Document not found! Please try again

Kinetics of Absorption of Nitric Oxide in Aqueous FeII-EDTA Solution

Jun 25, 1987 - Nippon Chemical Consultants, Inc. Report 15, Jan 1980. Pohlmann, H. P.; Meyer, D. H.; Leipold, H. A. Prepr-Am. Chem. Roffia, P.; Tonti,...
1 downloads 0 Views 549KB Size
Ind. Eng. Chem. Res. 1988,27, 770-774

770

Hobbs, C.; Van't Hof, H. (Celanese Corporation) Can. Patent 1038 403, 1978. Nippon Chemical Consultants, Inc. Report 15, Jan 1980. Pohlmann, H. P.; Meyer, D. H.; Leipold, H. A. Prepr-Am. Chem. SOC.,Diu. Pet. Chem. 1983, 28, 1085. Roffia, P.; Tonti, S. Montedipe, Internal Report, 1983; Bollate, Italy. Roffia, P.; Calini, P.; Tonti, S. Oxid. Commun. 1985/1986, 8(1-2), 167. Roffia, P.; Tonti, S.; Tancorra, R. Chem. Znd. (Milan)1980,62, 500.

Serguchev,Yu. A,; Beletskaya, I. P. Russ. Chem. Rev. (Engl. Transl.) 1980, 1119. Sheldon, R. A.; Kochi, J. K. Metal Catalyzed Oxidations of Organic Compounds; Academic: New York, 1981. Waters, W. A. Discuss. Faraday SOC.1968,46, 158.

Received for review December 9, 1986 Revised manuscript received June 25, 1987 Accepted November 24, 1987

Kinetics of Absorption of Nitric Oxide in Aqueous FeII-EDTA Solution Li Huasheng* and Fang Wenchi Chemical Engineering Department, Zheng Zhou Institute of Technology, HeNan Province, China

The rates of absorption of nitric oxide in aqueous solution of Fe"-EDTA chelate were measured by using a stirred vessel with a free flat gas-liquid interface. The experimental results were analyzed with the chemical absorption theory based on the film model. The second-order forward rate constants for the reaction between nitric oxide and Fe'LEDTA in aqueous solution were calculated and correlated as a function of temperature and ionic strength of the solution. The chemical equilibrium constants for the reaction were determined from the measurement of the total solubility of nitric oxide in aqueous Fe*'-EDTA solution. The exhaust gases emitted from nitric acid plants have caused concern because of the high NO, concentration and the great exhaust quantity. How to decrease NO, concentration in exhaust gases has become a problem that scientists have been trying to solve. So far, though many methods have been proposed, none can be considered perfect. The complexing absorption method of NO, in aqueous FeeEDTA solution is a promixing method because of very fast absorption. However, as absorption proceeds, the existence of a little O2 and NOz oxidizes a part of the Fen-EDTA into Fern-EDTA which is of lower complexing activity. So, how to reduce Fe"'-EDTA to Fe'I-EDTA is an important problem in such a NO, removal process. Many methods of reducing Fe"'-EDTA have been reported. Because they have their own disadvantages, there is along way to go before commercialization can be realized. Based on our previous research on removing H2Sin coal gases of ammonia plants by using Fe"'-EDTA solution, this paper presents that H2S can be used as a reducing agent for Fe"'-EDTA in solution after removing NO,. Thus, NO, and H2S are not only removed, but elemental S and NO are also recovered. Removing NO, and removing H2S form a closed, circulating system. The main reactions are

NO

+ FeII-EDTA

= Fe"(N0)EDTA

control this process balance becomes a key problem in realizing this combined proceM for removing H2Sand NO,. The purpose of this paper is to establish the kinetic equation of absorption of NO in aqueous Fe"-EDTA solution.

Chemical Absorption Mechanism The complexing reaction 1is a fast, reversible reaction, second order with respect to the forward reaction (first order in NO and FeLEDTA, respectively) and first order with respect to the reverse reaction (Kustin et al., 1966; Sada and Kumazawa, 1980). The complete analytical solution for the enhancement factor for this (2,1)-orderreaction can't be obtained. It is only possible to obtain a particular solution under a certain particular condition. Hatta (1957), Danckwerts (1970), and Teramoto et al. (1978) presented enhancement factor expressions for a similar system, but the scope of the use of these expressions was restricted by their assumed conditions. For a general, (2,1)-order, fast, reversible reaction (4) AM + 4)= E(l) based on the film model, the material balance equation for each component in the liquid film can be written as DA

(d2A/dX2)= K2AB - K-1E

(5)

DB

(d2B/dX2) = KzAB - K-IE

(6)

DE

(d2E/dX2) = K-IE - KQAB

(7)

(1)

2N02 + 2Fe"-EDTA = Fe"(N0)EDTA + FeIII-EDTA + NO,- (NO, removal process) (2) H2S + Fe"'-EDTA = Fe"-EDTA + S + 2H+ (H2Sremoval process) (3) Obviously, compared with processes for separately removing H2S and NO,, this combined method can omit two regeneration processes, which makes energy consumption, cost, and equipment investment decrease greatly. So, this combined method is no doubt the most economical. It is easy to imagine that, if the combined method were operated, considerable process coordination would be required. The two processes would influence each other, if either of them acts too fast. Thus, researching the kinetics of absorption of NO in aqueous Fe"-EDTA solution to 0888-5885/88/2627-0770$01.50/0

The boundary conditions are dB/dx = dE/dx = 0

x = 0; A = Ai, x = x,; -DA

(8)

(dA/dx) = (U - Xo)Ro, (9) E = Eo B = Bo, Furthermore, the chemical equilibrium of reaction 4 is established in the bulk of the liquid, so the following condition should be fulfilled (Zhu, 1981): K = Eo/A$o (10) For such fast reactions, the authors considered that the diffusion of component B into the liquid film from the bulk 0 1988 American Chemical Society

Ind. Eng. Chem. Res., Vol. 27, No. 5, 1988 771 liquid is relatively slow; thus, its interfacial concentration Bi is much lower than its bulk concentration Bo. Because the main reaction region between A and B is near the interface, it is easy to imagine that concentration B of component B which takes part in the reaction is closer to its interfacial concentration Bi (a little more than Bi, practically). On the basis of such consideration, the above differential equations can be solved. The kinetic equations of absorption for fast (2,1)-order reversible reactions are obtained as NA = PK,"(Ai - A,) (11) where

5=

[

Pm-P

P-1

]'I2

(14)

Om = 14- (KBorE)/(l + KAjrE/rB)

(15)

rj = Dj/DA qE

+ rBqB ( j = B, E)

=

(16) (17) (18)

The above equations can be relevantly simplified, when they are used in different cases. (1)In the initial stage of absorption, the concentration of component A in the bulk liquid is very low. That means that A, is approximately equal to zero. So the above equations can be simplified to

NA = PKLOAi P = Ytl/tanh (rtJ

t1=(-)Pm - P Pm- 1

I /

Figure 1. Stirred vessel: 1, gas inlet; 2, gas outlet; 3, gas-phase stirrer; 4,liquid-phase stirrer; 5, liquid inlet; 6, liquid outlet; 7, baffle; 8, water jacket.

+ -(Ao/Ai) Pm-1 6 - 1 p' = 1

n

(19) (20)

lJ2

(2) In the general case of the fast reaction, YE > 3. Thus, tanh ( ~ t=) 1. When A. is equal to zero, the expression for the enhancement factor P can be given by

-

(3)The forward reaction rate is much greater than the reverse reaction rate, that is, K m. When A. equals zero, the above equations can be simplified to the kinetic equations of absorption for a second-order irreversible reaction: NA = PKLOAj (23) P = w/tanh (77) (24)

This result is identical with Hikita and Asai's (1964) kinetic equations of absorption with second-order irreversible reactions, which were directly derived from the absorption process mechanism.

Experimental Section The reactor, 8 cm in diameter, used in the experiment is the double-mixed contactor with jacket (see Figure 1). The absorber was operated continuously with respect to

the gas phase and batchwise with respect to the liquid phase. The cylindrical absorption vessel had four symmetrically located baffles in the liquid phase. Two stirrers, driven by two separate motors, were used to agitate the gas and liquid phases. The stirrers were fan turbines with four flat blades. This liquid stirring blade was gradient in shape with equal sides, which didn't destroy the planar gas-liquid interface. The liquid-phase and gas-phase stirrers were operated at constant speeds of 80 and 300 rpm, respectively. The liquid phase was an aqueous solution of FeLEDTA. The Fe"-EDTA solutions were prepared by adding equimolar amounts of FeS04 and EDTA-2Na (ethylenediaminetetraacetic acid disodium salt) to distilled water, and the pH of the solution was adjusted by NaOH. The concentration of FeII-EDTA ranged from 0.01 to 0.05 M. The solute gas NO from a cylinder was diluted by N2, saturated with water vapor at the temperature of the apparatus, and fed into the absorber. The feed concentrations of NO were varied from 45 to 410 ppm. The concentrations of NO in the gas phase at the inlet and the outlet of the absorber were determined by gas chromatography with a FPD. The absorption rates of NO were calculated from the difference between inlet and outlet concentration and the total gas flow rate.

Predictions of Mass-Transfer Coefficients and Physical Properties The liquid-side mass-transfer coefficient, KLO, was determined by measuring the rate of physical absorption of pure C 0 2 in water at 25 "C and correlated to the liquidphase stirring speed nL by KL,Co2"= 9.293 X 10-7nLo.79 (26) Under the experimental conditions for the dilute COP absorption in aqueous solutions of KOH, the overall mass-transfer coefficient, KG, was found to be independent of KOH concentration, that is, KG = k,. The gas-side mass-transfer coefficient was correlated to the gas-phase stirring speed nG by KG,CO2= 2.423 x 10-3nGo.56 (27) The values of KL and KG for NO were estimated by use of the relation that kL and kc are proportional to D2I3, respectively. The liquid-phase diffusivities of NO and FeLEDTA and the gas-phase diffusivity of NO were estimated by Wilke-Chang's (1955) equation and Wilke-Lee's equation

772 Ind. Eng. Chem. Res., Vol. 27, No. 5, 1988 Table I. Physical Properties for Fe"(N0)EDTA System 10-2H, (atm.L) / T,"C Eo,mol/L I, mol/L mol io9& m2/s 2.531 25 0.01 0.05 5.179 2.513 0.02 0.10 5.179 25 2.492 0.03 0.15 5.179 25 5.179 2.486 0.04 0.20 25 2.415 0.05 0.25 5.179 25 1.896 0.03 0.15 4.363 15 3.130 0.03 0.15 6.005 35 7.377 4.822 0.03 0.15 55 6.756 0.03 0.15 8.164 75

io9&, m2/s

iog&, m2/s

105kL,m/s

102kG, mol/(m2. atm-s)

0.679 0.679 0.679 0.679 0.679 0.514 0.868 1.319 1.866

0.633 0.633 0.633 0.633 0.633 0.479 0.809 1.229 1.738

3.512 3.495 3.476 3.470 3.404 2.674 4.447 6.819 9.583

6.570 6.570 6.570 6.570 6.570 6.178 6.970 7.808 8.664

1 0 4 ~L/mol , 153 153 153 153 153 218 78.2 53.1 25.2

34.5 7

u 2

w

,Y

1

-5.9

1

I

L

-58

-7.5

Figure 2. Effect of temperature on equilibrium constant.

(1955), respectively. The solubility of NO in the liquid absorbent was evaluated by the method of van Krevelen and Hoftijzer (1948): log (Ai/Aiw) = 4ke111 + k d 2 )

k , = i+

+ i- + i,

(28)

(29)

where i+(Fe2+)is 0.049, i-(SOd2-)is 0.022, and i,(NO) is -0.1825. In the present work, the chemical equilibrium constant for reaction 1 was determined by measuring the total solubility of NO in aqueous Fe"-EDTA solutions. The experimental procedure was as follows. A known quantity of Fel*-EDTA solution (usually about 60 cm3) was introduced into the absorption cell placed in a water bath. The NO gas with a certain concentration was supplied to the absorption cell, and the solution in the cell was agitated by a magnetic stirrer. The outlet gas-phase concentrations of NO were determined continuouslyuntil equilibrium was established. The amount of gas dissolved, i.e., the total solubility of NO, was found from the total gas flow rate, inlet gas concentration, absorption time, and the change of outlet gas-phase concentration with time. The values of the chemical equilibrium constant K for reaction 1were calculated from the measurements of the total solubility AT of NO by using the following equation: (30) The predicted values of the physical properties are listed in Table I.

Results and Discussion Figure 2 represents the effect of temperature on the equilibrium constant as an Arrhenius plot, where the K values at about 0.15 mol/L are plotted against the reciprocal of absolute temperature 1/T on semilogarithmic coordinates; the values of K obtained in the present work were well correlated by the following equation: log K = 8124(1/T) + 5.333 (31)

-71 -57 -6.3 - 5 9 log A ,

;mol/ 1

>

Figure 3. Absorption rates of NO in aqueous solutions of Fe"EDTA at pH 7 and 25 "C.

35

3.0

10

i . . . . . ;

2.88

3.08 T.]