Anal. Chem. 2000, 72, 4205-4211
Sampling of Trace Volatile Metal(loid) Compounds in Ambient Air Using Polymer Bags: A Convenient Method Karsten Haas and Jo 1 rg Feldmann*
Department of Chemistry, University of Aberdeen, Meston Walk, Old Aberdeen, AB24 3UE, Aberdeen, Scotland, U.K.
The sampling of volatile metal(loid) compounds (VOMs) such as hydrides, methylated, and permethylated species of arsenic, antimony, and tin is described using Tedlar bags. Advantages as well as limitations and constraints are discussed and compared to other widely used sampling techniques within this area, namely, stainless steel canisters, cryotrapping, and solid adsorbent cartridges. To prove the suitability of Tedlar bags for the sampling of volatile metal(loid) compounds, series of stability tests have been run using both laboratory synthetic and real samples analyzed periodically after increasing periods of storage. The samples have been stored in the dark at 20 °C and at 50 °C. Various volatile arsenic species (AsH3, MeAsH2, Me2AsH, Me3As), tin species (SnH4, MeSnH3, Me2SnH2, Me3SnH, Me4Sn, BuSnH3), and antimony species (SbH3, MeSbH2, Me2SbH, Me3Sb) have been generated using hydride generation methodology and mixed with moisturized air. Three static gaseous atmospheres with concentrations of 0.3-18 ng/L for the various compounds have been generated in Tedlar bags, and the stability of the VOMs has been monitored over a period of 5 weeks. Sewage sludge digester gas samples have been stored only at 20 °C for a period of 48 h. Cryotrapping GC/ICPMS has been used for the determination of the VOMs with a relative standard deviation of 5% for 100 pg. After 8 h, the recovery rate of all the compounds in the air atmospheres was better than 95% at 20 and 50 °C, whereas the recovery after 24 h was found to be between 81 and 99% for all VOMs at 20 and 50 °C except for Me3Sb and Me3As. These species show a loss between 48 and 73% at both temperatures. After 5 weeks at 20 °C, a loss of only 25-50% for arsine and stibine and the above-mentioned tin compounds was determined. Only Me3Sb, Me3Bi, and Me2Te were present in the digester gas sample. After 24 h, losses of 44, 10, and 12%, respectively, could be determined. Given these results, Tedlar bags could even be used, with some limitations, for long-term sampling of air containing traces of VOMs. The loss is more pronounced at higher temperatures. Sampling is the first step within an analytical investigation. Clearly, the process of sampling can have a strong influence on the correctness and quality of any analytical results. Factors such * Corresponding author: (e-mail)
[email protected]. 10.1021/ac000313c CCC: $19.00 Published on Web 08/04/2000
© 2000 American Chemical Society
as representativeness, homogeneity, and stability have to be considered.1 Many sampling devices using different strategies have been employed for the analysis of volatile organic compounds (VOCs). Several articles give a detailed overview of all possible methods described in the literature, under which certain are fairly general whereas others are focused on light hydrocarbons.2-5 Volatile metal(loid) compounds (VOMs) have been identified in various anthropogenic gases such as landfill gas and sewage sludge digester gas.6 In terms of their physical and chemical properties, volatile metal(loid) compounds and volatile organic compounds cannot be totally lumped together. Most VOMs are thermodynamically unstable, thus being prone to degradation of any kind. This feature is important for the choice of an appropriate sampling technique. Possible phenomena causing analyte loss are diffusion, oxidation, hydrolysis, photodecomposition, adsorption, absorption, and heterogeneous surface-catalyzed breakdown. Volatile compounds in general can be sampled using a variety of equipment and techniques. Whole air sampling can be carried out using balloons, in the simplest form, cylinders with inlet and exit valves, internally polished stainless steel canisters, or plastic bags made out of an inert polymer material. Polymer bags (usually Teflon or Tedlar) and stainless steel containers are the most widely used sampling vessels.7,8 Sophisticated coating techniques have increased the field of application for canisters as well as the range of analytes that can be collected.9 Sampling methods involving preconcentration of the analytes use liquid absorbents, cryotrapping,10-12 adsorbent cartridges,13 or impregnated surfaces or fibers (solid-phase microextraction) to (1) Osberghaus, U.; Helmers, E. In Sampling and sample preparation: A practical guide for analytical chemists; Stoeppler, Ed.; Springer: Berlin, 1997; p 72. (2) Dewulf, J.; van Langenhove, H. J. Chromatogr., A 1999, 843, 163-177. (3) Camel, V.; Caude, M. J. Chromatogr., A 1995, 710, 3-19. (4) Helmig, D. J. Chromatogr., A 1999, 843, 129-146. (5) DesTombe, K.; Verma, D. K.; Stewart, L.; Reczek, E. B. Am. Ind. Hyg. Assoc. J. 1991, 52, 136-144. (6) Feldmann, J.; Hirner, A. V. Int. J. Environ. Anal. Chem. 1995, 60, 339359. (7) Schweigkofler, M.; Niessner, R. Environ. Sci. Technol. 1999, 33, 36803685. (8) Wang, Y.; Raihala, T. S.; Jackman, A. P.; John, R. S. Environ. Sci. Technol. 1996, 30, 3115-3117. (9) Pittwell, L. In Encyclopedia of Analytical Chemistry; Townsend, A., Ed.; Academic Press: London, 1995; p 4515. (10) Hirner, A. V.; Feldmann, J.; Goguel, R.; Rapsomanikis, S.; Fischer, R.; Andreae, M. O. Appl. Organomet. Chem. 1994, 8, 65-69. (11) Pecheyran, C.; Quetel, C. R.; Lecuyer, F. M. M.; Donard, O. F. X. Anal. Chem. 1998, 70, 2639-2645.
Analytical Chemistry, Vol. 72, No. 17, September 1, 2000 4205
collect the analytes. Whole air sampling is attractive because of the following features: no risk of breakthrough for the analytes as is the case for trapping techniques, no effect of moisture upon sampling, and the possibility of multiple sample analysis. On the other hand, sampling involving cryotrapping and adsorbents allows large volumes of air to be sampled. Since it is necessary to transport and carry the pump as well as the cryogenic liquid container, these techniques are quite inconvenient to handle. So far, the method most used for sampling VOMs has been cryotrapping. Feldmann and Hirner6 as well as Pecheyran et al.11 used a chromatographic packing (SP-2100 10% on Supelcoport (60/80 mesh) in a U-shaped glass tube immersed in liquid nitrogen as the cryogenic liquid, for the sampling of VOMs in human breath,14 in headspace of a microbial culture,15 in landfill gas,16,17 in sewage gas,18 and in urban air.19 The use of liquid nitrogen on-site, carrying a pump as well as the necessary power supply, is not very convenient and the large amounts of liquid nitrogen needed during a sampling trip represent a major hazard during the transport. Practical experience unfortunately highlights plugging problems when sampling atmospheres with high levels of humidity. Cryotrapping can be performed with or without adsorbents being used. Depending on the trapping temperature, the use of adsorbents may not be necessary, with the advantage that moderate volatilization temperatures can be applied, so that the labile VOCs do not suffer the risk of thermal degradation.20 Usually, a second cryotrap is needed (cryofocusing) in order to allow narrow bands to enter the GC columns and thus to enhance the resolution. Stainless steel containers are officially used in the United States Environmental Protection Agency canister method TO-1421 where VOCs are monitored in urban air. It has to be mentioned that this method describes the use of the canisters in the passive mode (i.e., vacuum filling) to eliminate any introduction of contaminants originating from the pump. Solvents or absorbing liquids cannot be used for the sampling of volatile metal(loid) compounds because the interactions needed for the analyte to be caught in the absorbent are far too high so that a change of the chemical form of the VOMs would be inevitable. For the same reason, adsorbents are not suitable for such unstable compounds. Irreversible adsorption, degradation during the desorption process, and buildup of artifacts22,23 are possible negative results associated with using materials such as Tenax and Porapak, especially when they are stored at ambient (12) Amouroux, D.; Tessier, E.; Pecheyran, C.; Donard, O. F. X. Anal. Chim. Acta 1998, 377, 241-254. (13) Woolfenden, E. Air Waste Manage. Assoc. 1997, 47, 20-36. (14) Feldmann, J.; Riechmann, T.; Hirner, A. V. Fresenius J. Anal. Chem. 1996, 354, 620-623. (15) Andrewes, P.; Cullen, W. R.; Feldmann, J.; Koch, I.; Polishchuk, E.; Reimer, K. J. Appl. Organomet. Chem. 1998, 12, 827-842. (16) Feldmann, J.; Cullen, W. R. Environ. Sci. Technol. 1997, 31, 2125-2129. (17) Feldmann, J.; Koch, I.; Cullen, W. R. Analyst 1998, 123, 815-820. (18) Feldmann, J.; Kleimann, J. Korresp. Abwasser 1997, 44, 99-104. (19) Pecheyran, C.; Lalere, B.; Donard, O. F. X. Environ. Sci. Technol. 2000, 34, 27-32. (20) Schmidbauer, N.; Oehme, M. J. High Resolut. Chromatogr. 1985, 8, 404406. (21) Evans, G. F.; Lumplein, T. A.; Smith, D. L.; Somerville, M. C. J. Air Waste Manage. Assoc. 1992, 42, 1319-1323. (22) Falter, R.; Hintelmann, H.; Quevauviller, P. Chemosphere 1999, 39, 10391049. (23) Cao, X. L.; Hewitt, N. J. Chromatogr., A 1994, 688, 368-374.
4206
Analytical Chemistry, Vol. 72, No. 17, September 1, 2000
temperatures. U.S. EPA method TO-17 is based on sampling onto solid adsorbents.24 Adsorbents might be tailored for a small number of VOMs, as, for example, sequential sampling for the determination of volatile Hg species using a noble metal trap in series with an activated-carbon trap.25 Tedlar bags are inexpensive and easy to use and are suitable for large screening studies. The aim of this study was therefore to test the use of Tedlar bags for the sampling of both air and process gases containing volatile organometal(loid) compounds. The compounds chosen occur in the environment and, in particular, in landfill and sewage gas up to concentrations of 84 ng/L, and they can be reproducibly synthesized in the laboratory in appropriate concentrations. They can then be transferred into Tedlar bags and sampled over a period of time. For a realistic simulation, concentrations in trace amounts (0.3-18 ng/L) have been generated. EXPERIMENTAL SECTION Reagents and Standard Solutions. A few selected compounds of As, Sn, and Sb were chosen to investigate their stability at different temperatures, namely, at 20 °C and at 50 °C. The following chloride salts served as precursors for the generation of the volatile hydrides: AsCl3 (99,9%, Hopkin & Williams), MeAsCl2 (99%, Vichem Chemicals), Me2AsCl (98%, Strem Chemicals, Newburyport, MA), SnCl4 (99,99%, BDH, Poole, England), Me2SnCl2 (97%, Avocado Research Chemicals, Heysham, England), Me3SnCl (95%, Sigma-Aldrich, Steinheim, Germany), Me4Sn (95% Aldrich, Milwaukee, WI) and n-Bu-SnCl3 (95%, Aldrich), and Me3AsO and Me3SbCl2 (both donated by Prof. W. R. Cullen, University of British Columbia, Vancouver). For As and Sb, acidified (HNO3, AnalaR from BDH) aqueous standard stock solutions (10 µg/mL as As, Sb, 1% HNO3) using deionized water (Elga UHQ II; Bucks, England) were prepared, whereas methanol (BDH, GPR grade, 99.5%) was the solvent used for the Sn standards (1000 µg/mL as Sn). Acidified aqueous standard working multispecies standards (1% HNO3) were prepared for each element and the concentrations are the following: (all as metals): 10 ng/mL for As and Sn; 50 ng/mL for Sb. Generation of Static Gas Standards. A standard hydride generation batch reaction, using an alkaline NaBH4 solution as the reducing agent, was used for the generation of the gaseous compounds, which were subsequently purged and collected in a Tedlar bag (Supelco, Bellafonte, PA). Since gas standards were generated containing hydrides whose experimental production conditions differ, mainly in the pH of the reaction medium, two consecutive batch reactions were n carried out to produce one Tedlar bag containing a multielement-multispecies standard. Four different Sb species (SbH3, MeSbH2, Me2SbH, Me3Sb) were formed with only Me3SbCl2 being used as a precursor following hydride generation due to rearrangement and demethylation reactions.26,27 These reactions are most reproducible at pH 7, whereas As and Sn species are most efficiently and reproducibly generated at pH 1-2. (24) McClenny, W. A.; Colon, M. J. Chromatogr., A 1998, 813, 101-111. (25) Sommar, J.; Feng, X.; Lindqvist, O. Appl. Organomet. Chem. 1999, 13, 441445. (26) Dodd, M.; Grundy, S. L.; Reimer, K. J.; Cullen, W. R. Appl. Organomet. Chem. 1992, 6, 2, 207-211. (27) Koch, I.; Feldmann, J.; Lintschinger, J.; Serves, S. V.; Cullen, W. R.; Reimer, K. J. Appl. Organomet. Chem. 1998, 12, 2, 129-136.
Procedure. To a 50-mL gas wash bottle, used as a batch reactor, were added 4-mL portions of the 10 ng/mL As and Sn working solutions, then 1 mL of hydrochloric acid (1 M) (BDH), and finally deionized water to a final volume of 10 mL to reach pH 1. A constant stream of air was purged through the solution for 5 min with a flow rate of 100 mL/min in order to generate a water-saturated atmosphere in the Tedlar bag. A total of 2 mL of a 2% NaBH4 solution (99% purity, Acros Organics, Pittsburgh, PA) was added and purged for another 5 min with air into a 5-L Tedlar bag. This should simulate the concentration of VOMs, which have been found in real samples such as landfill gas or sewage sludge digester gas.28 The transfer was accomplished by using a Swagelok union (Nylon) and a piece of appropriate Tygon tubing (l ) 2 cm) to connect the glass tubing of the reaction vessel to the valve connecting piece. Then the Tedlar bag was detached from the first reaction vessel and connected to the second, which had the same features, now containing 2 mL of Sb standard solution and also topped up to 10 mL with deionized water. The same sequence of venting and purging was applied again. The yield of these reactions was higher than 99%. Analysis of the residual solution revealed no detectable As, Sn, and Sb. The theoretical concentration of the gaseous compounds in the water-saturated air atmosphere were finally 10 ng/L for all the As and Sn compounds, except, MeSnH3, which was added in concentrations of ∼0.3 ng/L in order to check for a concentration effect. If one calculates the corresponding concentrations of the various Sb species relative to the total amount of Sb used (50 ng/L), the following results are obtained (given as an average for the room-temperature experiments): 8, 18, 16, and 9 ng/L for the different stibines, respectively (MeXSbH3-X; X ) 0, 1, 2, 3). For the experimental series carried out at 20 °C, triplicates of the gas standards were prepared; duplicates were prepared for the experiments at 50 °C. Real Samples. Since significantly less sample volume is required if a capillary column is used, it would be preferable to collect the gaseous samples in inert plastic Tedlar bags, provided that the analytes are stable over a relatively long period of time, in order to carry out the measurements flexibly. Sample volumes of 10 L were sampled from the sewage sludge digester tank in a local sewage treatment plant (Inverurie, North of Scotland Water Authority, Inverurie, Aberdeenshire, Scotland). The sewage treatment plant is fed by ∼35 000 households, and its catchment area is free of any metal-processing industry. Only Me3Sb, Me3Bi, and Me2Te were present in the digester gas sample, and the concentrations were estimated to 5 ng/L for Me3Sb and 18 ng/L for Me3Bi. Analytical Procedure. The analytical method of capillary GC/ ICPMS was realized by the on-line coupling of chromatographic separation carried out in a commercial GC oven (GC 95, Ai Cambridge Ltd., London, England) followed by transient multielement detection using a Spectromass 2000 argon ICPMS (Spectro UK Analytical Intruments, Halesowen, England). The ICPMS was run under standard operating conditions (coolant, 15 L/min; auxiliary, 1.8 L/min; nebulizer flow, 0.8-1.0 L/min) with the only modification being that the usually applied analog detection system was changed into pulse counting mode. This instrumental modification significantly improved the background (28) Feldmann, J.; Krupp, E. M.; Glindemann, D.; Hirner, A. V.; Cullen, W. R. Appl. Organomet. Chem. 1999, 13, 739-748.
Figure 1. Schematic of analytical device of CT-capillary GC/ICPMS for the determination of volatile metal(loid) compounds: (1) Fusedsilica capillary cryotrap; (2) CP Sil-5CB Ultimetal cryofocusing trap.
level and stability since only very short dwell times are required for the small peak widths observed in capillary gas chromatography. The gas chromatographic separation for both the air standards and the real samples was run using He at a flow rate of 2.5 mL/min and using a temperature program (70 °C, 1 min, 30 °C/min, 160 °C, 5 min). The chromatographic column used was a CP-Sil 5CB Ultimetall capillary (25 m × 0.53 mm i.d., d ) 1 µm) from Chrompack (Chrompack UK Ltd., London, England). A gas sample of 10 mL in the case of the standard atmospheres and 100 mL for the digester gas samples, taken from the Tedlar bag through the integrated septum, was injected by means of a gastight glass syringe through a six-port stainless steel switching valve (rotor material, polyaryl ethyl ketone/PTFE composite) (Valco, 4C6WE, 1/16 in.; Valco Europe, Schenkon, Switzerland) and was then cryotrapped on a fused-silica capillary (l ) 40 cm) emerged in liquid nitrogen. A trapping temperature of -80 °C, realized in a mixture of acetone and liquid nitrogen was used for the digester gas samples since the high methane content would have lead to blocking of the capillary at liquid nitrogen temperatures. After switching the valve into the inject position, the liquid nitrogen of the cryotrap was removed and the analytes were volatilized and transported with the carrier gas flow. A second cryotrap, the first few centimeters of the column, immersed in liquid nitrogen serves as a cryofocusing trap where the analytes sit on the column as a short sample plug which significantly enhances resolution by reducing the peak width. The analytes were then eluted out of the capillary and subsequently into the plasma in the sequence of their boiling points. To achieve an overall detector response time compatible with the signal and the width of the peaks, the dwell time for each element being measured in the peak hopping mode should be optimized prior to each analysis, depending on the number of elements to be monitored. Dwell times of 50 ms for each mass have been used. Transfer lines (Ultimetal tubing, 0.53 mm i.d., methyldeactivated) were installed from the six-port valve to the cryotrap inlet and from the cryotrap outlet to the top of the capillary column. Figure 1 shows a schematic of the setup. Connections between the different types of tubing and the capillary as well as the coupling of the capillary to the ICPMS torch were realized using appropriate Swagelok (Nylon) unions or reducers as well as PEEK tubing material for ultimate sealing and mechanical support. The gases are transferred into the plasma through a heated transfer tube. Transient signals for ions with m/z values of 75 (As), 120 Analytical Chemistry, Vol. 72, No. 17, September 1, 2000
4207
Figure 2. Chromatograms showing the separation performance of the capillary column. Simultaneous measurement of m/z 75 (As), 103 (Rh as an internal standard), 120 (Sn), and 121 (Sb). Species detected: AsH3 (1), MeAsH2 (2), Me2AsH (3), Me3As (4), SnH4 (5), Me2SnH2 (6), Me3SnH (7), Me4Sn (8), BuSnH3 (9), SbH3 (10), MeSbH2 (11), Me2SbH (12), and Me3Sb (13); 100 pg of each species was injected except for 10-13, which are 80, 180, 160, and 90 pg, respectively.
(Sn), and 121 (Sb) were recorded during the whole chromatographic run as well as m/z 103 (Rh), which is continuously introduced as a nebulized solution and which serves as a continuous internal standard. A simultaneous chromatogram is illustrated in Figure 2. Due to this, plasma effects caused by matrix gases or signals caused by interfering ions can be recognized. In addition to that, the Rh, as is usually the case for an IS, takes into account instrumental drift and instability. Identification was achieved purely by matching the retention times of the analytes with those of the individual standards. The GC/ICPMS is calibrated with freshly prepared gas standards and has exhibited a dynamic range between 0.1 and 600 pg (R2 > 0.9). For 10 mL of a 10 ng/L As standard, i.e., for 100 pg injected, relative standard deviations of 5% were determined. Due to the methane/carbon dioxide matrix of the process gas, the values for the real sample are afflicted with greater uncertainty and lead to RSD values of 5-40% for three replicates. To compare the concentrations and to establish a stability curve, relative sensitivity factors were determined prior to each measurement with Rh used as continuous internal standard. Experimental Setup. The stability of volatile organometal(loid) compounds in air standards and in sewage sludge digester gas was tested out by monitoring the concentrations of the different analytes in the Tedlar bags over a period of 8 weeks for the bags stored at 20 °C and over 5 weeks for those stored at 50 °C. The sewage sludge digester gas samples were stored for a period of 48 h. All samples were stored in the dark. The stability was tested in a water-saturated air atmosphere stored at 20 °C 4208 Analytical Chemistry, Vol. 72, No. 17, September 1, 2000
and at 50 °C and in the digester gas sample stored at 20˚C. After an intensive monitoring during the first 4 days, where the time intervals during the individual measurements ranged from 2 h at the beginning up to 24 h after 2 days, the time intervals increasingly became bigger and reached several days at the end of the observation period. For each Tedlar bag within one individual series, three replicate injections were carried out each time the stability was monitored; e.g., one data point in the final stability graph represents the average of nine individual values (three measurements for each of three gas bags) for the experiments at 20 °C. The test mixtures stored at 50 °C as well as the digester gas sample resulted in six data points since only two duplicate bags were produced. After 60-days storage, the bags were emptied and the adsorbed metal species were washed out by a combination of methanol (5 mL) followed by 5 mL of 3% HNO3 solution. The solutions were analyzed by the given hydride generation procedure in order to investigate the amounts of metal(loid) species absorbed onto the bags surface. RESULTS AND DISCUSSION The gas standards produced in-house can be used for the stability testing of volatile organometal(loid) compounds. Although slight concentration differences (10 ( 1.2 ng/L) between the replicate Tedlar bags exist, the recoveries of most of the compounds show a remarkable closeness. Consequently, replicate recovery values can be given and increase the confidence of the analytical results. Apparently, the
Figure 3. Average recovery rate of SbH3 and Me3Sb stored in Tedlar bags in moisturized air in the dark. Data represent the mean values determined for three replicate gas bags, which are sampled at each time 3-fold containing (8 ng/L SbH3 and 9 ng/L Me3Sb).
small differences in initial concentration do not affect the stability significantly. For example, the three data sets in terms of recovery for stibine stored at 20 °C match well (the averages for the three Tedlar bags exhibit a standard deviation of only 1-5%) and the average recovery and the error bars are plotted in Figure 3. It is equally important to get a low spread of results if repeated injections are performed. RSD values ranged from 1 to 18%, depending on the compounds and the closeness to the limit of detection. As a general feature, the following phenomenon could be observed for all gaseous air standards: each concentration reached a maximum after 6-24 h with varying degrees in showing this effect depending on the analyte compound. It can be assumed that improper mixing of the compounds, i.e., unequal distribution in the total volume of the gas bag, occurred during the early hours. This behavior can possibly be explained by the hydride generation procedure. The performance of a hydride generation system is influenced by the rate of formation of the hydrides. A relatively high rate of production occurs immediately after the introduction of the reducing agent and this rate then decreases with time. The relatively low flow rate of the purge gas implies the establishment of a laminar flow regime which in turn hinders the VOMs from being properly mixed. Tedlar bags stored at 50 °C did not show any improvement as far as this mixing effect is concerned. The enhanced diffusion coefficient of the gas molecules as well as the increased kinetic energy that normally leads to more molecule interactions and better distribution seem to have no effect. Common to all compounds is a relatively fast decrease in concentration during the first days of storage and a certain leveling out after this initial period where the compounds seem to be degraded but much more slowly than during the first days. A possible explanation is adsorption or condensation of the less volatile hydrides on the walls of the Tedlar bags (Figure 3). This adsorption effect should not affect the recovery rate to the same extent when working under the same conditions but with higher concentration of the gases. The stability curves of two methylated tin hydrides in different concentrations are shown in Figure 4. The behavior of 0.3 ng of Sn as MeSnH3 per liter showed a similar stability curve to 10 ng/L Sn as Me3SnH per liter. This shows that the adsorption effect is not the major factor of the depletion of volatile tins in this concentration range.
Figure 4. Average recovery rate of MeSnH3 and Me3SnH stored in Tedlar bags in moisturized air in the dark. Data represent the mean values determined for three replicate gas bags, which were analyzed three times at each time. Concentration of the volatile tin compounds: 0.3 ng/L MeSnH3 and 10 ng/L Me3SnH.
Although an initial decrease in the concentration was recorded, there is no doubt that the standard atmospheres were stable for a few days. Recovery rates of up to 90% for virtually all analytes investigated were much better than had been expected from their thermodynamic stability data (Table 1). This in turn would lead to the assumption that a direct sampling technique, such as collecting ambient air or processes gases, is possible. A sample could be taken into the laboratory and analyzed thereafter within a day. By cryotrapping the sample gases on site, the risk of speciation change or loss is virtually nonexistent, but as mentioned earlier, this technique is quite inconvenient and one cryotrap was necessary for each individual analysis whereas samples in Tedlar bags can be subdivided or analyzed repeatedly. Table 1 gives an overview of some recovery rates found after different storage times. It is interesting to ascertain that, after 8 h of storage at 20 °C, virtually the whole gas sample’s integrity is not affected at all. There is even a slightly higher recovery than 100% for the gas standards which may be due to the mixing effect postulated earlier. This time window of 8 h gives the analyst a certain flexibility and allows the analysis of samples from rather remote areas which necessitate a greater delay between sampling and the analysis in the laboratory. Even if the analysis is not possible on the same day, recovery rates of on average 90% are found after 24-h storage at 20 °C for both the air standards and the real sample. Only relatively unstable compounds such as Me3Sb show a significant loss of ∼26% in the standards and of 44% in the digester gas. In addition, a temperature effect is already significant for Me3As and Me3Sb and can be expected for other VOMs such as Me3Bi. This result has to be considered when samples are transported in dark bags in the car under sunny weather conditions. Fields of application include the monitoring of various types of gases such as landfill and sewage gas, car exhaust, or urban air. Air samples can be taken inside buildings such as battery factories in order to monitor workplace atmospheres29 or outside for the detection of anthropogenic contaminants. Although Tedlar is considered to be inert with good nonsorbing properties, various groups could confirm the loss of (29) Hetland, S.; Martinsen, I.; Radziuk, B.; Thomassen, Y. Anal. Sci. 1991, 7, 1029-1032.
Analytical Chemistry, Vol. 72, No. 17, September 1, 2000
4209
Table 1. Average Recovery of Volatile Organometal(loid) Compounds in Moisturized Air Stored in Tedlar Bags in the Darka sewage sludge digester gas
moisturized air atmosphere 8h
24 h
5 wks
compd
20 °C
50 °C
20 °C
50 °C
20 °C
50 °C
AsH3 MeAsH2 Me2AsH Me3As SbH3 MeSbH2 Me2SbH Me3Sb Me3Bi SnH4 MeSnH3b Me2SnH2 Me3SnH Me4Sn MBT-H
101.7 ( 6.4 102.0 ( 2.4 118.7 ( 4.0
87.6 ( 1.0 89.9 ( 4.3 95.9 ( 1.8 115.1 ( 27.0 90.3 ( 7.5 91.0 ( 2.4 99.2 ( 11.5 83.5 ( 10.3
96.9 ( 7.1 90.2 ( 4.9 99.9 ( 17.3
86.1 ( 5.6 73.0 ( 1.8 25.8 ( 3.1
87.2 ( 3.3 96.7 ( 6.8 91.5 ( 3.9 73.6 ( 7.9
95.7 ( 2.6 93.9 ( 1.6 80.5 ( 10.2 69.2 ( 46.5 86.8 ( 8.9 87.8 ( 3.4 76.5 ( 21.5 48.3 ( 34.5
73.1 ( 2.7 83.6 ( 2.0 19.8 ( 4.5 2.2 ( 0.3
46.8 ( 9.5 55.6 ( 0.6 25.3 ( 3.2 14.6 ( 4.1 38.8 ( 4.3 17.1 ( 1.7 1.0 ( 0.4 0.0 ( 0.0
95.0 ( 1.9 78.9 ( 12.7 98.8 ( 1.0 104.2 ( 3.2 98.9 ( 1.4 101.1 ( 5.2
85.5 ( 5.2 82.3 ( 2.5 135.3 ( 32.9 81.4 ( 11.0 85.6 ( 3.6 91.5 ( 4.6
97.2 ( 2.7 73.3 ( 3.0 86.2 ( 0.8 86.1 ( 12.0 84.0 ( 13.0 94.2 ( 4.3
78.0 ( 1.9 75.2 ( 3.7 154.8 ( 31.5 49.1 ( 3.2 64.8 ( 3.2 55.9 ( 5.2
48.9 ( 4.5 94.5 ( 12.1 69.2 ( 0.6 71.2 ( 7.3 69.2 ( 3.1 38.7 ( 2.1
102.6 ( 1.9 109.9 ( 5.0 109.6 ( 1.7 104.9 ( 8.8 65.5 ( 4.2 99.6 ( 5.4 94.9 ( 9.2 110.4 ( 16.8 98.6 ( 2.3 104.9 ( 4.2 107.7 ( 2.4
landfill gas41
8h
24 h
6 wks
21 wks
20 °C
20 °C
20 °C
20 °C
19
