249
Anal. Chem. 1981, 53, 249-253
Volatile Metal-Chelate Sample Introduction for Inductively Coupled Plasma-Atomic Emission Spectrometry Marilyn S. Black Consultant, Indian Valley Trail, Atlanta, Georgia 3034 1
Richard F. Browner' School of Chemistry, Georgia Institute of Technology, Atlanta, Georgia 30332
A technique for inductively coupled plasma (ICP) sample Introduction is presented by which certain metals are converted to volatile Pdiketonates and the resulting metal-chs late vapor is carried by argon into a stabilized inductively coupled argon plasma. The feasibillty of directly reacting Pdiketone ligands and certain metals with minimal sample preparation is discussed with qualltatlve data supporting the direct chelation reaction of Fe, Zn, Co, Mn, and Cr in various matrices (Le., pure metal, National Bureau of Standards (NBS) bovine liver, human blood serum). Quantitative data for the determlnatlons of Fe, Zn, and Cr in bovine liver and in blood serum are presented and compared wlth values obtained by alternative analytical methods. Advantages and limitations of this technique over existing means of ICP Sampie introduction are discussed.
Since the early reports of successful inductively coupled plasma (ICP) analysis (1,2) the technique has rapidly become established as a trace analysis tool in the geological, metallurgical, agricultural, biological, and environmental sciences. The basic work on parametric optimization and instruments development has been thoroughly reviewed (3-8) as have the available applications data (9-13). It is clear that the ICP offers benefits over other methods for trace analysis of various metals in environmental and biological samples. In these samples, natural levels of many elements are inherently low, and often only small amounts of sample are available to the analyst. The capacity of ICP spectrometry for low detection limits, simultaneous multielement analysis, and relative freedom from matrix interferences are of major relevance. Identifiable weaknesses to ICP spectrometry when applied to trace analysis result primarily from current limitations in sample preparation and sample introduction to the ICP, rather than from deficiencies inherent in the ICP itself. Extensive sample pretreatments (wet digestion, fusions, preconcentrations, etc.) are time consuming and possible sources of sample comtamination. The possibility of introducing microsamples to the ICP by graphite (14) or tantalum-furnace (15) volatilization has been reported with improvements in detection limits compared to continuous aqueous sample introduction. Direct introduction of sample to the plasma in a graphite cup (16) may also improve detection capability, but quantitative aspects of organic samples have not yet been fully investigated. The purpose of this study is to evaluate the feasibility of selectively removing metal species from the sample matrix through a direct chelation reaction with fluorinated P-diketones and to use the ICP as a spectroscopic detector for the resulting volatile metal chelate vapor. This approach offers significant advantages over existing techniques, including elimination of extensive sample preparation, use of small sample (solid and liquid), rapid reaction and hence short analysis time, containment of sample reaction in a closed 0003-2700/81/0353-0249$01 .OO/O
atmosphere, and capability of multielement analysis. The gas chromatography (GC) of metals by formation of volatile metal chelates has been successfully demonstrated for various metals and has been applied to numerous matrices (17,18). The technique has been extensively used for Be and Cr determinations with very low detection limits (19-21). However, the chromatographic (GC)/extraction technique has disadvantages including necessity of involved sample preparation, the use of an electron capture detector (ECD) which is nonspecific for the metal and is subject to contamination responses, and the inability to complete multielement analysis for a wide range of metals due to the instability of some metal chelates on GC columns. There have been several reports of directly chelating certain metals from a matrix without prior digestions. One study, in particular, used the concept to quantitatively determine nanogram quantities of Cr in blood plasma (22, 23). The disadvantages of GC/ECD have generated some interest in the use of spectroscopic detectors for the GC of metal chelates. Black et al. (24) successfully employed a microwave-excited plasma as a GC detector for the determination of Cr as its trifluoroacetylacetonate, and Wolf et al. (25) reported the use of an atomic absorption (AA) as a detector for various Cr 0-diketonates. In these studies, complex sample preparations were used and direct chelations were not attempted. This present study involving the use of an ICP as a spectroscopic detector for direct chelation reactions offers a different approach for quantitative metal determinations in a variety of matrices.
EXPERIMENTAL SECTION Instrumental. A Plasmatherm Model 2500-E radio frequency generator, 2.5 kW input, 27.12 MHz fixed frequency, and a
Plasmatherm concentric Model T-1.0 quartz torch were used to generate and stabilize an argon plasma. A spherical mirror with vertical and horizontal micrometer adjustments was used to focus radiation from the plasma onto a Jarrell-Ash 0.5 Ebert monochromator equipped with a RCA 1P28 photomultiplier tube, Keithley picoammeter, and a 100-MV Sargent Welch recorder. The ICP was operated at 1.75 kW with a coolant flow of 17.5 L/min and auxiliary argon flow of 1.2 L/min. The sample introduction argon flow rate varied from 0.9 to 1.5 L/min depending on the specific chelate ligand and sample being introduced. Sample emission signals were monitored 12 mm above the generator coil. The sample vapor introduction system is illustrated in Figure 1. A high temperature, Teflon-lined, stainless steel six-port valve with zero volume fittings (Valco Instruments) was used to direct the flow of argon sample introduction gas. The 10-mL reaction vessel was used to contain the sample being analyzed and the reacting chelate ligand. The vessel was fabricated from borosilicate glass with a removable ground glass top and Teflon stopcocks on the side arms. All connecting lines were 3 mm 0.d. or 6 mm 0.d. Teflon with Teflon Swagelock unions and fittings. The vessel, valve, and lines were wrapped in Variac-controlledheating tape and maintained at 120 "C. As shown in Figure 1, with the valve set to position A the sample introduction Ar gas passed directly 0 1981 American Chemical Soclety
250
ANALYTICAL CHEMISTRY, VOL.
Plasma Torch
53, NO. 2, FEBRUARY 1981
Reaction Flask
n n
Valve
Position A
Position B
Flgure I. Schematic of instrumental ICP and gas flow system.
through the sample vessel to the plasma torch. With the valve set to position B, the sample argon gas went to the plasma torch without passing through the vessel, while auxiliary Ar was used to purge the vessel. Chelate ligand vapors were used to purge the vessel before reactions were effected. A Varian Model 3700 research gas chromatograph equipped with a 69Nielectron capture detector and 1-MW recorder was used for the chromatographic studies. A 0.75-m, 3-mm i.d. Teflon column packed with 10% Se-30 (Applied Science Laboratories) on 80/100 Gas Chrom Q (Alltech Associates) was prepared for GC separations. Chromatographic conditions were as follows: column temperature, 130 "C; injection temperature, 140 "C; detector temperature, 250 O C ; carrier gas flow rate, 42 mL/min. On-column injection was used. Reagents. Reagent grade chemicals were used throughout the study except where otherwise indicated. The chelate ligands trifluoroacetylacetone [H(tfa)] and hexafluoroacetylacetone [H(hfa)] (PCR chemicals, Gainesville, FL) were distilled prior to use and stored in silanized borosilicate glass bottles at 4 "C. Samples of solid metal chelates were prepared by use of previously reported methods (17,23). Standard solutions of metal chelates were prepared by dissolving weighed amounts of pure metal chelates in Nanograde benzene or hexane and diluting volumetrically. Chelation Reaction Procedure. Direct chelation reaction/ICP studies were made with pure metals and/or metal salts, NBS bovine liver (SRM 1577),and pooled samples of human blood serum. Two different approaches were used for the chelation reactions. In the first procedure, the reaction was effected directly in the sample vessel described previously (Figure 1).The sample to be analyzed was placed in the sample vessel, chelate ligand added, and the system closed and heated at 120 OC for 5-10 min. The stopcocks of the vessel were then opened and the resulting vapor was flushed into the plasma with the Ar sample introduction gas. The reaction vessel was purged before the reaction with the appropriate chelate ligand. The second procedure involved the
use of sealed ampules for chelation reactions. Sample and chelate ligands were placed in a 2- or 5-mL reaction ampule. The contents were frozen with dry ice and the ampules sealed with an oxymethane torch. The ampules were placed in an oven [110 OC for H(hfa) reactions and 130 "C for H(tfa)] and heated for various reaction times. After cooling, the ampule contents were refrozen with dry ice and then the ampule opened and placed into the heated reaction vessel of the ICP flow system. After allowing the ampule and contents to reheat and volatilize (10 min), the vessel stopcocks were opened and the metal chelate vapors flushed into the plasma with Ar sample introduction gas. The glass ampules were preleached with chelate ligands at elevated temperatures prior to use; this was extremely important in removing contaminating metals. Samples that were to be analyzed by gas chromatography were reacted with ligand/benzene solution in a sealed ampule (23,24). The bovine liver, as required for the GC analysis, was digested with nitric and sulfuric acids in closed systems as described by Black and Sievers (24). ICP Analysis. The following atomic lines were used in this study: Fe, 259.9 nm; Zn, 213.9 nm; Co, 238.9 nm; Cr, 267.7 nm; Mn, 257.6 nm. The specific atomic line was initially isolated while monitoring the emission spectrum of a liquid sdlution containing the element of interest. Once the atomic line had been selected, the nebulizer was taken out of the system and the heated sample introduction line was connected from the reaction vessel directly to the plasma torch. The plasma was initiated with the coolant and auxiliary gases and allowed to stabilizefor 5 min. Neat chelate ligand (as appropriate for the particular reaction) was placed in the reaction vessel and heated for 5 min. With the sample valve in position A, the Ar sample introduction gas was passed through the reaction vessel, flushing with it the chelate-ligand vapor. The Ar/chelate-ligand vapor mixture was gradually bled into the plasma with needle valve control. While bleeding in this vapor, the Ar auxiliary flow to the torch was steadily adjusted to 1.2 L/min in order to stabilize the plasma. Optimum stabilization was achieved with an auxiliary flow rate of 1.2 L/min and a sample introduction flow rate of -0.9 L/min for H(tfa) solvent and 1.5 L/min for H(hfa) solvent. These sample introduction gas flow rates varied within hO.1 L/min depending on the sample matrix reacted. Once the plasma had stabilized with chelate vapor introduction from the reaction vessel, the emission signal was monitored for background metal present in the system. The sample valve was then switched to position B so that the sample introduction Ar gas would bypass the reaction vessel and go directly to the plasma. The sample/chelate ligand reaction was then carried out in the reaction vessel by one of the two procedures previously discussed. When the reaction was complete, the valve was switched back to position A in order to allow the Ar sample gas to purge the reaction vapors directly into the plasma. When H(tfa) chelate ligand was used, satisfactory plasma stability was maintained by using the direct switching procedure. However, with the more volatile H(hfa) chelate ligand, the plasma often extinguished. N
RESULTS AND DISCUSSION The reactivity of acetylacetone toward alkali and alkaline earth elements has been established, but little data are available concerning direct reactivity of the fluorocarbon fl-diketones with respect to most elements. Preliminary studies both in this and in other laboratories have shown the feasibility of directly reacting certain metals with some fluorinated p-diketones to form volatile metal complexes (22). Qualitative evaluations of the reactivity of Fe, Cr, Zn, Co, and Mn metals toward the fluorinated 0-diketones, H(tfa) and H(hfa), are presented in Table I. These metals were chosen for initial studies from the list of elements which are known to be biologically, nutritionally, and environmentally significant. The chelate ligands, H(tfa) and H(hfa), were chosen for evaluation because of their high volatilities as reactant species and also because of both the excellent thermal and solvolytic stabilities and high volatilities of their known metal chelate complexes. Sealed ampule experiments with finely powdered metals and chelate ligands showed that immediate
ANALYTICAL CHEMISTRY, VOL. 53, NO. 2, FEBRUARY 1981 Table I. Reaction Observations of Various Metals with H(tfa) and H(hfa) in Sealed Ampules metal H(tfa) H(tfa) + heat H(hfa) Fe Zn
a
red solution, red pptC cloudy solution, white ppt co red solution, red-brown ppt Mn yellow-brown solution, yellow pptC Cr yellowbrown solutiona Slow reaction rate. Moderate reaction rate.
red ppt white ppt
251
H(hfa) + heat red ppt white ppt
brown ppt
red solution, red ppt cloudy solution, white ppta red-brown solution
yellow ppt
yellow-brown solutionC
yellow ppt
red-brown ppt
violet ppt yellow solutiona Rapid reaction rate.
green PPt
Table 11. A Comparison of ICP Detection Limits Obtained by Various Methods detection limits for ICP/chelate for ICP-continuous in graphite relative. b nebulization. furnace.d absolute.a metal wavelength, nm ng aid‘ ng/d rigid' 0.086 5 0.002 0.086 Fe 259.9 0.42 2 0.001 0.42 Zn 213.9 0.085 0.7 0.01 0.085 Mn 251.6 0.67 3 0.02 co 238.9 0.67 0.95 1 0.01 Cr 267.7 0.95 Based on 1 d sample size and quantitaExperimentally determined from pure trifluoroacetylacetonate complexes. tive extraction from matrix, Reference 29. Reference 30.
1
Zn c O ~
,Mn
Flgue 3. Reaction profile of Fe in bovine liver with H(tfa) ligand (259.9 nm). Wvdength-
Figure 2. ICP scans of various metals reacted with H(tfa)ligand: Zn, 213.9 nm: Co, 238.9 nm; Mn, 257.6 nm: Fe, 259.9 nm; Cr, 267.7 nm. reaction invariably occurred with metal/ligand pairs. Reaction was apparent from the immediate start of metal dissolution, which was also accompanied by the formation of highly colored solutions. In all cases, reaction was accelerated by heating the ampules to 110-130 “C. The applicability of these direct reactions as a means of metal sample introduction for the ICP was first investigated by directly reacting the powdered metals and chelate ligands a t elevated temperatures in the reaction flask and flushing the product vapor into the plasma. Significant continuous atomic emission signals were obtained for each of the five metals when reacted with each chelate ligand. Figure 2 presents scans obtained during the reactions of the metals, individually, with H(tfa) ligand. Direct Reaction w i t h Biological Samples. An evaluation of the applicability of the direct reaction procedure to more complex matrices was carried out. Qualitative studies were made for the reaction of NBS bovine liver and samples of human blood serum with H(tfa) ligand. After a sample of bovine liver and chelate ligand was reacted in the reaction vessel, the Fe, Zn, and Cr atomic lines were individually scanned to ascertain whether volatile metal chelate introduction into the plasma was taking place. Significant signals were obtained for the three metal chelate complexes. With an excess of H(tfa) ligand added to the reaction flask, the Fe atomic emission signal was monitored to follow the reaction rate. The resulting profile (Figure 3), obtained after allowing
Wavelength
-
Figure 4. ICP scans of Fe (259.9 nm), Zn (213.9 nm), and Cr (267.7 nm) in human blood serum by reaction with H(tfa) ligand.
an equilibration period of 2 min in the heated reaction vessel before flushing the product vapor into the plasma, indicated a rapid chelation of Fe from bovine liver with subsequent metal chelate volatilization. Maximum Fe signal was obtained in 6 min. This emission profile indicates that after a selected optimum reaction period, the resulting volatile metal chelate could be introduced as a “plug” sample to the plasma, resulting in a sharp emission profile. A sharp emission profile was obtained for Fe in bovine liver after allowing an 8-min reaction time prior to flushing the vapors into the plasma. Emission signals obtained from the volatile Fe, Zn, and Cr chelates, formed by the direct reaction of human blood serum and H(tfa), are shown in Figure 4. A different sample was used for each elemental emission scan, reacting 1.0 mL of serum for Cr analysis and 0.5 mL for Fe and Zn analysis with a twofold volume excess of H(tfa) in a sealed ampule. This particular serum sample was found to contain 0.84 pg of Fe,
252
ANALYTICAL CHEMISTRY, VOL. 53, NO. 2, FEBRUARY 1981 -
T a b l ~ I I I . Quantitative Determinations of Fe, Zn,and hNm B Q V h LhZtX meta1
NBSn
Fe 270 t 20 Zn 130 + 1 0
Cr
0.088 t
concn found, pg/g chelation/ICP b
Gcc
252 k 25 265 t 5 112+ 1 5 132 k 3 0.012 0.022 ?: 0.010 0.035 t 0.003
Five determinations (five Certificate of analysis. different samples), Six determinations (three samples, two aliquots of each). Table IV. Fe and Zn Concentrations in Human Blood Serum
metal Fe Zn
ICP/chelate, mean k SD bg/dL)" 68.4 62.3
k
t
5.2 6.5
CDC, mean i SD (rg/dL) 70.2 i 2.33 61.0 k 4.37
a Standard deviation of six determinations. erence 28.
Ref-
0.62 Mg of Zn, and 12 ng of Cr. Quantitative Analysis by Direct Chelate Reaction. The success of the qualitative studies for metals in organic matrices, using direct chelation followed by ICP analysis, led to an evaluation of the quantitative efficacy of the technique. Detection capability of the ICP system for Fe, Cu, Zn, Cr, and Mn as their volatile trifluoroacetylacetonate complexes was evaluated. The chelate/ICP detection limits, as presented in Table 11, represent that amount of metal as the chelate complex required to produce a signal twice that of the standard deviation of the background signal. Quantitative determinations of Fe, Zn, and Cr in NBS bovine liver are presented in Table 111. Sealed ampule reactions with bovine liver and H(tfa) were used for metal chelation prior to introduction to the ICP system. Calibration was by standard additions technique, and a separate sample was used for each metal determination. For comparative GC determinations an acid digestion of the bovine liver was carried out, followed by chelation and liquid extraction of aliquots of the acid digest. The agreement between the chelate/ICP values for Fe and Zn, those obtained by GC and certified NBS values, were acceptable and indicative of the quantitative applicability of the direct chelation/ICP technique for those elements. The mean Fe and Zn values as determined by ICP were 93.3% and 86.2%, respectively, of the NBS mean values. The agreement between the NBS Cr value and those obtained by chelate/ICP and GC were not as close. However, the determination of Cr in biological materials is historically a difficult task. Chromium is normally present in extremely low quantities so that determinations are made in the lowest detection range of available techniques. In addition, analysis capability of a method may depend on the particular Cr species present in the biological sample. Previous investigations have indicated a relationship between the actual form of the Cr complex, sample preparation, and quantitative determination (26,27). A more thorough study of Cr analysis by ICP, involving improved detection capability, sample reaction, and Cr speciation, may gave further insight into analysis discrepancies. Data for Fe, Zn, and Cr concentrations in human blood serum were obtained by direct reaction of serum with excess H(tfa) in a heated ampule, followed by ICP determination. These results are presented in Tables IV and V. The Fe and Zn serum determinations were made by using a special sample collected and prepared by the Center for Disease Control
~
~~~
Table V. Cr Concentrations in Human Blood Serum w e r m w by- Dire& W a t & / I C P sample A
B
C D1 D2 D3 D4
D5
amt Cr found, nglg
+ SD *..
mean
12.3 3.6 9.4
6.7 5.2 5.7 5.5 7.2
... ...
II
6.1 i 0.8
(CDC) in Atlanta, GA. Mean values for Fe and Zn in the same serum sample as reported by the Nutritional Biochemistry Laboratory of CDC (28)are also listed in Table N.Agreement between the ICP/chelate values and the mean CDC values obtained by various methods was excellent. The chromium values of Table V were determined from various pooled serum samples also obtained from CDC. A relative standard deviation of 14% was achieved for five Cr determinations in one sample. Comparison Cr values were not available. These values may be slightly higher than some investigators are presently finding in human serum; however, special precautions were not taken during sample collection and storage to specifically avoid Cr contamination. Trace metal determinations in complex matrices using the chelation/ICP approach has presented certain advantages over existing procedures: (1)extensive sample pretreatment and separation may be avoided; (2) acid digestion or dry ashing of the sample may not be necessary in many instances, avoiding potential sources of sample contamination or loss; (3) formation and volatilization of volatile metal chelates from the sample matrix will allow a degree of selectivity over other potentially interfering metals present in the matrix; (4) the technique, as with hydride or mercury generation techniques, provides an efficient preconcentration step and so allows the possibility of lower limits of detection; ( 5 ) the introduction of samples in the vapor phase bypasses the problems associated with pneumatic nebulization; (6) rapid, multielement analysis may be possible if the ICP is coupled to a directreading spectrometer. The technique clearly also has certain limitations in its present state of development. Precision is variable depending on the specific element and matrix, although this may be improved by further instrumental modification and further study of reaction parameters. Calibration requires the use of a standard additions technique, which can be troublesome when very low levels of elements are to be determined. The extent of chelate/metal extraction has not yet been determined in a wide variety of sample matrices and so the range of sample to which this approach is applicable is not yet known. Finally, the complete range of elements which will react and volatilize quantitatively is not yet determined, but from existing metal P-diketonate stability data it can be assumed that the technique will be applicable mainly to certain transition elements. Nevertheless, for many samples the volatile chelate/ICP approach may be a valuable supplement to conventional pneumatic nebulization sample introduction, where this has proved to be inadequate.
ACKNOWLEDGMENT The authors gratefully acknowledge the Center for Disease Control, Nutritional Biochemistry Division, Atlanta, GA, for supplying the serum samples and Andrew Boorn for his expert advice on ICP atomic emission spectrometry during the course of this work. This work was supported by the National Science Foundation under Grant No. CHE77-07618.
Anal. Chem. 1981, 53, 253-258
LITERATURE CITED Greenfield, S.; Jones, I.V.; Beny, C. T. Analyst (London) 1964, 8 9 , 713. Wendt, R.; Fassel, V. A. Anal. Chem. 1965, 37, 920. Kirkbright, G. F. Analyst (London) 1971, 96, 609. Fassel, V. A.; Kniseley, R. N. Anal. Chem. 1974, 46, lllOA, 1155A. Boumans, P. W. J. M. Phlllps Tech. Rev. 1974, 34, 305. Winefordner. J. D.: FitzaeraM, J. J.: Omeretto, N. A.m . / . Smctrosc. . 1975, 29, 369. Barnes, R. M. Anal. Chem. Fund. Rev. 1976, 48, 106R. Boumans. P. W. J. M. Opt. Pura Apl. 1978, 1 1 , 143. Locke, J. Anal. Chlm. Acta 1980, 113, 3. Dahlquist, R. L.; Knoll, J. W. Appl. Spectrosc. 1976, 32, 1. Barnes. R. M.. Ed. “ADDliCatiOnS of Inductiielv Cou~ledPlasmas to Emission Spectroscopy’:; Franklin Institute Priss: kiladelphla, PA,
1978. Abercrombie, F. N.; Silvester, M. D.; Cruz, R. B. A&. Chem. Ser. 1979, No. 172, 10. b a s , W. J.; Fassel, V.; Grabau, F.; Kniseley, R. N.; Sutherbnd, W. L. Adv. Chem. Ser. 1979, No. 172, 91. Gunn, A. M.; Millard, D. L.; Kirkbright, G. F. Analyst (London) 1976, 103, 1066. Nixon, D. E.; Fassel, V. A.; Kniseley, R. N. Anal. Chem. 1974, 46,
253
(18) Guichon, G.;Pommier, C. “Gas Chromatography in Inorganics and Organometallics”; Ann Arbor Science Publlshers: Ann Arbor, MI, 1973. (19) Ross, W. D.; Sievers, R. E. Sixth International Symposium on Gas Chromatography, Rome, Italy, 1966. (20) E h t r a u t , K. J.; &!est, D. J.; Ross, W. D.; Slevers, R. E. Anal. Chem. 1971, 43, 2003. (21) Black, M. S.;Slevers, R. E. Anal. Chem. 1973, 45. 1733. (22) Sievers, R. E.; Conndly, J. W.; Ross, W. D. J. Gas Chromatogr. 1967 5,241. (23) Hansen, L. C.; Scribner, W. 0.;Gilbert, T. W.; Sievers, R. E. Anal. Chem. 1971, 43, 349. (24) Black, M. S.;Slevers, R. E. Anal. Chem. 1976. 48, 1872. (25) Wolf, W. R. Anal. Chem. 1976, 48, 1717. (26) Wolf W. R. J. Chrornatogr. 1977, 134, 159. (27) Wolf, W. R.; Mertz, W.; Masirmi, R. J. Agric. FocdChem. 1974, 22, 1037. US. (28) Carter, Richard J. “Summary Report: Trace Metals S w e y I”; Department of Health, Education, and Welfare, Center for Disease Control: Atlanta, GA.Jan 1977. (29) Fassel, V. A; Kniseley, R. N. Anal. Chem. 1974, 46, lllOA. (30) “The Guide to Techniques and Applications of Atomic Spectroscopy”; Perkin-Elmer Corp: Norwak. CT.
210. Salin, E. D.; Horlick, G. Anal. Chem. 1979, 51, 2284. Moshier, R. W.; Sievers, R. E. “Gas Chromatography of Metal Chelates”; Pergamon Press: New York, 1965.
RECEIVED for review April 30, 1980. Accepted October 27, 1980.
Analysis of a Workplace Air Particulate Sample by Synchronous Luminescence and Room-Temperature Phosphorescence Tuan Vo-Dinh,’ R. 8. Gammage, and P. R. Martinez’ Health and Safety Research Division, Oak RUge National Laboratory, Oak Ridge, Tennessee 37830
An analysis of a XAD-2 resin extract of a particulate air sample collected at an industrial environment was conducted by use of two simple spectroscopic methods performed at ambient temperature, the synchronous luminescence and room-temperature phosphorescence techniques. Results of the anaiyds of 13 polynuclear aromatic compounds including anthracene, benzo[ a Ipyrene, benzo[8 Ipyrene, 2,3-benzofluorene, chrysene, 1,2,5,&dlbenzanthracene, dibenzthiophene, fiuoranthene, fluorene, phenanthrene, peryiene, pyrene, and tetracene were reported.
The identification and quantification of polynuclear aromatic (PNA) compounds are of extreme importance because many of these compounds are potentially carcinogenic and occur frequently in the environment (1). The development of practical and cost-effective analytical techniques is necessary in order to monitor these compounds routinely. Recently, we have investigated the potential of two relatively new techniques for the analysis of complexes mixtures. These two techniques are the synchronous excitation method (2) and the room-temperature phosphorescence technique (3). This paper reports the analysis of an environmental sample, an air particulate extract from an industrial site, by use of these two analytical techniques. The synchronous luminescence (SL) excitation method is a methodology that can be applied to either fluorescence, e.g., synchronous fluorimetry (SF),or phosphorescence, e.g., synchronous phosphorimetry (SP). On the other hand, the room-temperature phosphorescence (RTP) Present address: A l u m i n u m Company of America, Alcoa,
TN.
0003-2700/81/0353-0253$01 .OO/O
technique is based exclusively on the detection of phosphorescence emission. The SL method, first suggested as a fingerprinting tool (4), gained interest among analytical spectroscopists as a technique for multicomponent analysis (2-9). Recent development of the technique has been reviewed (10). The advantages as well as the limitations of this method have been the topic of several reports (11,12). One of the goals of this study was to evaluate the capabilities as well as the limitations of the SL method practically and realistically by analyzing a complex environmental sample. Since most real-life samples are extremely complex systems,this assay represents a most demanding test for an analytical technique. The analysis deals with a liquid chromatographic (LC) fraction of XAD-2 resin extracts from a particulate air sample collected a t a workplace environment according to the scheme developed for the source assessment sampling system (SASS). The SASS scheme with its associated analytical procedures is a sampling and analytical approach developed by the U.S. Environmental Protection Agency (EPA) for conducting environmental source assessments of the feed, product, and waste streams associated with industrial and energy-related processes (13). Our main emphasis was to gauge how effectively a rapid, easy-to-use and cost-effective technique can be applied for the characterization of PNA compounds. The second method employed in this study is the R T P technique. This technique is also relatively new and is presently under investigation by many workers (14-24 and other references herein). A review of the RTP method is given in ref 24. Recently, the R T P technique was applied to the analysis of a Synthoil sample (25). In this work, the RTP data are given to compare the results with those obtained by the synchronous fluorescence method. Due to the complementary 0 1981 American Chemical Society