Ind. Eng. Chem. Res. 1991,30, 1693-1699 Ruckenstein, E.; Vaidyanathan, A. S.; Youngquist, G. Y. Sorption by Solids with Bidisperse Pore Structures. Chem. Eng. Sci. 1971,26, 1305. Sahimi, M.; Tsotsis, T. T. A Percolation Model of Catalyst Deactivation by Site Coverage and Pore Blockage. J. Catal. 1985,96, 552. Sharratt, P. N.; Mann, R. Some Observations on the Variation of Tortuosity with Thiele Modulus and Pore Size Distribution. Chem. Eng. Sci. 1987,42, 1565. Smith, J. M. Chemical Engineering Kinetics, 3rd ed.; McGraw-Hill: New York, NY, 1981;pp 486-487. Spry, J. C.; Sawyer, W. H. Configurational Diffusion Effects in Catalytic Demetallization. Paper presented at the 68th Annual AIChE meeting, Los Angeles, Nov 16-20, 1975;paper 30c. Stewart, W. E.; Johnson, M. F. L. Pore Structure and Gaseous Diffusion in Solid Catalysts. J. Catal. 1965, 4, 248. Ternan, M.; Rackwood, R. H.; Buchanan, R. M.; Parsons, B. I. Preparation of High Porosity Catalysts. Can. J. Chem. Eng. 1982, 60, 33. Tischer, R. E. Preparation of Bimodal Aluminas and Molybdena/ Alumina Extrudates. J. Catal. 1981, 72, 255. Tischer, R. E.; Narain, N. K.; Stiegel, G. J.; Cillo, D. L. Large-Pore
1693
J. Catal. Ni-Mo/AloOl Catalysts for Coal-Liquids Upgrading. ._ 1985,95, 4bs:
Tsai, C. H. Restrictive Diffusion in Hydroprocessing Catalysts. Ph.D. Dissertation, University of Utah, 1990. Tsai, C. H.; Massoth,-F. E.; Lee,-S. Y.; Seader, J. D. Effects of Solvent and Solute Configuration on Restrictive Diffusion in Hydrotreating Catalysta. Ind. Eng. Chem. Res. 1991, 30, 22. Wakao. N.: Smith, J. M. Diffusion in Catalyst Pellets. Chem. Eng. Sci. '1962, 17, 825. Wakao, N.; Smith, J. M. Diffusion and Reaction in Porous Catalysts. Ind. Eng. Chem. Fundam. 1964.3. 123. Ware, R. i.; Wei, J. Catalytic Hydrodemetallation of Nickel Porphyrins 1. Porphyrin Structure and Reactivity. J. Catal. 1985, 93,loo. Weisz, P. B.; Schwartz, A. B. Diffusivity of Porous-Oxide-Gel-Derived Catalyst Particles. J. Catal. 1962, 1 , 399. Wheeler, A. Reaction Rates and Selectivity. Catalysis; Emmett, P. H., Ed.; Reinhold: New York, 1955;Vol. 11, Chapter 2.
Received for review February 25, 1991 Revised manuscript received May 21, 1991 Accepted May 28, 1991
Deuterium Tracer Study on the Effect of CO on the Selective Hydrogenation of Acetylene over Pd/A1203 Yeung H. P a r k and Geoffrey L. Price* Department of Chemical Engineering, Louisiana State University, Baton Rouge, Louisiana 70803
The effect of carbon monoxide on the selectivity to ethylene during the selective hydrogenation of acetylene over Pd/A1,03 has been studied by isotopic tracer techniques. CO enhanced the total amount of deuterium exchange during acetylene deuteration while reducing the hydrogenation rate and increasing the selectivity to ethylene. For ethylene deuteration, CO suppressed the deuterium exchange reaction without affecting the reaction rate. Acetylene results are consistant with the mechanistic argument that the probability of hydrogenating adsorbed vinyl decreases when CO is added, which is caused by the displacement of surface hydrogen. However, for ethylene hydrogenation, CO increases the probabilities of ethylene desorption and adsorbed ethyl hydrogenation, which is consistant with the interpretation that CO displaces ethylene. For industrial processes for removal of trace acetylene from ethylene, it is suggested that the selectivity increase induced by CO is largely due to CO displacement of ethylene.
Introduction The selective hydrogenation of acetylene is a commercial process for removal of traces of acetylene from ethylene streams. The high selectivity of Pd catalysts for this reaction is the key to the success of this process, so understanding the nature of the selective behavior of Pd is important in improving the process. Previous studies have shown that carbon monoxide may be added to the process stream to increase the selectivity of the catalyst, but the mechanism has not been well established. McGown et al. (1977,1978) suggested that the selectivity enhancement by carbon monoxide may be attributed to carbon monoxide adsorption in preference to ethylene on metal sites or to a reduction in the amount of ethylene that tends to a,a adsorb. On the other hand, Al-Ammar and Webb (1978a,b, 1979) found that CO prevents the dissociative adsorption of hydrocarbons but does not affect a secondary type of adsorption, which they consider to be relevant to hydrogenation. Therefore, blockage of the hydrogen sites was considered to be responsible for the selectivity increase. Leviness et al. (1984) and Weiss et al. (1984) suggested from reaction studies that ethylene hydrogenation takes place mainly on the support and that CO displaces hy-
* T o whom correspondence should be addressed.
drogen on single, geometrically isolated sites that are left unoccupied after acetylene adsorption, resulting in a selectivity increase. In this paper, the effect of CO on the selective hydrogenation of acetylene has been studied by using isotopic tracer methods. The effect of CO on the product distribution in the deuteration of acetylene and ethylene has been assessed with the method developed by Kemball (1956), and the interaction of CO with ethylene was investigated by displacement experimenb. Experimental Section A. Materials. A 1% Pd/A1203catalyst was prepared by impregnation of Pd(NH3)4(NOJ2 on y-alumina (Linde 503). A 0.2M Pd solution was added to dried alumina with just enough distilled water to make a slurry. Water was boiled off on a hot plate, and then the catalyst was dried at 393 K overnight and calcined a t 573 K for 4 h in a muffle furnace. Pd dispersion of the catalyst (after H2 reduction) was determined by chemisorption of carbon monoxide in a static adsorption apparatus and found to be 45.6% Hydrogen (99.999%) and helium (99.999%)were each passed through Supelpure (Supelco) purifier to remove trace oxygen. Acetylene (99.6%) was purified by bulbto-bulb distillation using liquid nitrogen and stored in a
0888-5885/91/2630-l693$02.50/00 1991 American Chemical Society
1694 Ind. Eng. Chem. Res., Vol. 30, No. 8, 1991
RECIRCUUTION PUMP
r CC OETECTOR
1c.l FLOW MODULE
I
MS I
I I d l
I
r-
41
GC OVEN
Figure 1. Experimental apparatus.
storage bulb connected to the batch reactor apparatus. Ethylene, carbon monoxide, and deuterium gases (CP grade) were used directly from the lecture bottle. B. Analytical Details. Reaction products were analyzed with a versatile mass spectrometry system. The mass spectrometer was a UTI lOOC quadrupole with 2-200 AMU mass range, electron impact ion source operating at 70 ev, electron multiplier detector, and turbo molecular pumped vacuum system. A jet separator roughed with an independent mechanical vacuum pump provided the interface to atmospheric pressure. A cross-pattern leak valve was used to split off carrier flow rates in exceess of 30 cm3/min. The jet separator removed the bulk of the H2 and He from the carrier gases. The mass spectrometer could be operated as a rough GC/MS system by switching a chromatographic column in series with the jet separator by means of a six-port chromatographic sampling valve. For CzH2,C2H4, and C,H, separation, a 2-m Poropak NS column thermostated at room temperature was used. C. Apparatus. The primary apparatus for this study, though physically a single unit, can be arranged in two basic configurations. One configuration is a batch recirculation reactor for deuteration experiments, and the other, a flow system for displacement experiments (Figure 1). The batch reactor system (690 mL) is made of Pyrex glass with a U-tube reactor containing a catalyst bed. Reactants were circulated through the catalyst bed by means of a magnetically operated piston pump with check-valve arrangement. The entire system could be evacuated by means of a rotary mechanical pump. System pressures were monitored with digital-readout diaphram type pressure gauges. A three-way stopcock arrangement allowed the reactor to be isolated from the remainder of the system during reactant mixing. Reactor sampling was accomplished by evacuating a sample loop on a six-port chromatographic sampling valve and expanding the reaction products into the sampling loop. Switching the sampling valve diverted the sample into the chromatographic column followed by the mass spectrometer. In flow configuration, a three-way-stopcock arrangement allowed the reactor to be attached either to the recirculating batch reactor for catalyst pretreatment and reactant adsorption or to a continuous flow of He carrier gas for the displacement experiments. The effluent from the reactor could be directed through the chromatographic column to the MS, or it could bypass the column and go directly to the MS.
The reactor temperature was controlled by an electric furnace or a heating-cooling block and digital temperature controller. The heating-cooling block, made of copper, was used for low-temperature reaction or adsorption, and it has rod type heaters imbedded in it for heating and a copper cooling coil surrounding it. Compressed air, cooled with liquefied nitrogen, was used as the cooling medium. He carrier gases could be independently controlled and monitored with mass flow controllers and digital readout. D. Procedures. 1. Deuteration Reactions. For deuteration experiments, 100 mg (acetylene reaction) or 10 mg (ethylene reaction) of Pd/Al2O8 (40-60 mesh particles) was packed in the reactor and reduced at 773 K for 1 h in the recirculation system filled with 53 kPa H2(or D2)and 53 kPa He with a liquid nitrogen trap in the circulating loop. The catalyst was reduced for 2 h if CO had been adsorbed in a previous experiment. When the acetylene deuteration reaction was observed, 1.3 kPa C2H2,1.3 kPa D2, and 103 kPa He were mixed and the reaction was initiated and followed at 323 K. CO was used to modify these base reaction conditions by either (1) adding additionally 0.13 kPa CO along with the initial C2H2/D2/Hemixture (comixed) or (2) preadsorbing CO by circulating 0.13 kPa CO in 1.3 kPa He over the catalyst at 323 K. In this preadsorption step, the CO pressure corresponds to 9.2 CO molecules/Pd atom in the catalyst so that only a small amount of the CO actually adsorbs on the catalyst. Following preadsorption of CO, the reactor section was bypassed, but the unadsorbed CO was not evacuated. Instead, 1.3 kPa C2H2,1.3 kPa D2,and an additional 102 kPa He were mixed with the remaining CO, and then the catalyst was brought back into the circulating loop at 323 K to initiate the reaction. Samples of reaction product were periodically expanded into a sampling loop and injected into the chromatographic column for separation followed by analysis on the MS. When ethylene deuteration was observed, 13 kPa C2H4, 13 kPa D2,and 80 kPa He were mixed and the reaction was initiated and followed a t 223 K. As in the acetylene deuteration reaction, CO was either added to the base reaction mixture (comixed) or preadsorbed on the catalyst. Procedures were the same as those described above except that the CO pressure was 13.3 Pa in the CO/C2H4/D2/He mixture while 1.3 Pa CO was used in the preadsorption experiment. Note that these CO pressures are 2-3 orders of magnitude lower than those used in the CzHzreaction. The 13.3 Pa CO corresponds to 0.9 CO/Pd, and 1.3 Pa CO corresponds to 0.1 CO/Pd. 2. Adsorption Study. In order to test whether CO can displace C2H4 by competitative adsorption, a displacement experiment was performed. For this experiment, 100 mg of 200-325 mesh catalyst particles were initially reduced in the same way as the deuteration reactions with the system in circulation mode. After reduction, the catalyst was flushed with He gas at 773 K for 30 min to remove adsorbed H,, and it was then cooled to the adsorption temperature (173 K)under He flow. A mixture of 79 Pa C2H4in 105 kPa He was circulated over the catalyst at 173 K for 6 min, and then the apparatus was switched to flow mode and the catalyst was flushed with He for 6 min to remove weakly bound C2Hk CO was then injected onto the catalyst as successive pulses, and the effluent from the reactor was continuously monitored by the mass spectrometer. Calculations Two types of calculations outside the scope of simple procedures were performed. First, a technique based upon the original description by Kemball(1956) using Kemball’s
Ind. Eng. Chem. Res., Vol. 30,No. 8, 1991 1695 Table I. Effect of CO on Acetylene Deuteration (1.3 kPa CrH,, 1.3 kPa D,, 323 K, 100 mg of Catalyst) ethylene D distribution parameters* do dl d2 d3 db Ma P Q s Run 1 (Without CO) experimental 0.0 0.151 0.636 0.180 0.033 2.095 calculated* 0.0 0.152 0.636 0.182 0.030 0.59 1.00 0.80 experimental calculated*
0.0 0.0
0.202 0.202
experimental calculated*
0.0 0.0
0.187 0.188
run 1 run 2 run 3
do 0.744 0.699 0.726
Run 2 (With 0.13 P a CO, Comixed, CO/Pd = 9.2) 0.569 0.191 0.038 2.065 0.569 0.191 0.038 0.54 1.00
conversion/ %
SclHIC
53.6
90.0
59.8
93.1
0.71
Run 3 (With 0.13 P a CO, Preadsorbed, CO/Pd = 9.2) 0.597 0.178 0.038 2.067 48.5 96.7 0.598 0.182 0.032 0.57 1.00 0.74 calculated deuterium distribution* acetylene vinyl (surface species) dl d2 Ma do dl d2 d3 Ma 0.002 0.742 0.220 0.038 1.296 0.218 0.038 0.294 0.248 0.054 0.356 0.0 0.699 0.248 0.054 1.357 0.230 0.044 0.318 0.002 0.725 0.228 0.044 1.313
OAverage D atoms/molecule. *From Kemball's method (see Table 111). c S ~ z=b Ycz~,/(Yc2~ + YcZ%).
parameters has been adapted for computer calculations. We have derived the 8 H/D balance equations based upon the 3 parameters p, q, and s and 1 overall H/D balance equation for acetylene deuteration and 18 H/D balance equations with 4 parameters p, q, r, and s for ethylene deuteration. Since both of these mathematical systems have more equations and constraints than unknown parameters, we have set up a least-squares objective function based upon the sum of squares of the difference between the observed deuterium distribution and the computed distribution. The objective function was then minimized by using a standard nonlinear optimization package. Second, in order to obtain accurate deuterium distributions in CzHz,CzHr,and C,H, from MS measurements, we have used experimentally determined fragmentation patterns for each of the possible isotopic species taken from the literature (Dibeler and Mohler, 1954 Amenomiya and Pottie, 1968;Mohler and Dibeler, 1947). For C2H2there are 3 isotopic species (do, d,, and d z ) ,while for C2H4and CzHBthere are 5 and 7 species, respectively. We have monitored 5 masses (24-28 AMU) for C2H2, 8 masses (25-32 AMU) for CzH4, and 12 masses (25-36 AMU) for C2He We have set up a least-squares objective function from the sum of squares of the difference between the observed and computed spectra and minimized the function with nonnegativity constraints on the mole fraction of each deuterated species by using a standard nonlinear optimization package.
Rssults A. Acetylene Deuteration. Experiments were performed without CO, with CO mixed with reactants (comixed), and with CO contacted with the catalyst prior to introduction of other components (preadsorbed). Conversion versus time for each of these c a m is given in Figure 2,while the selectivity to ethylene for each case is given in Figure 3. We observed a strong effect in overall reaction rate when CO was added, but there is little difference whether CO is comixed or preadsorbed on the catalyst. CzH4 selectivity is not a strong function of conversion, as in Figure 3,but CO does appear to modify the selectivity and the order of CO addition (comixed or preadsorbed) has a minor but observable effect on the selectivity. Samples depicted in Figures 2 and 3 have also been subjected to deuterium content analysis, and the ethylene distributions are detailed in Table I. We have found that the deuterium product distribution is virtually independ-
403010-
0
100
200
600 700
300 400 500 Time (minutes)
Figure 2. CO effect on the rate of acetylene deuteration over 100 mg of Pd/A1203at 323 K. (P H )o = 1.3 P a ; (P& = 1.3 P a . Run 1, without C O run 2,with 0.13Ua CO comixed; run 3,with 0.13 P a CO preadsorbed. ...
D
Y
.
Run 3 \ -
-
1
95-
E. 90-
2
! "1i B
-
85
P
U
Run 1-
5 7
0 10 20 30 40 50 60 70 BO 90 100
Conversion (X)
Figure 3. Ethylene selectivity in acetylene deuteration over Pd/ A1203. Experimental conditions and run types as in Figure 2.
ent of conversion, so these values have been reported in the vicinity of 50% conversion. In all cases, a dominance of the d2 ethylene species was observed but broadening of the distribution at the expense of the dz intensity was notable when CO was comixed or preadsorbed. B. Ethylene Deuteration. Similar experiments were performed for the ethylene deuteration reaction. The rate of the ethylene reaction is several orders of magnitude faster than the acetylene reaction so that reaction conditions (most notably the reaction temperature) had to be adjusted to allow the observation of reasonable rates. Figure 4 compares the three cases, no CO, 13.3 Pa CO comixed, and 1.3 Pa CO preadsorbed. There is virtually no observable effect on the conversion versus time curve
1696 Ind. Eng. Chem. Res., Vol. 30, No. 8, 1991 Table 11. Effect of CO on Ethylene Deuteration (13 kPa C& ethylene dl d2 d3 d4 do
13 kPa D* at 223 K over Pd/Alumina (10 mg))
dl
Run 1 0.091 0.266 0.088 0.269
d2
ethane d3
d4
d6
0.264 0.271
de
Ma
experimental calculatedb
0.160 0.144
0.006 0.053
0.0 0.013
0.116 0.115
0.067 0.037
0.028 0.008
0.0 0.0
1.83
experimental calculated*
0.157 0.143
0.0 0.052
Run 2 (With 13.3 Pa CO, Comixed, CO/Pa = 0.9) 0.0 0.0 0.082 0.191 0.337 0.127 0.013 0.002 0.039 0.213 0.337 0.142
0.079 0.047
0.027 0.11
0.0 0.001
2.02
experimental calculatedb
0.202 0.202
0.0 0.047
0.0 0.008
0.51 0.027
0.0
0.0 0.0
1.86
0.004
0.0 0.002
Run 3 (With 1.3 Pa CO, Preadsorbed, CO/Pd = 0.1)
run 1 run 2 run 3
0.0 0.097 0.132 0.408 0.001 0.016 0.159 0.415 Kemball's parametersb
0.110 0.119
P
P
r
S
0.88 0.84 0.67
0.55 0.71 0.83
0.78 0.71 0.62
0.58 0.71 0.83
overall conversion/ % ' 13.9 13.1 14.0
aAverage number of deuterium atoms per molecule. bFrom Kemball's method (see Table 111).
fRun
-35w
T3025-
;
2015-
0
0
5
10
15
20
25
30
0
400
Figure 4. CO effect on the rate of ethylene deuteration over 10 mg of Pd/Al2O9 at 223 K. (Pc )o = 13 kPa; (P,), = 13 kPa. Run 1, without CO; run 2, with 13.!%a CO comixed; run 3, with 1.3 Pa CO preadsorbed.
for ethylene deuteration when CO is preadsorbed at 1.3 Pa. At 13.3 Pa comixed CO, however, we see a strong poisoning of the reaction at high conversions, but the initial rate is only slightly affected. The deuterium distributions of the ethylene reactant and ethane product are given in Table 11. The most abundant product species was dl ethane, which agrees with the result of Bond et al. (1966)at similar conditions. With added CO, the most abundant species changed to dz ethane and the distribution became narrower. Also note that 1.3 Pa of CO preadsorbed on the catalyst affected the breadth of the ethane distribution to a greater extent than 13.3 Pa of comixed CO without a significant change in the reaction rate (Figure 4). C. Ethylene Displacement Study. Displacement experiments show that CO adsorption can proceed on an ethylene-precovered surface (Figure 5). Ethylene and ethane products were observed when CO pulses were added repeatedly to the precovered surface.
Discussion A quantitative analysis of the product distribution of deuterated species from the acetylene or ethylene reaction is possible by a method developed by Kemball(l956) that is based on a general mechanism for hydrogenation. A brief description of the method is as follows. According to the general mechanism (Table 111),acetylene hydrogenation proceeds through adsorbed intermediates C2X2(a)and CzX3(a)where X = H or D. The C2X4(a)species is assumed to be replaced rapidly by
800
1200
1600
2000
Time (minutes)
Time (minutes)
Figure 5. Displacement of preadeorbed ethylene by CO pulses. The temperature for ethylene adsorption and CO displacement equals 193 K. The response of CO has been divided by 1OOO. Table 111. General Mechanism for Deuteration and Acetylene and Ethylene (Adapted from Bond and Wells, 1963)O Acetylene Deuteration Overall Reaction: X X CzXz F? C2X3(a) CzX4(a)(X = H or D)
-
-- ---
reaction X + C2X2(a) C2X3(a) C2X2(a)+ Hf C2X2H(a) C2X2(a)+ Dt C2XZD(a) C2X3(a) C2X2(a)+ X X + C2X3(a) C2X4(a) C2X3(a)+ Hf C2X3H(a) C2X3(a)+ D' C,X3D(a)
probability 1
1-9 9 1-P P 1-s S
Ethylene Deuteration Overall Reaction:
c2x4 2 C2X4(a)
X
s CzXs(a)
----
-
reaction C2X4(a) C2X, C2X4(a) C2X6(a) C2X4(a)+ H C2X4H(a) C2X4(a)+ D C2X4D(a) CzX6(a) C2X4(a) C2X&a) CzXe(a) C2X6(a)+ H C2X6H(a) C2X6(a)+ D C2X6D(a)
X
CzXs (X = H or D) probabilityb 1-P P 1-9 Q r 1-r 1-s S
"The source of atoms marked with a dagger is unspecified; they may come from an adsorbed atomic species or by disproportionation with other adsorbed hydrocarbons. bBond (Bond and Wells, 1963) originally referred to these values as p*, 9*, r*, and s* instead of p, 9, r, and s, respectively.
H Z ( g ) t 2: C2H2(g)
HCiCH t H
I
!
t 2HC-CH2
t
1.00
2H
t 2t
b 8
-
Ind. Eng. Chem. Res., Vol. 30, No. 8, 1991 1697 I
0.80
HCiCH
I b'b
HC-CH2
I
t
(3)
2t
9
H C i C H t H2C-CH2
I H2C=CH2
I
I
I
11
C2H,(g)
a
t
0 20-
'
(5)
t HCZCH?
I HC=CH2 t H
I
I
S
0.1Oj
___ 0 ,
aDiSprOpQrfionation of v i n y l : 21
0 30-
HCiCH t H
I
prob, l-p
(4a)
p r ~ b ,p
(4b)
1
t
b b
HzC-CH2
t
I
I
Figure 6. Reaction mechanism for acetylene hydrogenation over Pd/A1208(adapted from Bond (Bond and Wells, 1963) and a review by Webb (1978)).
gas-phase CzH2. Three isotopically distinguishable acetylenes and six isotopically distinguishable surface vinyls would be formed in acetylene deuteration as a result of these molecular interconversions, and in the steady-state treatment, the probabilities of these interconversions are defined as p , q, and s (refer to Table 111) and the H/D balances for each adsorbed species are set up with these parameters to result in a series of simultaneous equations. The nine simultaneous equations thus obtained may be solved for values of p , q, and s to obtain a satisfactory fit between the calculated and observed deuterium distributions. Bond and Wells (1966) solved the equations by using a graphical technique, but we have been able to set up a least-squares minimization routine to solve these equations. In ethylene deuteration, similar relationships are established by using four probability parameters p , q, r, and s, which we have also obtained by least-squares minimization. The relative amounts of the various surface species may also be obtained by this calculation. Table I contains a synopsis of these calculations for the acetylene deuteration reaction with and without CO. 'Talculated and experimental distributions show excellent agreement and computed parameters p , q, and s are similar to parameters computed by Bond and Wells (1966). The computed parameters p , q, and s did not show a trend with conversion for the three cases (no CO, comixed, and preadsorbed). Only minor effects are noted as a function of conversion and the presence of CO, and the contacting method has only a slight effect. Our primary consideration is how CO affects the distribution of surface species and the ultimate mode by which the rates of reaction and the Selectivityto ethylene are affected by CO. Clearly, the overall deuterium content of acetylene goes up when CO is added (see calculated values for M in Table I; experimental values are similar), and there is a corresponding increase in the deuterium content of the adsorbed vinyl species (Table I). However, there is simultaneously a drop in the deuterium content of the ethylene product when CO is added (Table I). These observations may be explained as follows. According to Bond's mechanism (Bond and Wells, 1963; and a review by Webb, 1978) for acetylene hydrogenation over Pd (see also Figure 6 ) , hydrogen (deuterium) adsorption is irreversible and ethylene formation proceeds primarily by vinyl disproportionation (reaction 4, Figure 6), suggesting that the hydrogen which is formed during vinyl disproportionation is not desorbed. Notice that when D2 is used as the gas-phase reactant, this disproportionation provides a source of H atoms which is the reason that the ethylene
I
10
20
I
.
30 40 Conversion (x)
50
60
Figure 7. Kemball's parameters in ethylene deuteration over Pd/ A120,. Experimental conditions a~ in Figure 4.
product is not exclusively d2 ethylene. If hydrogen adsorption is irreversible as Bond (Bond and Wells, 1963) suggests, hydrogen atoms liberated by vinyl reversal must either combine with acetylene or vinyl. If hydrogen adsorption was readily reversible, we would expect to achieve average ethylene deuterium levels greater than 2, which is not observed. When CO is added (either before or during the reaction), we have obtained a decline in the p value so vinyl reversal (reaction 4a in Figure 6, probability 1- p ) is enhanced over vinyl hydrogenation (reaction 4b in Figure 6, probability p ) . This may be accomplished if adsorbed hydrogen produced by vinyl disproportionation is displayced by CO. However, the fact that the reversal reaction is favored ensures a higher population of H (which comes from the hydrogen-rich radicals) than D, and we expect a decline in the overall deuterium content of the ethylene as observed. This postulate is also consistent with the observation that s, which is the probability of incorporating deuterium in vinyl hydrogenation, decreases, which has also been observed (Table I). One problem we have encountered is that we would like to be able to distinguish whether the primary effect of CO addition on the selectivity to ethylene during acetylene hydrogenation is to reduce the amount of adsorbed hydrogen or the reactivity or amount of adsorbed ethylene. Though these results show that CO reduces surface hydrogen, the acetylene hydrogenatin mechanism by Bond (Bond and Wells, 1963; and a review by Webb, 1978) does not allow us to make a direct observation regarding possible effects due to ethylene displacement or the blockage of ethylene reactivity because one of the fundamental assumptions in the mechanism is that desorption of adsorbed ethylene is certain (probability = 1.0) and that further reactions of ethylene are impossible. This requires the selectivity to ethylene to be loo%, which of course is not quite true. Therefore, the relative effects on adsorbed hydrogen and ethylene cannot be deduced by using this mechanism. We have considered extending Bond's theory to allow a small probability of ethylene hydrogenation, but the equations become very complex, and we have been unable to accurately measure the ethane deuterium distribution because it arises in such low concentrations (the selectivity to ethylene is indeed high). Therefore, we have turned to measuring the effect of CO on the hydrogenation of ethylene directly and hope to be able to use these data in further discussions regarding the selective hydrogenation of the acetylene. The calculation results for ethylene deuteration are given in Table 11. Unlike the case for acetylene deuteration, Kemball's parameters (extracted from both the ethane and ethylene deuterium distributions) are strongly dependent upon the degree of conversion, as in Figure 7. Deuterium distributions are difficult to obtain in the low conversion
1698 Ind. Eng. Chem. Res., Vol. 30, No. 8, 1991
region because the amount of ethane is not large enough to be analyzed accurately. High conversions of ethylene tend to give randomly scrambled products so that all reactions extrapolate toward the same parameters. Therefore, we have chosen to compare parameters near 14% conversion, which is the lowest practical conversion for which reliable data could be extracted. In ethylene deuteration, the calculated product distribution also shows reasonable agreement with the experimental distributions. The values of Kemball’s parameters showed large changes with CO, reflecting the corresponding changes in the product distribution. We also observe that preadsorbed CO has the greatest effect on Kemball’s parameters, even at the lowest partial pressure of CO. Table I1 shows that the values of p and r dropped with addition of CO, while those of q and s increased. These results are equivalent to an increase in the probability of desorption of ethylene, 1+ p , a decrease in the probability of the reversal of ethyl, r, and increases in the probabilities of deuterium incorporation, q and s. In other words, when CO is added, ethylene desorption is favored, ethyl reversal is suppressed, and hydrogenation depends less on surface hydrogen, which comes from the hydrogen-rich adsorbed hydrocarbon species. With a decrease in adsorbed ethylene, further deuterium exchange between ethylene and ethyl would decrease. With the decrease of ethyl reversal, the amount of surface hydrogen would decrease and the deuteration reaction would depend more on deuterium from the gas phase. Therefore, the product from the deuteration of ethylene, dz ethane, is expected to be predominant and the deuterium distribution in ethane should become narrower, which agrees with the observed changes in the experimental deuterium distribution. The observation that the probability of ethylene desorption is enhanced in the presence of CO can be explained by two possible alternatives. The first possibility is direct displacement of adsorbed ethylene by CO during the reaction, and the second possibility is a decrease in the availability of surface hydrogen for the conversion of adsorbed ethylene in which case it would be more likely that ethylene desorbs rather than reacts. If this latter case is the main reason that CO enhances ethylene desorption, we should observe a parallel decrease in the probability of ethyl hydrogenation, 1- r, since the source of hydrogen in the two additional steps are the same. (The parameters q and s, which represent the source of hydrogen, were similar in each case, as seen in Table II). But the observed result is the opposite; an increase in the probability of ethyl hydrogenation was observed. Furthermore, the observed overall ethylene deuterium distribution cannot be explained by the hydrogen displacement theory. The ethylene deuterium distribution depends on the relative rates of the deuterium exchange reaction and the direct hydrogenation (deuteration) reaction. If the surface concentration of hydrogen is reduced, the adsorbed hydrocarbon intermediate CzXs(a) would tend to undergo the reverse reaction rather than the forward hydrogenation reaction so that the degree of deuterium exchange would increase. This is similar to the observations already made for acetylene deuteration. However, for ethylene deuteration, CO caused a decrease in the overall deuterium content of ethylene. Therefore, it is concluded that direct displacement of ethylene by CO is the main reason for the reduction in surface ethylene. The ethylene displacement experiment showed directly that CO can displace ethylene preadsorbed on Pd (Figure 5), and therefore the alternative that CO directly displaces adsorbed ethylene is feasible.
Our results also show that preadsorbed CO is more effective in altering the ethane deuterium distribution than comixed CO, even though a smaller amount of preadsorbed CO was used. The relatively small amount of preadsorbed CO (0.1 CO/Pd), however, did not significantly alter the hydrogenation rate. These observations are easily explained. Preadsorbed CO (0.1 CO/Pd) simply dilutes the surface concentration of ethylene. Since the hydrogenation reaction involves one surface hydrocarbon species and surface hydrogen, we expect a reduction in surface hydrocarbons to linearly affect the rate of the hydrogenation reaction. Exchange occurs primarily via disproportination of two surface hydrocarbons, so we expect a greater effect on the exchange reaction as observed. In the preadsorption case, CO does not have to compete with ethylene, while, in the comixed case, it must displace ethylene. In Figure 4,we see that the initial rate of the ethylene reaction is similar in all cases. However, the rate of the reaction falls rapidly in the comixed case as CO coverage increases via successful competition with ethylene for surface sites. With these results in hand, we turn our attention to the application of these theories to the selective removal of acetylene from raw ethylene streams and the effect of CO under conditions nearer to the actual conditions used industrially. First, our studies have been performed at subambient temperatures, while real processes operate at about 333-353 K. However, Bond (Bond et al., 1966) has shown that the effect of temperature on both the acetylene and ethylene hydrogenation reactions are similar. Higher temperatures generally promote conversion of acetylene to vinyl or ethylene to ethyl rather than cause a complete shift in reaction mechanism. Therefore, we cautiously suggest that the results we have obtained are roughly applicable to industrial temperatures. Also, we must consider that we have studied the acetylene reaction at high partial pressures of acetylene compared to industrial conditions where the actual process stream might be of the order 0.35% acetylene, 0.40% Ha, and the balance ethylene with the reaction proceeding nominally at atmospheric pressure. Under these conditions, nearer the actual conditions, ethylene may successfully compete with acetylene for Pd sites (Guczi et al., 1979), while our conditions virtually exclude this possibility. Now consider the overall conversion processes that probably occur under real conditions:
C2H&)
2H
C2Hda)
To form C2H6 from CzH2,the ethylene intermediate that is produced via single hydrogenation of acetylene undergoes a second hydrogenation, or acetylene is hydrogenated after a single adsorption to CJ&. Another source of CzH6is via hydrogenation of ethylene, which competes for Pd sites. McGown et al. (1977, 1978) suggested that CzH6comes mainly from gas-phase C2H4 when studying the ethylene-rich mixture, 14% C2H2,57% C2H4,29% Hp However, others such as Guczi et al. (1979) and Al-Ammar and Webb (1979) argue that the main route to CzH6 is direct hydrogenation of C2Hzlbut their experiments were done with acetylene-rich mixtures. Another proposed route to C2H6 formulation was given by Sarkany et al. (1984) as C2H4 hydrogenation on support sites of aged catalysts. Our results suggest that CO displaces ethylene or blocks adsorption of ethylene, which favors the theory that the primary effect of CO is to slow the formation of
Ind. Eng. Chem. Res., Vol. 30,No. 8, 1991 1699
C2H, by reducing adsorbed ethylene, not by halting direct hydrogenation of acetylene to ethane through the ethylidyne species. Conversely, Al-Ammar and Webb (1978a,b, 1979) suggested the primary effect of CO on ethylene hydrogenation is to block hydrogen adsorption sites. They postulated that ethylene adsorbs dissociatively on metal sites and molecularly on other sites, possibly carbonaceous overlayers, as a secondary adsorption. This secondary adsorption was thought to be involved in hydrogenation and was found to be independent of CO preadsorption. However, our CzHl displacement experiments show that adsorbed ethylene can be displaced as ethylene by CO. This observation indicates that ethylene adsorbs on the metal surface associatively to some extent, if we assume that CO adsorbs only on metal sites. Our deuterium tracer experiments also showed that carbon monoxide displaces ethylene that is directly involved in the hydrogenation, and these observations cannot be explained with the secondary adsorption theory. An effect on hydrogen adsorption sites was also suggested in the study by Leviness et al. (1984) and Weiss et al. (1984), who argued that CO effectively competes with hydrogen atoms for ethylene hydrogenation, while reducing the rate of acetylene hydrogenation. The possibility that CO blocks ethylene adsorption was excluded since the primary sites for ethylene adsorption during the selective hydrogenation of acetylene was thought to be support sites that are activated by the accumulation of surface polymer. We cannot make a direct comparison of our results with this study since we have studied fresh catalysts. Acknowledgment We appreciate the financial support of the Exxon Foundation.
Literature Cited Al-Ammar, A. S.; Webb, G. Hydrogenation of Acetylene over Supported Metal Catalysts Part 1.-Adsorption of [ l e ] Acetylene and [“C]Ethylene on Silica Supported Rhodium, Iridium and Palladium and Alumina Supported Palladium. J. Chem. SOC.Faraday Trans. I 1978a, 74, 195. AI-Ammar, A. S.; Webb, G. Hydrogenation of Acetylene over Supported Metal Catalysts Part 2.-[W] Tracer Study of Deactivation
Phenomena. J. Chem. SOC.,Faraday Trans. 1 , 1978b,74,657. Al-Ammar, A. S.; Webb, G. Hydrogenation of Acetylene over Supported Metal Catalysts Part 3.-[“C] Tracer Studies of the Effects of Added Ethylene and Carbon Monoxide on the Reaction Catalyzed by Silica-Supported Palladium, Rhodium and Iridium, J. Chem. SOC.,Faraday Trans. 1 1979,75, 1900. Amenomiya, Y.;Pottie, R. F. Mass Spectra of Some Deuterated Ethanes. I. The Effect of Ionizing Volgate. Can. J. Chem. 1968, 76,1735. Bond, G. C.; Wells, P. B. Mechanism of the Hydrogenation of Unsaturated Hydrocarbons on Transition Metal catalysts. Adv. Catal. 1963,15,91. Bond, G. C.; Wells, P. B. Hydrogenation of Acetylene IV. The Reaction of Acetylene with Deuterium Catalyzed by Alumina-Supported Rhodium, Palladium, Iridium, and Platinum. J. Catal. 1966,6,397. Bond, G. C.; Philipson, J. J.; Wells, P. B.; Winterbottom, J. M. Hydrogenation of Olefins Part 3.-Reaction of Ethylene and Propylene with Deuterium over Alumina-Supported Palladium and Rhodium. Trans. Faraday SOC.1966,62,443. Dibeler, V. H.; Mohler, F. L.; deHemptinne, M. Mass Spectra of the Deuteroethylenes. J. Res. Bur. Stand. 1954,53, 107. Guczi, L.; LaPierre, R. B.; Weiss, A. H.; Biron, E. Acetylene Deuteration in the Presence of [“C] Ethylene. J. Catal. 1979,60,83. Kemball, C. The Deuteration and Exchange of Ethylene on Evaporated Metal Catalysts at Low Temperatures. J. Chem. SOC.1956, 735. Leviness, S.; Nair, V.; Weiss, A. H.; Schay, Z.; Guczi, L. Acetylene Hydrogenation Selectivity Control on PdCu/Alz03 Catalysts. J. Mol. Catal. 1984,25,131. McGown, W. T.; Kemball, C.; Whan, D. A.; Scurrel, M. S. Hydrogenation of Acetylene in Excess Ethylene on an Alumina Supported Palladium Catalyst in a Static System. J. Chem. SOC., Faraday Trans. 1 1977,73,632. McGown, W. T.; Kemball, C.; Whan, D. A. Hydrogenation of Acetylene in Excess Ethylene on an Alumina-Supported Palladium Catalyst at Atmospheric Pressure in a Spinning Basket Reactor. J. Catal. 1978,51,173. Mohler, F.L.; Dibeler, V. H. Mass Spectrometer Analysis of C2H2, C2D2and CzDH. Phys. Rev. 1947, 72, 158. Sarkany, A.; Guczi, L.; Weiss, A. H. On the Aging Phenomenon in Palladium Catalyzed Acetylene Hydrogenation. Appl. Catal. 1984,10, 369. Webb, G. Catalytic Hydrogenation. In Comprehensive Chemical Kinetics: Bamford, C. H., Tipper, C. F. H., Eds.; Elsevier: Amsterdam, 1978; Vol. 20. Weiss, A. H.; Leviness, S.; Nair, V.; Guczi, L.; Sarkany, A.; Schay, Z. The Effect of Pd Dispersion in Acetylene Selective Hydrogenation. Int. Congr. Catal. 1984,8,591.
Receiued for reoiew November 19, 1990 Accepted March 12, 1991
