CIHM
Microfiche
Series
({Monographs)

ICMH

Collection de
microfiches
(monographles)

Canadian Institute for Historical Microreproductions / Institut canadien de microreproductions historiques

f^^^rr

Technical and Bibliographic Notes / Notes techniques et bibliographiques

The Institute has attempted to obtain the best original
copy available for filming. Features of this copy which
may be bibliographically unique, which may alter any of
the images in the reproduction, or which may
significantly change the usual method of filming are
checked below.

D
D

Coloured covers /
Couverture de couleur

Covers damaged /
Couverture endommag6e

Covers restored and/or laminated /
Couverture restauree et/ou pelliculee

Cover title missing / Le titre de couverture manque

Coloured maps / Cartes g6ographiques en couleur

Coloured ink (i.e. other than blue or black) /
Encre de couleur (i.e. autre que bleue ou noire)

Coloured plates and/or illustrations /
Planches et/ou illustrations en couleur

Bound with other material /
Relie avec d'autres documents

Only edition available /
Seule edition disponible

Tight binding may cause shadows or distortion along
interior margin / La reliure serr6e peut causer de
I'ombre ou de la distorsion le long de la marge
int^rieure.

Blank leaves added during restorations may appear
within the text. Whenever possible, these have been
omitted from filming / II se peut que certaines pages
blanches ajoutees lors d'une restauration
apparaissent dans le texte, mais, lorsque cela etait
possible, ces pages n'ont pas ete fiim^es.

Additional comments /
Commentaires suppl6mentaires:

L'Institut a microfilm^ la meilleur exemplaire qu'il lui a
6\6 possible de se procurer. Les details de cet exem-
plaire qui sont peut-§tre uniques du point de vue bibli-
ographique, qui peuvent modifier une image reproduite,
ou qui peuvent exiger une modification dans la m6tho-
de normale de filmage sont indiqu^s ci-dessous.

Coloured pages / Pages de couleur

I I Pages damaged / Pages endommag6es

Pages restored and/or laminated /
Pages restaur^es et/ou pellicul^es

Pages discoloured, stained or foxed /
Pages d6color6es, tachetees ou piqu§es

I I Pages detached / Pages d6tach6es

I y\ Showthrough / Transparence

I I Quality of print varies /

Quality in^gale de I'impression

Includes supplementary material /
Comprend du materiel suppl^mentairc

Pages wholly or partially obscured by .naia slips,
tissues, etc., have been refilmed to ensure the best
possible image / Les pages totalement ou
partiellement obscurcies par un feuillet d'errata, une
pelure, etc., ont 6t6 film^es a nouveau de fa?on k
obtenir la meilleure image possible.

Opposing pages with varying colouration or
discolourations are filmed twice to ensure the best
possible image / Les pages s'opposant ayant des
colorations variables ou des decolorations sont
film^es deux fois afin d'obtenir la meilleure image
possible.

This Item is filmed at the reduction ratio checktd below /

Ce document est filme au taux de reduction Indiqui ci-dessous.

10x

14x

18x

22x

26x

30x

12x

16x

20x

24x

28x

32x

*<•—*

.lti-?J^'

^M^

^^^m^ %i^jk fe^vMk-

TlY^r^l^f^J^^ISa^^^^^lPr

The copy filmed here has been reproduced thanks
to the generosity of:

L'exemplaire fiim^ fut reproduit grace i la
gAnArositA de:

National Library of Canada

Bibliotheque nationale du Canada

The images appearing here are the best quality
possible considering the condition and legibility
of the original copy and in keeping with the
filming contract specifications.

Las images suivantes ont iti reproduites avec le
plus grand soin, compte tenu de la condition et
de la nettetA de l'exemplaire lilmi. et en
conformity avec les conditions du contrat de
filmaga.

Original copies in printed paper covers are filmed
beginning with the front cover and ending on
the last page with a printed or illustrated impres-
sion, or the back cover when appropriate. All
other original copies are filmed beginning on the
first page with a printed or illustrated impres-
sion, and ending on the last page with a printed
or illustrated impression.

Les exemplaires originaux dont la couverture en
papier est imprimAe sont filmAs en commencant
par le premier plat et en terminant soit par la
derniire page qui comporte une empreinte
d'impression ou d'illustration, soit par le second
plat, salon le cas. Tous les autres exemplaires
originaux sont filmAs en commandant par \t
premiere page qui comporte une empreinte
d'impression ou d'illustration et en terminant par
la derniAre page qui comporte une telle
empreinte.

The last recorded frame on each microfiche
shall contain the symbol —»• {meaning "CON-
TINUED"), or the symbol V (meaning "END"),
whichever applies.

Un des symboles suivants apparaitra sur la
derniire image de cheque microfiche, selon Is
cas: le symbole "^ signifie "A SUIVRE". le
symbole V signifie "FIN".

Maps, plates, charts, etc., may be filmed at
different reduction ratios. Those too large to be
entirely included in one exposure are filmed
beginning in the upper left hand corner, left to
right and top to bottom, as many frames as
required. The following diagrams illustrate the
method:

Les cartes, planches, tableaux, etc.. peuvent etre
filmAs A des taux de reduction diffirents.
Lorsque le document est trop grand pour etre
reproduit en un seul cliche, il est filmA i partir
de Tangle supArieur gauche, de gauche d droite.
et de haut en bas, en prenant le nombre
d'images n^cessaire. Les diagrammes suivants
illustrent la m^thode.

IK j'T'-rr.:«ar i •?'^0r3W''^.=5:-'jL'^"vv5rr-?T*T>'i:?'ri5s::^s^^

jA^>-i

•: 'a.'SFi^^rtsKi

MICROCOPY RESOLUTION TEST CHART

(ANSI nnd ISO TEST CHART No. 2l

1.0

I.I

111^

lii

|3.2

III —

U£

II ii

1.25

^ /APPLIED IM^^GE Inc

Sr '6^3 East Mai" Street

— J5 Rocnesfer, N«* 'o'k U609 uSA

-^ ^ne) 482 ■ J300 -Phone

^= (?'6) 288 - ^989 - ra«

DOMINION OF CANADA

THE HONORARY ADVISORY COUNCIL FOR SCIENTIFIC
AND. INDUSTRIAL RESEARCH

REPORT No. 2

THE RECOVERY OF
VAPOURS FROM GASES

Particularly of Benzene and Toluene
from Coal Gas

HAROLD S. DAVIS. M.A.. Ph. D..

and

MARY DAVIDSON DAVIS, B.A.

Canaba

Published by the authority of the Sub-Conunittee of the Privy
Council for Scientific and Industrial Research

OTTAWA, 1918

DOMINION OF CANADA

THE HONORARY ADVISORY COUNCIL FOR SCIENTIFIC
AND INDUSTRIAL RESEARCH

REPORT No. 2

THE RECOVERY OF
VAPOURS FROM GASES

Particularly of Benzene and Toluene
from G)al Gas

HAROLD S. DAVIS, M.A., Ph. D..

and

MARY DAVIDSON DAVIS, B.A.

Canaba

Published by the authority of the Sub-Gmimittee of the Privy
Council for Scientific and Industrial Research

OTTAWA, 1918

TABLE OF CONTENTS.

PAGE.

I Prefack

II iNTRODtCTION

Destructive Distillation of Coal

Recovery of " Light Oils " from Coal Gas 4

The Relation between Maximum Enrichment at 1 the

Composition of the Oil ^

The Extraction of Toluene only from Coal Gas 9

III A New Method for the Quantitative Estimation of

Vapours in Gases • 1"

Theory of Method ^^

Experimental Part

Efficiency of Method ^^

The Estimation of the Benzene Content of Coal Gas .... 13
Contents of the Rich and Poor Washing Oils 14

IV Conclusion

V List of Previous Publications on the Subject by the
Authors

00937797

I. PREl ACE.

This bulletin is not intended to be a <=°'"P;f f"\'^'^„J'",\*i^,°S
"benzol recovery." It embodies the main resu ts oJ researcn
cSout for two sessions, 1916-18. in the University of Man'toba.
wTth some reference to practical testscarried out on the Light Od
Recoverv Plant of the toronto Chemical Co., at Sault Stc. Mane.
oSt lilostof the results of this rese.'ch have a ready b^nput^^
Zuh fuller details in the papers listed on page IS, but it is nopea in
rhil^Setiulo Resent fhe conclusions and their significance in a

^"' We" wish to"acknowledge the substantial grant given by the

Adviso^ Council to aid the prosecution of this work during the

:^«7qi 7 1 H This crant was used to secut* as a research assistant

S Donal^i mIS?. who has been intimately associate with

the work.

II. INTRODUCTION.

The number of problems arising from the tremendous impetus
given the pr(3uction of benzene and toluene since the year 1914. led to
the investigations outlined in this paper. r,^,:„„ -«■

Benzene is a colourless liquid boiling at about 80° and freezing at
oKoMt rs" wtth a soecific gravity of 0-87. On account of its theo-
?et°cal and pmairaSporta^ce. it holds a place almost unique among
ocean ccl^poun^^^^ It is the basic substance from which the greater
narfof organic compounds constituting what is known as the aromatic
series may be considered to be derived Many o these compounds
such arphenol and aniline are actual y «y"*hes«ed from it on a
commercial scale, while great quantities ^'^ ,"«f J^j'^f^^f'^p^;.
facture of high explosives. It makes an excellent "\otor tuei. par
ttuSrlv when certain suitable organic liquids are dissolved in it.
Cerine iu freezing pomt and also preventing the deposition of carbon
IrthTc^Hnders cLmercial fuels such as " bengas are formed m

^^'^ Tohiene is also a liquid, of a somewhat higher boiling point and
lowe7free'^ng poin? than benzene. It is used • . large quantities m the
manufacture of the high explosive, trinitrotouene.

The inCTease in the number of industries conceriied with th^
manufactir"of these substances since the beginning of the war ha ^1^
to an overproduction of benzene in America; the supply of toluene,

'""Tesl^Si^^liSin^ce^^^^^^^^^^^ and toluene, are mainly obtained
.s by-products from the d-' tructive distillation of coal.

THE DESTRUCTIVE DISTILLATION OF COAL.

When coal is heated in a closed space with f^J e^J=|"ded. it
undergoes decomposition and a great many chemical substances are
formed At ordinary temperatures some of these are solids, others
v^A^' a«^ manv irases Aereat deal of solid material remains in the
'ZninTco':^^^^ke''^^ml.tur. of other substances formed

4521(V-1|

in ereat volume is drawn of! from the coke ovens. The ««"»tfJ Pf/* °S
hifgt. mixture is composed of such permanent k;-*».^ "^^Jj'^^^^^^^t
carbSn monoxide, but it also contams smalljuant.t^ o a JUng-
number of liquids and solids of various boilmg pomts. As it is cooiw,
Mrtial condensation takes place, and one or more complex solutions
JSiar At Cr^ iary temjLatures. at least two solutions separate
oSrone com^ Urgely^water. known "^^^rnmontacalliquor^he
oSir madHrX grelt number of organic substances and known by

'""^ 'lTwo!,rdX''aTundamental mistake, however, to -upm^ that
some subices completely condense into the ».<? "Jjon wh. e^o^^^^^^
pass along in the gas. Rather, each substance distributes itself in a
certain ratio between the gaseous and liquid phascs.

About 1-2 p.c. by volume of the gases from V^e coke ovens is
com^ < 1 co'mpou^ds of the aromatic -ries. ch.efly the hydro-
carboi^such as Ix^nzene and toluene. They are the chiet con
stituents of what is technically called the ight oil

Of course a certain amount of these substances stays in the tar^
especiaUy if it is «parated from the gas at a low temperature; inde^
Srmanv vears tar was their chief source. But by far the greater
cuamityonhe light oils remains in the coal gas. even after it has been
SKth watir or otherwise purified. The ratio of the amounts in

*'^ ^xL^loun'tTea^h of'^hele vapours in the gas is far below it.
»aturi?ron concentration, i.e. the concentration '" «lH}!'b"""^j;*JhJS
mire solid or liquid substance at that temperature. To separate tnese
Saooms from ?he gas. two main methods have been employed :-
vapours iromtneg.^^ and compressed so that the concentration of

the vapours is brought above saturation and they separate out a

"''"'?The vapours are absorbed into a high boiling oil. from which
thev are subsequently separated by »<-stillation.

^The first method has never been a success commercially: the
latter is the one universally employed.

THE RECOVERY OF " LIGHT OILS " FROM COAL GAS.

The theory of the washing process may perhaps be more readily
understo^ if we consider the extraction of one vapour only, e.g.

^Tonir a\"bS:ofgSTtrrfiStrpressure and tem.perature.
surround'^ by a pu?ew^^^^^^^ HQ^id- The bcr.zene vapour in the gas
wm d^LoWe in the oil until the vapour p.cssure of benzene fron. the
Tu^ounding oil is equal to that in the gas. So the maximum content
of Wene which can be absorbed by the oil from the gas is that which
riv^a va^ur p^^^ -f benzene from the solution equal to the vapour
weSure Sf beSzene in the gas. Consider a counter-current washing
svSrsuchasIs represented in Fig. 1. for washing the gas con-
dniously with oi . The gas passing through the system meets the oi
travS.^ he opposite direction. Fresh oil takes out the last traces of
bSSne f rom the gas. since such oil has no vapour pressure of benzene^
The on thereby acquires a vapour pressure of benzene proportional to

ur«i«.t. vaa with a h Kher apour preMure of beniene until it couia
eave witK v^Zr p SSre oTb*n«ne equal to that in the rich ga.
Intering -nie Sn« i« •epa'atcd by distillation from the oil and th.

•^"'^tI: :;r^o5^-£. of 'r^ene fr.n the ga. there are

*"°';'^7oru«rtw«n ga. and oil in the ^.rubber. n.u.t be

-'^TlVr^e^rBt^o^r'Lrr.t^^

quantity of l^nzene. carried from the washer, per «.cond. .hall be
equal to the quantity brought in by the gas.

<v'veralimDortant points should be noted here:—
L Iffust tCt amount of oil is passed through the w^v^erj which
is necessaW to completely remove all the^^^^^^
thp vanour oressure of benzene from the oi Tiergmj^i "•»-

fSSy'tfS ■■ rich oil ") will be »u»' - .V^f " r'™ if teU ofl
ill the benzene will have been washed fiVi.. the poor ^as. u l«» on

V "^t^n t^cirrt:';%Tte'^;n»U'tieveT,''w^
^'iXlg" in thi'^r PnS'thete is noAsufficiej, oil to remove
it Finally, if more oil than necessary is used, all the J^eMene win
fc removed from the poor gas. but the vapour pressure from the rich

^■'^"i" inri^dli^^'Vo'fompSet^^^^^^^^ an the benzene from

the gas the iow^^^ shouTdX soi^ewhat greater than the amount
thLretlcally necessary, so that the vapour pressure of bfn"n«"lJ£^
Srill always be grSier than its vapour pressure from the oil- Jhis
SLiTentoT vapour pressure will tend to dissolve the benzene fromjhe
faSo the oil The more nearly perfect the wash.n , system the smaller

this gradient need be. On the other hand, a very inefficient washing
system will extract all the benzene if sufficient oil is passed through the
scrubbers. However, since the chief cost in lx;nzene recovery is
concerned with the separation of the benzene from the rich oil, it is
desirable to use as little washing oil as possible.

3. In order to completely remove the benzene from the gas, the
poor oil which enters the scrubbers must be completely free from
benzene, for no amount of washing will ever lower the vapour pressure
in the poor gas below that of the poor oil with which it is finally washed.
The presence of benzene in the poor oil means that the steam dis-
tillation is inefficient, and does not completely separate the benzene
from the washing oil. Such inefficiency may cause a very serious
loss to the plant, for the fraction of benzene lost will be the ratio of the
vapour pressure of benzene from the poor oil to that in the rich gas.
Later in this paper, methods of measuring this ratio will be described.

THE RELATION' BETWEEN MAXIMUM ENRICHMENT AND THE
COMPOSITION OF THE OIL.

From what has preceded it w''.l be plain that the rnaximum
enrichment of benzene which the oil can assume is the quantity which
will give a vapour pressure of benzene equal to the vapour pressure of
benzene in the gas. Hence the maximum enrichment will depend
upon the content of the gas and the pressure upon it. In order to
determine the other factors governing this maximum enrichment, we
carried out a series of experiments in which we measured the vapour
pressure of benzene from solutions in washing oil, using diflferent con-
centrations and different temperatures. The results obtained may be
summarized as follows : —

1. The vapour pressure of benzene at 33 • 2° from two solutions in
oil, one of approximately double the concentration of the other, was
directly proportional to the concentration That is to say, the vapour
pressure of benzene from a solution in oil, and the concentration of
benzene in the oil are related according to Henry's Law for the
solubility of gases in liquids— a result of great theoretical and prac-
tical importance.

2. The ratio of the vapour pressure of benzene from a solution of
definite concentration to its saturation pressure was independent of
the temperature. (Von Babo's Law.)

It is thus possible to calculate the vapour pressure of benzene from
a solution of any particular concentration at any temperature, since
the vapour pressure of pure benzene is known at that temperature and
also the ratio borne to it by the vapour pressure from a 1 p.c. solution
in oil.

The amount of fresh washing oil required for any particular rate
of flow of gas depends only on the temperature of washing. The
benzene content of the gas does not affect it; for if this content is
increased, then the maximum content in the rich oil is increased in the
same proportion.

Now the efficiency of an oil for absorbing benzene from gases is
determined by the magnitude of the vapour pressure of benzene from its
solution in the oil, or to look at it from another point of view, the
efficiency of an oil is determined by its power to lower the vapour

pressure of benzene when dissolved in it. For dilute solutions in
common solvents, this power of lowering the vapour pressure is in-
versely proportional to the molecular weight of the dissolved sub-
stance, so that if the molecular concentration of the solution is known,
the proportional lowering of the vapour pressure of the solvent can be
calculated from a well-known formula. Since benzene and the
washing oil used were soluble in all proportions we suspected that
this relation might hold for very concentrated solutions; results con-
firmed this supposition. • . u ^.

The average molecular weight of the oil determined by the
freezing point method was found to be 205. From this, the vapour
pressure from a 99 p.c. solution of oil in benzene, i.e. a 1 p.c solution
of benzene in oil, was calculated to be 0263, in excellent agreement
with the experimental result 026. •, u ^•

The determination of the molecular weight of an oil when dis-
solved in benzene seems therefore to afford a good method for testing
the efficiency of the oil for absorbing benzene or other vapours from
gases; the lower the molecular weight, the greater the efficiency.

A few tests were made on absorbing oils used at a light-oil
recovery plant, in order to ascertain if this method of standardization
would hold. In general it was found that oils of low molecular
weight, 200-230, were good absorbers in practice, while an oil with
a much higher molecular weight, 280, had been rejected as inefficient.
It is hoped that research may lead to the development of washing
media with lower molecular weights than those of the present absorbers,
which at the same time will have the desired qualities of fluidity,

boiling point, etc. i r tu

So far we have considered the removal of one vapour only trom tne
gas. Similar considerations apply to the removal of a mixture of
vapours such as those which make up the " light oils.

In considering the process of absorption of light oils into washing
oil, account must be taken of the following premises, experimental
confirmation of which is cited further on in the paper: —

1. Each separate vapour contained in the gas dissolves in the
washing oil, independently of the other vapours present, until equili-
brium is reached between its vapour pressure in the gas and its vapour
pressure from the oil solution.

2. \\ hf-n washing of moderate efficiency is employed, equilibrium
between the oi' and each of the light oil vapours is reached much more
quickh- than lias often been supposed. . . u

Of course each of the constituents of the gas, other than the
light oils, also dissolves in the washing medium until corresponding
equilibrium is reached, but the amount dissolved is so small that its
effect may be neglected.

Now, although the oil constantly absorbs vapours from the gas,
equilibrium between the vapour pressures of each substance m the gas
and from the oil being closely maintained at every stage, it does not
follow that the oil becomes saturated with every vapour at the same
stage of washing. Rather, the stage at which the oil becomes satur-
ated with respect to any one vapour is determined by the saturation
pressure of the pure liquid of the same composition as the vapour in
<luestion.

Consider now the absorption of benzene and toluene vapoura
from the gas into the washing oil. The benzene content in the rich
oil will be at its maximum when it. gives a vapour pressure of benzene
equal to that in the rich gas. At this stage the dissolved toluene
only gives a pressure from the oil equal to a fraction of the toluene
pressure in the rich gas. This fraction can be shown to be equal to the
ratio of the saturation pressure of pure toluene to that of pure benzene,
so that when the oil is saturated with benzene it is only ^ saturated
with toluene. Any attempt to obtain a higher percentage of toluene
in the rich oil will result in loss of benzene, for some will pass now
through unabsorbed.

If by the maximum enrichment of the washing oil is meant
the maximum quantity of light oil which it can contain, without any
light oil passing through unabsorbed, then it is reached when the
vapour pressure from the washing medium of the lowest boiling com-
pound in the light oil, in this case benzene, is equal to its pressure in
the rich gas.

From the equation for the lowering of the vapour pressure
referred to above, using an oil with an average molecular weight
205, the following calculations can be made for a gas with a vapour
pressure of benzene of ■ 6 cm. and of toluene of ■ 15 cm. : —

1. The possible concentration in oil at 26° for benzene is 2 • 3 p.c.
and for toluene 2-3 p.c.

2. When the oil is saturated with benzene, it will have taken up
only 0-68 p.c. of toluene.

The addition of the toluene (mol. wt. 92) will also lower tho
the average molecular weight of the solution in which the benzene is
dissolved, thereby increasing the possible concentration of benzene.

Similar considerations apply to the small quantities of high-boil-
ing compounds which exist in the gas and which are soluble in the oil.
These factors would result in a steady slow increase in the amount of
the total hydrocarbons absorbed, over a comparatively long period.

To test these conclusions, the following experiments were carried
out :^-

1. Two flasks were filled with known weights of absorbing oil, and
illuminating gas was passed through them for a period of two months,
the gain in weight being recorded daily.

2. Two.flasks filled with absorbing oil were connected in series to
the gas supply, and the gain in weight of each recorded at certain
equal intervals.

In the case of 1 the following results were obtained : —

(a) The curves plotted to represent the gain in weight of the
flasks rose quite sharply at first, then gradually fell off. At the end of
the time the greatest gain in weight of the oil was 5 p.c, with the curve
still rising.

(b) After the curves began to fall off, due undoubtedly to the
fact that the benzene was no longet being absorbed from the gas, they
began to fluctuate with changes in atmospheric pressure. A low
pressure would lessen the concentration of benzene in the gas, so that
the -^il would give up some of its benzene to the gas. When this
amount was greater than the amount of higher-boiling compounds still
being absorbed from the gas, the result was a loss in weight of the oil.

The same result would follow a rise in temperature of the absorbing

flasks.

In the case of 2 the results were:—

(a) The second flask did not begin to gain in weight until there
was a perceptible vapour pressure of the light oils from the first flask.

(b) When the pressure on the second flask was increased, an
immediate gain in weight was recorded. When the extra pressure
was taken off, the weight immediately fell off again.

To sum up, our experiments showed : —

1. Equilibrium was closely maintained between the vapour pres-
sure from the oil and that from the gas, although the washing was
not particularly efficient.

2. The oil continued to gain in .veight long after it had reached
its maximum benzene content, and this increase in weight continued
ovei a very long period.

THE EXTRACTION OF TOLUENE ONLY FROM COAL GAS.

The shortage of toluene and the overproduction of benzene have
brought up in this country the question raised some time ago in
Europe as to whether it is possible to obtain all the toluene from the
gas without extracting all the benzene. As has been pointed out
already, it is theoretically possible to extract any of the constituents
from the gas, provided the proper amount of oil is used.

In any washing system the absorption process is really selective,
some of the constituents being wholly extracted, others only partially.
The rich oil is always saturated with respect to the latter. Since the
vapour pressure of pure toluene is considerably less than the vapour
pressure of pure benzene, it follows that much less oil is required to
completely remove the toluene than to completely remove the ben-
zene. It is only necessary to have that quantity of oil which will
give a vapour pressure of toluene equal to that in the rich gas. This
quantity of oil will also remove all the xylenes and other high-boiling
constituents, but will only partially extract the benzene.

Experiments conducted at absorption plants, however, have
shown that if the flow of oil is much below that required to nearly
extract all the benzene, the toluene yield of the plant is decreased.
The following reasons may be given for this decrease in the yield of
toluene : —

1. If the scrubbing system is inefficient, a considerable excess of
oil over the amount theoretically required is necessary in order _to
provide a solution gradient for the toluene, as previously described for
benzene.

2. The steam stills for separating the light oils may not be
efficient and may leave a certain vapour pressure of toluene in the poor
oil. As has already been pointed out in the case of benzene, the
proportion of toluene lost will be in the direct proportion of the vapour
pressure of toluene from the poor oil to that in the rich gas. Now, the
time occupied by any quantity of oil in passing down the steam still is
fixed, as is the amount of heating and the quantity of steam to which it
is exposed. We may therefore assume that as the amount of toluene
in the rich oil is decreased, the amount left in the poor oil will also be

decreased. Hence, with an excess of oil, the poor oil coming from the
still will contain but little toluene, so that the amount left in the poor
gas will be correspondingly small. The result is a comparatively high
yield of toluene. The steam distillation still is therefore a very
important part of a light-oil recovery plant; nothing can compen-
sate for its inefficiency.

III. A NEW METHOD FOR THE QUANTITATIVE ESTI-
MATION OF VAPOURS IN GASES.

A great need has been felt, by those engaged in the commercial
production of light oils, for methods of gas analysis requiring only
small samples and giving a rapid estimation of the content of these
vapours either collectively or individually. Such methods would make
it possible to find the conditions of production necessary to obtain the
maximum concentration of each aromatic substance, and would also
permit the efificiency of absorption processes to be checked at every
point.

Methods at present employed require the absorption of the light-
oil vapours from the gas over a comparatively long period, so that the
results represent the average values for this period. Such a method
is wholly unsuitable for checking the efficiency of the different parts
of the washing system, since conditions such as the rate of flow of the
gas, its composition, the temperatures of the washing towers, etc., are
continually changing.

We have developed a new method for the quantitative estimation
of vapours in gases, which requires only a s lall sample of gas, 150-350
cc and which can be carried to completioii m ^-l hr.

THEORY OF METHOD.

It is a generally accepted principle that the vapour pressure from a
liquid is independent of the kind of gas above it, provided the gas is
inert. Deviatiijiis from this law are well known, but it holds with
surprising accuracy in th^ case of a mixture of benzene and air at
atmospheric pressure.

Consider tw.v clos-fl flasks, as in Fig. 2, connected by a mano-
meter and filled with r 'r at atmospheric pressure. If now a small
sealed bulb conta.ning a v'i., i'j liquid be broken in each, the liquid
will partially evaporate, j\A 'i the temperatures of the flasks remain
the same, the same additicnal y^ssure will be developed in each, so
that the manometer connect i:^ 'hem will register no difference in
pressure. Ev en if the tempcratinx^ of the flasks do vary, no difference
in pressure will be recorded urtil there is a relative diffe^^ence in
temperature between them.

Now, suppose that one of the flasks had contained a certain
quantity of the \ apour of the volatile liquid corresponding to a pressure
less tnan the saturation nressuiC. Wli ..» ihe small bulb of liquid was
broken in this ono, the liquid wo'iki nor ~iud all its vapour pressure to
the pressure already ir the flark, fo ■ p-xri of that was already due to its
vapor. It would add only the amount of pressure necessary to bring
its pressure up to SwiKa.ion, i. ' si'i.yj the total saturation pressure

was added to the pure air - the other flask, the manometer connecting
the two would register a pressure equal to the pressure of vapour in the
original gas. The following points are of especial importance : —

1. Each of the sealed bulbs must contain considerably more
liquid than is required, at the temperature cf the experiment, to
saturate the atmosphere of the flask into which it is broken with the
vapour in question.

2. It is essential that the substances remaining in the bottoms of
the flasks after the bulbs are broken shall have the same composition
and shall give the same pressure of the vapour in question, and that
this pressure shall be greater than the original pressure of the - apour in
the gas. ... '

3. A solid can be used in the bulbs in place of a liquid, if the
substance exists as a solid at the temperatu-e of the experiment.

4. A solution containing the liquid of thv vapour to be measured
can be used in the bulbs provided there is sufficient of this liquid to
give a pressure greater than the pressure in the gas. A solution will offer
advantages over the pure liquid if less total pressure is developed from
it, so that the danger of leakage from the flasks is decreased. Further, if
a solution is used which gives partial pressures of two or more separate
substances, and if these substances were contained in t!ie original gas
as vapours, then ihe differential pressure developed between the two
flasks will be equai to the sum of the original partial pressures of these
vapours in the gas.

5. The partial pressure of any particular vapour in a sample
of gas is independent of the temperature of the gas, provided that
the total pressure on the gas remains constant while the volume can
change with the temperature, and provided the vapour remains always
unsaturated and obeys the simple gas laws. An apparatus con-
structed on this principle will, therefore, measure a definite quantity:
viz., the pressure of the vapour in a gas at atmospheric pressure. This
measurement can easily be reduced to standard pressure.

6. Though the partial pressure of a vapour i? independent of
the temperature, the actual weight of the vapour contained in unit
volume of the gas depends on the temperature. For one vapour, this
weight may be calculated from the partial pressure, on the assumption
that the vapour gives the same partial pressure as it would if it were a
true gas at that temperature and molecular concentration. In the case
of two or more vapours, the actual weight of each vapour cannot be to
calculated, unless the relative proportions of the vapours are also
known. This, however, does not impair the usefulness of the method
for comparative tests on the total quantities of several vapours, e.g.
light oils, in different samples of gas.

EXPERIMENTAL P.\RT.

The apparatus used is of the form shown in Fig. 2. Details of
the apparatus can be obtained by reference to Paper No. 2 listed on
page 15.

The essential features are : —

1. Two compensating flasks connected by a manometer.

2. That these flasks be air tight, but that they can be opened to
atmospheric pressure.

, Tha, .hen .h. «ask, .,. clo»d the bulb o. liquid (or »lid)
kiS »Wle the manon-eter remams outside.

W X"a* *: 'b* Mel with the desired liquid (or solid,.
b Estate the flask w|i* '^ J,lS.: as'the'"op.?S;
\jeHe^tcr.X^ra„fo\a'.Ser"prr^^^^

cock* We have found -"f 8« J;\<'„';X°"g^s sampto- The

'-sti'pte?f™rfr;t"f:^^fusb^^

klS;:.?hVlStirs°i'n waStui« Si'lorsutal, samples o,
gas. j2

the stop-cock in the manometer, equalize ^ i* j
^^ferte^rSly^ron-r^ie .a, .ide 6r.t. a.
the pressure wiU develop the Other >«ay. 5 ^.^ ^nd

drawing warm air through each side^ ^^^^ ^^H^^ ,^hi h

Considerable care must be ^a* ;" ^^ }" "^Jl » laboratory tech-
■must have thin bottoms A person w.th^^^^^^^ .^ ^ ^^ ,i

ffiis'rtr/'i^-eSSeV^^^^^^^^^^^ -^^^^^"°^ ^^^ '^^^" '"

^-^SSe L^JS^in^fs can ^^rn^de ."i. su^^^^^^^^^ a room

required. efficiency of the method.

in order to check the -^"^ J^^ netlir w^^^^^^^^
which different known Quantities of benzene in ^^^^^ ^j ^^^

S:.r^d SurS^vatef r7hir£n\Vi-n^s being within

'■' ^W^hen, however the air contained to^^^^^^^^
the benzene, it was fo""^ that the toUiene p y ^^^ ^.(^^,^^^,^1
liquid benzene in the bottoni of the HasK. ^ ^ ^^ pressure of

p?essure recorded was not the Pr^««^^^«^°^ Sved out. If sufficient
Lnzene plus the pressure «/ /^l^X^n" jS out, and thediflfer-
S Xure^n rrp'retfe rh^^m^of the benzene and toluene
pressures in the original samp e ^l a.r^ following liquids

„ Jir.r='absra^"r«S ?,SXon.e„t o. mu^.a.ng .as,

In the first case, the toluene^and xyk^ tr^^^^^^ ^^^ ^^1^^„^

the benzene. In the ^°"^ ca«.. the pr^^^^^^^^^
Sh^gTrndXt'^rriSate^^^ P-iple. In the third

Ssealfthree vapours are estimated in this way.

THK ESTIMATION OF THE BENZENE CONTENT OF COAL 0A8.

It has already been pointed -^J^^^^^^^frDi^l^^^^^^^

SS'MX^ilSS^^-^^ I^nzene

1. The toluene and "ther light oils ms^ . ^^^^ ^j^^ ^^^^

so that their pressures are almost entirely rem

while, at the same time, these dissolved substances lower the vapour
pressure of the benzene. ....

2. A less serious error is caused by the gases other than the hght
oils dissolving in the liquid benzene.

In order to eliminate these sources of error, we have conducted
investigations on gases by the Differential Pressure Method, using
bulbs filled with solid benzene, and immersing the apparatus in a bath
below the freezing point of benzene, S • 48°. It is well known that the
solubility of permanent gases in a solid, such as frozen benzene, is
vanishingly small, so that the second source of error mentioned above
is c( mpletely removed. .

Again, the vapour pressure from the solid benzene is, at any hxed
temperature, independent of the solution by which it is surrounded,
so that while the solid benzene is present at equilibrium there can be
no lowering of the vapour pressure. , . , - . ...

The experiments were highly successful. It was found possible
to determine the benzene content of a gas in this manner with con-
siderable accuracy, even when fairly large quantities of toluene and
xylene were present in the gas.

It is quite easy to keep a bath for the apparatus at a temperature
near 0° with ice and water. The chief difficulty we experienced was in
freezing the small bulbs of benzene, for they could be enormously
supercooled. The best method is to freeze them in the escaping
stre.ir of liquid and gas from a carbon dioxide cylinder. Once
frozen, they are of course quite stable for several degrees above zero.

As far as we know this is the only method by which the benzene
content only can be found accurately in a small sample of gas which
contains also its homologues.

CONTENTS OF THE RICH AND POOR WASHING OILS.

In order to measure the vapour pressures of the light oils from the
rich and poor washing oils, we have used two methods: —

1. A small amount of air was shaken up with a large amount of
washing oil until equilbrium was reached. The quantity of oil must
be so large that its concentration of light oils is not appreciably
affected by the amount which evapor tes into the air. The air was
then drawn into a Differential Pressure Apparatus and analysed for
its total light-oil content. . ..... j u iu

2. About 5 gm. of the oil was sealed up in a thin-bottomed bulb.
The bulb was then placed in a Differential Pressure Apparatus whose
flasks had a small capacity. After equilibrium was reached, the bulb
was broken and the pressure of light oils from the washing oil measured

By means of these measurements, and those previously described
in this paper, a close check can be kept on the efficiency of the washing
system, and on the steam distillation stills of a light-oil recovery plant,
and this in general will lead to an increased production.

IV. CONCLUSION.

It is possible that the methods outlined in this paper may be of
use in other fields besides light-oil recovery. In natural processes and

in many industries it often happens that large quantities of gas
mixtures are produced which contain the vapours of valuable liquids.
Thus, natural gas often contains, as vapours, quantities of the sub-
stances which constitute gasolene, while methyl alcohol and acetone
vapours exist in the gases given ofT during the destructive distillation of
wood. The method may also prove useful for the quick estimation of
the water content of air.

Finally, it is hoped that by this method of analysis it will be
possible to obtain more precise knowledge concerning the effect of
conditions of coking on the production of Hght oils. We have done
some preliminary work on the variation of the light-oil content in
illuminating gas with the different stages of coking, but extensive
research with good equipment is needed in this field from which so
much information of great economic value is still to be obtained.

V. PREVIOUS PUBLICATIONS ON THE SUBJECT BY THE

AUTHORS.

1 . The Extraction of Aromatic Hydrocarbons from Gases by means

of Liquid Absorbents, by Harold S. Davis. University of Mani-
toba publications, 1917.

2 . A new Method for the Estimation of Vapours in Gases, by Harold

S. Davis and Mary Davidson Davis. J. Ind. and Eng. Chem.
(In Press).

3. The Application of the Differential Pressure Method to the

Estimation of the Benzene and of the total Light Oil Content
of Gases, by Harold S. Davis, Mary D. Davis and Donald G.
MacGregor. J. Ind. and Ei.g. Chem. (In Press).

4. Studies on the Absorption of Light Oils from Gases, by Harold S.

Davis and Mary D. Davis. J. Ind. and Eng. Chem. (In Press).

5. "A Process for the Quantitative Estimation of Vapours in Gases."

U.S.A. Patent No. 1,272,922, July 16,1918.

Clifton, Colchester Co.,
Nova Scotia.

mtm