In?. J. Heat Mass Transfer. Vol. 5, pp. 429-433. Pergamon Press 1962. Printed in Great Britain.

PIONEER PAPERS IN CONVECTIVE MASS TRANSFER

5. W. G. WHITMAN: The two-film theory of gas absorption, Chemical and Metallurgical
Engineering 29, 146-148 (1923). Reprinted with permission from Chemical Engineering, Copy-
right 1923, McGraw-Hill Publishing Co.
Editors Foreword
In the preface to their book Absorption and Extraction (McGraw-Hill, 1951), Sherwood and Pigford state: Methods
of applying the Whitman two-film theory to various design problems have been extended and refined, but it is
curious that after 28 years the theory itself has never been adequately checked experimentally. Whatever may be
meant by an adequate check in this case, their remark at least shows the central position which the Whitman idea
has occupied in the thinking of chemical engineers. Now that films have been replaced by boundary layers in
the mass transfer literature, it is interesting to note that Whitman was quite clear in 1923 that the film of definite
thickness was a fiction, as witness his phrase: actually no such sharp demarcation exists.
D.B.S.

A Preliminary Experimental Conjirmation of

THE TWO-FILM THEORY OF GAS ABSORPTION*

It Seems to Explain Satisfactorily the Well-Recognized

Dlflerences of Absorption Rate for Varying Concentrations
WALTER G. WHITMAN
Assistant Professor of Chemical Engineering, Massachusetts Institute of Technology

THE various theories for the mechanism of gas dW

absorption proposed within the past 8 years __ = coefficient times driving potential (1)
de
agree on the general form for the absorption where
equation. Essentially these all assume that the W = amount of gas absorbed
rate of absorption is proportional to a potential
factor multiplied by certain coefficients which 0 = time for this absorption.
are dependent upon the construction and size of A considerable divergence is noted, however,
the absorbing apparatus and the operating con- in formulations of the potential factor. In all
ditions. cases, this driving potential is proportional to the
difference between actual conditions and condi-
* Few subjects are creating more interest both tions at equilibrium-i.e. it is proportional to
theoretically and practically than gas absorption. For the distance from equilibrium. Certain writers
years entirely empirical in its applications and even yet
preponderantly so, there have been developed several [Lewis,J. Znd. Eng. Chem., vol. 8, p. 825 (1916);
theories that seem to shed some light on its mechanism. Whitman and Keats, J. Znd. Eng. Chem., vol. 14,
This paper clears up some of the fog around the potential p. 185 (1922)] have expressed the potential as
factor. It will contribute in helping to put this unit the difference between the partial pressure of
process on a substantial basis. One of the slogans should
perhaps be, No more monstrosities as absorption
solute in the gas and the partial pressure of solute
towers! No more of the old formula, Lets make it a exerted by the liquid. Others [Donnan and
foot bigger in diameter and 5 ft higher just for good luck. Masson, J. Sot. Chem. Znd., vol. 39, p. 236~
429
430 PIONEER PAPERS IN CONVECTIVE MASS TRANSFER

(1920); Van Arsdel, Chem. & Met., vol. 23, general. From the work on humidification and
p. 1115 (1920), and vol. 28, p. 889 (1923)] dehumidification it was shown that the rate of
picture the potential as a difference between diffusion is controlled by two surface films, an
the concentration of solute in the gas (converted exterior gas film surrounding the liquid. and a
into concentration in the liquid by the solubility liquid film on the surface of the liquid. The
relations~p) and the concentration in the liquid. relative importance of these two films varied with
These two concepts may be expressed as follows : the experimental conditions, the liquid film
resistance being eliminated in. humidification
Driving potential = p. - pi = p (2) processes and amounting to 75 per cent of the
Driving potential = c, - cl = c (3) total for certain dehumidi~Gation runs. Further-
more, the effect of such variables as liquid and
where 1~~= the partial pressure of gas velocities on the resistances of the separate
solute exerted by the films differed to a considerable extent.
liquid, and In the article referred to, no attempt was made
co = the concentration of to point out the significance of the two-film
liquid which would be in theory in absorption problems. It is the purpose
equilibrium with the gas. of this paper to present this view of the absorp-
tion process and illustrate its application.
The two different formulations give the same Fig. 1 shows a liquid in contact with gas from
result for the specific case of absorption at which the solute is being absorbed. The gas and
constant temperature using a solute which liquid films at the boundary can be indicated as
obeys Henrys law-i.e., p = kc. Under these having a definite thickness, although actually no
conditions, Ap = kdc, and either expression such sharp demarcation exists. Conditions at the
would be permissible. outside of the gas film (1) are the same as in the
Consideration of the physical signi~GanGe of main body of gas, while those at the inside of the
these terms shows the cause of this divergence. liquid film (3) are the same as in the main body
In all cases (except where a slow chemical of liquid. The gas and liquid at the boundary
reaction is involved) the rate of absorption is between the two films (2) are in equilibrium.
controlled by the rate of diffusion through the Absorption occurs therefore through two fiims
surface iilms at the gas-liquid boundary. The in series. Diffusion through the gas film should
first concept pictures diffusion through a gas
film, actuated by a difference in the partial
pressure of the solute at the outside of the film
(in the main body of gas) and the inside of the
film (in eq~~brium with the liquid). The other
concept considers diffusion through a liquid film
on the surface of the main body of liquid with
diffusion controlled by a difference in concentra-
tion between the outside of the film (in equili-
brium with the gas) and the inside (the true
liquid Concentration). Actually, the choice
between these two theories has been made in
the past by selecting the one which gave most
consistent results in the experiments under
investigation at the time.
In 1922, Keats and the author published
experiments (Whitman and Keats, J. lnd. Eng.
Chem., vol. 14, p. 185) contrasting the processes
of humidification and dehumidification, and
comparing heat transfer with absorption in FIG. 1. Contact of gas and liquid phase.
 PIONEER PAPERS IN CONVECTIVE MASS TRANSFER 431

be determined by a partial pressure gradient of Such a case is illustrated in the absorption of

the solute-i.e., (by pI - pz) and through the hydrochloric acid. A series of absorption
liquid film by a concentration gradient (c, - c& experiments with this gas were made at a constant
The following equations represent the process: temperature of gas and liquid of 30 deg. C. The
pressure solubility relations at this tem~rature
are shown in Fig. 2 on co-ordinate and in Fig. 3
(4) on semi-log plots.
It is important to note that the partial pressure
where k, = the coefficient of diffusion through
of HCl over aqueous solutions of the acid is
the gas film, and
negligible up to concentrations of approximately
k, = the coefficient of diffusion through 250 g/l, but that it rises rapidly with increase in
the liquid film. acid concentration above this range.
Since for acid concentrations up to about
The values of k, and k, will, of course, be 250 g/l the partial pressure of HCI is negligible, it
dependent on experimental conditions. follows that acids the surface concentration of
Comparison of this equation with the two which is below this figure will absorb gas as
previously proposed : rapidly as it can reach the surface-i.e. rate of
absorption is determined solely by rate of
dW diffusion through the gas film and the equation
z = KAP, - P3)
for absorption rate becomes d W/dt = K,,p,.
dW Furthermore, Kg is in this case identical with k,.
-- = KG% - c3) However, at high concentrations of acid, the
df?
equilibrium pressure on the surface becomes
(where I& and iu, are overall coefficients) ex- large and also changes rapidly with the concen-
plains why neither of the latter have broad tration. Consequently the absorption of a small
application. Theoretically equations 5 and 6 amount of acid into the surface brings the surface
should apply only when the concentration is into substantial equilibrium with the gas, thereby
directly proportional to the pressure. Under other preventing absorption until the absorbed acid
conditions it is usually not permissible to has diffused through the liquid film into the
use an overall coefficient KP or I& with an overall interior. However, owing to the shape of c-p
potential (pl - p3) or (cl - ~3, since there are curve, the available concentration gradient is
two separate potential factors involved and there slight and consequently the rate of diffusion
is no direct proportionality between them. slow. This explains the well-recognized fact
Many cases cannot be handled by such that for the same pressure gradient the rate of
arbitrary simphfications. For example, if the absorption into the dilute towers of a hydro-
deviation from Henrys law or if the temperature chloric acid system is far more rapid than in
range is considerable, a simplification based on those towers containing strong acid.
direct proportionality betweenp and c would be Table 1 presents data obtained in absorption
unjustified and the overall coefficients Kc or KD runs under different concentrations and pres-
would have no significance. sures, and the coefficients which should be

Table 1
=- --- -c.-- p
dW
Run zi p, p3 AP Cl G AC &I KG
c-------mmHg------- f------g/l-
1 41.0 225 55 170 425 378 47 0.24 0.87
2 24.0 41 0.3 407 3: 204 164 0.59 0,146
3 24.1 41 0 41.0 9 359 0.59 om7

zr--- ~- __._~._ I__ ~--

432 PIONEER PAPERS IN CONVECTIVE MASS TRANSFER

obtained by assuming either a gas or a liquid

diffusion alone.
Comparison of runs Nos. 1 and 2 shows that
K, increases nearly two and one-half fold in the p2 = 156 mm
latter case. The values for K, vary even more,
that for the first run being twelvefold that for and the corresponding concentration, c2, from
run 3. It will be noted that the values of K, for Fig. 3, is 412 g/l.
runs 2 and 3 are identical, although those for
Kc differ more than twofold. The conditions of W 41.0
k, = -e (c2 - es) = 412 _ 378 = 1.2
these runs represent a specific limiting case

400 2.8

I" 360 2.6

: 320 2.2

s 280 I.8
I
z 240 1.4
0)
; 200 I.0
:
g 160 0.6

G
8 120 0.2
> 0
80 -0.2

40 -0,6

0
-1.0
240 320 400 480 I80 220 260 300 340 380 420 460

g HCI/C g HCl/L

FIGS. 2 and 3.
FIG. 2. (At left)-Pressure-solubility relationship at 30 deg. C. FIG. 3
(At right)-Semi-log plot of Fig. 2. The crosses in Fig. 3 represent data by
Dolezalek, Z. ph. C/I., vol. 26, p. 334 (1898). and the circles are
data by Bates and Kirschman, J. Am. Chewz. Sot., vol. 41, 1897 (1919).

where treatment by the pressure potential alone It is now possible to check these figures in a
is permissible. general way by applying them to run 2.
These data can be treated from the two-film
k
concept as follows. From equations 4 and 5 2,$~=PL~~2Lz
k, . c2 - c3 c2 - 204
kc Pl - P2
(71 Referring again to Fig. 3 for corresponding
kD c2 - (3
values of pz and c2, the equation is satisfied by
and pz = 0.6 mm and c2 == 224 g/l.
Therefore,
K, _ Pl -- P2
4 Pl - P3
(*) dW
-=/&(c,-c3)=1.2x20=24g/h,
d8
By assuming that the value of K, from run 3
(where the liquid is so dilute that the back checking with the observed value in Table 1.
pressure pn is practically zero) equals k,, run 1 It is recognized that the data presented in this
can be used to calculate k,. illustration are insufficient to prove definitely the
 PIONEER PAPERS IN CONVECTIVE MASS TRANSFER 333

truth of the two-film theory, and further work the relationship between concentration and
has been started along the same lines. It does, pressure is approximately a direct proportion.
however, show clearly that the methods formerly For the more complicated problems, the two-
proposed are entirely incapable of handling film theory seems fundamentally sound and the
cases of this nature, and that the concept of a preliminary experiments have afforded checks
single gas or liquid film is permissible only when as to its validity.