Geological and inorganic materials - ACS Publications - American

May 18, 1982 - (5N) Maeda, S. U. S. US 4,329,874 (Cl. 73-190CV;GO1K17/00), 18 May. 1982. (6N) Springer .... (East) 142,760 (Cl. G01N27/58), 09 Jul 198...
0 downloads 0 Views 4MB Size
Anal. Chem. 1983, 55, 4QR-56R (5M) Lunskii, M. Kh. Zavod. Lab. 1982, 48 (5), 17-20 (Russ); Chem. Abstr. 1982, 97,58151. Calorlmetry (IN) Curry, R. N. Oil Gas J. 1981, 79 (32), 95-6, 98-9. (2N) Howard, R. L. Proc. Int. Sch. Hydrocarbon Meas. 1981, 56, 105-8. (3N) Kude, W. B.; Youngbauer, D. L.; Pearman, A. N. J. Eur. Pat. Appl. 31,145 (Ci. GOIN25/22), 01 Jul 1981. (4N) Maeda, S. U. S. US 4,329,873 (Cl. 73-19OCV; GOIN25/30), 18 May 1982. (5N) Maeda, S. U. S. US 4,329,874 (Cl. 73-19OCV; G01K17/00), 18 May 1982. (EN) Springer, T. A.; Norris, C. G.; McCoy, R. D. Anal. Instrum. 1980, 18, 109-17. (7N) Szonntagh, E. L. Brit. UK Pat. Appl. GB 2,074,728 (Cl. GOlN25122), 04 Nov 1981. (EN) Van Rossum, G. J.; Benes, G. J. GWF, Gas-Wasserfach: GaslErdgas 1981, 122(1), 12-19 (Ger); Chem. Abstr. 1981, 94,159346. (9N) Vlllalobos, R. ISA Trans. 1982, 21 (2), 93-100. (10N) Watson, J. W.; White, F. A. Oil Qas J. 1982, 80 (14), 217-18, 220, 225. (11N) Williams, R. A. Proc. Int. Sch. Hydrocarbon Meas. 1981, 56, 518-22. Density and Speclflc Gravlty (1P) Hanklnson, R. W.; Coker, T. A.; Thomson, G. H. Hydrocarbon Process ., Int. Ed. 1982, 61 (4), 207-8. (2P) Kahmann, A. R. Proc. Int. Sch. Hydrocarbon Meas. 1981, 56,501-4. (3P) Lewis, H. E. Proc. Int. Sch. Hydrocarbon Meas. 1981, 56, 234-9. (4P) Parrish, W. R.; Pollin, A. G.; Schmidt. T. W. Proc. Annu. Conv.-Gas Process. Assoc. 1982, 61, 164-70. (5P) Rozentsvaig, A. K.; Grevtsov, V. M. Neftepromysl. Del0 1982, (4), 24-6 (Russ); Chem. Abstr. 1982, 97, 41238. (6P) Rybalkin, V. I.; Labinov, S. D.; Zhurba, A. S. Neftepererab. Neftekhim. (Moscow) 1981, (6), 47-50 (Russ); Chem. Abstr. 1981, 95,143612. (7P) Siegwarth, J. D.; LaBrecque, J. F. NBS Tech. Note (US.)1981, 1035. (8P) TerBush. D. J. Proc. Int. Sch. Hydrocarbon Meas. 1981, 56,230-3. Samp II ng ( l a ) Astryanln, S . N. Gazov. Prom-st. 1981, (12), 20-1 (Russ); Chem. Abstr. 1982, 96, 125762. (2Q) Carpenter, R. L.; Newton, G. J.; Cheng, Y. S.; Barr, E. B.; Yeh, H. C. Report Dec 1980, LMF-84, pp 384-6 Avail. NTIS; E R A . 6(13), 19188. (30) Drake, C. F. Oper. Sect. Proc.-Am. Gas Assoc. 1980, T175-Tl86. (4Q) Neulander, C. K.; Walmet, G. E.; Zarchy, A. S.Report 1980, ANL-8062, pp 380-7; E.R.,A. 8(6), 6830. (5Q) Phillips, J. B. Proc. Int. Sch. Hydrocarbon Meas. 1981, 56, 107-11.

(8Q) Schepers, H. H.; Kilmer, J. W.; Bernos, J. Proc., Annu. Conv.-Gas Process. Assoc. 1982, 6 1 , 1-8. (7Q) Stepanek, A.; Wittenberg, E.; Waloschek, V. Czech. CS 189,073 (Cl. G01N1/22), 15 Dec 1961. (8Q) Welker, T. F. Proc. Int. Sch. Hydrocarbon Meas. 1981, 5 6 , 531-5. (9Q) Welsch, T.; Engewald, W.; Huenerbein, G.; Apel, G. Chem. Tech. (Leipzlg) 1981, 33 (3), 149-51 (Ger); Chem. Abstr. 1981, 95,64697.

Mlscellaneous (IR) Abdurakhmanov, A. A.; Abbasov, A. A.; Askerov, A. B.; Veliyulin, E. Yu.; Imanov, L. M. Azerb. Khim. Zh. 1980, (l), 147-52 (Russ); Chem. Abstr. 1981, 94, 124230. (2R) Caffey, W. R. Proc. Int. Sch. Hydrocarbon Meas. 1981, 56, 132-4. (3R) Dliler, D. E. J. Chem. Eng. Data 1982, 27(3), 240-3. (4R) Diller, D. E.; Chang, R. F. Appl. Spectrosc. 1980, 34 (4), 411-14. (5R) Erickson, M. D.; Frazier, S. E.; Sparacino, C. M. Fuel 1981, 60 (3), 263-6. (6R) Eubank, P. T.; Hall, K. R.; Holste, J. C.; Scheloske, J. J. Proc., Annu. Conv.-Gas Process. ASSOC.1980, 59, 18-30. (7R) Fleischmann, D.; Schwab, H. Ger. (East) 148,503 (CI. GOlN31/06), 11 Feb 1981. (8R) Haas, W. J. Jr.; Eckels, D. E.; Kniseley, R. N.; Fassel, V. A. Report 1981, IS-M-327 Avail. NTIS; E R A . 6(16), 23032. (9R) Keady, M. J. Proc. Int. Sch. Hydrocarbon Meas. 1981, 56, 334-5. (10R) KLpka, H. VDI-Ber. 1980, 363, 127-31 (Ger); Chem. Abstr. 1981, 94,211248. (11R) Lloyd, D. W, 011 Oas J . 1981, 79 (38), 126-8, (12R) Loree. T. R.; Radziemski, L. J. Plasma Chem. Plasma Process. 1981, 1 (3), 271-9. (13R) Meyer, B.; Goetze, R.; Eidner, D.; Mueller, R.; Mottitschka, W.; Roessler, D. Ger. (East) 142,760 (CI. GOlN27/58), 09 Jul 1980. (14R) Moore, B. J. Report 1979, PB-80-142300, 119 pp Avail. NTIS; E.R.A. 8(2), 1969. (15R) Nersesova, N. A. Izv. Vyssh. Uchebn. Zaved., Neft Gaz 1981, 24 (12), 52-5 (Russ); Chem. Abstr. 1982, 96, 145575. (16R) Reid, R. C.; Shanes, L. M.; Virk, P. S. Report Dec 1979, PB-80210685, 81 pp Avail. NTIS; E.R.A. 6 (12), 16521. (17R) Sallet, D. W.; Wu, K. F. Report Apr 1980, PB-80-189053, 111 pp Avall. NTIS; E.R.A. 6 (lo), 13402. (18R) Tilley, H. C. Proc. Int. Sch. Hydrocarbon Meas. 1980, 55, 330-1. (19R) Wllliamson, E. Proc. Int. Sch. Hydrocarbon Meas. 1981, 56, 324-5. Standards (1s) ASTM Standards, Philadelphla, PA, Part 26 (1982). (2s) Curry, R. N. OilGas J. 1981, 79(44), 121-4. (3s) Helke, T. GWF, Gas-Wasserfach; GaslErdgas 1982, 123 (3), 119-21 (Ger); Chem. Abstr. 1982, 96, 183901. (45) Nagakura, R. Nippon Gasu Kyokaishi 1980, 33 (E), 26-35 (Japan); Chem. Abstr. 1981, 94,88778.

Industrial Hygiene Richard G. Melcher The Dow Chemical Company, Michigan Applied Science and Technology Laboratories, Analytical Laboratory, 574 Building, Midland, Michigan 48640

A. INTRODUCTION This review covers a period from approximately 1978 through 1982. Earlier references are included in some cases to give a more complete picture of the technique being discussed. There has been an exponential growth in industrial hygiene over the period, and no attempt was made to cover the entire field of industrial hygiene or even the analytical aspects of industrial hygiene chemistry. For some chemicals which have come under close scrutiny, such as vinyl chloride, benzene, acrylonitrile, and formaldehyde, there are many dozens of references in the literature and anyone interested in a complete survey for a specific chemical would best be served by running a specific literature search. The intent of this review was to cover some basic concepts of the most widely used technique in personal monitoring. One in-depth review for a class of compounds, the diisocyanates, is given since, in addition to showing a multidirectional analytical approach, it may give a sense of appreciation for the degree of effort expended in the development of methods for trace amounts of reactive chemicals. To ensure a safe working environment in the laboratory and production plants, it is often necessary to determine trace 40 R

0003-2700/83/0355-40R$08.50/0

quantities of organic chemicals in the work atmosphere. For a successful program it is necessary to have good communication between the toxicologist who must assess the toxicity of the compound, the industrial hygienist who must investigate the hazard of exposure to personnel, and the analytical chemist who must determine the concentration of the compound in the environment and in biological samples for metabolism and pharmacokinetic studies. The factors which affect the collection and determination of trace quantities can be quite complex, and each specialist involved in method development, sampling, or analytical measurements must have a basic understanding of the total effort. It is beyond the scope of this review to discuss the complexities of toxicological, industrial hygiene, or analytical techniques. The main emphasis will be on collection with solid sorbents, either using ump and tube or diffusional sampling techniques, followed \y thermal desorption or solvent desorption. This includes a majority of the samples presently being collected and new methods being developed. Solid sorbent sampling tubes and dosimeters are convenient to use, can concentrate trace contaminants, and can be used for area samples as well as for employee breathing-zone samples. 0 1983 American

Chemical Society

INDUSTRIAL HYGIENE

R M r d 0. M.lsh.r is a Research Assock ate in the Analytical Labaalay of The Cow Chemlcai Company. Mkidland. MI. He recBived his B.S. degree in ehemisby horn Vaiparalso Univerrny and M.S. degree in analytical chemistry horn Purdm UniversHy. He joined The DOWChemical Company in 1962 and has been involved in moiecular spCboscopy. nuclear magnetlc ,esonanCe. and chromatOgraphic separati~ns. In 1972, he organized an analytical laboratw wimin me Dow U.S. Area Indushl~iHvgiene Depamnem and has since been i""OivBd i" i"duslrhi hygiene and environmental m e w s research and Mtveiopmenl in Environmental OToup ot the Analyticai LBbMalory. Mr. Melcher has authored or coauihwed 20 publications in the field ot trace chemicals in ak and has received Dow Chemical's Vernon A. Slenger award lor his etfms in lhis area. He is a member ot the American Chemical Society. American Industrial Hygiene As~o~IBlion, and Sigma Xi and is the present chairman 01 ASTM SutCommMee D22.04 on Sampling and Analysis of Workplace Atmospheres.

These techniques have proven valuable, not only to the industrial hygienist because they are easy to use and transport but also to the analytical chemist who finds the analytical procedures straightforward and adaptable to a wide range of compounds. In the early development of solid sorbent sampling techniques, a 10-min sampling period was usually adequate to establish the concentration of a compound in air in an area. This technology was quickly followed with emphasis toward portable personal sampling systems which could be used for up to full shifts, to determine the "time-weighted-average" (TWA) exposures. Recent emphasis has been in examinin and understanding parameters which affect collection a n i recovery and to develop validation criteria to establish reliability of the methods. Although methods for many new compounds can be extrapolated by comparison to existing methods, difficulties with collection, recovery, stability, or analysis often arise which require further sophistication of the technology. Various techniques have been developed to extend the weful range of available solid sorbents and new specialized sorbents have been developed to solve specific problems. The development of specific and highly sensitive gas chromatographic detectors, liquid chromatographic systems, ion chromatography, and other new analytical instrumentation has enabled a greater flexibility in the collection parameters and choice of sorbent/solvent systems. The uses of chemical reactions and derivatization also play an increasingly important part in solid sorbent methods for compounds which are unstable or difficult to analyze. Some of the recent developments in technique, sorbents, and instrumentation which appear to have broad application are discussed in this review. Some specificexamples are given in each area to illustrate the concepts, however, diligent surveillance of the literature will be necessary to keep pace in this rapidly growing field. Several reviews on personal monitoring are available in the literature (AI-2) along with a number of books containing sampling and analytical methodology (A3-8). NIOSH has published a compilation of over 500 monitoring methods for 737 analytes in seven volumes (with an eighth volume in preparation), and they are presently available from National Technical Information Service (NTIS), Springfield, VA ( A S 15). Several reviews on air pollution monitoring are also available (A16181 which may be useful since many techniques are similar.

B. PUMP AND TUBE (SOLVENT DESORPTION) Solvent desorption of collected compounds from solid sorbents is the most commonly used technique. The procedure is relatively simple; once the compound is desorbed, the extract can be analyzed by gas chromatography or other standard analytical techniques. Parts-per-million concentrations in air are usually determined, although parts-per-billion sensitivity can be obtained for some compounds by using large sample volumes and high-sensitivity detectors. Charcoal is the most widely used sorbent while silica gel, alumina, porous polymers, and various gas chromatographic packings are used for specialized applications. Most methods require a sampling pump

to pull the air through the sorbent tube, and small batteryoperated personal pumps are available. Pumpleas dosimeters have recently been developed; they rely on the diffusion or permeation of compounds into a chamber containing a sorbent and will be discussed in the "Chemical Dosimeter" section. Collection and Recovery. In designing the development of methods based on solid sorbents, it is desirable to have some guidelines to predict the collection (breakthrough) of various compounds. Some equations have been developed to theoretically calculate the breakthrough or to relate the collection efficiences to adsorption breakthrough curves or to other properties of the compouuds (Bl-8). The choice of a sorbent is often a compromise between the collection properties and the desorption properties for a particular chemical. For example, charcoal is an excellent sorbent for many compounds, but the recovery of many compounds is poor. Various physical and environmental factors affect collection and recovery and are discussed in the literature (89-13). Several techniques for determining breakthrough are also discussed (B14-17). A detailed description of techniques for preparing simulated atmospheres for determining breakthrough and for validating methods will he given in a separate section of this review. Theoretical treatment of hreakthrough has been suggested (B18)in an attempt to accurately determine concentrations in air even though a large amount of the compound is found on the back-up section. Breakthrough can be related to the concentration of the compound in air, and for charcoal, the empirical Freundlicb isotherm appears to apply ( B 7 3 1 9 ) . This equation takes the form log TB= log a + b log c A straight line results if the log of the breakthrough time TB is plotted against the log of the concentration, C. The line will have the slope b and intercept of log a. Equations developed for the collection of amines on silica gel predict the amount of a compound collected is independent of the inlet concentration and flow rate (B3). A theoretical approach to the collection of compounds on gas chromatographic packings has also been reported ( 8 2 0 ) . For the recovery of compounds from solid sorbents, the desorption efficiency is the most significant of the factors defining the sorptionaolvent desorption system. The most common technique in determining desorption efficiency is to inject the compound or a solution of the compound directly into the solid sorbent (BlO,821-23). The mixture is allowed t o stand overnight and then desorbed and analyzed. Gases and highly volatile compounds are usually introduced as a mixture in air or nitrogen from a plastic bag or cylinder. The percent recovery can be determined in the same manner as the desorption efficiency except a known flow of air is pulled through the sorbent a t the same time to determine the effect of the atmosphere. The closer the test atmosphere is to the actual air to be sampled, the more accurate the recovery factor will be. For many systems the desorption efficiency can be written in terms of an equilibrium constant; it is dependent on the ratio of solvent to sorbent for the distribution of the compound between the two phases. An equation has been derived by Dommer and Melcher (B24) which relates the desorption efficiency to the volume of solvent and the amount of sorbent. The equation assumes the system is in equilibrium and can be approached from either direction. That is, the same desorption efficiency should be obtained when the compound is initially in the solvent or the solid phase. This has been shown to apply to most organic compounds in the concentration range of inerest in industrial hygiene analyses. Posner (B25) further developed this concept and derived several other equations for calculatingthe desorption efticiency for any solid and liquid ratio and for determining the ratio necessary to obtain a desired desorption efficiency. It should be noted that the partition ratio a t equilibrium predicts the optimum desorption efficiency attainable, and other experiments may be necessary to detect nonequilibrium situations. After a solvent sorbent system is selected and tested, using the phase equiibrium technique, direct injections of the test compound are made into collection tubes with and without air being pulled through. If the desorption efficiencies as determined by direct injection are considerably lower than phase equilibrium values, interaction or reaction on the sorbent surface is indicated. If the total recovery from the simulated air ANALYTICAL CHEMISTRY, VOL. 55, NO. 5, APRIL 1983

41 R

INDUSTRIAL HYGIENE

collection is lower than the direct injection efficiency, even though no breakthrough has occurred, hydrolysis, oxidation, or another reaction may be indicated. Krajewski (B23) compared three methods of determining desorption efficiencies to dynamically prepared standards and found the phase equilibrium data somewhat higher in most cases. Evans and Horstman (B26) also showed that the recovery of styrene from dynamically samples tubes was 18% lower than the direct spike method. They later showed that this was not due to a storage problem since they found no significant change in the recovery of styrene after 6 days (B27). Other researchers @IO, B28-29) have reported the effects of coadsorbed compounds on the desorption efficiency. Posner and Okenfuss (B12) studied the effects of the presence of other compounds on the desorption efficiency and reported no effect with nonpolar compounds; however, a major effect was observed with mixtures of polar compounds. Pozzoli et al. (B30) further expanded the concept of the phase equilibrium method by proposing a double elution technique to estimate the desorption efficiency directly on samples collected in the field by performing two successive elutions with the same volume of CS2. The double elution technique was recommended to circumvent environmental conditions that may affect desorption efficiency during sampling and any effect of coadsorbed compounds. Recovery from Charcoal. Since charcoal is such a good sorbent and is readily available, the solution to some sampling problems is to find a way to increase the recovery of the desired compound from charcoal. One way is by increasing the solvent/sorbent ratio as discussed in the phase equilibrium section. Two other approaches are the use of mixed solvents and the two-phase solvent system. In general, polar compounds usually show low recoveries from charcoal. By adding several percent of a polar solvent to the carbon disulfide desorbent solvent, recovery is often improved by 10 to 20%. Methanol is the most frequently added polar solvent, and it is usually effective as long as it does not interfere with the gas chromatography. If the methanol/carbon disulfide system is used, the samples should be run within 4 h since methanol reacts with carbon disulfide in the presence of charcoal (B9). In some cases ethanol, butanol, 2-propanol, or acetone has been added to carbon disulfide to increase desorption efficiency. One or two percent acetone in carbon disulfide has been used to increase the recovery of acrylonitrile from charcoal; however, much larger amounts can be used if needed for other compounds since acetone is completely miscible with carbon disulfide. Posner (B13) suggested using methylene chloride with 5% methanol to improve and eliminate variability of desorption efficiencies for polar compound mixtures, and Johansen and Wendelboe (B31) suggested dimethylformamide which can be backflushed before it elutes from the GC column. The mixed solvent technique has limited use for complex mixtures since it is more difficult to chromatograph, precludes determination of the polar solvent added, and may cause additional interference to other compounds present. A two-phase system has been developed by Langvardt and Melcher (B32)which is capable of measuring both polar and nonpolar organic solvents present simultaneously in work environments. The charcoal collection tubes are desorbed with a 50/50 mixture of carbon disulfide and water. After desorption, the water and carbon sulfide layers are analyzed separately. The high recoveries of the polar compounds are attributed to their partitioning into the aqueous phase after desorption from charcoal by carbon disulfide. Not only does the partitioning eliminate interferences of some polar and nonpolar combinations but the partition coefficients give additional qualitative information. Other modifications using an acidic or basic aqueous phase, or other immiscible solvents such as methanol/carbon disulfide, are presently being studied. Once the compounds are desorbed, a more complex extraction/separation scheme can be used before analysis, if necessary (B33). Validation. As can be seen from the discussion, recovery cannot always be predicted from desorption efficiencies and as a result a validation protocol should be followed for each chemical to determine what effect the various parameters have for a specific situation (B9-11,B34-36). Many of the parameters which affect collection have been studied in detail, but the number of variables encountered in actual samples is often too great to rely solely on these extrapolations for 42R

ANALYTICAL CHEMISTRY, VOL. 55, NO. 5, APRIL 1983

untested compounds. High concentration of collected compounds on the backup section of a collection tube can be used to detect deteriorative effects on the collection efficiency. However, if these factors are recognized before or during sampling, modification may be made in the collection procedure which will reduce the number of invalid samples. A false indication of breakthrough may be caused by migration of the compounds collected on the front section to the back-up section over an extended storage period. The exact protocol and amount of testing necessary depend on the extent of the sampling project. Methods for limited surveys, where the sampling conditions and concentrations are defined and relatively constant, can be validated for that specific situation. Long-term or ongoing surveys which involve a variety of compounds, geographical locations, temperatures, humidities, and storage times require a greater collection and recovery data base as well as a continuing quality control program. If additional sampling variables are subsequently introduced, additional data should be obtained. During the development and validation of industrial hygiene air sampling methods in the laboratory, various parameters of collection, storage, and analysis are studied to develop a valid method. Although some environmental parameters such as humidity are studied in the laboratory, it is not feasible to study the environmental variables which could cause difficulties for all sampling situations. Judicious on-site testing during method development and after a successful validation will either uncover unpredictable circumstances or increase confidence. One technique used in on-site testing is comparison to an established (but possibly inconvenient) alternate method. The alternate procedure may be a time-averaging technique such as an impinger, a continuous monitoring technique such as a portable infrared, or use of periodic short-term or instantaneous grab samples. “Field spiking” is also a useful technique where a duplicate sampling system is setup in an on-site area. One of the collection tubes is then spiked with an appropriate amount of the test compound. Identical treatment and analysis of the duplicate tubes will indicate the unknown concentration in the area as well as the recovery factor from the spiked tube by difference. While field spiking may not reveal the specific causes of method problems, it will indicate when a problem exists. Field spiking helps the industrial hygienist avoid unpredictable error, provides a verification of quality, and increases defensibility of monitoring data. Chapmann et al. (B37) described a field spiking technique for quality control purpose during plant surveys. When interpreting field spike data, there is a modified statistical approach to consider. Before a decision can be made whether or not the spike recovery is acceptable, information about the precision of the experiment is necessary. Since a precision factor is involved in the collection and analysis of both the field sample and the spike field sample, the significancein the recovery of the spike depends on the combined error from both samples. In order to determine if the spike recovery indicates a problem, the range of recovery expected from combining the two errors must be determined. The ratio of the amount spiked compared to the amount in the area sample plays a significant part in the statistical treatment. Therefore, when using field spiking for method validation, the ratio should be close to the range of 0.5 to 3 and a minimum of duplicate side-by-side (two spiked and two field) samples should be taken for a suitable statistical treatment. Various selective detectors for chromatography have been used to increase sensitivity and selectivity (B38-41), and infrared has been recommended for the analysis of hydrocarbons after extraction with Freon 113 fluorocarbon (B42). Kolb and Pospisil (B43) described a headspace analysis after desorption with benzyl alcohol and Magnante (B44)used X-ray absorption to determine compounds containing elements, such as halogens, phosphorus, arsenic, and sulfur, directly on the carbon tube. Various types of carbon have been tested for some specific applications (B23,B45-46) and treated carbon has been developed to preserve otherwise unstable compounds (B47). Many of the recent publications report comprehensive surveys using charcoal (B48-49), and probably the most intense evaluation of charcoal tube methods is in conjunction with passive dosimeter field and lab evaluations which will be referenced in a later section. Although CS2 is the most

INDUSTRIAL HYGIENE

common (desorption solvent, a number of alternate ciolvents such as lhexane ( B O ) , acetone (B51), acetonitrile (B38), methylene chloride (B13),dimethylformamide (B30),and Freon 113 (B42) have been used. Special drying tubes have also been designed to be used in front of the charcoal tube to reduce moisture effects and improve collection (B45). Silica Gel. Silica gel is the second most widely used adsorbent. Many compounds which cannot be recovered from charcoal, such as highly polar compounds (B52-54), phenols, amines (133, B55-56), high boiling compounds, and multifunctional compounds, often can be collected and recovered with silica gel. The biggest problem with silica gel is the adsorption of water which may cause desorption and loss of the collec1,ed compounds through frontal elution. The heating effect due to the heat of adsorption may cause polymerization or reaction of some compounds unless an inhibitor i3 added (B57). Often chemicals collected on silica gel require analysis by HPLC (B58,B59) or derivatization prior to analysis. The examples of derivatization will be discussed in a later section. Porouri Polymers. Alternative sorbents for the collection of polar organic compounds which are sensitive to hydrolysis are porous polymers such as the Chromosorb porous polymer column packings, Porapak porous polymer column packings, Tenax-GC column packing, and Amberlite XAD sorbent products. The use of various porous polymers sorbents and desorptioin solvents permits (often necessitates) gas chromatographic detectors other than the most commonly used flame detector or electron capture detector. The use of a selective detector can greatly simplify a difficult analytical problem by reducing interferences and background peaks. Many applications are discussed in the literature (B39, B40, B35, B38, B55, and B60). The use of porous polymers in thermal desorption systems will be discussed in the next section. Chromosorb 101 was used by Mann et al. (B61) for deterin air, Barnes et al. (B62) mining l,f!-dibromo-3-chloropropane collected 36 odor forming compounds on Chromosorb 103 and Tenax GC, Glaser and Woodfin (B63) used Chromosorb 106 to collect 2-nitropropane, and Thomas et al. (B64) collected chlordane with Chromosorb 102. The Porapak porous polymer packings are similar to Chromosorb porous polymers packings and Porapak T was used by Dillon (B65) to collect hexachlorocyclopentadiene, Porapak P was recommended by NIOSH (B66) for collection of diethylcarbamoyl chloride, and acetic anhydride was determined by Qazi and Vincent (B67) in the presence of acetic acid by collecting on Porapak N sorbent. XAD-2 polymeric sorbent is composed of single resin beads consisting of an agglomeration of numerous minute microspheres. A clue to the efficiency of collection while minimizing reactivity may be found in the structure of the resin. The porous structure is of an open-cell type so that water can readily penetrate the pores. During adsorption, the hydrophobic portion of the molecule is preferentially adsorbed on the hydrophobic polystyrene surface of the adsorbent through van de Waals attraction. The compounds being adsorbed do not penetrate substantially into the microsphere but remain adsorbed at the surface thus allowing the adsorbate to be rapidly eluted during the recovery step. XAD-2 resin has been used for a number of compounds (B68-71) and, for example, has been shown valuable for the collection of reactive or an0 thiophosphates (B72). The reactive compounds, O,&diethylphos~phorochloridothioate (DEPCT) and monoethylphosphorodichloridothioate (MEPCT) were collected on XAD-2 resin, and no loss in recovery was observed after humid air (95% R.H.) was passed through. Hydrolysis is observed for these compounds in atmospheres with high humidity, but once collection has occurred, the resin tends to stabilize the compoundls and reproducible recoveries are obtained even after 7 dqys of storage. Langhorst and Nestrick (B6O)developed a method for the collection of the chlorobenzene series on XAD-2 resin followed by analysis with a photoionization detector. In addition, XAD-7 has been used to sample epichlorohydrin and ethylene chlorohydrin (B73)and XAII-4 has been used to collect 5,5,5-tributyl phosphorotrithioate and dibutyl disulfide (B74). Although Tenax GC is widely used in thermal desorption it has also been used in a number of solvent desorption methods (B75, B76) and in combination with polyurethane foam for pesticides and semivolatile organic compounds (B77).

Miscellaneous sorbents have been used for some specialized applications. Langvardt and Melcher (B78) used alumina to collect alkanolamines and Giam et al. devised a method for collection and separation by selective elution of polychlorinated biphenyls and phthalates by using Florisil (B79). Polyurethane foam has been used alone or in cambination with filters or other sorbents for the collection of pesticides and other high boiling compounds (B77,B80-82). A chemically bonded packing was also shown by Melcher et al. (B17)to be excellent sorbents for pesticides and high boiling compounds. The recovery of these compounds is very poor from sorbents such as charcoal, alumina, and silica gel. A small tube containing bonded sorbent can be used for over 8 h for the collection of chlorpyrifos with no breakthrough and 95% recovery. Another advantage of the bonded packing is its low retention of water vapor and volatile solvents. Solvent vapors readily pass through the bonded solvent without increasing the breakthrough of chlorpyrifos. This property reduces interference and allows collection of the solvent vapors on a subsequent charcoal tube in series. Gold et al. (B83)found 13X molecular sieves to be a good sorbent for acrolein as well as for other aldehydes and alcohols. Suzuki and Imai (B84) also use molecular sieves to collect acrolein and subsequently determined it by fluorimetry with o-aminobiphenyl. Another technique is to coat a solid sorbent with acid (B3)or base (B85, B86) to collect basic and acidic compounds.

C. PUMP AND TUBE (THERMAL DESORPTION) With the ever increasing need for the determination of trace organic contaminants in air, the development, testing, and selection of solid sorbents for collection and preconcentration have become very important. Preconcentration constitutes an essential step in the determination of these contaminants which often occur in sub-part-per-million concentrations and in complex mixtures in air. A number of sorbents, such as activated carbon, silica gel, alumina, and porous polymers currently being employed in sampling ambient air and workplace atmospheres, have been discussed above. The above examples using solid sorbents for air sampling involved removing the sorbent from the collection tube, desorbing with a suitable solvent, and analyzing the extract by gas chromatography or other analytical technique. A different approach which is increasing in use, particularly for sub-part-per-millilon concentrations, is thermal desorption of the collected sample. Evaluation of Sorbents. In the thermal desorption technique, samples are collected by pulling air through a tube containing a thermally stable sorbent bed. Instead of removing the sorbent and extracting with a solvent for analysis, the tube is heated and the adsorbent compounds are purged directly into a gas chromatograph. Thermal desorption eliminates use of solvents and other handling operations and is more sensitive than solvent desorption techniques, and the collection tubes are reusable. The main advantage of this technique is the high sensitivity obtained since the total sample collected in 1-3 L of air can be injected at one time. Sensitivity is in the low parts-per-billion range for most compounds. Another iinportant advantage, especially for complex mixtures, is the absence of the solvent peak which is a major interference whien low levels are being determined. One disadvantage of thermal desorption is the “one-shot” nature of the analysis. Multiple samples must be taken in order to run duplicate analyses and/or to examine by more than one technique. An unpredicted interference or instrumental miscue results in a lolst sample. However, with the advancement of computer data handling and storage, this becomes less of a problem. One desorption apparatus, to be discussed later, allows multiple injections of the same sample. A number of factors must be considered when selecting a sorbent for use with thermal desorption: (1)suitable collection properties, (2) thermal stability with repeakd use, (3) low background contaminants during desorption, (4) minimal decomposition or reaction during collection, storage or analysis. A wide variety of sorbents have been evaluated which include: activated charcoal and synthetic carbons; porous polymers such as Tenax, Chromosorb series, Porpack series, and XAD series; liquid phase coated GC packings and bonded GC packings. The parameters which are often used in evaluating the collection properties of sorbents are the retention volume (peak maxima) and the breakthrough volume (first detectable ANALYTICAL CHEMISTRY, VOL. 55, NO. 5, APRIL 1983

43 R

INDUSTRIAL HYGIENE

loss). There are three techniques which can be used for determining the retention volume or breakthrough volume: 1. Temperature Extrapolation. The sorbent to be tested is packed into a collection tube and connected inside a gas chromatograph similar to a GC column. A compound is injected onto the column and the GC retention volume is taken as the product of the flow rate and elution time. Retention volumes are determined at several temperatures and the log of the retention volume is plotted vs. the reciprocal of the temperature (l/K), This plot is then extrapolated to determine the retention volume at ambient temperatures. 2. Disappearances of Vapor During Purging. Collection tubes are loaded with a measured quantity of a compound and then purged with known volumes of air. The tubes are then desorbed and the amount lost is determined by comparison to unpurged tubes. 3. Measurement of Breakthrough. The breakthrough of compounds under sampling conditions are monitored by analyzing the effluent from the tube directly using a GC detector or by attaching a back-up tube which is changed periodically and analyzed. The breakthrough volumes determined by all of these methods are senerallv in agreement. Evaluation of various sorbents for