- Email: [email protected]
The absorption of chlorine from air by solutions of olefins and iodine in carbon tetrachloride
Chemiesl~&&se&r
Sdence, IBLIB,Vol. 2, PP. 247 to %a.
The absorption
Peupmoa Press Ltd.
of chlorine
from air by solutions
and iodine in carbon G. School of Chemical Engineering,
N.&W.
of olefins
tetrachloride
H. ROPER
University
of Technology,
Broadway,
Sydney, N.S.W.
(Beceioed 1 Septetnber 1953) Summary-The STEPHENS-D~ORRIS diiccolumn isusedtostudytheinfluenceof chemicalreaction upon the rates of absorption of chlorine from air by solutions of either cyclohexene or oleic acid with iodine in carbon tetrachloride. The increase in the liquid film coefhcient due to chemical reaction between chlorine and cyclohexene was found to be independent of the concentration of cyclohexene, but no satisfactory correlation of the experimental results could be obtained. The fractional increase in the liquid film coefficient due to reaction between chlorine and oleic acid varied from zero to 9, and was found to be
+-I=89
1
(CC+ 0*00005)cB0.5 ci
where 4 is the ratio of the liquid fihn coefficient to the coefficient in the absence of reaction, CC is the concentration of the catalyst (iodine), cg is the concentration of oleic acid, and ci is the concentration of the dissolved chlorine at the gas-liquid interface. The units are lb. moles, feet and hours. The film theory is unable to predict the effect of the chemical reaction on the liquid film coefficient. R&sumB-Au moyen de la colonne B disques de STEPHENS-MORRIS,l’auteur Btudie l’influence d’une reaction chimique sur la vitesse d’absorption du chlore dans l’air au moyen de solutions d’iode dans Ccl, contenant soit du cyclohextne soit de l’acide oleique. L’accroissement du coefficient de transfert relatif au liquide accompagnant la r4action chimique entre le chlore et le cyclohexene est independant de la concentration de ce dernier, sans qu’il ait Bti possible de trouver une correlation satisfaisante entre les divers r&ultats expkrimentaux. L’accroissement partiel du coefficient provenant de la reaction entre le chlore et l’acide ol&ique a varie entre 0 et 9 et peut ttre represent6 par :
4-1=30
(CC. + 0*00005)cBo.5 4
ou 4 est le rapport entre les deux valeurs du coefficient avec ou sans &action, cc est la concentration du catalyseur (iode), cg, la concentration de l’acide oleique, Ci, la concentration en Cl, dissous $ l’interface gaz-liquide. Unites : lb. - moles, pieds, heures. La theorie du film de passage ne permet pas de prkdire l’effet de la reaction chimique sur le coefficient de transfert relatif au liquide.
INTRODUCTION
There has been little previous investigation on a laboratory scale into continuous absorption of a soluble gas followed by chemical reaction in the liquid. Absorption in a wetted-wall column, although continuous, does not resemble absorption in a commercial packed tower sufficiently closely to enable reliable predictions to be made of the mass transfer coefficients in the commercial tower. In a packed absorber the liquor is mixed each time it flows from one piece of packing to
another. STEPHENS and MORRIS [6] designed a disc column which is suitable for small scale laboratory experiments and which gives for physical absorption results similar to those obtained in packed towers. The disc column has been used by STEPHENS and MORRIS to study the absorption of chlorine by solutions of ferrous and ferric chlorides. With a total iron concentration equivalent to 28.7 lb. of FeCls per cu. ft., the ratio of the ferrous chloride concentration (c,) to the concentration
247
G. Xl. ROPER : The
absorptionof chlorinefrom au
dissolved, but unreacted, chlorine at the gasliquid interface (ci) was varied from zero to 700. The fractional increase in the liquid film coefficient due to chemical reaction was found to be 4 -
1 = 0.75
[I :
0’a8
640 l./(g. mol.) (sec.) or 87 x lo6 cu. ft./(lb. mol.) (hr.). EXPERIMENTAL PROCEDURE The apparatus and procedure have been previously described fully [a], [6]. The absorption unit is a STEPHENS-MORRISdisc column whose major dimensions are :
0)
The absorption of chlorine from air by solutions Z-ethyl hexene-1 and iodine in carbon tetrachloride has been studied using a similar disc column (8). The ratio of the concentration of the olefin (c,) to the concentration of dissolved chlorine at the interface (ci) was varied from zero to 800. The increase in the liquid film coefficient due to chemical reaction was found to be
(2) L
Ci
J
where b varied linearly with the concentration of the catalyst (iodine) and was taken to be proportional to the specific reaction rate for the addition of chlorine to 2-ethyl hexene-1 under the conditions of the absorption. In the absence of iodine, b was 0.08 and was 1.66 when the iodine concentration was 0.001 lb. moles per cu. ft. The kinetics of the addition, in carbon tetrachloride, of chlorine to 2-ethyl hexene-1, to oleic acid, and to cyclohexene have been investigated [41- The kinetics of the addition to 2-ethyl hexene-1 and to oleic acid may be expressed
dc_4 _-.- = ka cA dt
Number of discs Disc diameter (mean) Disc thickness (mean) Tube diameter Mean perimeter for liquid flow Equivalent diameter for gas flow Free space (dry) Absorption surface (dry)
81 1.475 0.45 2.50 o-122 0.052
cm cm cm ft. ft.
89%
0.184 ft.
The discs are threaded edgeways on a 2 mm glass rod and successive discs are maintained at right-angles. The overall coefficients were calculated from the equation KG = N_,JA Ap
(5)
and were corrected for the resistance in the gas phase in order to evaluate the liquid film coefficients. The gas film coefficients were calculated from kP ~-G Em -
0.15 FlS
~G-o~4
f3cG-03
63)
GM
and the liquid film coefficients in the absence of a reaction from
cB
Hk; = 00041 P.7 (7) where c_~ and cB refer to the concentration of the chlorine and olefin, respectively. The value For each olefin (i.e., oleic acid or cyclohexene) of the specific reaction rate IE,, at 70’F is 155 the first series of runs were made without using The concentration of the l./(g. mol.) (sec.) or 9 x lOa cu. ft./(lb. mol.) the iodine catalyst. (hr.). The kinetics of the addition to cyclohexene physically dissolved chlorine in the liquor leaving may be expressed the discs was determined from the concentration of chlorine in a hydrochloric acid solution placed k c 2 in the liquor run-off tube. The total concentration dt -PA of chlorine in the liquor was determined by the the reaction being second order with respect to decrease in the olefin concentration after sufficient the chlorine concentration and zero order with time had been allowed for all the chlorine to respect to the cyclohexene concentration when react. There was no smell of hydrogen chloride there is an excess of cyclohcxene in the liquor. from the liquor, indicating that there was no The value of the specific reaction rate, at WF, is substitution.
dc, -
246
The sbaorptionof chlorine fromsir
G.H.RoPE~:
by sohiiom of cyclohcxene and iodine in carbon Mrochloridc. Table 1. Abaorprion I4iquor
_-
I&m No.
I
- _Temp. OF.
1 2 8 4 5 6 7 8 9 10
79.1 122 84.5 78.0 70-6 78.8 68.5 48-l 118 185
108CA
_in
82 68 70 71 65 66 78 78 78 78
.-
-_
out
free
59 54 52 54 58 52 66 55 56 68
0.82 0.80 oml 0.42 0.81 0.86
I
108 cg
10' ci
O-65 1.21 O-51 I.18 1.59 0.72 2.01 146 0.65 I-00
2.52 2.58 6.9 6.6 10-8 10.6 8.1 8.2 8.9 8.1
o-57 1.0 0.49 l-0 l-2 o-71 0.78 om 0.41 l-25
lG8NA liquor
IIkL
_-
With iodine in the liquor, the presence of the interhalogen compound prevented the use of the method for determining the concentration of the physically dissolved chlorine in the liquor leaving the discs. The ratio of the physically dissolved chlorine to the total amount of chlorine absorbed was estimated from a graph in which this ratio was plotted as a function of NA/cir in the absence of a catalyst. RESULTS AND DISCUSSION Absorption
GUtI
% ‘CI, mean
106cc
total
_-
T
C.onstanl air rate = O-700 c.f.m.
by solutions of yclo?mene
The experimental data for this absorption are given in Table 1. Six runs were made without catalyst at liquor rates from 70.6 to 122 lb./(hr.) (ft.), and four runs were. made at liquor rates from 48-l to 185 lb./(hr.) (ft.) with kn iodine concentration of 065 x lOwa lb. moles per cu. ft. The addition of a small amount of cyclohexene to the liquor passing over the discs increased the liquid film coefficient, but further additions did not increase the liquid film coefficient (Runs 8, 6). This agrees with the kinetics expressed by equation (4) in which the rate of disappearance of chlorine is independent of the concentration of cyclohexene provided that there is sufficient olefin present to react with all the chlorine. The catalysed absorption has the same characteristics which suggests that the rate of the catalysed
0 0 0 0 0 0
0.65 0.65 0*65 0.65
044
0.98 040 0.84 1.01 Oa8 o-77 0.47 040 1.15
0.068 0.146 o-051 O-108 0.187 0.065 0.168 0.076 o*oQ4 0.288
O*lW O-129 0.101 O-064 0*081 0.074 o-124 0.092 0.187 o-141
0.181 o-171 o-127 0.108 OaQ8 0.089 0*168 O-116 o-190 0.188
reaction is similarly independent of the concentration of cyclohexene. No satisfactory correlation of the data could be obtained. Based on either the film theory or the penetration theory, an elementary treatment of absorption followed by second-order disappearance of the gaseous solute indicates that the liquid fin: coefficient should increase with increasing values of ci. The data of Table 1 (runs 1, 8, 4, 5, 6 and runs 9, 10) show that ci has the opposite effect on the coefficient. Comparable results have been obtained by PEACEMAN [2, p. 1181 for the desorption of chlorine from water by air. He found that the film theory fails badly in explaining his experimental results. The most significant aspect of the failure is that the film theory predicts that the liquid film coefficient of decreases with increasing concentration chlorine, while it was found experimentally that the liquid film coefficient remains constant for low concentrations and then increases sharply with concentration. Absorption
by solutions of okic acid
The experimental data for this absorption are given in Table 2. Ten runs were made without using catalyst at liquor rates from 50-2 to 241 Ib./(hr.) (ft.) and seven runs were made at liquor rates from 54.5 to 118 Ib./(hr.) (ft.) and with iodine concentrations from O-09 x lOma to 0.96 x 1Om8lb. moles per cu. ft.
249
G. H. ROPER : The absorption
Table2.
Absorptionhi soldons of oh& acid and iodine in carbon telmchloridc.
Run No.
--
1
Temp. OF.
Rae
1.5 )./(hr.) vt.)
1 2 a 4 5 6 7 8 9 10 11 12 18 14 15 16 17
Et56 60*6 58.3 61.1 241 218 262 71.4 60.4 56-2 118 71.5 65.6 61.6 54.5 80.2 90.2
1
108i C
-l-1108 _-
cg
in
out
free
total
65 62 63 64 71 70 73 69 64 64 69 69 69 68 69 72 70
56 59 54 68 68 70 65 56 56 64 62 51 56 54 54 68 64
0.87 0.14 0.11 0.27 o-51 0.69 O-25 0-1s 0.18 094
2.14 8.2 4.7 16.7 2-l 18-O 5.0 16.7 3.8 18-5 18.8 5-o l-9 20.2 1.8 8.8 4-o 20.9 9.0 18.7 2.6 8.2 1.0 4.0 2.4 a-4 1-Q 11.2 0.28 15.8 2.2 14.8 5-o 6.4
Table 8.
r
lo*
CB
1o*q
Abso
1o'cc
tc*cc
108q
Couaiani air rate = 0*7QOc.] ‘m.
zi
108NA
KC-4
out
in
fm
&?a
1.9 1-Y 0.6 2.3 2.2 3.1 1.0 o-7 1-o 0.4 l-4 o-4 0.7 o-5 0.4 1-O 1.2
2-a I.6 0.7 2.7 3-l 4.5 1.4 O-8 1.2 0.4 l-9 0.4 0.8 o-5 o-4 1.8 1.5
0.38 0.85 0.13 0.87 1.13 l-88 0.47 0.16 0.29 o-55 0.46 0.09 o-19 0.14 0.019 0.22 0.55
0.4 o-4 o-1 0.5 1-l 1.6 o-5 o-1 0.2
0 0 0 0 0
0 0 0 0
0 o-45 o-22 0.45 0.22 O-96 o-09 O-96
0.6
88.6
66.6 58.8 61-l 241 71.4 80.4 50.2 118 71.5 65.6 61-6 54.5 80-2 90.2
2-l 16.7 18-O 16-7 18.5 8.8 2O.Q 18-1 2.6 4-o 3.4 11.2 15.a 14-a 6.4
2.2 l-5 O-8 2.9 l-7 0.9 l-2 4.6 1.5 0.89 0.76 O-58 o-24 1.5 o-9
0 0 0 0 0 0 0 0 0.45 O-22 0.45 0.22 0.96 0.09 '0.96
ZfkL’
0.091 0.078 0.066 O-078 O-191 0.081 0.076 O-064 O-116 0.080 o-077 0.078 O-060 O-088 0.096
O-119 O-182 0.154 O-106 O-42 O-148 o-20 o-102 0.25 0.21 0.20 o-28 0.68 0.22 044
250
0.119 O-182 o-154 0.106 O-42 O-81 O-88 0.198 o-200 o-102 0.25 o-21 0.20 0.28 0.68 0.22 044
gaseous solute with one component of the liquor is determined by two dimensionless groups : M or kacBDA/(kL’)a ; and q or D,ci/D,c,. The
ionof chlorine&IJsolutionsof okic acid.
Hh
0.092 0.134 0.115 0.086 0.242 O-204 0.228 0~118 0*138 O-088 O-165 0.146 o-141 O-156 o-27 0.152 O-28
4M
T
_-
!l
film 1 2 8 4 5 8 9 10 11 12 li3 14 15 16 17
% --
2.2 1.5 0.8 2.9 1.7 ii.1 l-4 0.9 l-2 4.6 l-5 0.39 0.76 O-58 0.24 l-5 O-9
The fihn theory of gas absorption has been examined by PFACEMAN [2,pp. 206-811. The effect of an irreversible second-order reaction of the
Run No:
of chlorine from air
1.7 6.6 7.0 6.6 2.4 4-o 6.6 7.4 4-B 6.2 8-l 11 28 7.8 13
2.3 0.20 0.099 0.89 o-20 0.23 O-18 o-55 1.3 o-22 o-50 o-11 0.085 0.23 o-31
0.8 l-5 1.3 0.5 l-2 0.8 1.6 0.6 1.2 l-6 1.6 2.2 9 1*5 8.6
8 4 2
Pm. -0-5 5 8 l-4 4
4
8 6 17 8 8
2 0.8 5 2 9 4 a
G. H. ROPER
:
The absorptionof chlorinefrom ait
theoretics1 treatment of PEACEMANis limited by the assumption that none of the gaesous solute is unreacted in the main body of the liquor. This requirement is approximated in those runs in which the concentration of the free (unreacted) chlorine in the liquor leaving the discs is small compared to the total concentration of absorbed chlorine. In Table 8 there are given the recalculated data from the runs using oleic acid in which the ratio 4 (= kL/kL’) is predicted by the theory to be greater than 8 ; the fraction of chlorine unreacted is expected to be small for these runs. The results previously published using Z-ethyl hexene-1 (8) have been similarly re-calculated and are given in Table 4. The diffusivities of the solutes, for use in calculating the dimensionless groups, were estimated by the method of ARNOLD Tabk 4.
[l]. The calculated values of the fractional increase in the liquid film coefficient (4 - 1) were read from the graphical solution of PILWEMAN. PERRY and PIGFORD [7] have published a solution for unsteady-state absorption and secondorder chemical reaction (penetration theory). By assuming equal diffusivities, the effect of the chemical reaction is determined by two groups, a dimensionless time k,c$ and c,.Jc? For certain limiting cases, the effect of unequal diffusivities is given by replacing c,Jci by D&B/DA+ In evaluating rj from the graphs of PERRY and PIGFORD, it was assumed that the effect of D, and D, is the same as in the limiting eases. PEACEMAN[p. 1971 has shown that the time, t, is t = 40,/8.18
(k&‘)S
Absorption of chlorinebg soluiitmaof P-clhylheamzc-‘l
d-1 Run No.
IOBCB
dM
1o=q
ZZkL
‘I
108cc
cxpt.
-_
--
2 2 4 5 0 7 8 9 10 11 12 13 14 15 16 17 IS IS 20 21 22 23 24 26 27 28
4.7 4.8 3.9 81 30 5.1 4.2 5.6 5.8 7.1 9.1 6.2 5.0 18 34 32 29 27 21 18 12 11 11 13 12 9.5
2.5 2.5 2.3 6.4 6.3 2.6 5.8 5.8 6.2 6.8 11 9.4 3.4 19 81 30 28 28 12 11 6.9
6-0 2.2 2.4 8.7 8.0
4.9 4.6 3.1 0.96 2.2 3.4 3.2 2.0 I.1 6.45 0.19 I.5 2.3 0.5a 0.12 0.37 1.2 2.2 0.10 I.2 0.75 I.4 2.4 0.36 2.4 2.5
-I-
1.9 I.7 I.4 0*0.54 0.124 I.1 1.3 0.77 0.33 0.11 0.037 0.41 0.80 0.073 t-b0062 0*020 0.073 0.14 0*0031 0.12 0.11 0.22 0.37 0.048 oa3 0.46
0.116 c 0*120 0.125 0.26 0.147 0.146 0.158 0.188 0.25 0.33 0.83 0.28 0.22 0.61 2.0 I.8 0.61 0.53 3.8 0.92 0.67 0.55 0.27 od4 0.46 0.56
0 0 0
0 0 0
0.20 0.20 0.20 020 0.50 0*56 0.50 I.04 I.04 1.04 1.04 I.04 0.77 0.77 0.42 0.42 0 0 0.42 0.77
0.2 0.3 0.8 1.7 0.6 0.5 0.5 I.0 I.6 2.5 7.7 2.0 I.3 5.5
---
4 4
3 4 8
11 29 22 12 7 7 5 4
19 I?5 5.4 4.5 21 4.6 8.1 2.4 0.6 I.9 1.8 2.4
3 2 --
251
fib
Pm. 0.5 0.6 0.8 2 2 0.3 0.8 I.3 3 0.9 3 8 1.2 14
G. H. ROPFAI:
and the other dimensionless 4k,,cc,D,/8.14 (k,‘)a, or kM/8.14.
group
The absorption
of chlorine
from air
is thus
For both the film theory and the penetration theory, the calculated values of (+ - 1) are of the correct order. Both theories fail in that they predict that M is an important dimensionless group. If M were important, the correlations of the previous paper [8, Figs. 5, 6 and 71 would depend on the liquid film coefficient, kL’, and hence on the liquor rate. Similarly, the points plotted in Fig. 1 of this paper would be scattered according to liquor rate. The correlation of STEPHENS Wig.1. The fractional increase in the liquid fllm coeRicient and MORHIS for the absorption of chlorine by due to the Chemical readionofchlorine in solutions of oleic ferrous chloride solutions similarly shows that acid and iodine in carbon tetrachloride. 3f is not important as their correlation of (4 - 1) CONCLUSION against cB/ci is independent of liquor rate.* The reaction kinetics for the addition of chlorine to oleic acid in carbon tetrachloride are, within experimental error, the same as the kinetics for the addition to 2-ethyl hexene-1. Equation (2) for the absorption of chlorine by solutions of S-ethyl hexene-1 and iodine in carbon tetrachloride may be re-written
O’S
I
Q-l=89
(8)
In Fig. 1, the data from Table 1 has been plotted according to equation (8). That the data for absorption by solutions of oleic acid is in excellent agreement with the data for absorption by solutions of 2-ethyl hexene-1 is shown by the grouping of the experimental points about the solid line representing equation (8). 0 V +
rc=O cc = 0.09 x 10” lb. mol/ft.* CC = 0.22 x lo4 lb. mol/ft.* CC = O-45 x 10-8 lb. mol/ft.*
A
CC = 04t3 x 104
x
The film theory has been previously criticized because it fails to predict the magnitude of the absorption coefficients in the absence of reaction, because it predicts that these coefficients are proportional to the diffusivity of the gaseous solute, and because it fails to explain the effect of concentration on the rates of desorption of chlorine from water [2], [5]. The film theory fails to explain the effect of the olefin concentration on the rates of absorption of chlorine by solutions of cyclohexene and iodine in carbon tetrachloride. The theory predicts that the fractional increase in the liquid film coefficient for the absorption of chlorine by solutions of a-ethyl hexene-1 or oleic acid with iodine in carbon tetrachloride is affected by the liquor rate, whereas the experimental values of the fractional increase are independent of the liquor rate. ACKNOWLEDGMENT
The author wishes to acknowledge the helpful advice received from Professor J. P. BAXTER.
lb. mol/ft.*
NOTATION A b
* “ In 8 recent paper, VAN K~EVELENand HWTWER [8] have reported 8 gmphical design method for gas-liquid cA re8CtC3IF3. Their method is based on the film theory, but is CB inapplicable to a disc column as the group M is used. The disc column more closely resembles a commercial packed CC tower than does either of their experiment81 units, and hence their design method should be used with considerable Ci caution for commercial towers.”
262
= surface arc8 for mass transfer, sq. ft. = a parsmeter defined by equation (2).
I concentration of chlorine, lb. moles per cu. ft. = log-mean
concentration
of the liquid
phase
re-
actant, lb. moles per cu. ft.
concentration of catalyst (iodine), lb. moles per cu. ft. = concentration of dissolved, but unreacted, chlorine at the gas-liquid interface, lb. moles per cu. f’t.
=
G. I-I. RoPEa : The absorption
DA DB
GM H
ka kb kc kL
kL’
= molecular diffusivity of chlorine, sq. ft. per hr. = molecular diffusivity of liquid phase reactant, sq. ft. per hr. = mass velocity of the gas stream, lb. moles/(hr.) (sq. ft.). = Henry’s lnw coefficient, lb. moles/(cu. ft.) (atm.). = specific reaction rate defined by equation (8), cu. ft./(lb. mole) (hr.). = specific reaction rate deiined by equation (4). cu. ft./(lb. mole) (hr.). = gas illm coefficient of mass transfer, lb. moles/(hr.) (sq. ft.) (atm.). = liquid Aim coefficient of mass transfer, ft. per hr. = liquid Alm coefllcient of mass transfer in the absence of reaction, ft. per hr.
of chlorine from air
= overall coeiliclent of mass transfer, lb. moles/(hr.) (sq. ft.) (atm.). M = a dimensionless group, kc cn DA/(kL’)g. = the rate of absorption of chlorine, lb. moles per NA hr. pBM = the mean partial pressure of air in the gas illm, ntm. AP = the logarithmic mean between the driving forces at each end of the column, atm. = a dimensionless group, DA ci/Dg ~2. 4 &G = the Reynolds’ number for flow in the gas stream. &%G = the Schmidt number for the gas phase. = time hr. t = the rate of flow of the liquid layer, lb./(ft.) (hr.). r = the ratio of the liquid Ahn coefficients = kL/kL’. 4 KG
REFFCRENCES [I]ARNOLD, J. II.; J. Amer. [2] [8] [4] [5] [6] [7] [8]
Chem. Sot. 1989 52 8987. PEACIZBAAN, D. W.; Thesis in Chem.Eng., M.I.T., 1951. ROPER, G. H.; Chem. Eng. Sci. 1958 2 18. ROPER, G. H.; Chem. Eng. Sci. 1958 2 27. Sarmwoo~, T. K. and Pnxoan, R. L.; Absorption and Extraction, 2nd ed. p. 58, McGraw STEPHENS, E. J. and MORRIS, G. A.; Chem. Eng. Progress, 1951 47 282. PERRY, R. and PIGFORD, R. L.; Ind. Eng. Chem., 1958 45 1247. KBEVELEN, D. W. VAN and HOFIYZER, P. J.; Chem. Eng. Sci., 1958 2 145.
a58
Hill, New York, 1952.
Sdence, IBLIB,Vol. 2, PP. 247 to %a.
The absorption
Peupmoa Press Ltd.
of chlorine
from air by solutions
and iodine in carbon G. School of Chemical Engineering,
N.&W.
of olefins
tetrachloride
H. ROPER
University
of Technology,
Broadway,
Sydney, N.S.W.
(Beceioed 1 Septetnber 1953) Summary-The STEPHENS-D~ORRIS diiccolumn isusedtostudytheinfluenceof chemicalreaction upon the rates of absorption of chlorine from air by solutions of either cyclohexene or oleic acid with iodine in carbon tetrachloride. The increase in the liquid film coefhcient due to chemical reaction between chlorine and cyclohexene was found to be independent of the concentration of cyclohexene, but no satisfactory correlation of the experimental results could be obtained. The fractional increase in the liquid film coefficient due to reaction between chlorine and oleic acid varied from zero to 9, and was found to be
+-I=89
1
(CC+ 0*00005)cB0.5 ci
where 4 is the ratio of the liquid fihn coefficient to the coefficient in the absence of reaction, CC is the concentration of the catalyst (iodine), cg is the concentration of oleic acid, and ci is the concentration of the dissolved chlorine at the gas-liquid interface. The units are lb. moles, feet and hours. The film theory is unable to predict the effect of the chemical reaction on the liquid film coefficient. R&sumB-Au moyen de la colonne B disques de STEPHENS-MORRIS,l’auteur Btudie l’influence d’une reaction chimique sur la vitesse d’absorption du chlore dans l’air au moyen de solutions d’iode dans Ccl, contenant soit du cyclohextne soit de l’acide oleique. L’accroissement du coefficient de transfert relatif au liquide accompagnant la r4action chimique entre le chlore et le cyclohexene est independant de la concentration de ce dernier, sans qu’il ait Bti possible de trouver une correlation satisfaisante entre les divers r&ultats expkrimentaux. L’accroissement partiel du coefficient provenant de la reaction entre le chlore et l’acide ol&ique a varie entre 0 et 9 et peut ttre represent6 par :
4-1=30
(CC. + 0*00005)cBo.5 4
ou 4 est le rapport entre les deux valeurs du coefficient avec ou sans &action, cc est la concentration du catalyseur (iode), cg, la concentration de l’acide oleique, Ci, la concentration en Cl, dissous $ l’interface gaz-liquide. Unites : lb. - moles, pieds, heures. La theorie du film de passage ne permet pas de prkdire l’effet de la reaction chimique sur le coefficient de transfert relatif au liquide.
INTRODUCTION
There has been little previous investigation on a laboratory scale into continuous absorption of a soluble gas followed by chemical reaction in the liquid. Absorption in a wetted-wall column, although continuous, does not resemble absorption in a commercial packed tower sufficiently closely to enable reliable predictions to be made of the mass transfer coefficients in the commercial tower. In a packed absorber the liquor is mixed each time it flows from one piece of packing to
another. STEPHENS and MORRIS [6] designed a disc column which is suitable for small scale laboratory experiments and which gives for physical absorption results similar to those obtained in packed towers. The disc column has been used by STEPHENS and MORRIS to study the absorption of chlorine by solutions of ferrous and ferric chlorides. With a total iron concentration equivalent to 28.7 lb. of FeCls per cu. ft., the ratio of the ferrous chloride concentration (c,) to the concentration
247
G. Xl. ROPER : The
absorptionof chlorinefrom au
dissolved, but unreacted, chlorine at the gasliquid interface (ci) was varied from zero to 700. The fractional increase in the liquid film coefficient due to chemical reaction was found to be 4 -
1 = 0.75
[I :
0’a8
640 l./(g. mol.) (sec.) or 87 x lo6 cu. ft./(lb. mol.) (hr.). EXPERIMENTAL PROCEDURE The apparatus and procedure have been previously described fully [a], [6]. The absorption unit is a STEPHENS-MORRISdisc column whose major dimensions are :
0)
The absorption of chlorine from air by solutions Z-ethyl hexene-1 and iodine in carbon tetrachloride has been studied using a similar disc column (8). The ratio of the concentration of the olefin (c,) to the concentration of dissolved chlorine at the interface (ci) was varied from zero to 800. The increase in the liquid film coefficient due to chemical reaction was found to be
(2) L
Ci
J
where b varied linearly with the concentration of the catalyst (iodine) and was taken to be proportional to the specific reaction rate for the addition of chlorine to 2-ethyl hexene-1 under the conditions of the absorption. In the absence of iodine, b was 0.08 and was 1.66 when the iodine concentration was 0.001 lb. moles per cu. ft. The kinetics of the addition, in carbon tetrachloride, of chlorine to 2-ethyl hexene-1, to oleic acid, and to cyclohexene have been investigated [41- The kinetics of the addition to 2-ethyl hexene-1 and to oleic acid may be expressed
dc_4 _-.- = ka cA dt
Number of discs Disc diameter (mean) Disc thickness (mean) Tube diameter Mean perimeter for liquid flow Equivalent diameter for gas flow Free space (dry) Absorption surface (dry)
81 1.475 0.45 2.50 o-122 0.052
cm cm cm ft. ft.
89%
0.184 ft.
The discs are threaded edgeways on a 2 mm glass rod and successive discs are maintained at right-angles. The overall coefficients were calculated from the equation KG = N_,JA Ap
(5)
and were corrected for the resistance in the gas phase in order to evaluate the liquid film coefficients. The gas film coefficients were calculated from kP ~-G Em -
0.15 FlS
~G-o~4
f3cG-03
63)
GM
and the liquid film coefficients in the absence of a reaction from
cB
Hk; = 00041 P.7 (7) where c_~ and cB refer to the concentration of the chlorine and olefin, respectively. The value For each olefin (i.e., oleic acid or cyclohexene) of the specific reaction rate IE,, at 70’F is 155 the first series of runs were made without using The concentration of the l./(g. mol.) (sec.) or 9 x lOa cu. ft./(lb. mol.) the iodine catalyst. (hr.). The kinetics of the addition to cyclohexene physically dissolved chlorine in the liquor leaving may be expressed the discs was determined from the concentration of chlorine in a hydrochloric acid solution placed k c 2 in the liquor run-off tube. The total concentration dt -PA of chlorine in the liquor was determined by the the reaction being second order with respect to decrease in the olefin concentration after sufficient the chlorine concentration and zero order with time had been allowed for all the chlorine to respect to the cyclohexene concentration when react. There was no smell of hydrogen chloride there is an excess of cyclohcxene in the liquor. from the liquor, indicating that there was no The value of the specific reaction rate, at WF, is substitution.
dc, -
246
The sbaorptionof chlorine fromsir
G.H.RoPE~:
by sohiiom of cyclohcxene and iodine in carbon Mrochloridc. Table 1. Abaorprion I4iquor
_-
I&m No.
I
- _Temp. OF.
1 2 8 4 5 6 7 8 9 10
79.1 122 84.5 78.0 70-6 78.8 68.5 48-l 118 185
108CA
_in
82 68 70 71 65 66 78 78 78 78
.-
-_
out
free
59 54 52 54 58 52 66 55 56 68
0.82 0.80 oml 0.42 0.81 0.86
I
108 cg
10' ci
O-65 1.21 O-51 I.18 1.59 0.72 2.01 146 0.65 I-00
2.52 2.58 6.9 6.6 10-8 10.6 8.1 8.2 8.9 8.1
o-57 1.0 0.49 l-0 l-2 o-71 0.78 om 0.41 l-25
lG8NA liquor
IIkL
_-
With iodine in the liquor, the presence of the interhalogen compound prevented the use of the method for determining the concentration of the physically dissolved chlorine in the liquor leaving the discs. The ratio of the physically dissolved chlorine to the total amount of chlorine absorbed was estimated from a graph in which this ratio was plotted as a function of NA/cir in the absence of a catalyst. RESULTS AND DISCUSSION Absorption
GUtI
% ‘CI, mean
106cc
total
_-
T
C.onstanl air rate = O-700 c.f.m.
by solutions of yclo?mene
The experimental data for this absorption are given in Table 1. Six runs were made without catalyst at liquor rates from 70.6 to 122 lb./(hr.) (ft.), and four runs were. made at liquor rates from 48-l to 185 lb./(hr.) (ft.) with kn iodine concentration of 065 x lOwa lb. moles per cu. ft. The addition of a small amount of cyclohexene to the liquor passing over the discs increased the liquid film coefficient, but further additions did not increase the liquid film coefficient (Runs 8, 6). This agrees with the kinetics expressed by equation (4) in which the rate of disappearance of chlorine is independent of the concentration of cyclohexene provided that there is sufficient olefin present to react with all the chlorine. The catalysed absorption has the same characteristics which suggests that the rate of the catalysed
0 0 0 0 0 0
0.65 0.65 0*65 0.65
044
0.98 040 0.84 1.01 Oa8 o-77 0.47 040 1.15
0.068 0.146 o-051 O-108 0.187 0.065 0.168 0.076 o*oQ4 0.288
O*lW O-129 0.101 O-064 0*081 0.074 o-124 0.092 0.187 o-141
0.181 o-171 o-127 0.108 OaQ8 0.089 0*168 O-116 o-190 0.188
reaction is similarly independent of the concentration of cyclohexene. No satisfactory correlation of the data could be obtained. Based on either the film theory or the penetration theory, an elementary treatment of absorption followed by second-order disappearance of the gaseous solute indicates that the liquid fin: coefficient should increase with increasing values of ci. The data of Table 1 (runs 1, 8, 4, 5, 6 and runs 9, 10) show that ci has the opposite effect on the coefficient. Comparable results have been obtained by PEACEMAN [2, p. 1181 for the desorption of chlorine from water by air. He found that the film theory fails badly in explaining his experimental results. The most significant aspect of the failure is that the film theory predicts that the liquid film coefficient of decreases with increasing concentration chlorine, while it was found experimentally that the liquid film coefficient remains constant for low concentrations and then increases sharply with concentration. Absorption
by solutions of okic acid
The experimental data for this absorption are given in Table 2. Ten runs were made without using catalyst at liquor rates from 50-2 to 241 Ib./(hr.) (ft.) and seven runs were made at liquor rates from 54.5 to 118 Ib./(hr.) (ft.) and with iodine concentrations from O-09 x lOma to 0.96 x 1Om8lb. moles per cu. ft.
249
G. H. ROPER : The absorption
Table2.
Absorptionhi soldons of oh& acid and iodine in carbon telmchloridc.
Run No.
--
1
Temp. OF.
Rae
1.5 )./(hr.) vt.)
1 2 a 4 5 6 7 8 9 10 11 12 18 14 15 16 17
Et56 60*6 58.3 61.1 241 218 262 71.4 60.4 56-2 118 71.5 65.6 61.6 54.5 80.2 90.2
1
108i C
-l-1108 _-
cg
in
out
free
total
65 62 63 64 71 70 73 69 64 64 69 69 69 68 69 72 70
56 59 54 68 68 70 65 56 56 64 62 51 56 54 54 68 64
0.87 0.14 0.11 0.27 o-51 0.69 O-25 0-1s 0.18 094
2.14 8.2 4.7 16.7 2-l 18-O 5.0 16.7 3.8 18-5 18.8 5-o l-9 20.2 1.8 8.8 4-o 20.9 9.0 18.7 2.6 8.2 1.0 4.0 2.4 a-4 1-Q 11.2 0.28 15.8 2.2 14.8 5-o 6.4
Table 8.
r
lo*
CB
1o*q
Abso
1o'cc
tc*cc
108q
Couaiani air rate = 0*7QOc.] ‘m.
zi
108NA
KC-4
out
in
fm
&?a
1.9 1-Y 0.6 2.3 2.2 3.1 1.0 o-7 1-o 0.4 l-4 o-4 0.7 o-5 0.4 1-O 1.2
2-a I.6 0.7 2.7 3-l 4.5 1.4 O-8 1.2 0.4 l-9 0.4 0.8 o-5 o-4 1.8 1.5
0.38 0.85 0.13 0.87 1.13 l-88 0.47 0.16 0.29 o-55 0.46 0.09 o-19 0.14 0.019 0.22 0.55
0.4 o-4 o-1 0.5 1-l 1.6 o-5 o-1 0.2
0 0 0 0 0
0 0 0 0
0 o-45 o-22 0.45 0.22 O-96 o-09 O-96
0.6
88.6
66.6 58.8 61-l 241 71.4 80.4 50.2 118 71.5 65.6 61-6 54.5 80-2 90.2
2-l 16.7 18-O 16-7 18.5 8.8 2O.Q 18-1 2.6 4-o 3.4 11.2 15.a 14-a 6.4
2.2 l-5 O-8 2.9 l-7 0.9 l-2 4.6 1.5 0.89 0.76 O-58 o-24 1.5 o-9
0 0 0 0 0 0 0 0 0.45 O-22 0.45 0.22 0.96 0.09 '0.96
ZfkL’
0.091 0.078 0.066 O-078 O-191 0.081 0.076 O-064 O-116 0.080 o-077 0.078 O-060 O-088 0.096
O-119 O-182 0.154 O-106 O-42 O-148 o-20 o-102 0.25 0.21 0.20 o-28 0.68 0.22 044
250
0.119 O-182 o-154 0.106 O-42 O-81 O-88 0.198 o-200 o-102 0.25 o-21 0.20 0.28 0.68 0.22 044
gaseous solute with one component of the liquor is determined by two dimensionless groups : M or kacBDA/(kL’)a ; and q or D,ci/D,c,. The
ionof chlorine&IJsolutionsof okic acid.
Hh
0.092 0.134 0.115 0.086 0.242 O-204 0.228 0~118 0*138 O-088 O-165 0.146 o-141 O-156 o-27 0.152 O-28
4M
T
_-
!l
film 1 2 8 4 5 8 9 10 11 12 li3 14 15 16 17
% --
2.2 1.5 0.8 2.9 1.7 ii.1 l-4 0.9 l-2 4.6 l-5 0.39 0.76 O-58 0.24 l-5 O-9
The fihn theory of gas absorption has been examined by PFACEMAN [2,pp. 206-811. The effect of an irreversible second-order reaction of the
Run No:
of chlorine from air
1.7 6.6 7.0 6.6 2.4 4-o 6.6 7.4 4-B 6.2 8-l 11 28 7.8 13
2.3 0.20 0.099 0.89 o-20 0.23 O-18 o-55 1.3 o-22 o-50 o-11 0.085 0.23 o-31
0.8 l-5 1.3 0.5 l-2 0.8 1.6 0.6 1.2 l-6 1.6 2.2 9 1*5 8.6
8 4 2
Pm. -0-5 5 8 l-4 4
4
8 6 17 8 8
2 0.8 5 2 9 4 a
G. H. ROPER
:
The absorptionof chlorinefrom ait
theoretics1 treatment of PEACEMANis limited by the assumption that none of the gaesous solute is unreacted in the main body of the liquor. This requirement is approximated in those runs in which the concentration of the free (unreacted) chlorine in the liquor leaving the discs is small compared to the total concentration of absorbed chlorine. In Table 8 there are given the recalculated data from the runs using oleic acid in which the ratio 4 (= kL/kL’) is predicted by the theory to be greater than 8 ; the fraction of chlorine unreacted is expected to be small for these runs. The results previously published using Z-ethyl hexene-1 (8) have been similarly re-calculated and are given in Table 4. The diffusivities of the solutes, for use in calculating the dimensionless groups, were estimated by the method of ARNOLD Tabk 4.
[l]. The calculated values of the fractional increase in the liquid film coefficient (4 - 1) were read from the graphical solution of PILWEMAN. PERRY and PIGFORD [7] have published a solution for unsteady-state absorption and secondorder chemical reaction (penetration theory). By assuming equal diffusivities, the effect of the chemical reaction is determined by two groups, a dimensionless time k,c$ and c,.Jc? For certain limiting cases, the effect of unequal diffusivities is given by replacing c,Jci by D&B/DA+ In evaluating rj from the graphs of PERRY and PIGFORD, it was assumed that the effect of D, and D, is the same as in the limiting eases. PEACEMAN[p. 1971 has shown that the time, t, is t = 40,/8.18
(k&‘)S
Absorption of chlorinebg soluiitmaof P-clhylheamzc-‘l
d-1 Run No.
IOBCB
dM
1o=q
ZZkL
‘I
108cc
cxpt.
-_
--
2 2 4 5 0 7 8 9 10 11 12 13 14 15 16 17 IS IS 20 21 22 23 24 26 27 28
4.7 4.8 3.9 81 30 5.1 4.2 5.6 5.8 7.1 9.1 6.2 5.0 18 34 32 29 27 21 18 12 11 11 13 12 9.5
2.5 2.5 2.3 6.4 6.3 2.6 5.8 5.8 6.2 6.8 11 9.4 3.4 19 81 30 28 28 12 11 6.9
6-0 2.2 2.4 8.7 8.0
4.9 4.6 3.1 0.96 2.2 3.4 3.2 2.0 I.1 6.45 0.19 I.5 2.3 0.5a 0.12 0.37 1.2 2.2 0.10 I.2 0.75 I.4 2.4 0.36 2.4 2.5
-I-
1.9 I.7 I.4 0*0.54 0.124 I.1 1.3 0.77 0.33 0.11 0.037 0.41 0.80 0.073 t-b0062 0*020 0.073 0.14 0*0031 0.12 0.11 0.22 0.37 0.048 oa3 0.46
0.116 c 0*120 0.125 0.26 0.147 0.146 0.158 0.188 0.25 0.33 0.83 0.28 0.22 0.61 2.0 I.8 0.61 0.53 3.8 0.92 0.67 0.55 0.27 od4 0.46 0.56
0 0 0
0 0 0
0.20 0.20 0.20 020 0.50 0*56 0.50 I.04 I.04 1.04 1.04 I.04 0.77 0.77 0.42 0.42 0 0 0.42 0.77
0.2 0.3 0.8 1.7 0.6 0.5 0.5 I.0 I.6 2.5 7.7 2.0 I.3 5.5
---
4 4
3 4 8
11 29 22 12 7 7 5 4
19 I?5 5.4 4.5 21 4.6 8.1 2.4 0.6 I.9 1.8 2.4
3 2 --
251
fib
Pm. 0.5 0.6 0.8 2 2 0.3 0.8 I.3 3 0.9 3 8 1.2 14
G. H. ROPFAI:
and the other dimensionless 4k,,cc,D,/8.14 (k,‘)a, or kM/8.14.
group
The absorption
of chlorine
from air
is thus
For both the film theory and the penetration theory, the calculated values of (+ - 1) are of the correct order. Both theories fail in that they predict that M is an important dimensionless group. If M were important, the correlations of the previous paper [8, Figs. 5, 6 and 71 would depend on the liquid film coefficient, kL’, and hence on the liquor rate. Similarly, the points plotted in Fig. 1 of this paper would be scattered according to liquor rate. The correlation of STEPHENS Wig.1. The fractional increase in the liquid fllm coeRicient and MORHIS for the absorption of chlorine by due to the Chemical readionofchlorine in solutions of oleic ferrous chloride solutions similarly shows that acid and iodine in carbon tetrachloride. 3f is not important as their correlation of (4 - 1) CONCLUSION against cB/ci is independent of liquor rate.* The reaction kinetics for the addition of chlorine to oleic acid in carbon tetrachloride are, within experimental error, the same as the kinetics for the addition to 2-ethyl hexene-1. Equation (2) for the absorption of chlorine by solutions of S-ethyl hexene-1 and iodine in carbon tetrachloride may be re-written
O’S
I
Q-l=89
(8)
In Fig. 1, the data from Table 1 has been plotted according to equation (8). That the data for absorption by solutions of oleic acid is in excellent agreement with the data for absorption by solutions of 2-ethyl hexene-1 is shown by the grouping of the experimental points about the solid line representing equation (8). 0 V +
rc=O cc = 0.09 x 10” lb. mol/ft.* CC = 0.22 x lo4 lb. mol/ft.* CC = O-45 x 10-8 lb. mol/ft.*
A
CC = 04t3 x 104
x
The film theory has been previously criticized because it fails to predict the magnitude of the absorption coefficients in the absence of reaction, because it predicts that these coefficients are proportional to the diffusivity of the gaseous solute, and because it fails to explain the effect of concentration on the rates of desorption of chlorine from water [2], [5]. The film theory fails to explain the effect of the olefin concentration on the rates of absorption of chlorine by solutions of cyclohexene and iodine in carbon tetrachloride. The theory predicts that the fractional increase in the liquid film coefficient for the absorption of chlorine by solutions of a-ethyl hexene-1 or oleic acid with iodine in carbon tetrachloride is affected by the liquor rate, whereas the experimental values of the fractional increase are independent of the liquor rate. ACKNOWLEDGMENT
The author wishes to acknowledge the helpful advice received from Professor J. P. BAXTER.
lb. mol/ft.*
NOTATION A b
* “ In 8 recent paper, VAN K~EVELENand HWTWER [8] have reported 8 gmphical design method for gas-liquid cA re8CtC3IF3. Their method is based on the film theory, but is CB inapplicable to a disc column as the group M is used. The disc column more closely resembles a commercial packed CC tower than does either of their experiment81 units, and hence their design method should be used with considerable Ci caution for commercial towers.”
262
= surface arc8 for mass transfer, sq. ft. = a parsmeter defined by equation (2).
I concentration of chlorine, lb. moles per cu. ft. = log-mean
concentration
of the liquid
phase
re-
actant, lb. moles per cu. ft.
concentration of catalyst (iodine), lb. moles per cu. ft. = concentration of dissolved, but unreacted, chlorine at the gas-liquid interface, lb. moles per cu. f’t.
=
G. I-I. RoPEa : The absorption
DA DB
GM H
ka kb kc kL
kL’
= molecular diffusivity of chlorine, sq. ft. per hr. = molecular diffusivity of liquid phase reactant, sq. ft. per hr. = mass velocity of the gas stream, lb. moles/(hr.) (sq. ft.). = Henry’s lnw coefficient, lb. moles/(cu. ft.) (atm.). = specific reaction rate defined by equation (8), cu. ft./(lb. mole) (hr.). = specific reaction rate deiined by equation (4). cu. ft./(lb. mole) (hr.). = gas illm coefficient of mass transfer, lb. moles/(hr.) (sq. ft.) (atm.). = liquid Aim coefficient of mass transfer, ft. per hr. = liquid Alm coefllcient of mass transfer in the absence of reaction, ft. per hr.
of chlorine from air
= overall coeiliclent of mass transfer, lb. moles/(hr.) (sq. ft.) (atm.). M = a dimensionless group, kc cn DA/(kL’)g. = the rate of absorption of chlorine, lb. moles per NA hr. pBM = the mean partial pressure of air in the gas illm, ntm. AP = the logarithmic mean between the driving forces at each end of the column, atm. = a dimensionless group, DA ci/Dg ~2. 4 &G = the Reynolds’ number for flow in the gas stream. &%G = the Schmidt number for the gas phase. = time hr. t = the rate of flow of the liquid layer, lb./(ft.) (hr.). r = the ratio of the liquid Ahn coefficients = kL/kL’. 4 KG
REFFCRENCES [I]ARNOLD, J. II.; J. Amer. [2] [8] [4] [5] [6] [7] [8]
Chem. Sot. 1989 52 8987. PEACIZBAAN, D. W.; Thesis in Chem.Eng., M.I.T., 1951. ROPER, G. H.; Chem. Eng. Sci. 1958 2 18. ROPER, G. H.; Chem. Eng. Sci. 1958 2 27. Sarmwoo~, T. K. and Pnxoan, R. L.; Absorption and Extraction, 2nd ed. p. 58, McGraw STEPHENS, E. J. and MORRIS, G. A.; Chem. Eng. Progress, 1951 47 282. PERRY, R. and PIGFORD, R. L.; Ind. Eng. Chem., 1958 45 1247. KBEVELEN, D. W. VAN and HOFIYZER, P. J.; Chem. Eng. Sci., 1958 2 145.
a58
Hill, New York, 1952.