Environ. Sci. Technol. 2001, 35, 2543-2547
Carbonyl Sulfide Derived from Catalytic Oxidation of Carbon Disulfide over Atmospheric Particles LIN WANG, FENG ZHANG, AND JIANMIN CHEN* Department of Environmental Science & Engineering, Fudan University, Shanghai, 200433, Peoples’ Republic of China
The formation of carbonyl sulfide (COS) by catalytic oxidization of carbon disulfide (CS2) over atmospheric particle catalysts was explored through FT-IR, MS (mass spectrometry), and a fixed-bed stainless steel reactor. Also the crystallizing conditions and specific surfaces (areas) of the catalysts were investigated by means of X-ray diffraction (XRD) and BET. Some oxides such as CaO, Fe2O3, Al2O3, and SiO2 were investigated under the conditions similar to the atmospheric particles as a comparison. The results showed that atmospheric particles and the oxide catalysts exhibited considerable oxidizing activity for CS2 at ambient temperature to form COS. Elemental sulfur as well as COS was one of the main products; even CO2 could be produced by a secondary reaction on some catalysts. Among the catalysts, CaO showed the strongest catalytic activity for oxidizing CS2. The catalytic activities of Fe2O3 and Al2O3 decreased considerably as compared with CaO, and SiO2 had the weakest catalytic activity. Atmospheric particles’ catalytic activity is between Fe2O3’s and Al2O3’s. The atmospheric particles we collected mainly consist of Ca(Al2Si2O8)‚4H2O, which is also the main component of cement. COS, the main product, is formed by the catalytic oxidizing reaction of CS2 with adsorbed “molecular” oxygen over the catalysts’ surfaces. The concentration of adsorbed oxygen over catalysts’ surfaces may be the key factor contributed to the oxidizing activities. This paper first revealed that CS2 could be catalytically oxidized over atmospheric particles to form COS. It induced that this reaction may be another important source of atmospheric COS from CS2.
Introduction Atmospheric particles or aerosols play an important role in affecting human health (1-3) and many aspects of earth environment (4-7). Recently, heterogeneous interactions between gaseous molecules with wet or dry aerosol particles surfaces have gained considerable interest since they have the potential to significantly alter the gas-phase chemistry of the atmosphere (8). Many works have been done on aerosol’s profound impacts on hydrocarbons (9), NO2 (10, 11), S(IV) (12, 13), and volatile components (14). However, we unexpectedly found that atmospheric particles have a strong activity for oxidizing CS2, even at ambient temperature, while we studied catalytic reaction on peroxodisulfated zirconia solid superacid catalyst (15). The main product is carbonyl sulfide (COS) accompanied by elemental sulfur. * Corresponding author phone: +86-21-65642521; fax: +86-2165642080; e-mail:
[email protected]. 10.1021/es0017763 CCC: $20.00 Published on Web 05/17/2001
2001 American Chemical Society
FIGURE 1. FT-IR cell. As we know, in our atmosphere, carbonyl sulfide is the predominant sulfur-bearing compound (16, 17) that has greatly affected the earth’s radiation balance and climate (18-20). In the Earth’s troposphere, the concentration of COS ranged from 0.3 to 0.5 ppb (16). While relatively inert in the troposphere (21), COS was transported into the stratosphere where, through photolysis (22), it may provide a major source of sulfate for the formation of the stratospheric sulfate layer. Thus, the origin of COS is of particular interest (22, 23). It is well accepted that the atmospheric reaction of CS2 with HO• is an important source of COS (24, 25) as well as the dominant sink of CS2 in the atmosphere, about 30% of COS comes from this reaction. A large fraction of atmospheric COS, perhaps as much as 50%, may result from anthropogenic activities related mainly to fuel processing and consumption (25). This paper investigated the catalytic oxidation of CS2 over atmospheric particle samples and some oxides by means of FT-IR, XRD, BET, MS, and a fixed-bed stainless steel reactor. The new reaction of CS2 over atmospheric particles to form COS reported in this paper may be another important source of atmospheric COS.
Experimental Section Reagents. The atmospheric particle samples were collected on the roof of an office building on Fudan University campus, Shanghai, with a collection rate of 80 L/h. Samples were obtained once a week from May to June 2000. The sampling was 48 h in duration. To prepare the Fe(OH)3, a dilute 10% Fe(NO3)3 solution was titrated by an ammonium hydroxide solution to a final pH of 8-9 under vigorously stirring conditions. Afterward, the precipitate was filtered and thoroughly washed with deionized water to remove the nitrate ions and then dried at 110 °C for 24 h. The dried solid was then calcined in air for 3 h at 500 °C. In this manner, the Fe2O3 catalyst was made. Other catalysts and reactants such as CS2, SiO2, CaO, and Al2O3 were purchased and were analytically pure. Reactor. The FT-IR study was performed in a self-designed FT-IR reactor as shown in Figure 1. The catalytic reaction of CS2 was performed in a continuous-flow and fixed-bed stainless steel reactor under ambient pressure. The diagram of the apparatus is shown in Figure 2. Experimental Procedure. FT-IR Study. For FT-IR study, all samples were used as self-supporting wafers of 0.50 g. VOL. 35, NO. 12, 2001 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
2543
FIGURE 2. Skeleton for catalytic reaction of CS2. Each one was sealed on the heating inlet and activated at 300 °C under the vacuum of 1.0 × 10-2 Pa for 3 h. Then saturated CS2 vapor was introduced into the IR cell at room temperature. At this moment, the IR cell was used as a recirculating reactor in which the gaseous content is uniform under the closed system of 6.0 × 104 Pa. Afterward, the sample was heated to a certain reaction temperature. The gaseous products during the reaction were monitored by FT-IR. Catalytic Reaction of CS2 under Ambient Pressure. The catalytic reaction of CS2 was performed at 250 °C under ambient pressure. The amount of catalyst was loaded up to 1.0 g. The catalyst was first activated at 400 °C under carrier gas flow for 8 h and cooled to 250 °C under a certain carrier gas. Then, the gaseous CS2, which was attained by means of bubbling the carrier gas into liquid CS2 at the temperature of 0 °C (ice bath), flowed over the catalysts. The mixture of carrier gas and CS2 was fed at a rate of WHSV 0.5 h-1. Afterward, the reaction was carried out at the reaction temperature, 250 °C. The carrier gas was high-purity helium, ordinary nitrogen gas, or air at a rate of 80 mL/min. The gaseous products were in-situ analyzed by gas chromatography (GC). Helium was always the carrier gas through out the analytical column with the flow rate of 60 mL/min. Analytical Methods. GC Analysis. An 1102 GC equipped with a 2 m and 3 mm interdiameter Porapak Q column was used. The gaseous products were in-situ analyzed on this Porapak Q column with a He flow rate of 60 mL/min. MS was performed on a HP5973MSD mass spectrometer. FT-IR Analysis. FT-IR analysis was performed on a Nicolet Avatar 360 FT-IR spectrometer. The spectrometer was operated through a Digital PII computer and ESP software (Nicolet, USA). CaF2 or NaCl served as the windows of IR. XRD Patterns of Catalysts. XRD analysis was performed on a Rigaku D/MAX-II X-ray diffractometer with Cu KR. The working current was 80 mA at the voltage of 40 kV. The diffraction scanning range was from 10° to 120° at a rate of 4°/min. BET of Catalysts. The BET surface areas of catalysts were measured with an ASAP2010 specific-surface meter. The flow gas method and nitrogen adsorption at liquid nitrogen temperature were used.
Results and Discussion FT-IR Study. In the IR cell, the reaction was carried out at 40 °C over atmospheric particles. Figure 3 shows the various FT-IR spectra of gaseous products as a function of time. CS2 has the strongest absorption band at 1540.96 cm-1 as well as other four weaker absorption bands at 2323.45, 2319.21, 2192.11, and 2178.57 cm-1 separately (26). Once the catalytic reaction occurs, new absorption bands appear at 2069.42 and 2045.81 cm-1. With time passing by, the latter two bands’ absorbency increases, and new bands at 2924.61, 2904.13, 3085.62, and 3110.56 cm-1 appear. The appearances of these new bands are attributed to the formation of COS (26). As the reaction continues, the reaction becomes weaker and weaker, arriving at the state of homeostasis at last. Elemental 2544
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 35, NO. 12, 2001
FIGURE 3. IR spectra of gaseous products over atmospheric particles (at 40 °C).
FIGURE 4. IR spectra of gaseous products while the reaction was carried out at 40 °C for 20 h: (A) atmospheric particle samples, (B) Al2O3, (C) CaO, (D) Fe2O3, (E) SiO2. sulfur is apparently observed to deposit with slight yellow color on inner surface of the cell. Figure 4 shows the various FT-IR spectra of gaseous products in the IR cell while the reactions were carried out at 40 °C for 20 h over atmospheric particles and oxide catalysts. The production of COS was observed whatever catalyst was used. However, CO2 was also revealed to be one kind of product of the catalytic reaction over Al2O3 and CaO, while it had not been found in case of Fe2O3, SiO2, or atmospheric particles. When the former two catalysts were used respectively, the strongest absorption band of CO2 appears at 2361.31 cm-1 (26). Four weaker absorption bands of CO2 at a range of 3723-3595 cm-1 were observed after the reaction was carried out overnight. SO2 or SO3 never occurred during all the reactions explored. XRD Patterns and BET Analysis. XRD profiles of atmospheric particle samples and metal oxides are shown in Figure 5. The atmospheric particle samples collected are totally crystallized. It consisted mainly of Ca(Al2Si2O8)‚4H2O, which is also the main component of cement. Its strongest intensities of 2θ are 26.60°, 20.82°, and 27.92°, which correspond to its crystal faces at (100), (121), and (310). The correspondening standard diffraction card number is 20-452. Moreover, the diffraction peak of Ca2(Mn2Fe)(PO4)2‚2H2O can be identified in the XRD profile of atmospheric particle samples. Its standard diffraction card number is 10-390. However, the diffraction intensity of Ca2(Mn2Fe)(PO4)2‚2H2O at 27.50°, 13.84°, and 29.40° is much smaller that that of Ca(Al2Si2O8)‚ 4H2O. In the case of Al2O3, it mainly consists of η-Al2O3 with the standard diffraction card number 16-394. Its diffraction is at (401h 2h ), (400), and (312) crystal faces. XRD profile of CaO, which includes a small amount of the Ca(OH)2 diffraction, shows that the strongest diffraction of CaO is at (200), (220),
TABLE 1. BET and Crystal Phase sample
Al2O3
CaO
Fe2O3
SiO2
atmospheric particles
BET (m2‚g-1) crystal phase
124.6 η-Al2O3
9.24
9.77 hematite
206.0 amorphous
7.37 Ca(Al2Si2O8)‚4H2O
FIGURE 7. Conversion of CS2 versus time (at 250 °C): (9) air as the carrier gas, (2) ordinary nitrogen gas, (1) high-purity helium. FIGURE 5. XRD patterns: (A) atmospheric particle samples, (B) Fe2O3, (C) CaO, (D) Al2O3, (E) SiO2.
FIGURE 6. Absorbency of gaseous COS per square-meter surface of catalysts (at 2069.42 cm-1, 40 °C) versus time. and (111). Its standard diffraction card number is 37-1497. XRD profile of Fe2O3 is the same as ferric oxide in hematite, whose strongest diffraction is at (104), (110), and (116). SiO2 mainly consists of amorphous phase. The BET and XRD profiles of catalysts are shown in Table 1. The BET of atmospheric particles, which is 7.37 m2 g-1, is the smallest one. The BET specific surfaces of SiO2, Al2O3, CaO, and Fe2O3 are respectively 27.95, 16.91, 1.25, and 1.33 times that of atmospheric particles. Catalytic Activity Order. In Figure 6, we report the evolution of absorbency of COS per square-meter catalysts’ surface (at 2069.42 cm-1) versus time. After CS2 was introduced into the IR cell, the reaction temperature was attained at 40 °C. Among the catalysts, CaO showed the strongest catalytic activity for oxidizing CS2. The catalytic activities of Fe2O3 and Al2O3 decreased considerably as compared with CaO. SiO2 had the weakest catalytic activity. Atmospheric particles’ catalytic activity is between Fe2O3 and Al2O3. Above all, we conclude that the catalytic activity order of atmospheric particles and oxides is CaO > Fe2O3 > atmospheric particles sample > Al2O3 . SiO2 (at 40 °C). Catalytic Reaction of CS2 under Ambient Pressure. The catalytic activity of Al2O3 was representative as both the primary and the secondary reactions could be observed in
FIGURE 8. MS chromatogram. IR study. Yet, we chose Al2O3 to be studied by means of the continuous-flow and fixed-bed stainless steel reactor. The conversion of CS2 was measured, and its variation versus time was plotted for each sampling analysis. As shown in Figure 7, we compare these three curves, obtained from reaction on Al2O3 under different flow carriers (high-purity helium, nitrogen gas, or air that severed as the carrier gas of saturated CS2 vapor, respectively) at a rate of 80 mL/min. When air served as the carrier gas, the conversion of CS2 is practically 72% at the initial stage of reaction. It rapidly decreases in 1 h and tends to stabilize at 54%, which is almost l8% lower than that of its initial stage. When the gaseous products flowing out of the reactor were analyzed by GC, chromatograms showed three obvious peaks, including a peak of CS2. The gaseous products, which were condensed by liquid N2, were characterized by mass chromatography. The chromatogram is shown in Figure 8. There were three peaks at mass number 60, 44, and 32, which belonged respectively to COS, CO2, and O2 (remaining from the reaction). Thus, it was confirmed that COS is one of main products. At the same time, we also found yellow elemental sulfur deposited at the bottom of the reactor. VOL. 35, NO. 12, 2001 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
2545
When ordinary nitrogen gas (99.9%) was chosen as the flow gas that carried saturated CS2 vapor, the conversion is nearly 55% at first and then decreases to 5% in 3 h. It stabilized at 4% after 6 h. We also could observe the formation of COS, but the yield of COS is much smaller than that under the condition of air as flow gas. When high-purity helium (99.99%) was selected as the flow gas, no COS or conversion of CS2 was detected. It is obvious that CS2 conversion strongly depends on carrier’s composition. The conversion differs considerably from each reaction under variant carrier gas having different O2 content. There is no catalytic oxidation of CS2 under highpurity helium flow without any molecule of oxygen. However, the reaction is continuous at least up to 100 h under the flow of air or ordinary nitrogen, which has more or less molecules of oxygen. Hence, we take into account that oxygen is concerned with the catalytic oxidation of CS2 on Al2O3. Oxygen content is the key factor during this kind of oxidation. The decreasing of the CS2 conversion after the initial stage of the reaction might have nothing to do with catalyst deactivation. The higher CS2 conversion at the initial stage resulted from higher concentration of oxygen both in carrier gas and on the surface of catalyst. Once the adsorbed oxygen on the surface of catalysts is exhausted, only part of oxygen in the carrier gas participated in the surface of catalyst to lead to CS2 conversion. Mechanism Investigation. Liu et al. once reported IR studies of the adsorption of CS2 on γ-Al2O3 (27). They observed absorption bands of both COS and CS2 on the wafer of γ-Al2O3 when contacting CS2 with γ-Al2O3 at room temperature. Repeated dosing with CS2 results in diminishing extent of reaction on γ-Al2O3, confirming that this oxidizing character is associated with the finite amount of oxygen available on the adsorbent. The formation of COS is possible from the surface reaction:
CS2 + [O] f COS + [S]
(1)
According to the results shown in Figure 7, we confirmed that the reaction on catalysts, i.e., atmospheric particles and some metal oxides, is a kind of catalytic oxidation reaction. CS2 can react with active adsorbed oxygen on the surface of catalysts, which results in the formation of COS (eq 1). Besides, when the catalytic reaction of CS2 with ordinary nitrogen carrier gas was over, we immediately changed the carrier gas to air. The CS2 conversion showed a considerable increase. It indicated that the catalysts can activate oxygen in the carrier to active [O], thus resuming the reaction. Without oxygen, the reaction of CS2 on catalysts cannot be proceeded. In FT-IR, however, because of adsorbed oxygen existing on the surface of catalysts, CS2 can react with [O] until it is all consumed. At that time, the amount of COS product tends to be invariable. Moreover, the formation of CO2 on some oxides is possibly caused by a secondary overoxidation of COS:
COS + [O] f CO2 + [S]
(2)
In conclusion, the following model shown in Figure 9 can express the reaction mechanism: Step I delegates that more active [O] generated from the activation of catalysts in the carrier gas having oxygen. Step II delegates that elemental sulfur deposited on catalyst surface because of the introduced CS2 vapor reacting with [O] adsorbed on surface. Step III delegates that the reaction continued, resulting in produced COS and quantities of elemental sulfur, and tended to be invariable for adsorbed oxygen exhausted step by step. Step IV indicates that CO2, which is the overoxidation product of COS, was observed in 2546
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 35, NO. 12, 2001
FIGURE 9. Model of mechanism on catalytic oxidation. the reaction. Finally, the catalyst covered with elemental sulfur can be regenerated. The elemental sulfur could be moved away from the surface of catalyst through sublimation under the reaction condition. In fact, the amount of aggregated elemental sulfur was found to deposit in the tube connected to the bottom of the reactor. The fresh surface of regenerated catalyst might catch oxygen in the carrier gas and thus resumed the reaction in terms of step I. However, except for the reaction of adsorbed oxygen, whether the lattice oxygen on catalyst surface participates in the reaction is unsolved by this model. Besides, whether the catalytic activity is associated with catalyst surface acidbase site is of great interest. All these features need further study.
Acknowledgments We gratefully thank Professor Huiqi Hou and Associate Professor Xunxi Pan (Department of Environmental Science & Engineering, Fudan University) for their kind opinions about the work.
Literature Cited (1) Harrison, R. M.; Yin, Jianxin. Sci. Total Environ. 2000, 249, 85101. (2) Spurny, K. R. J. Aerosol Sci. 1996, 27, S473-S474. (3) Reichardt, T. Environ. Sci. Technol. 1995, 29 (8), A360-A364. (4) Friedlander, S. K.; Lippmann, M. Environ. Sci. Technol. 1994, 28 (3), A148-A150. (5) Hughes, L. S.; Allen, J. O.; Bhave, P.; Kleeman, M. J.; Cass, G. R.; Liu, D. Y.; Fergensor, D. F.; Morrical, B. D.; Prather, K. A. Environ. Sci. Technol. 2000, 34 (15), 3058-3068. (6) Kalberer, M.; Yu, J,; Cocker, D. R.; Flagan, R. C.; Seihfeld, J. H. Environ. Sci. Technol. 2000, 34 (23), 4894-4901. (7) Kreidenweis, S. M. Re. Geophys. 1995, 33, 775-784. (8) Baltensperger, U.; Ammann, M.; Kalberer, M.; Gaggeler, H. W. J. Aerosol Sci. 1996, 27, S651-S652. (9) Hoffmann, T.; Odum, J.; Bowman, F.; Collins, D.; Klockow, D. J. Aerosol Sci. 1996, 27, S233-S234. (10) Ammann, M.; Baltensperger, U.; Bochert, U. K.; Eichler, B.; Gaggeler, H. W.; Jost, D. T.; Turler, A.; Weber, A. P. J. Aerosol Sci. 1995, 26, 61-70. (11) Kalberer, M.; Ammann, M.; Tabor, K.; Parrat, Y.; Nyeki, S.; Weingartner, E.; Piguet, D.; Jost, D. T.; Turler, A.; Gaggeler, H. W.; Baltensperger, U. J. Aerosol Sci. 1996, 27, S669-S670. (12) Brandt, C.; Elding, L. I. Atmos. Environ. 1998, 32 (4), 797-800. (13) Brandt, C.; Eldik, R. V. Chem. Rev. 1995, 95, 119-190. (14) Bunz, H.; Koyro, M.; Mohler, O.; Saathoff, H. J. Aerosol Sci. 1996, 27, S653-S654. (15) Zhang, L.; Wang, L.; Chen, J. M. Chem. J. Chin. Univ. 2000, 21 (1), 116-119. (16) Hanst, P. L.; Speller, L. L.; Watts, D. M. J. Air Pollunt. Control Assoc. 1975, 25, 1220-1226. (17) Maroulis, P. J.; Torret, A. L.; Bandy, A. R. Geogphs. Res. Lett. 1977, 4, 510-512. (18) Pollack, J. B. Nature 1976, 263, 551-555. (19) Junge, C. E. J. Meteorol. 1961, 18, 81. (20) Pollack, J. B. J. Appl. Meteorol. 1976, 15, 247. (21) Prinn, R. G. Science 1973, 182, 1132.
(22) Mark, C.; Davis, D. D. Global Biogeochem. Cycles 1993, 7, 321337. (23) Thornton, D. C.; Bandy, A. R.; Blomquist, B. W. J. Geophys. Res. 1996, 101 (1), 1873-1881. (24) Sze, N. D. Nature 1979, 278, 731-732. (25) Thrco, R. P.; Whitten, R. C.; Toon, O. B. Nature 1980, 283, 283286.
(26) Standard IR Spectra; Sadtler Research Laboratories, Inc.: 1972. (27) Liu, C. L.; Chuang, T. T. J. Catal. 1972, 26, 474-475.
Received for review October 17, 2000. Revised manuscript received January 9, 2001. Accepted January 29, 2001. ES0017763
VOL. 35, NO. 12, 2001 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
2547
