Liquid-Side Mass Transfer Coefficients

in Packed Towers
KAKUSABURO ONDA, E I Z ~SADA, and YASUHIRO MURASE
University of Nogoya, Nagoya, Japan

The physical absorption of gas by water in a tower packed with Raschig rings has been For hydrogen a sample of 500 cc. was
investigated. The liquid-side mass transfer coefficient which was separated by dividing withdrawn by Swanson and Hulett's
the capacity coefficient by the wetted surface area is discussed from the standpoints of method (21) and analyzed by the explosion
the two-film and penetration theories. A new and simpler dimensionless group is pre- method.
sented which correlates about 90% of the data reported, including the author's own, As stated above, the gas used was of more
within an accuracy of f20%. than 99% purity, and therefore the gas-side
resistance could be assumed to be negligible.
T h e relation of t h e liquid-side mass about 80 cm. (2.6 ft.) long, with an inside The liquid-side capacity coefficient can
diameter of 6.0 cm. (2.4 in.), packed to a consequently be computed from
transfer coefficient t o parked-tower per-
height of 30 cm. ( I ft.) with 6- (%-in.), 8-
formance and physical properties of the
(%-in.) and 10-mm. (2/5-in.) ceramic kLa = { L / ( P1 {In
~ ) (C,
system is iniportant from the standpoint (1)
Raschig rings. The wet packing method
of packed-tower designs and interesting
was used to pack Raschig rings. Tap water - CJ/(Cs - CJI
from the viewpoint of the mechanism of was introduced from the head tank into the The saturated concentration C, at 25°C.
mass transfer. tower through the thermostat. was taken from the International Critical
M a n y investigators (7, 15, 20, 22) have The distributor was constructed as Tables*.
been studying t h e relation since Sher- follows: Seven glass tubes with an inside The results for carbon dioxide are shown
wood's (16) experimental equation was diameter of 4 mm. (x in.) were arranged in in Figures 2, 3, and 4. I n these cases it is
reported, b u t u p t o t h e present time t h e a position of an equilateral triangle. The known by previous reports ( 11, 16, 24) that
results seem t o be far from satisfactory. lower end of each distributor tube was set the gas velocity has no relation to the
I n this paper physical gas absorption as closely as possible on the top of the pack- liquid-side capacity Coefficient under its
by water in a tower packed with Raschig
rings is described, and kL is discussed.
ing t o avoid upper end effects.
The mass velocity of the liquid was
ascertained by the measurement of the
50 -
loading point, and here it was about
150 kg./(sq. m.)(hr.) [I1
(sq. ft.)(hr.)].
-
31 lb./
T h e relation of k L , which covers about quantity of water flowing from the tower To test the dumped packing method the
90% of the d a t a reported, including t h e bottom per unit time. The termperatures tower was repacked with 6-mm. Raschig
authors', within a 207, accuracy, was at the top and bottom of the tower were ring, and the capacity coefficient was
obtained. almost the same and were maintained a t measured. The reproducibility was good
25 f 1°C. (Figure 2).
EXPERIMENTS AND RESULTS T o measure the end effect samples were The capacity coefficient increases linearly
The absorption of pure carbon dioxide by taken from the lower tower end and the with the increase of liquid velocity; that is,
water was studied. T o ascertain the expo- funnel which was placed immediately below
nent of the Schmidt number, the absorption the support plate of the packing. kLa = c'L"' (2)
of pure hydrogen was also studied. Because For carbon dioxide a sample of 50 cc.
was introduced into 0.2N barium hydroxide where for 6-, 8 , and 10-mm. Raschig rings
the degree of purity of the gas was more c' and rn" were almost the same; c' =
than 99%, the gas-side resistance was solution and then back titrated with 0.1N
hydrogen chloride solution. The end effect 0.047, and m" = 0.72. These results agree
regarded as negligible in comparison with
for 6-, &, and 10-mm. Raschig rings was with those of previous investigators (16,25).
the liquid-side resistance. Figure 1 is the
equivalent t o a packing height of 3.6, 3.5, The results for hydrogen will be described
flow sheet of the equipment. The packed
tower consists of a single glass cylinder and 3.1 cm. respectively. later.

DERIVATION OF THE EQUATION OF 'k

It is obvious that kL and a depend on
packed-tower performances and physical
*Tabular material bas been deposited its dopu-
ment 5874 with the American Documentation
Institute Photoduplication Service. Library of
Congress: Washington 2.5, D. C., and may be obtained
Water for $2.50 for photoprints or $1.75 for 35-mm. mlcro-
film.

lo'
0
6

r 4
-
b

$2

10
#

6
4 + 69,0t 2 1
L (ks/#h)
Fig. 2. Relation of kLa vs. L in carbon
dioxide absorption by water; 6-mrn.
Fig. 1 . Schematic diagram of the absorption system. ceramic Raschig ring repacked.

Vol. 5, No. 2 A.1.Ch.E. Journal Page 235

 Fig. 5. Relation of uw/ar vs. L / a , r by
various investigators: 1. Fujita and
Sakuma (3), la. Niwa and Hashimoto (IZ),
2. Sherwood and Pigford (28), 3 and 3‘
L k%/m2 h) Shulman and DeGouff (29), 4a, 4b, Su, 5b
Fig. 3. Relation of kLa vs. L in carbon Fig. 4. Relation of kLa vs. L in carbon Weisman and Bonilla (26), 6 and 6’ Mayo,
dioxide absorption by water, 8-mm. ceramic dioxide absorption by water, 10-mm. Hunter and Nash ( l o ) , 7. Pratt (141,
Raschig ring. ceramic Raschig ring. 8. Grimley (4, 9. van Krevelen, el. al.
(ZZ), and 10. Hikita and Kataoka (6)
properties of fluids in different ways,
because the mass transfer coefficient
depends on the diffusion cocfficient and
film thickness or contact time, but the (kL[L’J/DL)will be more reasonabre. number became
contact area does not. From this principle, operating holdup Nu’ = k,h/(D,a,) (5)
To separate kL from kLa it is necessary
to evaluate a, but a t present a is a very
obscure value. For this reason it was
assumed that a is proportional to a,.
If it is possible within a reasonable error
to correlate kL separated by a, instead
of by a, it will be more convenient. The
replacement of a by a, means that
a = ku,. This assumption has been used
throughout the paper to separate kL from
kLa and has proved its usefulness within
reasonable error, at least for Raschig
rings and water. Many investigators have
presented the formulas for a, (Figure 5).
Among these formulas Fujita’s (3) was
selected because its form is more reason-
able, and its validity for water has been
confirmed by Hikita (6). Fujita’s formula
is

The total surface area for Raschig ring

was computed from a, = 4 . 7 / 0 , derived
from Figure 6.
Derivation of the formula from the standpoint
Fig. 6. Estimation of a, vs. D, by various investigators.
of two-film theory

Basic concept of kL
From the dimensional analysis and the
analogy of the heat transfer the following
equation was proposed for the gas
absorption in packed towers:

For this dimension of length previous

investigators have proposed various
lengths, for example, the tower diameter
(7) and the diameter of a sphere having
the same surface area as a packing (go),
and have eliminated L’ by Galilei’s
number (15, 22). The correlations with
various variables however are not always
good.
This dimensionless group was inter-
preted as follows. According to the two-
film theory the mass transfer coefficient
for the liquid film, is equal to D L / z ;
consequently k L / D L should be a reciprocal Fig. 7. Relation of kL(h/a,)/DLvs. L / ( a , p ) ;two broken lines indicate the region of &20T0
of the effective film thickness. If the error.

Pa g e 236 A.1.Ch.E. Journal June, 1959

 \d of the thin water layer flowing over
packings was considered. Generally, the
6
depth of thin layer flowing down over a
4 plain plate inclining 8' toward the hori-
zontal plane is derived theoretically as
2
D = [3pV/(Bpgsin e1)]1/3
The breadth of the plate may be con-
10 sidered as a,SZ (sin 0")/2, because
6 a,SZ sin 0" is the perpendicular projec-
tion of the total wetted surface area in
the tower.
The following formula is gained:
D = [3pV/a,pgS (sin 8') (sin B ' ' ) ] ~ ' ~
Putting a, instead of a, and omitting
(sin B')(sin e"), one gains the following
modified Nusselt number:
Nu' = kL.[PV / ( %P g s ) Y /DL (7)
where DL for carbon dioxide is calculated
by the formula of Wilke and Chang (27).
When one uses at instead of aurand omits
10-1 2 4 6 1 2 4 6 2 4 6 2 4 6- ,03
(sin #)(sin O"), a hypothetical depth of
'O (L/+) viscous flow over packings is taken in
the same way as if L/atp were used as
Fig. 8. Relation of k L ( p V/pgatS)"3/DLvs. L / ( a , p ) . Keys are the same as those in Figure 7. the Reynolds number of a liquid. If a,
Two broken lines indicate the region of &20% error. can be replaced by a, to correlate kL
within reasonable error, it will be much
The operating holdup was calculated 2 more convenient, and this paper will
from the experimental equation of have proved its effectiveness. By this
replacement one could correlate kL in
6 t a k e ( I S ) and a, was calculated from
Fujita's formula (3) as described above.
6take's equation is
h = 1 .295(D,L/p)""'"
*Jr b 210

3
8
6
the same equation, as described later.
Dependence of Equations (5) and (7) on
Reynolds number
(6) The modified Nusselt number in
.(Dp3gp2/p2)--0.14(u~D~) 4 Equation (5), kLh/(D,u,) vs. L/(u,p),
10' 2 4 6
is plotted in Figure 7 for the authors'
This equation is only partially correct,
and moreover it is very inconvenient to r/s DL1 experimental data as well as those of
Sherwood-Holloway
Fig. 10. Relation of k L ( p V / P ~ U ~ S ) ~ / ~ / D L , (16), Vivian-Whit-
vs. p / ( p D ~ )for the data in Figure 9 at ney (24) and Hikita et al. (5).The ranges
constant NRs= 3. of the conditions are as follows:
Name
Sherwood-Holloway
Vivian-Whitney
6
2
-
N
49 in.
8 ft.
Packing height
20
Tower diameter, in.

4 and 14
Hikita et al. 30 cm. ( 1 ft.) 2.8 and 5 (7 and 12.5 cm.)
The authors 30 cm. (1 ft.) 2.4 (6 cm.)

--- ---
Name Liquid velocity Gas velocity
Sherwood-Holloway 250 32,000 lb./(sq. ft.)(hr.) 36 1,320 lb./(sq. ft.(hr.)
Vivian-Whitney 970 16,000 lb./(sq. ft.)(hr.) 60 120 lb./(sq. ft.)(hr.)
Hikita et al. 380
78 -- 70,000 kg./(sq. m.)(hr.)
14,300 lb./(sq. ft.)(hr.)
20
4 -- 80 kg./(sq. m.)(hr.)
16 lb./(sq. ft.)(hr.)
1 2 4
(L/Ot,)
6

Fig. 9. Relation of k ~ ( p V / p g u ~ S ) vs.

8 1 0

~/~/D~
2 The authors 1,500
310 - 35,000 kg./(sq. m.)(hr.)
7,200 lb./(sq. ft.)(hr.)
50
11 - 150 kg./(sq. m.)(hr.)
31 lb./(sq. ft.)(hr.)

L / ( a , p ) in the absorption system of carbon

dioxide-water and hydrogen-water at 25°C. ; calculate h and a, from their experi- The agreement of the data in the &20%
tower diameter is 6 crn. and packed with mental equations at every point. error region however is rather poor;
6-mm. Raschig ring to the height of 30 cm. To avoid such inconvenience, the depth moreover it is most inconvenient to
calculate h and a, from experimental
Equations (3) and (6), as mentioned
above.
From Equation (7) kL[pV/(pgatX)ll'a/
DL vs. L/(u,p) is plotted in Figure 8
with the same data used as in Figure 7.
I n Figure 8 the agreement of the data is
fair, and the values are easy to calculate.
The straight line in Figure 8 represents

k L bV / (P Pt 8I ""DL (8)
Fig. 11. Relation of kL(p/utL)1fZ/DL1/2vs. L / ( u t p ); keys
are the same as those in Figure 7. = 0.44[L/(~~,p)]~'~*

Vol. 5, No. 2 A.1.Ch.E. Journal Page 237

 6
Dependence of Equation (7) o n the Schmidt (I/@M a t o 1 = c’ [Ll(a,P)I”’
number 4 (
The exponent of the Schmidt number [p29/(P2at3)>l“’(Dfa,)v‘
in Equation (4) has been discussed
frequently by many investigators. Some
of them (11, 16, 20, 25) suggest n = %,
but others (‘7, 29) support the value
n = g.To decide which exponent is
41
4%
4
4

1
2
The first term of the right-hand side of
Equation (15) is the Reynolds number
based on the total area of packings in
unit volume, and the second term is
Galilei’s number. This latter number was
right, the absorption of hydrogen by introduced in Equation (15) by the
10 100 2
consideration of the dependcnce of
-
+

water was carricd out in the same appa- ( a . a rI

ratus as described above with 6-mm. gravity acceleration. The third term
Fig. 13. Relation of kL(p/utZ)1/2/DL1’2 refers to dimensions of tower diameter
Raschig rings. vs. @,a,) at constant (Galilei) (0,= 1 in.)
DL for hydrogen,.a correction of the
and packing piece.
value used in Perry’s Hand Book, was From Equations (13) and (15) the
The exponent of the Reynolds nupber following is obtained:
selected; that is, DL = 6.30 X 10-6 in Equations (10) and (11) is in accord
sq. cm./sec. a t 25OC. N , , = 143. The with that of wetted-wall ( I , 9) and kL [p/(a,rJ> 11/2/J91,’/2
experimental data are plotted in Figure 9; liquid-rod (23) experiments. = c [ ~ / ( a f P ) l ” [ P 2 g s / ( P 2 ~ f J ) l n (16)
the ordinate kL[fi V/(pgatS)1’ /3/DL, and
the abscissa is L / ( a t p ) . Derivation of the formula from the standpoint
‘(Dtat)’
The data for carbon dioxide and of the penetration theory The constants and the exponents of
hydrogen in Figure 9 were calculated Basic concept of k L Equation (16) will now be determined
from the value a t the tower end, because from the experimental data and the
for hydrogen it was very difficult a t a In the packed tower, where the two results reported by others.
low Reynolds number to take the sample fluids flow countercurrcntly, the model
of 500 cc. from the funnel placed immedi- of the two-film theory can hardly be Dependence of Equation (16) on the
ately below the support plate. understood, even though the interface Reynolds number
From Figure 9 the values of kL[pV/ of the fluids is modified as an effective one. The exponent of the Reynolds number
(pga,S)]Ij3/DLvs. Schmidt number a t The penetration theory or its modifica- in Equation (16) was determined by the
N E a = 3 were plotted in Figure 10. tion, the surface renewal theory, seems authors’ experimental data and those of
(As the two lines in Figure 9 are almost to be preferable to the two-film theory. Sherwood-Holloway ( I 6 ) , Deed-Schutz-
parallel, the value of the Reynolds By Higbie’s penetration theory kL can Drew ( 2 ) , Vivian-Whitney (24), and
number is indifferent.) The slope of the be described as Hikita et al. (5) given in Figure 11. The
line in Figure 10 is 0.58; i t is clear that kL = 2 [ ~ ~ / ( ~ e ) ] i / z (12) range of the Reynolds number is from
the exponent of N s c = is preferable. Using an idea similar to that described 0.5 to 500, and it is found that the
in the two-film theory one can consider exponent m = 0. I n Figure 11 some of the
Results experimental data are omitted to avoid
the following dimensionless formula:
The correlation of kL for the Reynolds eonfdsion. (All data of each experiment
and Schmidt numbers becomes ~ L [ T ] ’ / ~ / D=L ~[Dimensionless]
/* (13) are plotted in Figure 14.
If T is taken to have a (:loser connection
kL[PV/(Pga,fi)l”3/DL with 0 in Equation (13), a more reasonable Dependence of Equation (16) o n Galilei’s
= o.021[L/(a,P)]o82 (9) dimensionless group can be obtained. number
However since the phenomena in the To determine the exponent of Galilei’s
.[P/(PDL)11/2 packed tower is very complex, a simplified number the values of kL[p/(afL)]1/2/0L1/*
Dividing both sides of Equation (9) by model can hardly bc presented; therefore a t constant Reynolds number were
(Re)1/3, one obtains the dimensional analysis has been used. plotted against each Galilei’s number
k , [F2/(P2g11‘/3/DL It may be considered reasonable that (Figure 12). From the slope of the plotted
the time during which the elementary line the exponent of Galilei’s number =
= 0.021[l/(a,p)]0
49 (10) surface of the liquid is exposed t o the gas 0.38. The range of Galilei’s number is from
[ P A P D L 1Ill2 depends upon the operating conditions, 2.5 x 104 to 1 . 1 x 107.
The left-hand side of Equation (10) is the physical prqxrties of the liquid, and Dependence of Equation (16) on the tower
the modified Sherwood number named by the characteristics of the tower and
packings; it can be assumed as dianzeter
Krevelen (22). From the data of Shervvood-Hollomay
By dividing both sides of Equation (16) (tower diameter 20 in.), Vivian-
(10) by the Schmidt number, one obtains 13y dimensional analysis various com- Whitney (24) (tower diameter 14 and
k , [ p / ( p g )]‘I3 = 0.021 ~hYW-4
lo 4 9 binations of dimensionless groups were 4 in.), and Hikita et al. (5) (tower diam-
(11) obtained, and from the many results eter 12.5 and 7 em.) the values of
reported the nest relation was selected. k L [ p / ( a f L ) /2/DL1;2
]l for 1-in. Raschig ring

I I IIIII I 1 I I I Ill1 I I I I IIIII I I I I 1

4 4 6 1 Z 4 5 1 0 2 4 6 1 d 2 4 6 ’
(L/arpl

Fig. 14. Relation of kL(p/afL ) ~ / ~ / ( D L ~vs.

/~GZ /u ) ~for
( &Op. ~)
Fig. 12. Relation of kL(p/afL)1;z/DL1’2vs. (Galilei) for the data reported which is thought to be purely physical; keys
the data in Figure 11 at constant Reynolds. are the same as those in Figure 7.

Page 238 A.1,,Ch.E. Journal June, 1959

are plotted against the values of (D,a,) dimension of length, and this number u = mean linear velocity in the
in Figure 13. From this figure it is clear covers about 90% of the data reported, tower, m.,/hr.
that the exponent of @,a,), p , equals 0. including the authors’, within an accuracy V = volumetric flow rate of liquid,
of f 2 0 % . cu. m./hr.
Results 4. From the standpoint of the pene- z = effective film thickness, m.
All results are given in Figure 14, tration theory a new and simpler dimen- 2 = height of packing layer, m.
where almost all the data of each experi- sionless formula is presented in Equation Greek Letters
ment reported are plotted; over 90% (17) which was applicable within a n
of the results lie within a &20% error. /J = viscosity of liquid, kg./(m.) (hr.)
accuracy of &20% to almost all the data
From the results discussed above p = density of liquid, kg./cu. m.
reported, as shown in Figure 14. 0 = time during which an element of
Equation (16) becomes 5. From both theories it is found that the liquid surface is exposed to
k, tPl(atL)11/2/DL1/2 kL depends on the Reynolds and Schmidt the gas, hr.
(17) numbers, as shown in Equation (19). O’, 0’’ = inclination angle of a plate to
= 0.013 [p”/(p“,”]”.”* The exponent of the Reynolds number
The left-hand side of Equation (16) is the horizontal
is in accord with that found in the experi- T = ratio of circumference of a circle
the result of the use of p/(a,L) as the ments by wetted wall and liquid rod; it is
dimension of time in Equation (13), but to its diameter
believed reasonable that a is proportional
the meaning of p/(a,L) can be considered t,o a,, where the proportional constant is LITERATURE CITED
as L/p = u, and a, = 4.7/0,; then independent of L and G within an accu- 1. Brota, W.. Chem.-Znq.-Tech., 26, 470
p/(a,L) a D,/u; that is, p/(a,L) is racy of f20%, a t least for Raschig rings ( 1954).
proportional to the apparent mean resi- and water. 2. Deed, D. W.. P. W. Schutz. arid T. B.
dence time over a packing. Drew; Znd. Eng. Chem., 39, 766 (1947).
ACKNOWLEDGMENT 3. Fujita, S., et al., Chem. Ens. (Japan),
RESULTS FROM THE TWO THEORIES 18, 64 (1954).
The authors are indebted to S. Hatta 4. Grimley, S. S., Trans. Znsl. Chem.
The exponent of Galilei’s number in of Tokyo University for his encouragement
Equation (17) may be assumed to be
then Equation (17) becomes
x.5; and also t o T. And6 for hydrogen-absorp-
tion experiments carried out as a part of
Engrs. (London), 23, 233 (1945).
5. Hikita, H., et al., Chem. Eng. (Japan),
20, 113 (1956).
his Bachelor of Science work and t o F. 6. Hikita, H., and T. Kataoka, ibid.,
kL [ P / ( P S ) 1 1 / 3 Otsubo for experiments on carbon dioxide 20, 528 (1956).
= 0.013[L/(~,p)]~” (18) absorption. Thanks also go to Mrs. M. L. 7. Kling, G., Chem.-1ng.-Tech., 25, 577
Mayland for her careful review of the
* [P/tDLP)l-1/2 English text. (1953).
The coincidence of Equation (11) and 8. Leva, Max, “Tower Packing and
Packed Tower Design,” p. 7 , United
Equation (18) is most satisfactory, and NOTATl0N
States Stoneware Co. (1953).
the correlation of kL from both theories a = effective surface area of pack- 9. Lyn, S., J. R. Straatemeier, and H.
becomes ings, sq. m./cu. m. Kramers, Chem. Eng. Sci., 4, 49 (1955).
a, = total surface area of packings, 10. Mayo, F., T. 0. Hunter, and A. W.
kL [ p i ( p g ) 11’~ = c w ( a t p ) Nash, J . SOC. Chem. Znd., 54, 375T
sq. m./cu. m.
[Il/(PDL)l-l/p
*
a, = wetted surface area of packings, (1935).
sq. m./cu. m. 11. Molstad, M. C., et al., Trans. Am.
.c = 0.01 -0.02 (19) Inst. Chem. Engrs., 38, 410 (1942).
The agreement of &20% in Figure 14 B = breadth of a plate, m.
12. Niwa, Hashimoto, thesis, Tokyo Inst.
corresponds to a value of c of 0.013. c, c’ = a constant Technol., Tokyo, Japan (1952).
The exponent of the Reynolds number C1, Cz, C. = concentration of liquid at 13. Otake, T., and K. Okada, Chem. Eng.
in Equation (19) is in accord with those the entrance, the exit of the (Japan), 17, 176 (1953).
found in the experiments by wetted tower, and at the saturation, 14. Pratt, H. R. C., Ind. Chemist, 26, 291
wall ( I , 9) and by liquid rod (B),and respectively, kg./ma. (1950).
D = depth of thin layer of viscous 15. Ramm, W. M., “Absorptionspronzesse
this seems to have an interesting signifi- in der chemischen Technik,” p. 229
cance for the investigation of the mecha- flow, m. , -

DA = diffusional coefficient of solute (1953).

nism of gas absorption in packed towers; 16. Sherwood, T. K., and F. A. L. Holloway,
moreover these coincidences indicate gas, sq. m./hr.
Trans, Am. Inst. Chem. Engrs., 36,
that the assumption made in this DaDer: D, == nominal diameter of a packing,
39 (1948).
(a = lca,) is not only convenient but also m. 17. Sherwood, T. K., and R. L. Pigford,
reasonable. Here the proportionally con- D, = diameter of the tower, m. “AbsorDtion and Extraction,” p. 226.
stant k is independent of L and G. G = mass flow rate of gas, kg./ McGrabHill, New York (1952‘).
If a were a function of G under its loading (sq. m.)(hr.) 18. Ibid., p. 228 (1952).
point, it is most natural that kLa would g = gravity acceleration, m./hr.2 19. Shulman, H. L., and J. J. DeGouff,
h = operating holdup. Ind. Eng. Chem., 44, 1915 (1952).
also be a function of G, but all the data 20. Shulman, H. L., et al., A.1.Ch.E.
reported prove that lcLa is independent k = proportional constant independ-
ent of gas and liquid velocity Journal, 1, 253 (1955).
of G(11, 16, $4). 21. Swanson, A. A., and G. A. Hulett,
kL = liquid-side mass transfer coeffi- J . Am. Chem. Soc., 37, 2490 (1915).
SUMMARY cient, m./hr. 22. van Krevelen, D. W., and P. J. Hof-
1. The absorption of carbon dioxide L = mass flow rate of liquid, kg./ tijzer. Chem. Ens. Progr.,- . 44, 529
and hydrogen by water in the tower (sq. m.)(hr.> (i948).
packed with Raschig rings was investi- L‘ = length, m. 23. Vielstich, W., Chem.-Zng.-Tech., 28,
gated. m,m’ = exponent of the Reynolds number 543 (1956).
2. It was proved within an accuracy of m“ = exponent of liquid velocity 24. Vivian, J. E., and R. P. Whitney,
&20% that kL is a function of the total n, n’ = exponent of the Schmidt number Chem. Eng. Progr., 43, 691 (1947).
surface area, as shown in Equations (11) Nu’ = modified Nusselt number, 25. Ibid., 44, 529 (1948).
26. Weisman, J., and C. F. Boriilla, Znd.
and (18). k L[ L W L Eng. Cham., 42, 1099 (1950).
3. From the standpoint of the two- p , p‘ = exponent of @ , a , ) 27. Wilke, C . R., and Pin Chang, A.I.Ch.E.
film theory a new modified Nusselt S = sectional area of empty tower, Journal, 1, 264 (1955).
number is presented which takes the sq. m. Manuscript received December 11, 1957: revision
apparent mean depth of liquid as a T = time, hr. received M a y 19, 1958; paper accepted October 6 , 1958.

Vol. 5, No. 2 A.1.Ch.E. Journal Page 239