Ind. Eng. Chem. Process Des. Dev. 1985, 2 4 , 149-152
149
fluid density, kg m-3
The fractional breakthrough times, determined from the solution charts, are listed in Table VI.
pf =
Nomenclature
L i t e r a t u r e Cited
Pb = bed density, kg m-3 Asher, J. W.; Campbell, M. L.; Epperly, W. R.; Robertson, J. L. Hydrocarbon Process. 1080, 48(1), 134. Balzli, M. W.; Liapls, A. I.; Rlppin, D. W. T. Trans. Inst. Chem. Eng. 1078. 58. 145. Basmadjiin, D. Ind. Eng. Chem. Process. D e s . D e v . 1080a, 19, 137. Basmadjlan, D. AIChE J. 108Ob, 2 6 , 625. Basmadjlan, D. J. Agric. Eng. Res. 1081, 2 6 , 251. Basmadjian, D. “The Adsorption Drylng of Gases and Liquids”, I n “Advances In Drying”, Vol. 3, Mujumdar, A. S., Ed.; Hemisphere: 1984. Carter, J. W.: Husain, H. Chem. Eng. Sci. 1074, 29, 267. Chi, C. W.; Lee, H. AIChESymp. Ser. 1080, 65, No. 96, 65. Collins, J. J. Chem. Eng. mag. Symp. Ser. 1087, 63,No. 74, 31. Cooney, D. 0.; Strusi, F. P. Ind. Eng. Chem. Fundam. 1072, 7 1 , 123. DeVault, D. J. Am. Chem. SOC. 1043, 65, 532. Fritz, W.; Schiunder, E. U. Chem. Eng. Sci. 1081a, 36, 721. FrRz, W.; Merk, W.; Schlunder. E. U. Chem. Eng. Sci. 108lb, 36, 731. Glueckauf. E. Discuss. Faradsy SOC. 1040, 7 , 12. Hiester, N. K.; Vermeulen. T. Chem. Eng. Prog. 1052a, 48. 505. Hlester, N. K.; Vermeulen, T. Document No. 3665, American Documentation Institute. Washington, DC, May 14, 1952b. Merk, W.; Fritz, W.; Schlwddunder, E. U. Chem. Eng. Sci. 1081, 36, 743. Clazle, R. N.; Vermeulen, T. Chem. Eng. J. 1080a, 19, Omatete, 0. 0.; 229. Omatete, 0. 0.; Clazie, R. N.; Vermeulen, T. Chem. Eng. J. 1080b, 19, 241. Perry, J. H., Ed. “Chemical Engineers’ Handbook”, 4th ed.;McGraw-Hill: New York, 1963. Santacesaria, E.; Morbidellie, M.; Danlse, P.;Mercenari, M.; Carra, S. Ind. Eng. Chem. Process Des. Dev. 1082a. 21, 440. Santacesaria, E.; Morbldellie. M.; Danise, P.; Mercenari, M.; Carra, S. Ind. Eng. Chem. Process Des. D e v . 1082b. 21, 446. Shen, J.; Smith, J. M.; Ind. Eng. Chem. Fundam. 1088, 7 , 1. Smith, J. M. private communication, University of Californla, Davis, 1981. Takeuchi, Y.; Wasai, T.; Suglnaka, S. J. Chem. Eng. Jpn. 1078, 1 7 , 458. Takeuchi, Y. private communication, Meiji Universlty, Japan, 1983. Takeuchi, Y. Kogyo Yosui 1078, No. 233, 4. Thomas, W. J.; Lombardi, J. L. Trans. Inst. Chem. Eng. 1071, 49, 240. Treybal, R. “Mass Transfer Operations”, 3rd ed.: McGraw-Hill: New York, 1080; Chapter 11.
D, = pore particle diffusivity, m2 s-l D, = overall “effective”pore diffusivity, m2 s-l D,= solid particle diffusivity, m2 s-l
D,, = overall solid particle diffusivity, m2 s-l f ( r ) = 2/(1 + r ) for O Ir I 1, r-1/2for r I 1 Cb = fluid carrier mass velocity, kg m-2 s-l kp = fluid film volumetric mass transfer coefficient, s-l LMn = length of mass transfer zone, m LES = length of equilibrium section, m LUB = length of unused bed, m NR= dimensionless parameter (Figure 2); see eq 7 q = adsorbate concentration, kg of solutefkg of sorbent q = ion concentration on resin, equivfkg of resin Aqq = solute content difference between end points of equilibrium curve, kg of solutefkg of sorbent r = dimensionless separation factor (Figure 2, eq 6) for adsorption along type I and desorption along type I11 equilibrium curve R = particle radius, m t = time, s TR = dimensionless parameter (Figure 2); see eq 8 u = superficial fluid velocity, m s-l V = bed volume, m3 W, = specific bed weight, kg of sorbentfkg of carrier treated W,* = specific bed weight under equilibrium conditions, kg of sorbentfkg of carrier treated Y = fluid phase solute concentration, kg of solutefkg of carrier Y’ = equivalent ion mass fraction Y / Ybt AY, = solute concentration difference between end points of equilibrium, kg of solutefkg of carrier AY/ AY, = fractional solute concentration z = distance from bed inlet, m t = bed void fraction
Received for review July 21, 1983 Accepted February 27, 1984
NO, Absorption by Ferrous Sulfate Solutions Serglo Boslo, Albert0 Ravella, Glovannl B. Saracco;
and Gluseppe Genon
Dipartimento di Scienza del Materiali e Ingegneria Chimica, Poiltecnico di Torino, Torino, Italia
+
A stirred-vessel absorber was utilized to study kinetics of reactions between FeSO, H,SO, aqueous solutions and nitrogen monoxide and dioxide. The reaction between FeSO, and NO was investigated in a wide range of liquid phase concentrations, and the data were fit by using the penetration theory. Subsequently experimental resutts for the absorption of NO, were correlated to those obtained with NO by the ratio of the rates of the two gasses absorbed in the same operating conditions. Then a calculation model of the NO, absorption rate, based on a simplified reaction scheme, is proposed. The salting-out parameter constant x(NO,), not available up to now in the literature, was evaluated.
of his absorption tests according to the penetration theory. This work aims to confirm the validity of the available correlations for evaluating monoxide kinetics in a wider range of operating conditions than has up to now been examined, and above all it is concerned with the study of dioxide absorption.
Introduction
The use of acidic solutions of ferrous sulfate for NOP scrubbing has two considerable advantages compared with other wet processes: very low operation cost with regard to reagents and easy regeneration (of exhaust solutions). The chemical reaction between NO and FeS04 is (1)
Experimental Section Apparatus. The absorption was carried out in a
was studied first by Kustin et al. (1966), then by Hikita et al. (1977), and by Sada et al. (1978). Kustin determined lz2 and K values a t 25 “C;Hikita interpreted the results
1500-cm3stirred vessel with a flat gas-liquid interface. It was provided with two stirrers: one for the gas and one for the liquid phase, and both phases were considered perfectly mixed. The stirring speed was between 14 and
k2
NO(g) + FeS04(aq) s Fe(NO)S04 k-1
OCIJW~O~I~I
124-0149~01.5010
0
1984 American Chemical Society
150
Ind. Eng. Chem. Process Des. Dev., Vol. 24, No. 1, 1985
i: 1 0 II
-3 q’ 9\ ”-:-yi1‘1
L
--------
!#I/
- - - -- - - - -
p%
Figure 1. Flow sheet for absorption tests: (1) cylinder with NOz in N,; (2) NO cylinder; (3) ball valves; (4) gas line thermostatic coil; (5) inlet flow meter; (6) outlet flow meter; (7) mercury gauge; (8) reactor; (9) storage tank; (10)pump of recycle; (11) glass stopcocks; (12) stirrer electric motor; (13) speed reducer and indicator; (14) ejector; (15) thermostatic bath; (16) gas analyzer.
160 rpm. The running was continuous with regard to the gas and discontinuous with regard to the liquid. In order to keep the concentration conditions almost constant in the liquid phase, this was recirculated through a 4-L storage tank. The absorbents were FeSOl solutions ranging from 0.25 to 1 mol/L containing 0.25 or 0.5 mol/L of H,S04. The tests with NO were carried out with technical grade gas at 1 atm and 25 OC. The NO absorption rates were determined by measuring the gas flow entering and leaving the reactor. For NO2 tests a N02-N2 mixture was used, calculating the absorption rates from the difference between NO2 concentrations in the gas flow at inlet and outlet. The NO2 concentrations in the gas flows were determined by absorbing the dioxide in H202solutions at 6 w t 5% and measuring the produced H+ by the potentiometric method (Streuli and Averell, 1970). The experimental equipment was modified for this purpose by connecting a 350-cm3 vessel containing H202to the place where the gas was dispersed by a porous diffuser. Taking into account the low NO, pressure in the mixtures, N204presence was disregarded. A scheme of experimental apparatus is reported in Figure 1.
I 7
ua I 5 I 3
2
IO 1 ow
ID0
Figure 2. Enhancement factor vs. Hatta modulus (solid lines indicate the theoretical values); points indicate [FeS04] and [H2S04], respectively: (0) 0.25, 0.5; ( 0 )0.5, 0.5; (A) 0.8, 0.5.
1 100
IO00
Figure 3. Enhancement factor vs. Hatta modulus (solid lines indicate the theoretical values); points indicate [FeS04] and [H2S04], respectively: ( 0 )0.25, 0.25; (m) 0.4, 0.5; (A) 0.75, 0.5; ( X ) 1, 0.25.
by measuring the N20 absorption in water at different stirring speeds. The results lead to the expression
kLNZww = 8.15 X 10-5n0.65 cm/s The liquid side mass transfer coefficient for NO in ferrous sulfate solutions can be obtained from kLNzO-r as
Results: NO The enhancement factor F was determined by measuring the absorption rate of NO and using the equation J = kF (Ai - Ao), The monoxide concentration in the liquid bulk, Ao, was assumed to be zero while the interface concentration, Ai, was evaluated from the relation
DA, and D N ~ O were calculated by the Wilke and Chang (1965) method and they are 2.26 X 10” cm2/s and 1.75 X 10” cm2/s, respectively. The same method leads to DB, = 2.16 X cm2/s. DA was evaluated by means of the equation used by Sada et al. (1978)
-DA- - 1 - 0.122[H2S04J- 0.291[FeS04J
where kSl and kn2 are the “salting-out parameters” for FeS04-H2S04solutions. The salting-out parameters depend on the dissolved ions and gases and they can be expressed as k, = x g + x, + x , Using the numerical values of Onda et al. (1970) Ai = Ai, x 10-(0.4076[FeSO~lfO.l533[H~O,])
(2)
Ai, = 1.88 X lop3mol/L at 25 OC (Lange)
The liquid side mass transfer coefficient was determined
(4) DAW The ratio v,/v was measured with Cannon-Penske capillary tubes at 25 OC. The experimental results were compared with the theoretical values of the enhancement factor and, as shown in Figures 2 and 3, they correspond closely to them. The theoretical curves F vs. Ha were drawn by using the solution of the differential balance equation proposed by Hikita et al. (1977) for the reaction A+R+E with the numerical values: k2 = 6.2 X lo5L/(mol s); K = 450 L/mol; DEIDA = 0.284; DE/D, = 1.
Ind. Eng. Chem. Process Des. Dev., Vol. 24, No. 1, 1985 b
1
,,
llDl
',' 2 I-;
''
74
25-
151
ow
iI c
i
I '
db
1
I
.I
I
1 *
2-
N
2
3
4
5
1
I
,y
Figure 5. Plot of z vs. ionic strength of solutions.
x,(NO,) = -0.34 was obtained. In the experimental conditions FA practically does not depend on Ai so we can suppose that FB is also not affected by Bi; if this were not so, FB would depend on pB. This assumption was confirmed by tests made varying the gas flow rate from 1to 10. These tests were planned in order to investigate the influence of mass transfer in the gas phase, too. We observed that z is independent of the gas flow rate and so must be FB. Besides it is possible to write expression 7 only if the ratio FBIFA' is constant with I. By extrapolating (6) to I = 0, we obtain zo = 25. By assuming Biw= 4.1 X mol/L (Andrew and Hanson, 1961), we obtain FB/FAr r 1 with a margin of error lower than 10%.
represents the ratio between the absorption rates of NO2 and NO in the same solution and at the same flow conditions, from gas phases where they are present at the same concentration. The experimental values of z obtained at constant concentration of FeS04 and H2S04are practically the same; since
is not affected by n, it is possible to conclude that FB is independent of n. In the range of Ha numbers considered, in fact, it is possible to verify that FA is constant with n less than a few percentage units. The i values obtained for every solution were plotted against the ionic strength of the solution (Figure 5). The results can be interpreted by a linear regression as log z = 0.1581 1.41 (6) The linear dependence of log z on I indicates that the ratio FBIFA' must be, with a good approximation, independent of the composition of the liquid phase, reading as follows
+
Conclusions The penetration theory and the correlations proposed in the literature for the physical parameters calculation enable us to evaluate with a satisfactory approximation the absorption rate of NO in FeS04/H2S04solutions. The absorption kinetics of NO2seems to be very similar to that of NO, from which it differs only in a factor equal to the ratio of physical solubilities. Therefore we could suppose, without discussing the mechanisms of nitrogen dioxide absorption and reaction in liquid phase, that the whole process takes place in two steps: in the first one NO2 is absorbed and transformed in NO with simultaneous oxidation of part of Fe2+to Fe3'; formally these phenomena lead to the balance equation NO2 + 2Fe2++ 2H+ F? NO
+ 2Fe3++ H 2 0
(9)
In the second step reaction 1 occurs. The first step is probably faster than complexation reaction 1,which has to be considered the controlling step of the overall reaction between NO2 and FeSO1. The calculation method proposed here for evaluating NOz absorption, that is the determination of JNo in relation with JNoand z , enables us to estimate witL good accuracy the performances of an absorption column for NO,. Like all the wet deNO, processes the absorption in FeS04 aqueous solutions is more effective with regard to dioxide than with regard to monoxide, since the solubility of the former is much greater. However, as far as we know, the acidic FeS04 solutions have two specific advantages:
Ind. Eng. Chem. Process Des. Dev. 1985, 24, 152-159
152
a lower cost of the reagents and a great facility for regeneration. Nomenclature A = nitrogen monoxide, NO B = nitrogen dioxide, NO2 E = complex, Fe(NO)S04 R = ferrous sulfate, FeS04 D = diffusivity coefficient, cmz/s F = enhancement factor, dimensionless I = ionic strength of solutions, mol L Ha = Hatta number, (kzRJIA/kL!) l 2 with film theory and with penetration theory J = absorption rate per unit interfacial area, mol/(cmz s) K = equilibrium constant for reaction 1, L/mol k , = salting out parameters, L/mol k2 = forward reaction rate constant for reaction 1, s-l kl = reverse reaction rate constant for reaction 1, s-l kL, k = liquid-side mass transfer coefficient in the absence of chemical reaction, cm/s n = stirrer speed, rpm p = partial pressure, atm t = exposure time, s z = empirical constant, L/mol z = defined by eq 5, dimensionless a = Bunsen absorption coefficient, dimensionless v = kinematic viscosity, mz/s
/
c = cation g = gas aq = aqueous 1 = at the interface 0 = in the bulk of solution w = water Superscripts
’ = referred to the same conditions of NOz - _- average
Registry No. NO, 10102-43-9;NOz, 10102-44-0; FeSO,, 7720-78-7;H2S04,7664-93-9.
Literature Cited Andrew, S. P. S.; Hanson, D. Chem. Eng. Sci. 1961, 14, 105. Hlkita, H.; Asai, S.; Ishlkawa, H.; Hirano, S. J. Chem. Eng. Jpn. 1977, 10, 120. Kustln. K.; Taub, I.A.; Weinstock, E. Inorg. Chem. 1966, 5 , 1079. Onda, K.; Sada, E.; Kobayashl, T.; Kito, S.; Ito, K. J. Chem. Eng. Jpn. 1970, 3 , 18, 137. Sada, E.;Kumazawa, H.; Tsuboi, N.; Kudo, I.; Kondo, T. Ind. Eng. Chem. Process Des. D e v . 1970, 17, 321. Streuli, C. A.; Averell, P. R. “Analytical Chemistry of Nitrogen and Its Compounds”, Part 1, Vol. 28 of “Chemical Analysis”; WUey-Interscience: New York, 1970; p 99. Wilke, C. R.; Chang, P. AIChE J. 1955, 1 , 264.
Subscripts a = anion
Received for review July 25, 1983 Accepted March 5, 1984
Kinetics and Mechanism of the Catalytic Hydrochlorlnation of Acetylene to Vinyl Chloride by Use of a Transient Response Technique Ajlt K. Ghosh and John B. Agnew* Department of Chemlcal EngineerlngsMonash University, Ckytm, Victoria, 3 168, Australla
Results are presented for an experimental study of the kinetics of acetylene hydrochlorinationto vinyl chloride over mercuric chloride catalyst on activated carbon by a transient response technique. Transients are produced by introducing step and pulse concentration changes in the feed stream to an internally recycled reactor. Elementary-step rate constants for a dynamic kinetlc model are estimated from measuredgas-phase concentration profiles. A break in the slope of the Arrhenius plot at a temperature of 140 OC is shown to be due to a change in reaction mechanism. Predicted transient behavior for a well-mixed reactor is closer to experimental data when the dynamic reaction is used than is the case when a steady-state kinetic model is employed.
Introduction Steady-state kinetic models for catalytic reactions are generally based on Langmuir-Hinshelwood rate expressions (Hougen and Watson, 1948). However, this method assumes the existence of a single rate-controlling step with the other steps in equilibrium; the concept is restrictive in the sense that instrinsic kinetics and reaction mechanisms may not be properly revealed. On the other hand, an elementary-step approach yields a more complete description as it more closely describes the microscopic events involved, providing a more reliable basis for extrapolation. The general availability of high-speed computers now
* Chemical Engineering Department, University of Adelaide, SA-5001,Australia. 0196-4305/85/1124-0152$01.50/0
permits the experimental kineticist and design engineer to utilize more complex kinetic expressions which describe fundamental catalytic processes more closely. The use of transient response methods in homogeneous catalysis is well documented. However, the application of such techniques to heterogeneous catalytic systems gathered momentum only during the mid-1970s. Yang et al. (1973) and Kobayashi and Kobayashi (1973) were among the first to implement the technique. Reactors ideally suitable for transient studies, various method of perturbation, and other criteria needed for better representation of transient data are discussed elsewhere (Kobayashi and Kobayashi, 1974; Bennett, 1976; Ghosh, 1981). This paper is concerned with the implementation of a transient response technique to study the kinetics of a complex reaction system in which all the gaseous compo0 1984 American Chemical Society
