Environ. Sci. Technol. 1999, 33, 1534-1537
Performance Evaluation Soil Samples for Volatile Organic Compounds Utilizing Solvent Encapsulation Technology JAMES DAHLGRAN* U.S. Department of Energy, Idaho National Engineering and Environmental Laboratory, Radiological and Environmental Sciences Laboratory, 850 Energy Drive, MS 4149, Idaho Falls, Idaho 83401 CURT THIES Thies Technology, 3720 Hampton, Suite 207, St. Louis, Missouri 63109-1438
A mixture of volatile organic compounds (VOCs) was encapsulated and mixed with a soil to produce a product suitable for use as a double blind source of VOCs in a soil performance evaluation sample. Two independent laboratories analyzed the standard encapsulated VOC/soil mixture for benzene, toluene, ethylbenzene, and xylene by using U.S. EPA SW-846 Method 5035 in conjunction with SW-846 Method 8020. One laboratory received the sample as a single blind standard, while the other laboratory received the sample as a double blind standard. The percent relative standard deviation (%RSD) for triplicate analyses ranged from 2 to 13%. The lowest %RSD was for m/p-xylene (2%) from the sample analyzed as a double blind sample. Analytical results from these pilot studies indicate that it is possible to prepare standard soil samples contaminated with known amounts of VOCs which will enable soil samples to be submitted to environmental analytical laboratories as a truly blind sample.
Introduction For most environmental analytical procedures, demonstrating proficiency of an analytical method is accomplished by utilizing known spiked samples, blanks, surrogate spikes, and appropriate performance evaluation standards. For analysis of volatile organic compounds (VOCs) in soils, the current performance evaluation standards are primarily methanolic solutions that contain target analytes and are spiked into a soil sample immediately prior to analysis. Demand for precise performance evaluation samples for VOCs in soil matrixes has stagnated due to lack of sample preparation technology. Current technology for preparing volatile performance evaluation samples utilize spiking clean soils with solvent standards (1) or vapor fortification methodologies developed by J. Hewitt (2) of the U.S. Army Cold Regions Research and Engineering Laboratory. Private sector companies typically provide analysts a dilute solution of analytes in a solvent, usually methanol. These methanol solutions are introduced either into the analytical technique (purge and trap or headspace analyses) or placed onto sand (3) (used to simulate soil) immediately prior to analysis. The * Corresponding author phone: (208)526-6243; fax: (208)526-2548; e-mail:
[email protected]. 1534
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 33, NO. 9, 1999
introduction of a methanol solution to a soil or analytical technique just prior to analysis does not adequately replicate soil sample handling procedures in the analytical laboratory. The availability of the liquid standards used to spike the simulated soil also provides the analysts with an inappropriate opportunity to analyze the spiking solution. Although vapor-fortified soils provide a means of examining spiked soils that are analogous to soils isolated from the environment, such standards are difficult to disseminate to analytical laboratories as double blind quality control standards. An ideal soil standard that contains VOCs should provide analyte concentrations across the concentration range of 5 µg/kg to 100 µg/g. Vapor fortification methodology can be customized to concentrations below 100 µg/g (4-6), but some target analytes may be lost due to the varying absorbtivity of vastly differing soil matrixes. Early attempts to prepare soil standards by spiking soils with water, which contained the target analytes, were unsatisfactory (7). To create a true VOCs in soil performance evaluation standard see the following: 1. The number and concentration of volatile target analytes of interest must be unknown to the analyst. 2. The target analytes must be provided to the analyst already incorporated into the soil matrix. 3. The volatile components must be protected from potential soil biological activity. 4. The standard must be stable over an extended period of time. 5. The performance evaluation standard must be accurately and reproducibly prepared so as to provide laboratories with a correct assessment of their analytical performance independent of analytical methodology applied. To meet these specifications, a new method of spiking native soil samples is required. The objective of this work is to demonstrate that microencapsulated VOCs provide a method of spiking soils and offer an improved means of assessing precision and accuracy of VOC analyses of contaminated soils reported by different analytical laboratories.
Experimental Section The first step in this study was to prepare microcapsules loaded with different VOC mixtures. The VOC mixtures encapsulated were a 1:1:1 (V:V:V) toluene:p-xylene:ethylbenzene mixture and unleaded gasoline. Microcapsules that contained these VOC mixtures were produced by a complex coacervation (8) encapsulation protocol. This encapsulation technology yields microcapsules with predominantly gelatin shells that, when dry, are excellent barriers to the transport of many VOCs. Complex coacervation encapsulation technology, originally developed to produce microcapsules for carbonless paper (9), is based on the interaction in aqueous media of gelatin with an oppositely charged polyelectrolyte to produce a colloid-rich phase that coats oil droplets dispersed in the system. Figure 1 illustrates schematically the steps of a typical complex coacervation encapsulation protocol. In the first step, the VOC mixture is dispersed in an aqueous gelatin solution to form an oil-in-water emulsion. This is normally done at 40-60 °C, well above the melting point of aqueous gelatin gels. The second step is to add a second polyelectrolyte to the system, such as gum arabic. When the pH of the system is adjusted to 4.0-4.5, the gelatin and second polyelectrolyte carry opposite charges. These oppositely charged molecules interact to form droplets of a liquid, colloid-rich phase called a complex coacervate (9). As shown in Figure 1, the droplets of complex coacervate adsorb on the surface of dispersed oil droplets thereby forming a thin coating around each dispersed droplet of VOC mixture in the system. This occurs spontaneously due to a lowering 10.1021/es9807845 CCC: $18.00
1999 American Chemical Society Published on Web 03/27/1999
FIGURE 1. Coacervation of solvents for addition to soil samples.
FIGURE 2. Typical steps of standard preparation process. of free surface energy. After each droplet of VOC mixture is coated with a film of liquid complex coacervate, the system is cooled (e.g., to 5 °C). This step gels, or begins to solidify, the liquid complex coacervate coating. Glutaraldehyde is added to the system in order to cross-link the gel structure. After this step, the capsules are isolated and dried to form a free-flow powder which, when blended with a soil, forms standard contaminated soil samples. Figure 2 summarizes the steps taken to produce microcapsules and form standard VOC-contaminated soil samples. The first VOC-contaminated soil sample was produced by mixing 1.099 g of dry microcapsules that contained a 1:1:1 (V:V:V) toluene:p-xylene:ethylbenzene mixture with 30.2 g of dry subsurface soil taken from an area surrounding the Idaho National Engineering and Environmental Laboratory (INEEL). The soil used was sieved through a 60-80 mesh screen. The soil/capsule mixture was tumbled in a small vial at room temperature for 20-25 min in order to ensure uniform mixing. VOCs in this preliminary soil standard were analyzed by the U.S. Department of Energy’s Radiological and Environmental Sciences Laboratory utilizing a Hewlett-Packard 5880 Series II flame ionization gas chromatograph. For the toluene,
ethylbenzene, and p-xylene mixture, the column was a 30 m × 0.53 mm J&W Scientific DB-624 capillary column. A subsequent experiment generated microcapsules loaded with an unleaded gasoline. A standard contaminated soil sample was produced by dispersing 10.0 g of microcapsules loaded with the unleaded gasoline into 193.0 g of dry subsurface soil. The soil/capsules mixture was tumbled in a 2 quart, stainless steel, twin shell blender for 6 h. VOCs in the gasoline in soil standard were characterized by a HewlettPackard 5880 Series II flame ionization gas chromatograph. For these analyzes, the column was a 30 m × 0.32 mm J&W Scientific DB-1 fused silica capillary. Each characterization analysis for the VOC content of a standard soil contaminated with the microcapsules was carried out by equilibrating 1 and 5 g samples of the contaminated soil in 10.0 mL of methanol at room temperature in a 100 mL volumetric flask for 30 min. The samples were gently swirled in order to ensure complete wetting of the soil with the solvent. In the case of gasoline in soil, the methanol solution obtained was assayed directly. Standards containing the 1:1:1 toluene/ethylbenzene/p-xylene microcapsules were diluted to volume with acetone prior to analysis. VOL. 33, NO. 9, 1999 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
1535
TABLE 1. Data from Simulated Contaminated Soil Sample That Contains Microcapsules Loaded with a 1:1:1 (V/V/V) Toluene/Ethylbenzene/p-Xylene Mixturea sample weight (g) 1.0 av std dev %RSD 5.0 av std dev %RSD a
TABLE 2. Gasoline in Soil Pilot Performance Evaluation Standard-Single Lab Analysisa toluene
ethylbenzene
sample 1
444 406 407
143 134 130
187 206 168
341 306 303
898 846 841
toluene
ethylbenzene
p-xylene
11.96 0.57 4.8
22.21 1.05 4.7
22.41 1.08 4.8
sample 2
12.00 0.62 5.1
22.78 1.16 5.1
23.02 1.17 5.1
367 402 416
131 130 133
192 198 205
267 286 280
792 793 828
sample 3
387 384 406
131 124 131
197 191 200
298 261 287
814 766 836
av std dev %RSD
402 20 5
132 5 4
194 11 6
292 22 8
824 36 4
Units are milligrams per gram of soil.
Results and Discussion VOC-loaded microcapsules, when mixed with a soil to create a standard, must be prepared in a reproducible manner and remain stable for prolonged periods. For a microcapsule loaded with a VOC mixture to be reproducible over time, the capsule shell or coating must have essentially zero permeability to the VOCs encapsulated. The shell must also be susceptible to water and/or methanol, since analytical methodology used to characterize VOC-contaminated soil involves these solvents. Dry microcapsules shells formed by complex coacervation are able to retain VOCs for prolonged periods as long as they are not subjected to high humidity storage conditions. The first objective was to demonstrate that a mixture of pure solvents could be encapsulated and retained. The solvent mixture encapsulated was a 1:1:1 (V/V/V) mixture of toluene, ethylbenzene, and p-xylene. Three batches of capsules were made in order to examine lot-to-lot reproducibility. Triplicate analyses of 2 g samples of each capsule batch established that the capsules were 95.8 wt % solvent (relative percent standard deviation; 1.0%). When 5.08 g of one capsule sample, fraction passing 60 mesh screen, was combined with 100.2 g of clean INEEL subsurface soil, the contaminated soil produced contained approximately 48.6 mg (estimated by calculation) of toluene/ethylbenzene/p-xylene mixture per gram of soil. To examine homogeneity and accuracy of the 1:1:1 toluene/ethylbenzene/p-xylene standard contaminated soil sample, three 1-g and three 5-g samples of contaminated soil were assayed. Table 1 summarizes the replicate analytical data obtained. Overall variability at one standard deviation was about 5% for all components at either the 1.0 or 5.0 g aliquot size. These data indicate that the concept of preparing a performance standard by utilizing microcapsules is realistic. The sample demonstrated adequate homogeneity at various sample sizes and exhibited relative agreement with the amount of material expected (estimated 48.6 vs 56.5 mg). A second pilot VOC sample involved the encapsulation of gasoline and preparation of soil sample contaminated with gasoline. Four batches of capsules loaded with gasoline were prepared. The mean gasoline loading (weight of gasoline per weight of microcapsules) of these four samples was 0.732 mg/g (standard deviation 0.062 mg/g). Three ∼2 g samples of one gasoline-loaded capsule sample subjected in triplicate to analysis contained 3274 µg toluene/gram capsules (std dev ) 248, %RSD ) 7.6), 4347 µg ethylbenzene/gram capsules (std dev ) 296, %RSD ) 6.8), and 14922 µg m- and p-xylene/ gram capsules (std dev 931, %RSD ) 6.2). Ten grams of gasoline-loaded capsules were combined with 193.0 g of soil to produce a pilot performance evaluation sample for analysis by commercial laboratories. Three approximately 20 g samples of this contaminated soil sample were placed into individual 40 mL, precleaned sampling vials 1536
9
m- and o-xylene p-xylene
benzene
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 33, NO. 9, 1999
a
Units are milligrams per gram of soil.
TABLE 3. Gasoline in Soil Pilot Performance Standard Sent as Double Blind to Second Lab for Analysisa m- and benzene toluene ethylbenzene o-xylene p-xylene sample 1
av std dev %RSD sample 1 confirmation
av std dev %RSD
2.5 U 2.5 U 2.5 U 2.5 U
120 120 99 100
150 150 120 110
270 270 230 250
750 760 610 680
NA NA NA NA
110 10 9 120 120 120 120
130 17 13 200 210 220 210
255 17 6 350 360 380 350
700 60 8 900 940 960 920
120 0
210 7 3
360 12 3
930 22 2
a Units are milligrams per gram of soil. U ) analyzed, not detected. NA ) not applicable.
and shipped to an environmental analytical laboratory for benzene, toluene, ethylbenzene, and xylenes (BTEX) analysis. Each sample was analyzed in triplicate utilizing U.S. EPA SW-846 Method 5035 (section 2.2.1) in conjunction with SW846 Method 8020. Table 2 contains results of these analyses. Another sample was sent as a double blind sample to a second analytical laboratory for analysis. Table 3 summarizes results of these analyses. Both laboratories utilized the same analytical methods. Samples arrived at the laboratories and were analyzed and reported within a 60 day time period. Results obtained independently by the two laboratories agree well, except for a disagreement concerning the quantitation of benzene. This apparent discrepancy in the data is unexplained at this time but may be due to an actual loss of benzene or an incorrect quantitation by either of the analytical laboratories. The analytical data obtained to date demonstrates that microcapsules can be used to prepare soil performance evaluation standards. These example performance evaluation soil standards can demonstrate the ability to measure the quality associated with the entire analytical methodology. This technology is capable of producing performance evaluation samples that contain volatile target organic analytes within well-defined concentration ranges. Variable concentration ranges for numerous target volatile analytes can be provided. The performance evaluation samples are amenable to many of the current U.S. EPA methodologies for the
analysis of environmentally significant VOCs in soils including analyses carried out by thermal desorption techniques. Current efforts are focused on reducing the concentration of target analytes to the 5-200 ppb range and to examine product storage stability. Once through this pilot examination, microcapsules will be created containing analytes including carbon tetrachloride, chloroform, trichloroethane, and tetrachloroethylene. To date, solvents miscible in water and with boiling points below 60 °C are not suitable candidates for this procedure. This potentially eliminates volatile organic compounds such as bromomethane, chloroethane, vinyl chloride, acetone, and methyl isobutyl ketone from these types of standards. At present, reference values for the target analytes are based on estimated values from calculations similar to those presented in this paper. As experience is gained from the process, reference values may be set from calculated values. Soil samples that contain microencapsulated VOCs have been submitted to the Mixed Analyte Performance Evaluation Program, operated by the U.S. Department of Energy, for round robin study utilizing contracted environmental analytical laboratories.
Literature Cited (1) Conversations with R. T. Corporation, Environmental Research Associates and Analytical Products Group. (2) Hewitt, A. D.; Clarence, L. G. Environ. Sci. Technol. 1995, 29, 769-773. (3) Environmental Research Associates Product Catalogue, 1997; p 16. (4) Hewitt, A. D. American Environmental Laboratory 1994; p 5. (5) Hewitt, A. D. J. AOAC Int. 1994, 77, 737. (6) Hewitt, A. D., Jenkins, T. F. Proceedings of Tenth Annual Waste Testing and Quality Assurance Symposium, Washington, DC, 1994. (7) Maskarinec, M. P.; Johnson, L. H.; Bayne, C. K. J. Assoc. Off. Anal. Chem. 1989, 72, 823. (8) Deasy, P. B. Microencapsulation and Related Drug Processes, Marcel Dekker: 1984. (9) Thies, C. Microencapsulation, Kirk-Othmer Encyclopedia of Chemical Technology, 4th ed.; 1995; Vol. 16, pp 628-651.
Received for review July 30, 1998. Revised manuscript received February 11, 1999. Accepted February 18, 1999. ES9807845
VOL. 33, NO. 9, 1999 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
1537
