Znd. Eng. Chem. R e s . 1987,26, 1901-1905 Kiviat, F. E.; Petrakis, L. J . Phys. Chem. 1973, 77, 1232. Kovach, S. M.; Castle, L. J.; Bennett, J. V. Znd. Eng. Chem. Prod. Res. Dev. 1978, 17, 62. Mieville, R. L.; Meyers, B. L. J . Catal. 1982, 74, 196. Muralidhar, G.; Massoth, F. E.; Shabtai, J. J . Catal. 1984, 85, 44. Nelson, H. C.; Lussier, R. J.; Still, M. E. Appl. Catal. 1983, 7, 113. Ocampo, A.; Schrodt, J. T.; Kovach, S. M. Ind. Eng. Chem. Prod. Res. Deu. 1978, 17, 56. Ratnasamy, P.; Knozinger, H. J . Catal. 1978, 54, 155. Ratnasamy, P.; Sharma, D. K.; Sharma, L. D. J . Phys. Chem. 1974, 78, 2069. Satterfield, C. N. AZChE J . 1975, 21, 209. Scaroni, A. W.; Jenkins, R. G.; Utrilla, J. R.; Walker, P. L., Jr. Fuel Process. Tech. 1984, 9. 103.
1901
Segawa, K.; Hall, W. K. J. Catal. 1982, 16, 133. Stohl, F. Y.; Stephens, H. P. Proc. 10th Ann. EPRZ Contractor's Conf. Coal Liquef. 1985, 1. Spencer, D. EPRZ J . 1987,12(1), 40. Tanabe, K. Solid Acids and Bases; Academic: New York, 1970; p
125.
Ternan, M.; Brown, J. R. Fuel 1982, 61, 1110. Thomas C. L. Catalytic Processes and Proven Catalysts; Academic: New York, 1970; p 168. Trimm, D. L. Appl. Catal. 1983,5, 263.
Received for review December 29, 1986 Revised manuscript received May 27, 1987 Accepted June 13, 1987
Conversion of CH4 into C2H2and C2H4 by the Chlorine-Catalyzed Oxidative-Pyrolysis (CCOP) Process. 1. Oxidative Pyrolysis of CH3Cl A. Granada, S. B. Karra, and S. M. Senkan" Department of Chemical Engineering, Illinois Institute of Technology, Chicago, Illinois 60616
The oxidative pyrolysis of CH3C1, representing the second stage in the chlorine-catalyzed oxidative pyrolysis (CCOP) of CH4,was studied in a flow reactor at about 980 "C and 0.68 atm. The presence of oxygen in the system decreased the extent of formation of carbonaceous deposits considerably without the formation of destructive flames. Under the reaction conditions studied, the formation of C2H2and CzH4with combined yields as high as 60% was observed, a t about 30% conversion of CH3C1. Conversion of methane into higher molecular weight hydrocarbons is of immense practical significance. Methane is available in large quantities in natural gas, thus constituting an important raw material for the synthesis of higher molecular weight hydrocarbons. Processes exist to convert methane into acetylene, ethylene, and hydrogen by using high-temperature pyrolysis. However, at the high temperatures needed for the thermal decomposition of methane, the yields of more valuable liquid and gaseous products are too low due to the formation of excessive amounts of carbonaceous deposits (see for example Back and Back (1983) and references therein). In an earlier patent, Gorin (1943) described a two-step chlorine-catalyzed process for the polymerization of methane. According to this process, methane is chlorinated first, forming chlorinated methanes (CM), followed by the pyrolysis of CM and formation of C2 and C2+ hydrocarbons and HC1 in the second step. The HC1 produced can either be converted into chlorine via the well-known Deacon reaction and recycled or can be used to oxychlorinate methane to form CMs, thus completing the catalytic cycle for chlorine. Recently, Benson (1980) patented a theoretical single-step process which involves the ignition of C12-CH4mixtures and formation of flames. Later, Weissman and Benson (1984) studied the kinetics of nonflame pyrolysis of CH3CI. As shown in these studies, although the decomposition temperatures of chlorinated methanes are considerably lower than those for methane, thus the destruction rates of useful products are lower, the formation of carbonaceous deposits, which include high molecular weight low vapor pressure hydrocarbons, tars, carbon, and soot, is still a
* To whom
correspondence should be addressed.
problem (Gorin, 1943; Weissman and Benson, 1984). This renders the direct pyrolysis of CMs unattractive for practical applications. It must be recognized that although the formation of carbonaceous deposits may represent a minor route, it has major consequences. Because of the accumulative nature of the deposits, the reactor and transportation lines ultimately plug up, and this results in major process inefficiencies. In this paper, we report on the results of experimental studies of the oxidative pyrolysis of CH3C1,representing the second stage in the chlorine-catalyzed oxidative pyrolysis (CCOP) of CHI developed recently (Senkan, 1987). Product distributions measured both in the presence and absence of oxygen clearly show that O2 effectively and efficiently ameliorates the problem of formation of high molecular weight carbonaceous deposits and allows the production of C2H2and C2H4 with high yields. A plausible reaction mechanism for the oxidative pyrolysis of CH3C1 is also discussed. Experimental Section The experimental facility used is shown in Figure 1. The reactor was a 2.1-cm-i.d. by 100-cm-long quartz tube, of which about 60 cm was placed in a three-zone Lindbergh furnace. The first zone of the furnace, which was about 15 cm long, was used to preheat the argon carrier gas. Small amounts of CH3Cl and CH3C1/ O2mixtures were introduced into the preheated argon carrier gas by using an air-cooled probe through radially directed injection holes. The amounts of gases injected were deliberately kept small to ensure the rapid heat-up of the reactants and to preserve the near-isothermal conditions during the experiments.
0888-5885/87/2626-1901$01.50/0 0 1987 American Chemical Society
1902 Ind. Eng. Chem. Res., Vol. 26, No. 9, 1987 Table I. Experimental Conditions Investigated 7' = 980 "C, P = 0.678 atm, u = 150 cm/s mean residence time range = 50-250 ms mole % species mixture A mixture B CH,Cl 7.32 7.47 0 2 2.05 Ar 90.6 92.5 1000
--
t
m 7 ID
950 c ID
900
Figure 1. Schematics of the experimental system
Temperature measurements were made by using chromel/alumel thermocouples, and they indicated the presence of near isothermal conditions during the experiments. Gases used were acquired from the Matheson Co. (Joliet, IL) and had the following reported purities: Ar, 99.999%; CH3C1,99.5% as liquid; and 02,99.6% extra dry. They were used directly from the cylinders, and their flow rates were regulated by the combined use of rotameters and needle valves. The needle valves were maintained under critical flow conditions to establish uncoupled flow rates. A mechanical vacuum pump was used to remove the carrier gas and the reaction products from the system. Mean gas flow velocities in the reactor were in the range 1-10 m/s, suggesting that laminar flow conditions were present. However, the deviation from ideal plug flow behavior would be in the range 10-15%, the same order of magnitude as the other experimental errors (Cathonnet et al., 1981). This was verified by the measurements of the species concentration profiles in the radial direction. Species concentrations were determined by withdrawing gases through a water-cooled quartz sampling probe positioned centrally a t the downstream end of the reactor, followed by gas analysis by on-line mass spectrometry. Details of the mass spectrometer system used have been presented previously (Chang et al., 1986) and thus will not be repeated here. Mass ions were generated by electron impact ionization, with electron energies of about 30 eV to minimize the fragmentation of species. Ion intensities at select mass numbers, corresponding to the species being monitored, were recorded and ratioed to the signal intensity for argon. Relative ion intensities subsequently were converted into absolute mole fractions via the use of calibration gas mixtures in a conventional way (Andersson et al., 1985). To minimize mass discrimination effects inherent in mass spectrometry, calibration gas mixtures were prepared with compositions similar to those present in the reactor during the experiments. We estimate our determination of the absolute mole fractions to be accurate within *15%. Species mole percent profiles were then obtained by moving the sampling probe along the reactor relative to the injection probe that was kept at a fixed position. Results and Discussion In this section, the product distributions for the oxidative pyrolysis of CH3Cl (mixture A) are presented, and they are compared to those obtained in the absence of oxygen (mixture B). The specific experimental conditions investigated are presented in Table I, under which the pyrolysis processes are expected to be dominated by gasphase kinetics, with minor contributions from surface-induced reactions (Weissman and Benson, 1984; Rotzoll, 1986).
15
28
25
38
35
Gistance a l o n g reactor, cm
Figure 2. Profiles for CH,C1 mole percent and temperature,
Before the quantitative results are discussed, however, a number of qualitative observations must be noted. First, in the CCOP process, no flame formation takes place because of the well-known flame inhibition characteristics of chlorine and chlorinated compounds (Chang and Senkan, 1985; Chang et al., 1987). Second, in the presence of O2 (mixture A), it is possible to conduct experiments for considerably long periods of time, i.e., 4-5 h, without the formation of any visible deposits at the exit of the transparent quartz reactor. In the absence of O2 (mixture B), however, dark carbonaceous deposits immediately form and render the quartz reactor opaque. Third, based on thermocouple measurements, the temperature levels in the reactor remain isothermal both in the presence and absence of oxygen. This suggests that both the pyrolysis and oxidative pyrolysis of CH&l are thermoneutral processes and that the flames are absent in the latter case. In all the experiments, the major species quantified, other than the reactants and argon, were C2H2,C2H4, C2H3C1,CH4,HC1, Hz,and CO. Minor species identified, but not quantified, were CsH6, HzO, COz, and HCHO. Trace levels of CH2Cl2also were noted in the absence of 02. In Figure 2 the mole percent profiles for CH3C1, both for mixtures A and B, are presented as a function of axial position along the reactor. As evident from Figure 2, the overall conversion of CH3Cl was about 32% for both mixtures studied. To ensure that the measurements represent intrinsic chemical kinetic changes that are free from mixing effects, data points are presented only beyond 15 cm from the point of injection of the reactants into the preheated argon carrier gas. As a visual guide, the data points have been connected by lines to indicate approximate trends. A representative temperature profile is also presented in Figure 2 to illustrate the extent of isother-
Ind. Eng. Chem. Res., Vol. 26, No. 9, 1987 1903 58
2.5
Mixture B
2.0
1.5 W c
L a a l
P
1.0
I
0.5
t I
I
I
25
30
35
I
a.a 15
20
25
30
35
20
15
Distance along r e a c t o r , cm
Distance along r e a c t o r , cm
Figure 3. Profiles for HCl, 02,and H2.
Ib
Figure 5. Yield profiles for C2H2,CZH4, CH4,and C2H3C1in mixture B.
I
Mixture A
28
c
0 O
"
0
8
15
28
25
30
35
Distance along r e a c t o r , cm
I
I
t- AL--J A
e l
15
I
I
A I A
1
28
25
38
35
Distance along r e a c t o r , cm
Figure 4. Yield profiles for C2H2,C2H4, CH,, C2H3C1,and CO in mixture A.
Figure 6. Yield profiles for UC in mixtures A and B.
mality present in the reactor. In Figure 3 the profiles for HCI, 02,and H2 are presented. For HC1, the mole percents were calculated from chlorine atom balances and by assuming that no chlorine is associated with the species that were not quantified by mass spectrometry. It is important to note that the conversion of O2 was quite low, less than about 10%. This supports the absence of flames and is also consistent with the formation of low levels of CO and the presence of only trace levels of C 0 2 and H20, which were not quantified. The mole percents for H2were calculated from hydrogen atom balances, since excessive mass discrimination in the mass spectrometer prevents its direct quantification. It is interesting to note that the amount of H2 calculated in the presence of oxygen was lower. This, however, is expected as oxygen promotes the formation of gaseous hydrocarbon products which are quantifiable. Consequently, the amount of Hz that would be unaccounted for, thus reported as H2, would be less in mixture A. In Figure 4, percent yield profiles for C2H2, C2H4, C2H3Cl,CH4,and CO are presented for mixture A. These profides were determined from the conversion of CH3C1and from the direct measurements of the concentrations of the
individual species. The corresponding yield profiles for mixture B are shown in Figure 5. As seen from the comparison of Figures 4 and 5, although the presence of oxygen in the system results in the formation of some CO, the yields for C2H4and C2Hzare minimally affected. In Figure 6 the calculated percent yield profiles for unaccounted carbon (UC) are presented for mixtures A and B. UC was defined as the percentage of carbon associated with CH3C1that was converted but which cannot be accounted for by the measurements of the major gaseous species noted above. According to this definition, it is reasonable to assume that UC should be proportional to the extent of formation of carbonaceous deposits. As evident from Figure 6, the yields for UC are significantly reduced in the presence of oxygen (mixture A), suggesting that the extent of formation of carbonaceous deposits is suppressed. Since the yields for C2H2,CzH4, C2H3C1,and CHI are nearly the same for both mixtures A and B (see Figures 4 and 5 ) , O2 appears to intercept some of the key hydrocarbon radical intermediates responsible for the further molecular weight growth and convert them into CO, thereby minimize the extent of formation of carbonaceous deposits. This is further supported by the data,
1904 Ind. Eng. Chem. Res., Vol. 26, No. 9, 1987
in which the difference in UC yield profiles between mixtures A and B are closely followed by the yield profile for CO (see Figure 4). Clearly, the uncertainties associated with UC are large due to the cumulative effects of the errors involved in the measurement of the individual species used in carbon balances. On the other hand, since UC profiles for both mixtures A and B are determined in a similar way, qualitative comparisons should be feasible. Nevertheless, the quantitative aspects of the process of formation and/or prevention of high molecular weight carbonaceous deposits should be made with caution. Quantification of the extent of formation of carbonaceous deposits will be the subject of a future investigation. A detailed chemical kinetic mechanism describing the oxidative pyrolysis of CH3C1is currently under development, and the results of this modeling work will be published shortly (Karra and Senkan, 1987a). However, our preliminary analysis of the CCOP process suggests that the pyrolysis of chloromethane starts with the well-established initiation reaction CHBCl+ CH3' C1' (1)
low, HC1 and C1 elimination channels would gain greater significance. An analysis of the chemically activated recombination reactions of CHzC1' and CH3' radicals using the quantized RRK (Rice-Ramspergw-Kassel) method will be presented elsewhere (Karra and Senkan, 198713). These radical combination reactions are then followed by the following, again pressure-dependent, unimolecular reactions leading to the formation of CzHzand C2H4: CzH3C1+ M * CzH2+ HC1+ M (13) C2H4C1' + M + C2H4 + C1' + M (14)
as well as by the route CH,Cl + O2 + CH2C1' + H02*
and forms one of the most important Cz radicals in the system, CzH3'. Similar destruction channels for CzHz would be too slow to be of any significance in the CCOP process. The primary reaction pathways available for C2H3' are its polymerization C2H3' + CzH2 + CHZCHCHCH' (19)
+
These reactions are followed by C1' + CH3Cl + CH,Cl'
(2)
+ HC1
(3) Once formed, HC1 undergoes the following fast reaction: CH3*+ HC1* CH4 + C1' (4)
regenerating C1' and forming CH4 as an inevitable byproduct of pyrolysis of CH3C1. Reaction 4 also rapidly consumes CH,', therefore rendering CH2C1' as the most important C1 radical in the system. The chemically activated recombination of CH2C1' as well as CHzC1' and CH,' then determines the major product distributions in the CCOP process. The reactions are CH2C1' + CH2C1' + [1,2-CzH4C12]* (5) CH3'
+ CHpCl' + [C2H,CI]*
(6)
where [ ]* denotes the chemically activated adduct. The
CH,' + CH3'=[C2H6]* reaction is likely to be unimportant because of the lower concentrations of the CH3' radicals in the CCOP process. The energized adducts [ 1,2-CzH4C12]*and [C2H5Cl]* then undergo the following parallel stabilization and decomposition reactions: M
[ 1,2-CzH4Cl2]*
1,2-CzH4Cl2 (stabilization) (7)
+ C1' [1,2-C2H4C12]* + C2H3C1 + HC1 [1,2-C2H,C12]* + C2H4C1' M
[C2H5C1]* CzH,C1 (stabilization) [C2H,C1]* + C2H4 + HC1 [CzH,Cl]* + CZH,'
+ C1'
(8)
(9)
(10) (11)
(12) As apparent from these reactions, gas density, indicated by the symbol M , has a significant impact on the nature of the product distributions in the CCOP process. For example, at higher pressures and/or lower temperatures where M is high, collisional stabilization of the chemically activated intermediates is enhanced; thus, the formation of recombination products would be favored. Conversely, at low pressures and/or higher temperatures where M is
1,2-CzH4C12+ M + C2H3Cl + HC1 + M
(15) CzH,Cl+ M + C2H4 + HCl + M (16) Reaction 13 is the major channel for the formation of CzHzand for the destruction of C2H3C1. The formation of CzH4 occurs primarily via reaction 11 and to a lesser extent by reactions 14 and 16. Ethylene also undergoes the destruction processes
+ HC1 C2H4 + CH2Cl' + C2H3' + CH3Cl C2H4 + C1'
CzH3'
(17) (18)
and to a lesser extent C2H3' + CzH4 + CHZCHCH2CH2'
(20) or its highly endothermic decomposition to acetylene C2H3' + CzHz + H'
(21)
The CH2CHCHzCHz'and CH2CHCHCH' radicals subsequently undergo dehydrogenation, hydrogenation, and further addition reactions with C2H3' and C2H2,cyclize, and ultimately result in the formation of high molecular weight carbonaceous deposits. Although the detailed chemical kinetic steps leading to the formation of solid products are not fully known at present, the process nevertheless is well-known to be extremely rapid (Trimm, 1983), and the C2H3' radical is believed to play a pivotal role at the inception stage (Weissman and Benson, 1984; Frenklach et al., 1984; Colket, 1987). In the presence of oxygen, the CzH3' radical has an additional reaction channel which effectively competes with the above processes: CzH3'
+ 0 2 F= HCOH + HCO'
(22)
This elementary reaction has only recently been isolated and studied (Park et al., 1984) and was shown t o be fast with no activation energy barrier. Since Oz intercepts the CzH3' radicals, it suppresses the rate of further molecular weight growth via reactions 19 and 20 and thus has a profound influence on the processes that ultimately result in the formation of high molecular weight carbonaceous deposits in the system. The HCOH and HCO' formed by reaction 22 subsequently are converted into CO via HCOH + C1' HCO' + HC1 HCOH + CH,CI* + HCO' + CH,C1 HCO' + M = CO
+ H'+
M
(23) (24) (25)
Ind. Eng. Chem. Res., Vol. 26, No. 9, 1987 1905 It should be recognized that although the use of O2results in the formation of carbon monoxide, CO is a gaseous product which can be handled much more easily than carbonaceous deposits. Furthermore, CO is an essential constituent of the "synthesis gas" and can itself be used to manufacture useful hydrocarbon products. As evident from the above reaction mechanism, although O2 interrupts the processes that ultimately lead to the formation of solid deposits, it does not directly interfere with the reactions responsible for the formation of ethylene and acetylene. This is clearly supported by the experimental measurements presented previously, in which the yields for C2H2and C2H4remained nearly the same both in the presence and absence of oxygen. I t is most important to note that the success of the CCOP process depends on the presence of the following fast reactions, the latter also being the major route for H2 formation: H'
+ CHSCl+ CH3' + HC1 H' + HC1 e H2 + C1'
(26) (27)
Reactions 26 and 27 efficiently remove the H radicals from the system and render the important combustion chain branching reaction H' + O2 + OH' + 0' (28) ineffective in building up the concentrations of 0' and OH' radicals (Chang and Senkan, 1985; Chang et al., 1987). Consequently the formation of flames and thus the destruction of CH3C1, CzH4,and C2H2are prevented. Before the onset of formation of higher molecular weight compounds or carbonaceous deposits, the two most important reactions competing for the CzH3' radicals are likely to be reactions 19 and 22. Therefore, a polymerization-to-oxidation ratio (POR) can be introduced to assess the potential for the formation of carbonaceous deposits in the system, as the ratio of the rates of these two reactions. By use of the most recent values reported for the rate parameters for reactions 19 and 22 (see Colket (1987) and Park et al. (1984), respectively), the following expression can be obtained: POR = 0.275 exp(-2125/T)[C2H2]/[021 (29) where T i s the temperature in K, and the square brackets denote concentrations. It should be noted that because of cancellation, [C2H3']does not appear in eq 29, and this allows the practical utility for POR. As seen from eq 29, higher temperatures, lower 02,and higher C2H2concentrations increase the POR and thus, the potential for the formation of carbonaceous deposits. Consequently, at the early stages of the oxidative pyrolysis of CH3C1,during which the concentration of CzH2is small, the potential for the formation of deposits is also small. However, as C2H2builds up in the system, this potential increases. Therefore, the concentration of 02 needed to suppress the formation of carbonaceous deposits is likely to be determined from the highest levels of C2H2expected to exist in the system. In conclusion, the oxidative pyrolysis of CH3C1,which represents the second stage in chlorine-catalyzed oxidative
pyrolysis (CCOP) of CH4, was shown to result in the formation of decreased levels of carbonaceous deposits, while allowing the formation of C2H2and C2H4 with combined yields as high as 60% at about 30% conversion of CH3C1. The CCOP process proceeds in a controllable manner, without the formation of flames because of the well-known flame inhibition characteristics of chlorine and chlorinated compounds. Preliminary chemical kinetic analysis of the process suggests that oxygen intercepts and converts the C2H3'radicals into CO, thereby decreasing the extent of polymerization of C2 species. Although some carbon monoxide is formed in the CCOP process, CO is a gaseous product which can be handled much more easily than carbonaceous deposits. In addition, CO itself can be used to synthesize valuable hydrocarbon products. Acknowledgment This research was supported, in part, by funds from the U S . Environmental Protection Agency, Grant R81254401-0, and the Illinois Institute of Technology, Industrial Waste Elimination Research Center, Project 8605. Registry No. CH,Cl, 74-87-3; CzH2,74-86-2; C2H4,74-85-1.
Literature Cited Andersson, L. L.; Christenson, B.; Hoglund, A.; Olsson, J. 0.;Rosengren, L. G. Prog. Astronaut. Aeronaut. 1985, 95, 164. Back, M. H.; Back, R. A. In Pyrolysis: Theory and Industrial Practice; Albright, L. F., Crynes, B. L., Corcoran, W. H., Eds.; Academic: New York, 1983. Benson, S. W. U S . Patent 4 199 533, 1980. Cathonnet, M.; Boettner, J. C.; James, H. In 18th Symposium (International) on Combustion; The Combustion Institute: Pittsburgh, 1981; p 903. Chang, W. D., Karra, S. B.; Senkan, S. M. Environ. Sci. Technol. 1986, 20, 1242. Chang, W. D.; Karra, S. B.; Senkan, S. M. Combust. Flame 1987, in press. Chang, W. D.; Senkan, S. M. Combust. Sci. Technol. 1985, 43, 49. Colket, M. B. 21st Symposium (International) on Combustion;The Combustion Institute: Pittsburgh, 1987, in press. Frenklach, M.; Clary, D. W.; Gardiner, W. C.; Stein, S. E. 20th Symposium (International) on Combustion; The Combustion Institute: Pittsburgh, 1984; p 887. Gorin, E. US.Patent 2 320 274, 1943. Karra, S. B.; Senkan, S. M., submitted for publication in AIChE J . 1987a. Karra, S. B.; Senkan, S. M., submitted for publication in Ind. Eng. Chem. Res. 1987b. Kondo, 0.;Saito, K.; Mukarami, I. Bull. Chem. SOC.Jpn. 1980, 53, 2133. Park, J. Y.; Heaven, M. C.; Gutman, D. Chem. Phys. Lett. 1984,104, 469. Rotzoll, G. Combust. Sci. Technol. 1986, 47, 275. Senkan, S. M. US.Patent 07/040853, 1987. Shilov, A. E.; Sabirova, R. D. Russ. J . Fiz. Khim. 1959, 33, 6. Trimm, D. L. In Pyrolysis: Theory and Industrial Practice; Albright, L. F., Crynes, B. L., Corcoran, W. H., Eds.; Academic: New York, 1983. Weissman, M.; Benson, S. Int. J . Chem. Kinet. 1984, 16, 307. Valeiras, H.; Gupta, A. K.; Senkan, S. M. Combust. Sci. Technol. 1984, 36, 123.
Received for review February 12, 1987 Revised manuscript received June 16, 1987 Accepted June 21, 1987
