Homogeneous catalytic oxidation of carbon ... - ACS Publications

Depts. of Chemistry and Engineering Technology, Western Kentucky University, Bowling Green, Ky. 42101. Carbon monoxide in gas mixtures can be rapidly ...
2 downloads 0 Views 242KB Size
COMMUNICATIONS

Homogeneous Catalytic Oxidation of Carbon Monoxide William G . Lloyd' and Donald R. Rowe Depts. of Chemistry and Engineering Technology, Western Kentucky University, Bowling Green, Ky. 42101

Carbon monoxide in gas mixtures can be rapidly and efficiently oxidized at ambient temperatures to carbon dioxide. An old analytical technique for the assay of carbon monoxide has been adapted to provide a fast catalytic process based on Pd(II)/Cu(II). Passage of a test mixture containing 2 x l o 4 ppm carbon monoxide through a n aqueous solution a t ambient temperature in a simple laboratory apparatus affords 9 7 z conversion t o carbon dioxide. H

T

he carbon monoxide output of modern technology is bringing ambient concentrations to levels which warrant serious concern with regard to public health. Clean nonurban air may have CO concentrations as low as the 0.09 ppm found at Point Barrow, Alaska (Brice and Roesler, 1966), but in major urban centers the 8-hour average C O concentration is often above 12 ppm (Larsen and Burke, 1969). Values for 8-hour averages have been found as high as 44 ppm, while transient concentrations of well over 100 pprn are known. Current criteria indicate that extended exposure to air containing 10 ppm CO is physiologically significant (Air Quality Criteria, 1970; Wolf, 1971). The oxidation of C O to COzis thermodynamically favored by more than 60 kcallmole. The problem is that of reducing the activation energy barrier to permit easy oxidation by atmospheric oxygen at ambient temperatures. We wish to report a facile catalytic oxidation of C O to COS,by means of dilute aqueous solutions of Pd(I1) salts. The reduction of Pd(I1) chloride in solution by CO was developed by Phillips (1894) as a method of detecting and assaying CO : CO

+ PdClz.2H20

CO?

+ Pd(0) + HzO + 2HC1

(1) This reaction has also been used for the assay of Pd(I1) salts (Ogburn and Brastow, 1933). In the presence of a reoxidant such as cupric chloride, palladium(0) is very readily reoxidized to palladium(I1) (Dragerwerke, 1941; Smidt et al., 1959): Pd(0)

+

+ 2C~C12

PdClz

+ Cu2C12

(2) The reoxidation of cuprous chloride to cupric chloride occurs very readily under mild conditions (Chaltykyan and Chytan, 1957; Jhaveri and Sharma, 1967): -+

+ 2HC1 + '/zO,

C~zClz

co + 1 / 2 0 ,

+

Table I. Gas Chromatographic Analyses" of Gas Mixture*

co; z

+

2C~C19 H20

(3) The sum of Reactions 1-3 is the desired conversion of carbon monoxide: +

through a n aqueous solution containing 0.020 M PdC12, 1.00 M CuC12, and 1.00 M Cu(N0J2. In each of three standard gas-washing bottles (250 ml capacity, fitted with coarse fritted glass diffusers) was placed 150 ml of catalyst solution, and the three bottles were connected in series downstream of a flowmeter connected to a cylinder containing the heliumcarbon monoxide mixture. The cylinder valve was set to permit a flow of 235 ml/min through the system. After 25 min of operation at 22"C, a portion of effluent gas was collected in a gas-sampling flask. A portion of untreated feed gas was then collected for comparison. The gas samples were analyzed by gas chromatograph, with a standard thermal conductivity instrument (Aerograph Model 202-1C) fitted with a 6 ft X '/id in. column of Porapak Q for carbon dioxide assay and a 6 ft X 1/4 in. column of molecular sieve 5A for carbon monoxide assay, both determinations run at a column temperature of 3 O O C . Assays were made in quadruplicate, with retention of the best three of each set of four determinations. These data (Table I) show the carbon monoxide content of the gas mixture t o be reduced from 2.0 to 0.06%, at the same time that a major CO? peak appears, as a result of passage through the catalyst system. Cross analysis of the untreated gas shows the presence of a small amount of CO?,about 0.05 Z, as an initial impurity. This preliminary study, with an effective gas-liquid contact time of about 1.5 sec, shows that a gas stream containing 2 X l o 4 ppm carbon monoxide can be treated to effect conversion of 9 7 x of the CO to CO?.The palladium catalyst is costly. However, its working concentration is low and its regeneration by Reaction 2 proceeds very readily, as is shown by this high conversion after the prior passage of more than enough CO to have reduced all of the initial charge of Pd(I1). Furthermore, both palladium metal and its salts appear to be without appreciable toxicity (Fassett and Irish, 1963). The present study has been made simply to determine if the process can work. In view of these encouraging results we are now undertaking a more detailed study of process variables, including contact time and concentrations of CO, oxygen, and other gases.

co*

z

Feed G a s

1.96 2 . 0 2 (av. 2 . 0 0 % ) 2.02

0,069 0 . 0 3 >(av. 0 , 0 5 0,046

Effluent gas

0,058 0 . 071 (av. 0 , 0 6%) 0.050

2.03 2 . 0 0 (av. 2 . 0 0 %) 1.97

(4)

Thus it appears possible in principle to scrub a stack gas containing CO and at least a little free oxygen, by use of an aqueous Pd(I1)-Cu(I1) catalyst solution, and thereby to convert C O to CO, at a negligibly low operating cost, The chemical feasibility of this system has been investigated by passing a stream of helium containing 2 x 104 ppm CO

COZ,

z)

of analysis: overall =k 0.021 as CO ( 8 D F ) low range i 0.010% as CO (4 D F ) a Using disk integrator; chromatographic conditions described in text. * Custom mixture of 2.00% CO in He, obtained from Air Reduction Corp.

SD

To whom correspondence should be addressed. Volume 5 , Number 11, November 1971

1133

Literature Cited “Air Quality Criteria for Carbon Monoxide,” National Air Pollution Control Administration, Publication AP-62, March 1970, pp 10-17. Brice, R. M., Roesler, J. F., J . Air Pollut. Contr. Ass. 16, 597 (1966). Chaltykyan, 0. A., Chytan, G. S., Nauch. Tr. Erecan. Cos. Unir;. Ser. Khim. Nauk 60 (4), 125 (1957); C A 53, 1091% (1959). Dragerwerke Heinr. u. Bernh, Drager im Lubeck, German Patent 713.791 (Oct. 23. 1941). Fassett, D. W., Irish, D. D.,’in “Industrial Hygiene and Toxicity,” Vol. 2, F. A. Patty, Ed., Interscience, New York,

N.Y.. 1963. D 1130. Jhaveri; A. S:,’Sharma, M. M., Chem. Eng. Sci. 22, 1 (1967). Larsen, R. I., Burke, H. W., “Ambient Carbon Monoxide Exposures,” presented at annual meeting, APCA, New York, N.Y., June 1969. Ogburn, S. C., Jr., Brastow, W. C., J . Amer. Chein. SOC.55, 1307 (1933). Phillips, F. C., Amer. Chem. J. 16, 225 (1894). Smidt, J., Hafner, W., Jira, R., Sedlmeier, J., Sieber, R., Ruttinger, R., Kojer, H., Angew. Chem. 71, 176 (1959). Wolf, P. C., ENVIRON. SCI. TECHNOL. 5, 212 (1971). Receicedfor reDiew Dec. 30, 1970. Accepted March 22, 1971.

Copper Contamination of Atmospheric Particulate Samples Collected with Gelman Hurricane Air Samplers Gerald L. Hoffman’ Dept. of Chemistry, University of Hawaii, Honolulu

96822

Robert A. Duce Graduate School of Oceanography, University of Rhode Island, Kingston, R.I.

rn Atmospheric particulate samples collected with Gelman Hurricane high-volume air samplers can be seriously contaminated with copper if special precautions are not taken to isolate the filter holder from the pump. This copper contamination can easily be observed when samples are collected in the unpolluted marine atmosphere. The source of the copper contamination is probably brush wear on the copper armature of the pump. Copper contamination may also occur with other types of high-volume samplers used to collect atmospheric particulate matter and may make much of the previous data o n atmospheric copper concentrations doubtful.

02881

only element that had higher concentrations in nonurban air than in urban air. Ludwig et al. (1970) point out that the high nonurban C u concentrations are probably due to mining and smelting of ores, as well as re-entrainment of soil particles. However, the average Cu and Fe concentrations reported at the 30 nonurban sites from all over the United States are approximately equal (-0.2 and 0.3 pg’m3, respectively). The average crustal ratio of C u to Fe is approximately (Mason, 1960). If one makes the conservative assumption that all the Fe in nonurban areas is from the soil and that the effect of local smelting and mining is largely removed in the national nonurban averages, it is apparent that the mean C u / F e ratio in the NASN nonurban particles is approximately 1000 times too high to be from crustal material. Experimen ta 1

C

ollection of atmospheric particulate matter for trace metal analysis has been accomplished for many years by various types of high-volume (hi vol) samplers. The major user of this sampler type, for collecting particulate matter in urban and nonurban areas of the United States, has been the National Air Sampling Network (USPHS, 1958, 1962, 1965, 1967, 1968). Several U S . companies manufacture this type of pump for the specific purpose of collecting atmospheric particulate matter on filters. A typical hi vol sampler is the Hurricane (Gelman Instrument Co., Ann Arbor, Mich.). It has come to our attention that a serious contamination of the filter with copper may occur when particulate matter is collected with Gelman Hurricane samplers. This contamination probably occurs with other types of hi vol samplers as well. Ludwig et al. (1970) have reviewed selected data collected by the NASN for the following elements: Cr, Cu, Fe, Mn, Ni, Pb, Sn, Ti, V, and Zn. Of these elements, only C u does not show some geographical variation. Furthermore, C u is the

~~

1

To whom correspondence should be addressed.

1134 Environmental Science & Technology

Sampling. In order to examine the possibility of pump contamination, the following studies were conducted during the summer of 1969. Two sets of samples were collected from a 20-m high tower at Bellows Air Force Station on the windward coast of Oahu, Hawaii. The Gelman Hurricane samplers were modified (i.e., the pump handles and leg supports for the motor were removed) and mounted in wooden shelters. The two sets of samples were obtained with different collection configurations. In one configuration (designated mode A ) the pump was used in the normal manner recommended by the manufacturer and used by the NASN (Le., the filter holder was attached directly to the pump). In the second configuration (designated mode B), the pump was attached to the filter holder by means of a 10-m length of 5-cm i.d. flexible plastic tubing, thus placing the filter a considerable distance away from the pump. Samples in modes A and B were not collected simultaneously in this study. The two sets of samples, however, were collected under similar meteorological conditions (Le., onshore trade winds with similar speed and low precipitation level). Each sample was collected during a period of 24 hr on Delbag Microsorban polystyrene filters, type 99 97. Results of the analyses of these samples for C u and N a are given in Table I. A further test of pump contamination was conducted at