation. .As pointed out by Mr. Sage , the
solubility fi gur es presented in his thesis had to be arrived
at by a method of approximation as he was then not exr e r i mentally
able to measure the change in liquid volume resulting f r om
dissolved gas. In order to make possible more a ccura te
c a lculation of gas solubility, apparatus was designed and
inst alled for directly me a suring the ch an~ e in volume of the
crude oil whi ch occurred as a result of dissolved gas. Other
changes were ma de in the equi; ment as first set u p by Mr . Sage,
which wer e desi gned to increase accura cy and s peed in carrying
out the experiments. The s e chang es will be discussed in the
main body of t h is re por t.
The solubility data presented in all cases is for a
temp er a ture of 100°F and pr essures ranging f r om zero to 2500
lbs. per sq. in. Th e follo wing representative crude oils were
used: Sugarland, Texas; Ya tes, Texas; Ventura, Calif.; Santa
Fe Spr ings, Calif.; Bartlesville, Okla.; Seminole, Okla.;
Lima, Ohio; and Bradford, Pa. These crude s as used were
s aturated wi th air a t a tmospheric pressur e.
 \ i::., I

Dry Santa Fe Springs natural g as was used in all cs.ses .

DETERMINATION OF SOLUBILITY
A definite weight of crude oil was pl a ced in a steel
cell cap ab le of withstanding pres sur es u p to 3000 lbs. per
s q. in. The temperature of the cell was maintained practic ally
constant, and measured Quantities of gas were pump ed into it.
For each successive portion of gas admitted the ·e·Quili briul
pr essure in the cell, and the volume of the liqu id :phase , were
r eco rd ed. The total volume of the cell, and the liQuid volume
being known, the free ga~ space over the crude oil cou ld be
determined by difference. Knowing this volume and the
compr es s ibility of the natural gas used, the Quantity of ga s
over the oil for each equilibrium point could be calculated.
By subtracting t h is figure from the Quantity of gas admitted
to the cell, the s olubility value for the natural gas was
determined.
METHOD OF PROCEDURE
A sample of the crude oil to be used in a solubility run
was weighed in a flask, a :p ortion of the samp le was then
poured into the solubility cell(Fig . 1), and the fl a sk r e-
wei ghed. The weight of sample t aken was dete rm ined by
difference. In order to know the volume which a definite
weight of crude oil would occu ~ y in the cell, af ter it had
come to tempe r a tur e, t h e gravity of the crudes used was
determined at 100 °F by means of a picnomete r. Knowing the
weight of crude t aken and the gravity at the temperature in
Question,, the initial volume of the oil sample in the cell
could be calcula ted. Afte r pl a cing the crude oil in the thick
walled steel cell, the to p of the cell was bolted on ti ght,
 Solubility Cell.

F ig.l

A- Gas inlet port

B- Micrometer for measuring liq_uid level
C- Scale for micrometer
D- Electric cont a ct for level indic a to r
E- ? ac }dng g l and for shaft
and the whole assembly pl aced in a rocker arm in an oil bath.
The gas line from the compressor wa s then connected by means
of a standa rd fitting( see Thesis of B.H.Sage). The oil level
in the bath wa s then raised by forcing in more oil, under air
pressure, from a reserve tank, until the cell was just covered.
The oil bath was provided with a steam coil and electric heaters.
After the bath had been brought a pp roximately to temper a ture
the electric heaters on the mercury regulator were adjusted
so as to maintain very nearly a constant temperature.
Fluctuations of temperature could be held to within 0.04 of
a degree Centi grade. At the same time that the oil bath wa s
coming to temp erature the electric heaters on the gas lines
were turned on, and they too brought to temperature. Line
temperatures were measured by means of thermocouples soldered
onto the steel tubing. Both lines and mercury traps (to be
discussed later) were heated to l00°C. The cell was g iven
time to come to thermal equilibrium, and then the liquid level
in the cell was measured. This was accom plished by means of a
s p ecial type of a ~ paratus to be re ~ orted on by B. H.Sag e in his
thesis for the Ph. D. degree. Briefly the method was as
follows: A very small wire, forming one side of a bridge
circuit, was attached to one end of a micrometer screw which
p assed thru the to p of the cell (see Fij.l). While the wire
was in the g as pha se a small current was passed thru it, and
the bridge brought into balance as shown by a galvanometer.
The micrometer screw was then lowered into the cell until a
sudden deflection of the galvanometer needle, resulting from
the bridg e circuit becoming unbalanced, indicated that the wire
was touching the liquid. The bridge circ ~ it was thrown out
of balance because the li Quid conducted he a t away fro m the
wire much faster than the gas did, there-by chang ing the
temp er a ture and resistance of the wi r e. By means of this
device the li quid level cou ld be measured to within a
thousandth of an inch. The volume of the liquid as determined
by measur ement agreed very closely with that calcu l a ted from
the weight and gravity of the crude oil.
Temperatures being correct the run could be started. In
order to obtain a solubility fi gure the q_uantity of gas
admitted to the cell had to be measured. This wa s accomplished
by drawing gas out of a calibrated reservoir held at constant
temp era ture, and observ ing t he fall in pressure. The met hod
was that outl i ned by F. P .Stapp iti h is thesis for the Masters
de gr ee. The reserv oir consisted of a steel cell pl aced in a
b a th of diethyleneglycol, the te m~ erature of which could be
closely maint a ined a t l00 °C. The pressure in the reservoir
was determined by pl acing wei ghts on a fluid pressure ba_lance.
As this type of eq_u i rnp ent was used throughout for measuring
pressures it wi ll be described here. The cell in which the
pressure was to be me a sured was provided with a mercury trap
simil ar in shap e to the conventi onal U tube, but made of steel.
Gas from the cell was in cont a ct with one side of the mercury
column while the other side was connected by means of a copp er
tube to the standa rd type of Crosby fluid pressure scale. This
whole system, including line and tester, was completely filled
with a light oil, g iving only a fluid medium between t he mercury
trap and the pre ssur e measur ing instrument. Whatever pressur e
was a nD lied to one side of the mercury U by the ga s in the cell
-- "
wa s directly tr ansmitted by means of a fluid system to the
balance. Th e instrument was so desi gned that the pressur e
was determined by pl aci ng wei ghts on a ba l an ce arm until the
beam was brought into b a l ance. One i mp ortant condition however
had to be met if the pressures read were to mean anyt h ing , and
that was that the volume of the system concerned must remain
constant. This was pr ovided for by piacing a pointed cont a ct
on the oil side of the mercury t r a p in such a way that, a s
the mercur y level chang ed, a cir cu it including an indicating
li ght was ma de or broken. Due to the compr essibility of the
oil, slight le aks in the fluid system, and at times a small
amount of gas in the lines, some means h a d to be provided for
adding more oil to the system in order that t he mercury column
could be held at · just the pr oper pl a ce. This oil was added
by means of a screw pl ung er,which was an integral part of t h e
Crosby instrument, and which when screwed down forced more oil
into the lines. Whenever a pr essur e reading was t aken the oil
in the line and trap was adjusted until the indic a ting light
just went out, there-b y g iving a constant volume on the cell
side of the trap . These mercur y t rap s were also electr ic a lly
h eated and thermally regulated so as to avoid condensation of
gas in the trap . Throughout the rest of this re por t whenever
pressure measurements are r eporte d it is to be understood that
they were taken in the manner described above. Two separate
balances were used; one covering the ra.nge zero to 3000 lbs.
per s q . in. and capable of being read to the nearest pound,
the other covering the rQnge zero to 300 lbs. per s q. in. and
capable of being read to the neare st tenth of a pound. The
former was use d for measuring the eq_uilibrium pressu r e in the
s a turation cell, a nd the latter for measuring t h e pressure in
 \ ' I

the reservoir. Valves pl a ced in the bottom of the me r cury

traps permitted one side of the mercury U to be shut off from
the other side. By closing this val ~ e the gas side of the
trap could be evacua ted without causing any trouble in the oil
side of the trap .
The gas measured out of the calibrated reservoir wa s t aken
into the gas compressor(Fig.2), which was a large steel cylinder
connected by means of a . throttle v alve to a pump which supp lied
mercury under pressure. This merc ury was used as a piston,
increasing the pressure on the gas as more mercury was forced
into the cylinder. The compressor used was the same in princi ple
as that reported in the thesis of B. H. Sage. Instead of using
high pressure air as the source of power for driving the me r cur y
into the gas cylina.er, a Worthington steam-driven plunger pump
was used. By regulating the steam pressure to the pump the
mercury system could be put under any desired pressure. This
pump .had the advantage that it could be run under practic a lly
st a lled conditions, there-by avoirling heavy i mpa ct pr essures
on the mercury lines when the throt t le valve was closed. As
in the compressor previously used, this one was provided with
electric light contacts in the side of the cylinder and a
carbon rod in the top for indicating the position of the mercury
in the compressor. The carbon rod being one side of a bridge
circuit contEdning an indicating galvanometer the exact position
of the mercury in the to p of the compressor could be de jermined.
The mercury was always brought to the same height in the small
tube in the head of the cylinder, there-by maintaining a constant
total volume for the solubility system. Also by this means the
free gas s pace in the top of the cylinder, after the gas had been
 \ ~I

Gas Compressor and Control Panel.

Fig.2
 \ ~ ,

forced out of the compressor into the solubility line and cell,
could be reduced to a negligible fr a ction of the total volume
of the system. In order to avoid condensation in the compressor
the mercury from the pump was heated by a steam jacket bef ore
enter·ing the compress ion chamber. The cylinder its elf was
electrically heated and lagged wi th insulating material. The
temperature was maintained at ab out 100 °C by means of an electric
relay and thermostat in the heating circuit. The temperature
was taken by means of a thermometer placed in a well in the head
of the compressor.
At the st art of a r..un the reservoir, compressor and lines
were evacuated; the cell however not being evacuated, as the
composition of the oil might have been changed slightly in so
doing. The reservoir was then filled with gas up to the desired
pressure. As previ ously stated gas was wi thdrawn from the
reservoir, and admitted to the compres s or forcing the mercury
present there thru a release valve into a storage bomb, which
was used to supply mercury to the pump . Air pressure of about
50 lbs. per sq. in. was maintained over the mercury in this
storage tank in order to force it quickly and easily into the
pump . For this reason the pressure in the reservoir cell could
not be drop ped much below seventy pounds. When the desired
amount of gas had been admitted to the compressor, the mercuny
release valve was closed and then the reservoir line shut off.
Mercury under pressure was then admitted to this closed system
com.pressing the gas in t he c;ylinder as the v ol ume became less
and less. A steel tube bourdon gage on the mercury inlet line
indicated roughly the pressure of the gas in the compressor.
When t his pressure was greater than the pressure in the solubility
 \ ..LV J

cell, which was known from the scale re ading, the valve to the
solubility line was o:r ened and the gas forced out until t h e
mercury in the compressor had risen to the shut-off point, as
shown by the galvanometer reading. The solubility line valve
was then closed, and the motor star ted which rocked the cell,
agitating the liquid within. This ca.used the pressure to fall,
as gas went into solution. The agitation was continued until
the pr es sure remained constant, as shown by the balance, at
which time it was assumed that equilibrium had been reached.
The time required to reach saturation was never more than five
minutes. The agitation was then sto pp ed, and the rocker arm
moved until a leveling glass indicated. that the cell was in a
vertical position. Several minutes were allowed for drainage
of the liquid on the inside cell walls, and then the liquid
level was measured in the manner described above. This reading
was t aken several times, and could be ma de to dui;: licate very
closely. The pressure reading was taken at a time when the
level indicator was just in contact with the liq_u id. This was
done in order that a small correction could be ma de in the total
volume of the cell due to the fact that the sha.ft for .the level
indicator occup ied s pace. From this readi ng and the re a ding
taken at the beginning of the run when no g as was present, the
p er cent change in liquid volume cou ld be calculated_. Once the
liquid volume had been d.eterm i ned for a particular i:i:.i::.:e ssurjL
reading, and the shaft correction deduc t ed from the tot a l volume
· of the cell, the free gas space over t h e li quid could_ be
calculated. After eac h e quilibri um pressure had been determined,
and the corres ponding liqu id level measured, more gas was
admitted in the same manner as above umtil the maximum pressure
 \ .J.. .J.. J

desired for t he run ha d been re a ch ed. At t h e end of t he run

oil wa s r emoved from the bath until the cell wa s uncovered,
and then the crude oil and ga s were blown out of the cell by
opening a plug in the bottom of the cell. This was done in
order to avoid blowing the cell down thru the gas inlet line,
as crude oil would then be car i ied over and de posited in tha t
system.
DETERMI NATI ON OF THE COMPRESSIBILITY OF NATURAL GAS USED
In order to arrive at a fi gure f or the solubility of t h e
natur a l gas in the crude oil it was n eces sary to determine
what ~ o r tion of the gas admitted to the cell was present in the
free s pace over the oil. By subtracting this amount of gas from
the tot a l gas measured in, the amount in the oj_l phase could be
determ ined. The g_uantity of ga s reg_uired at a given temperature
to bring one c.c. of free s pace in the cell up to a p articular
pressure was determined in the f ollowing way: A run s i milar· in
all res f ects to that carri ed out for a solubility determination
was made, with the exce ption that no crude was ylaced in the cell.
The q_uanti ty of gas admitted and the corres i- onding pr·es s lu·es
were reco r ded. Then a second run was made similar to the first
exce pt that a steel slug of accurately known volume was placed
in the cell. By plotting for each of these two runs the
quantity of gas admitted to the cell a gainst pressure in the
cell the amount of gas occupying the s:i;:ace corres p onding to the
volume of the steel slug could be determined directly by diff-
erence. From this a fi gure could be c a lcu lated for the q_uantity
of gas occupying a c.c. of s pace in the cell at a particu lar
tem1:er a ture and pressure. The tot a l volume of the cell and t h e
li Quid volume being known for each eq_uilibrium p oint of a
solubility r u:p. , t h e free gas s i. a ce over t"he crude oil could be
found by diff er ence. From this, and t he compr e s sibility of
the gas as determined above the tot a l Quantity of gas over t h e
oil could be -calcul a ted.. The difference between this figur e and
the tot a l amount of g a s admitted t o the cell was the amount
actua lly present in the crude oil phase. This met hod wh ile
giving a figure for gas solubility is not however entirely
accur a te owing to a peculi ar phenomenon wh ich t c..kes 1 l a ce du e
to solubility of gas in the crude oil. It ha s been ex ~ erimentally

observ ed t hat some of the li ght hydrocarbon constituents pr esent

at the st art in the crude oil, leave the oil :ph a se and ]:. a ss into
the varJ or _rhase when the crude is pl a ced in cont a ct with nat u r a l
gas under hi gh :pr es sure. The re a son for this phenomenon is no
doubt related to the f a ct that fugacity chang es wit h pres sure.
When t h is occurs t he com: osition of the ga s over t he oil would
be chang ed slightly arid therefo r e its compressibility factor
mi e;ht be chang ed. slightly. An even l arger difference between

the comr osition of the g as over the oil and the com ~ osition of
the g a s used for the compressibility determination resu lts from
the differenti a l solubility of t h e constituents making u p the
natur a l g as. The errors t hus introduced could not be evaluat e d,
and so were not t aken into account in the data as re ported.

SAMPLE ·CALCULATION
The data used is taken from a r un using dry S~ta Fe
Spring s natural g as and Lima crude oil. The symbol 11
W' specifies

the position of the mi~rometer screw used in determining the

lig_uid level.
Weight of crude oil placed in the cell 51. 02 gm.
Initial volume of crude oil 62 .7 5 c.c.
Total volume of cell when 'lr" is zero 138. 25 c.c.
 \ .A..'-' I

Saturation pr essure i n cell 19 64 l bs . per s q . in.

Total volume of gas adm it ted to cell 15,76 4 c.c. (7 0 °F)
Liquid volume as determined by. H"reading 74.33 c.c.
Volume correcti on for micrometer screw 0.33 c.c.
Volume of cell less correction 137.9 c.c.
Free gas space over li quid 137. 9 - 74 .3 = 63.6 c.c.
From compress i b ility data f or a pr essure of 196 4 lbs.
per sq. in. we find ,that 14,700 c.c. of gas (measure d
a t 70 F and 1 a t mosphere) occupy 87.69 c.c. of actual
s pace in the cell.
Therefore the amount of g as in the free s pace over
the crude oil = ~1:~~ X 63.59 = 10 , 660 c.c.
By difference t h e amount of ga s in the li q_u id
phase -:: 15 _, 7 6'± - 10, 660 :: 5104 c.c.
51G4
The solubility fi gur e is t hen 51 ; 02 = 100 c.c./gm.
When experssed in terms of cu. ft. per barrel of
initial cru de oil t h is figure is459 .3
The relative volume of the liqui d is -674.33
2 .7 5 = 1.184

DISCUS SI ON OF RES ULTS AND ? LOTS OF DATA

In the case of a crude oil of h i gh A. P .I. gravity (low
den~ity) par t of the more vol at ile hydrocarbon constituents,
a s mentioned above, may pass from the li qu id p.hase to t he gas
phase a s the pressure of n a tural gas ap plied ove r the crude
is increased. If one now assumes t hat in a solubility run
equilibrium is attained between a ll of the molecular s r ecies in
the gas :phase and t h e corresponding molecular s pe cies in the
liquid pha se, t h en it becomes import ant to.know t h e ratio of
t h e total volume of the cell to the initial vo ~ume of oi l taken.
The greater the f r ee gas s pace over t h e crude oil t h e larger
t h e quantity of a particul ar v ol a tile hydr oc arbon t hat will
 \ _,_-..:I

have to leave the liqu.id pha se in order to :pr oduce t he re qu ired

e qui librium partial pressur e in the g as pha se. As an aid in
s p ecifying star ting conditions, and for convenience in cor-
relating with other data, t h e rati o of total cell volume to
initi a l volume of oil t aken is recorded for each run. In all
cases t h is ratio was approximately 2.
Solubility values a r e g iven in terms of cu. ft. of gas
per barl' el of initial crude oil. (For uni ts see a pp end ix).
The relative volume of t h e li quid is t h e ratio of t he li quid
voJume fo r t h e particular satur a tion p oint, to t h e initial
volume of crude oil in t h e cell at t h e tem1iera ture of t h e run.
In some cases t h e i-. lots of solubility and .r el a tive volume
of liquid against :pr es sure resulted in straight lines, wh ile
in oth ers t h ere is a definite falling off in slope at t he
highest 9r essures. Because of the complexity of t h e system
both the ga s and the crude oil containing a number of diffe rent
constitu ents), and t he sc arcity of info rma tion r eg a r ding such
mixtures, it is at pr e s ent almost impossible to exi: lain t he
i rr egul ar ities which occur in the cur ves. As a first a p1ir ox-
ima tion one might say th a t the solubility of natur a l ga s in
crude oil is di re ctly proportional to pr essur e ove r t h e range
from zero to 2500 lbs. per s q . in. An interesting result is
obtained when solub ility is plotted a gainst change in volume of
t h e crude oil. It was found t hat, for t h e rather wide variety
of cru des studied, the change in volume of the li quid phase in
all c a ses, exce pt that of Seminole crude, wa s, wit h in 10%,
the same fo r cor res ponding s a tur ation v alues. That is to
say t hat fo r a definite quantity of gas dissolved in an e qual
volume of any one of the above mentioned crudes, t h e resulting
change in li quid volum e would be of t h e same order of
magnitude. Furthermore the data when plotted in this way
gave practi cally strai ght lines, showing that the chang e in
li quid volume was pro portional to the quantity of dissolved
g as.
Another inte r esting re l ation is that obt ained by plotting
ga s solubility, for a particular temp erat ure and pressure,
against t he gravity of the c rude oil. From t r is one sees
that natural gas is more soluble in the li ghter c rude oils,
that is oils with hi gher A. P .I. grav ity. Perhap s a better
relation than this could be obtained if it were possible to
plot solubility against average molecul ar weight of the
c rude oil.
j.
..
;..

I i ..

.~ j~- -
· •

'·

--,
I
~ - -i- - --
f--
i --..
r:-~

I !·:
Ii . ::j:·
--;--:-1-.-

- -..,.----
'

I .. . ' J.
1~-- t
lt:i
. .. .. 1, I
...

!Wi-6
~.:
r~
~,.j
·----
j ..
~
ti [
\).fi' £/ )J)'
 PREPAR ATION OF :;.' UUE METlIAKE FROM NATlJl' AL GAS

I NT HODUCT I N

Realizing t hat t he solubility of a na t ura l ga s in crude

oil was part l y a function of the com:1:;osi t ion of t he gas , it
was deemed a dvisable to make solubility runs using pure g ase s,
with the h ope t hat it wou ld be r os s ible to corr el at e t he data
for t he yu r e and mixed ga ses. Methane, being t he component
present i n larg est amount in most natural gases, wa s obviously
t h e first gas to be gin work with. Dif f ic u lty wa s encount ered
at the very st art, a s it was found imi: ossible to obtain methane
of suf f icient pur ity on the market. It was therefore necessary
to desi gn and set u p l ab orato ry a pr aratus c a pable of pr oduc i ng
methane cont a ining not more t han 0.3% of h eavier hydroc arb ons .
The met Lod _of pr ocedure used was determined l argely by the facts
t hat t he r re pared gas was needed at a pressure of at le a st 300
lbs. per sq_. in., and t hat wit h su ch pressures suff ici ent
quantities cou ld be stored in steel cy lind er s of sma ll actual
volume. Fr om t h e st and:coint of exri ense and convenience it was
deci ded to use nat ural g Bs a s t h e sour ce of methane, wh ich is
usually present in amounts ranging fr om 7 5;;ti to 85% 1• During t he
first st age of experiment at ion dry Santa Fe Spring s n a t ural gas
was used, but later on it was found desi rab le and feasible to
use ordihary city gas .
MET HOD OF PR OC ED~E

The g ener a l plan of procedur e was to condense ou t, by means

of h i gh pressure and low tem ~ era ture, practically all of the
constituents in t h e na tural gas h eavi er than ethane, and t hen
to se1iarat e the methane from the ethane and remaining heavier
hydr ocar bons by means of se lective adsorption on char coal.
 \ -- ,

The equi pment necessar y wa s, t he r efore, a gas pump fo r producing

h igh pressure, a low temr eratur e condensation trap, and charcoal
cells desi gned t o operate under pr essure. .At first the atte mpt
was ma de to us e th e compr es s or desi gned f or t he solubility work.
Br iefly the pr oce dur e was as f ollows: ])ry Santa :F'e Springs
natural gas wa s pump ed by means of the compress or into a steel
bomb, immersed in a b a th of a lcoh ol and soli d c arb on di oxide,
until the pr es su., e i n the bomb was ab out 2000 lbs. per sq. in.
At this pre s sure and a temperature of about -7 0 °C most of the
hydr oc arb ons h eavier t han pr opane wou ld condense out. The
resulting gas, consis t ing l ar gely of methane and •thane, wa s
then pass ed thru a char co a l cell operating und er a pr essure of
about 800 lbs. f er s q. in., and held at a temperature of ab out
10 ~. Her e, due to selective a dsorp tion, practic ally all of
the eth ane was r emoved, le av ing almost pu:r·e methane, cont aminated
ch i efly by nitrog en which was not removed by the ch ar co al.
This pr ocedure did n ot pr ove satisfacto r y for several
reasons. In t h e first pl ace t he compr essor wa s not ada pted for
this par ticu l ar type of continuous op e r ati on, and therefore t h e
mani pulations were slow and tedious. The we ar and te ar on the
compr essor v a lves was excessive, and furthermore t he c ompress or
could n ot be used for t h e solubility work whe n me t h ane wa s being
pre pared . .A change was also desi rab le wh ich wou ld make it
possible to pr oduc e meth ane cont ain ing less oxyge n and n ~ t r o gen.

The met h od of :pr oc edur e f inally decided up on is set forth in t he

flow s h e e t of Fi g .3. Gas was drawn from the city lines int o a
motor-driven sing l e st age compr essor, whi ch delivered t he g as
at a pre ssur e of ab out 300 lbs. per s q . in. The gas t hen passe d
thru a t rap filled with cotton, where entraine d oil par ticles
 VAC. PUMP
~IM D
A
~VAC. PUMP -----t======:;----1
£
/j

"""'
~ ~
~ l!h~JM .....
~ er ~
[Li [Li Trll I 11
~- .._. 1-

~IM
c
•

A-GAS COMPRESSOR F- CHARCOAL CELJ.S rLOW SliEET

tJ- EIVTRAINMEA/7 TRAP t;-METHANE STORACE TANK
C- l OW TEMP. TRAP H-AAIALYSIS RESERVOIR MeTl-/ANE
Q- 6AS .STORA(,£ TAIVK I- FREEMVT C/-IAMD&R
E- L !Ql.110 AIR 130MfJ U-MERCURY MANOMeTERS PRODUCTION
 from the compr essor were re rr oved, and into a steel bomb
immersed in a bath of alc ohol and solid c arbon dioxide . The
temperature and pre ssur e here were such t hat t h e heavy hydro-
carbons and mercaptans wou ld condense out. The Fartly stripped
gas was t h en pas sed into a steel bomb of about 900 c.c. v.olume
(Fi g . 4 ), immersed in li q_uid a i r , where it was compl ete l y
li q_uefied; gas be i ng admitted until the bomb was near l y filled.
The level of t he li quid inside t h e bomb could be followed by
obse r ving t h e boiling of t he li qu id air on t he outside of the
bomb. Th e gas inlet va lve was t h en closed, and whi le t he
bomb was still at li quid a ir temp eratur e t h e line to t he vacuum
pu.iTJp was opened, and most of t h e non-condensible gases were
removed from the bomb. Vacuum was ap plied until t he press ure
in t h e bomb fell to about 8 or 10 cm. of mercury. The vacuum
line wa s s hut off, the li q_uid a ir jacket removed, and the b omb
allowed to warm up , t h ere-by causing an i ncre a se in pr essure
in t h e bomb. When t h e pressure reached 500 l bs. p er ~q. in.,
gas was t hr ottled into t h e char co a l cell at such a rate as to
maintain const ant pressur e in t h e bomb. Th e tem perature of the
b omb was measured by means of a t h ermocou ~ le soldered to the
b omb, end when t h is indic a ted that t h e te mr erature had ris en, to
about -90°C t he line leading to t he charco a l c ell was closed.
This pr evented too much ethane from va por izing and passing
ove r jnto t h e cell. Li qu id air wa s again applied to the bomb,
and the cycle re pe a ted until analysis of t he outlet gas f r om trre
char co al cell s howed t hat t he charcoal had become nearly saturated.
In the char co a l c e ll, t h e gas vassed by means of a small tube
to t he bottom of t he cell, then up thru the charco a l to the top,
where it was drawn off and ; l a ced in a storage tank.
 ( 26 )

Fig. 4

A- Liquefying bomb
B- Control valves
C- Bourdon pressure gag e
D- Air thermostat containing gas analysis a pparatus
At the start the outlet v alve of the charcoal cell was closed,
and the pressure there-in increased a s ga s was admitted. When
this f ressure had built up to about 500 lbs. per sq. in. gas
was bled f1·om the top of the cell. The flow of ga s into and
out of the cell was so re gul ated that the cell pr essure was
maintained constant. : The highest pressure attainable in the
storage tank was that at which the charcoal cell was oper a ted.
Th e average life of the charcoal in a cell was equivalent
to an out put of from 15 to 20 cu. ft. of gas measured at
atmospheric pressure. Because the f TOcess wa s one of adsorption
the end-point of char co al usefulness was very sharp. The
charco al was re-activated by heating the cell to a te m1. erature
of ab out 220°C for an hour or more, and at the same time
maintaining low pressure by means of a vacuum pump. When at
t h is temperature the pressure had been reduced to a few milli-
meters of mercury t he heaters were turned off, and t he line to
the vacuum pump closed. City water was then circulated thru
the cooling coj_l until the temperature of the cell had fallen
nearly to room te mp erature. Water from t he refri geration tank
was then circulated and when the temperature of the cell had
been lowered to about 8°C it was ready for use.
AN ALYS IS OF ME'l'IIANE ?H ODUCED

In order to follow the oper a tion of the charcoal cell,

analysis of the outlet gas was taken at intervals. The analysis
wa s designed to g ive an indication of the per cent of hydro-
carbons, heavier than met hane, present. A sample of gas at
about atmos fh eric pressur e was stored in a reserv oi r(ab out 800
c.c. volume) to whi ch was connected a mercury manometer . The
 initi a l pr es sur e of t h e gas in the re se r voi r was re co r ded .
Gas was t he n pulled t hru an eva cuate d co ~~ e r coil i mmers ed i n
liqu id air at suc h a r a te t hat t h e pressur e did not rise above
8 mm.of mercur y. At t h is pr essur e ethane and h e avie r hydr o-
carbons wou ld condense ou t, t h e methane b eing removed by t h e
v a cuum pump . After about 90 c.c. of gas had been fa s s ed t hru
t h e co i l, t h e valve was closed and_ t h e final pressur e i.n t h e
reservoir reco r ded. The vacuum liump wa s allowed to r un for
about a minute long er, in order- to be sur e t hat a ll of t h e
met hane had b e en removed. '.l'h is was ind ic a. ted_ by t h e pr essur·e
fallin g to zero. Th e valv e s a t both ends of t h e freez-o u t
coil being now close d , t h e li qu id a ir j acket was r emoved, and
t h e c oil allowed to warm uy to t h e tem1_;e rature of t he t h ermostat .
In so doing t h e condensed hydr oc ar bons vapor ized, causing t h e
pressure to rise. The final pres sur e in t h e coil was re a d on
t h e differential manometer, and by knowing t h e ratio of the
volume of t h e reservoir to t h e volwne of t h e coil 1 t h e p er cent
of ethane and h i gher hydr ocarbons cou ld be c a lcul a ted.. Analyses
t aken at different ti mes dur i ng t he cycle of or er a tion ar e
g iven in t h e a pp endix. Th e a 1p ara tus j u st desc r ibed for mak ing
t h e ga s ana lysis was pl a c ed in an air t hermostat so t ha t t h e
temr er a tur e cou ld be h eld constant. Gas densities we r e me a sur ed
by me ans of t h e st andard type of Edwards gas density balance.
DESIGN OF CHAHCOAL CELLS
In order th at t h e charcoal cells cou ld be oper a ted under
h i gh pr es sur e it wa s necessar y t h at t h ey be made out of steel.
A furt h er r e qui r ement wa s t hat t h e cell be e qui pped with cooling
and h eating units for use during t h e adso rp tion and re a ctivation
cycles. A vacuum connection to t h e cell wa s also n ecessar y for
re mov a l of gases d uring t h e f eriod of charco al re-activation.
Taking t h ese f a c t ors in t o consideration the final de si gn of
t he cell wa s a s f ol l ows: Th e chamber cont a i ning t h e c har co a l
was made fro m Shelby seamless steel tub ing 3! in. ou tside
diameter, 2! in. in s ide diameter, and 24 in. long. The tube
was closed at t he bottom by means of a screw plug wh ich was
welded around t he ~d g e after being ti gh te ne d up . A fl ang e was
welded to t h e to; of t h e tube for use in bolting on t h e h e a d of
t h e cell. Inside t h e cell a small steel tube, extending to t h e
bottom of the cell, was welded to t h e ga s inlet port. When
comr leted, this hi gh pr essure cylinder wa s placed inside of a
t h in walled steel t ub e h aving about an inc h l arger diameter.
A cooling coil made of 3/16 in. co pper tubing wa s pl a ced ar ound
t h e inner cy lind er, and molten le a d pour ed i n until the annular
s pace between t he cell and out er tube wa s completely f ille d.
The lead on solidifying serv e d two pur poses. First it supp lied
a complete met a l path between t he cell and the cooling and
heating elements, there-by gr e a tly incre asi ng the heat co nduc-
tivity. Sec ondly it he ld the cell and coil fi rm ly in pl a ce
wi t h in the enclosing metal s he ll. Th e electric heater was wound
on t he butside of t h e met a l s h ell, and was insul t t ed from it by
means of asbestos pap er and insolut e cement. A t h ermome ter well
was pl a c e d in t h e cell head for use in taking temper a tur e s. Gas
lines were conne cted t o t h e cell by means of standard fitti ng s.
The head of t h e c ell was bolted on in order that it cou ld be
re moved whe n desired. Cold wate r for t h e cooling coil was
ci r culated, by means of a mot or-driven pump , from a storage
tank e q_u i pp ed with an electric refri e;er a tion unit, t hru t he coil
and back to the tank. The cool ing coil wa s also connected to the
tap water line so t hat ci ty wa ter cou ld be u sed duri ng t h e
preli minary cooling f eriod, and to t h e compr es se d a i r lin e
so t hat t h e coil cou ld be blown f r ee of wa t er dur i ng t h e
he a ting periods. The t wo cells were so ar range d that when
one was being used for methane :production, th e oth er cou ld.
be he a ted and re-activa ted. The char co a l cells with thei r
necess ary vac uum, gas, and water connections ar e s h own in
Fi g . 5 .

ACKNOWLEDGEMENT
This work was financed by the American Pet r oleum
Institute, und er their Research Project # 37, and wa s carried
out under t h e direction of Professor W.N.Lacey. Credit fo r
t h e desi gn of the ap :p a~c a tus us e d for t h e solubili t y
dete :r·m ina t ions is due to B. H. Sage, and for t h e desi c;n of t h e
charco a l cells to E.S. Hill. The aut h or wish es to ex press
h is thanks and a p:r;reciation to t hose named. above for t h e
or J; or tuni t y of working with t h em on t h is project.
 Fig. 5

A- Charc oa l cells
B- Low temperature condens ati on trap
C- Tank containing refrigeration unit for cooling water
D- Control valves on air and water lines to char coal cells
E- Metr ic valves for regulating flow of gas into and
out of char co a l cells
 APPENDIX

Solubilities are expressed in terms of cubic feet of

gas, measured at 60°F. and at a pressure of 14.73 lbs. per
sq_. in. , per barrel of 42 gal. of original oil.
The pressur es given are gag e pressures; but, since
air was not removed from the ap paratus, they ar e the partial
pressures of the g~ s being forced into the a pp aratus.
With the exce r tion of a few s pecially desi gned needle
valves, Metric v alves were used on all the high pressure
eq_ui pment.
The Santa Fe Springs g as used for these experiments
consisted app r oximately of 87% methane, 8% ethane, and 5%
pro pane.
The city g a s analysis was as follows: 83.3% methane,
15.8% ethane and higher hydrocarbons, and 0.9 % nitrogen.
The analyses given b e low give some idea of the
operation of the charcoal cells. The analysis figur e
gives only the per cent of ethane and heavier hydrocarbons
present with the methane. One cycle as used below corres1 onds
1

to one filling of the bomb immersed in liquid air and the

passage of this am ount of gas thrµ the charcoal cell.
Gas sample _from cell at end of 1st cycle 0.022%
Gas sample from cell at end of 2nd cycle 0.031%
Gas sampihe from cell at end of 3rd cycle 0.17 %
Gas sample from bomb at end of vaporization period 2.94%
Gas sample from bomb after it had been allowed to
warm u p to o 0 c. the pressure being ke f t at 400 lbs.
per sq_. in. by throttling out gas 18.7%
Gas sample fr om methane storage tank 0.13%