- Email: [email protected]
[31] Determination of selenium in biological matrices
[31]
SELENIUM IN BIOLOGICAL MATRICES
391
nickel added in concentration of 8/zghiter to 16 serum specimens averaged 97 - 3%; recovery of nickel added in concentration of 200/zg/kg (dry weight) to homogenates of 7 rat kidneys averaged 101 -+ 8%. 11-13 No interferences in the described procedures were observed by additions of Fe, Cu, or Zn (0.5 mM) or As, Au, Ba, Bi, Cd, Co, Cr, Hg, Mn, Pb, or V (50/zM) as soluble salts to serum or urine specimens. 12,~3Standard additions curves, prepared by additions of nickel to specimens of serum, whole blood, urine, and tissue homogenates, were consistently parallel to the slopes of the calibration lines, so that computations by the method of standard additions are unnecessary.
[31] D e t e r m i n a t i o n of S e l e n i u m in Biological M a t r i c e s By S. A. LEwis Introduction
There has been an increasing need for methods to determine selenium in biological matrices because of its toxicological and physiological importance and its potential carcinopreventive and anticarcinogenic roles. There has also been a demand for the determination of selenium species, both organic and inorganic. Many reviews of selenium methodologiesT M J. H. Watkinson, in "Selenium in Biomedicine" (O. H. Muth, J. E. Oldfield, and P. H. Weswig, eds.), p. 97. AVI Publ., Westport, Connecticut, 1967. 2 0 . E. Olson, I. S. Palmer, and E. I. Whitehead, Methods Biochem. Anal. 21, 39 (1973). 3 j. E. Alcino and J. A. Kowald, in "Organic Selenium Compounds: Their Chemistry and Biology" (D. L. Kalyman and W. H. H. Gunther, eds.), p. 1049. Wiley, New York, 1973. 4 W. C. Cooper, in "Selenium" (R. A. Zingaro and W. C. Cooper, eds.), p. 615. Van Nostrand-Reinhold, New York, 1974. 5 A. D. Shrendrikar, Sci. Total Environ. 3, 155 (1974). 6 0 . E. Olson, Proc. Symp. Selenium-Tellurium Environ., Univ. Notre Dame, Indiana May 11-13 (1976). 7 N. T. Crosby, Analyst (London) 102, 225 (1977). 8 T. M. Florence, Talanta 29, 345 (1982). 9 H. Robberecht and R. Van Grieken, Talanta 29, 823 (1982). ~0 S. E. Raptis, G. Kaiser, and G. Trig, Anal. Chem. 316, 105 (1983). II H. J. Robberecht and H. A. Deelstra, Talanta 31, 497 (1984). 12 L. Fishbein, Int. J. Environ. Anal. Chem. 17, 113 (1984). " G. T61g, "Selenium Analysis in Biological Materials: Trace Element Analytical Chemistry in Medicine and Biology," Vol. 3. de Gruyter, Berlin, Federal Republic of Germany, 1984. 14 M. Verlinden, H. Deelstra, and E. Adriaenssens, Talanta 28, 637 (1981). 15 S. A. Lewis and C. Veilion, in "Selenium" (M. Ihnat, eds.). CRC Press, Boca Raton, Florida, in press.
METHODS IN ENZYMOLOGY,VOL. 158
Copyright© 1988by AcademicPress, Inc. All rightsof reproductionin any form reserved.
392
ANALYSIS OF METALS
[3 1]
have been published and the reader is encouraged to scrutinize these. However, it is not the purpose of this chapter to provide a comprehensive review, but rather to cover a few methods that are used to determine selenium in specific biological matrices, emphasizing applicability but at the same time pointing out the potential pitfalls and limitations of these methods. Isotope dilution gas chromatography/mass spectrometry (GC/MS) and fluorometry will be discussed in detail. Using the classification scheme established by the International Federation of Clinical Chemistry, ~6 the GC/MS method can be considered a definitive technique representing the ultimate in quality. Isotope dilution GC/MS employs an ideal internal standard--another isotope of the same element--negating the need for quantitative sample preparation and external standardization. 17Definitive methods are generally considered to be so sophisticated that they are outside the realm of most laboratories. However, with the advent of small, simple-to-use, benchtop mass spectrometers coupled to capillary gas chromatographs, this technique is now both feasible and affordable for routine selenium analyses. Reference methods are those that generally can be considered as representing the best method available for routine use under carefully controlled conditions. The Association of Official Analytical Chemists' (AOAC) fluorometric method for the determination of selenium in plants and foods ~8can be classified as a reference method. This method has been adapted for many other biological matrices; however, these adaptations do not necessarily confer reference method status to these other sample types. Some routine methods may or may not have analytical biases or systematic errors. The following methods fall into this category and will be discussed in less detail: atomic absorption spectrometry using either hydride generation or electrothermal atomization (preferably with Zeeman background correction) and gas chromatography coupled to detection systems other than mass spectrometry. However, emphasis will be placed on methods that have been developed for the analysis of specific biological matrices for selenium using these techniques. In the discussion of all of these methods, the limit of detection, imprecision, and suitability for the determination of selenium in specific matrices will be examined. 16j. Bittner, R. Bolton, J. A. Boutwell, and P. M. G. Broughton, Clin. Chem. 22, 532 (1976). z7 C. Veillon and R. Alvarez, in "Metal Ion in Biological Systems" (H. Sigel, ed.). Dekker, New York, 1983. ~s "Official Manual of Analysis." Association Official Analytical Chemists, Washington, D.C,, 1984.
[31]
SELENIUM IN BIOLOGICAL MATRICES
393
Quality Assurance All aspects of any analytical method, from sampling through method validation, must be incorporated if a quality assurance program is to be successful. Sample collection and subsequent sample manipulation must be documented. Standards and reagents must be chosen carefully and the judicious use of sample and reagent blanks is mandatory. Contamination is not generally a concern with selenium determinations; however, analyte loss because of the volatility of selenium can present a problem especially for biological matrices. Sample containers of borosilicate glass are reported to cause less problems through adsorption and desorption than plastics.19,2° The ideal way to establish the validity of a method is by the analysis of samples by two or more methods that utilize different physical characteristics. However, this procedure is generally beyond the scope of most laboratories. In the author's opinion, the single most effective way of assuring the accuracy and measuring the imprecision of a method is by consistent use of well characterized, matrix-suitable quality control materials with similar analyte concentrations that are treated exactly like samples. A list of available and potentially available quality assurance materials of varying biological matrices is shown in Table I. However, in spite of the availability of these reference materials, many authors do not report the use of suitable reference materials or the establishment of imprecision. Isotope Dilution Gas Chromatography/Mass Spectrometry (GC/MS)
Introduction
Combined GC/MS has the advantages of shorter analysis time and less expensive equipment over other definitive methods such as other mass spectrometry techniques, and neutron activation analysis. This is especially true since the advent of benchtop capillary GC/MS systems. Another limitation is the lack of suitable chelates for many elements of biological interest; however, a suitable chelate and GC/MS method for selenium has been developed. 21 This method has been logically extended to a double-isotope technique for use with nutritional tracer studies--the ,9 j. R. Moody and R. M. Lindstrom, Anal. Chem. 49, 2264 (1977). 20 M. Ihnat, Y. Thomassen, M. S. Wolynetz, and C. Veillon, Acta Pharmacol. Toxicol. submitted for publication (1985). 2t D. C. Reamer and C. VeiUon, Anal. Chem. 53, 2166 (1981).
TABLE I BIOLOGICAL REFERENCE MATERIALSFOR Se
Source a Bowen NBS NBS NBS NBS NBS NBS NBS
NBS NBS NBS NBS NBS NYE NYE IAEA
IAEA IAEA IAEA
IAEA IAEA
IAEA IAEA IAEA
IAEA IAEA NPCC NIES NIES IUPAC d BCR BCR BCR
Reference material Bowen's kale SRM 1566 SRM 1577A SRM 1567 SRM 1568 SRM 1549 SRM 1572 SRM 2670
SRM 909 RM 8419 RM 1589 RM 8431 RM 50 105 108 MA-A-I MA-A-2 H-5 HH-1 H-4 A-11 MA-M2 A-13 V-10 H-8 V-9 TORT- 1 CRM-6 CRM-5 Seronorm e (Batch 1 0 3 ) CRM- 150 CRM-151 CRM-063
Matrix Kale Oyster tissue Bovine liver Wheat flour Rice flour Nonfat milk Citrus leaves Urine Normal Elevated Human serum b Bovine serum Bovine serum c Mixed diet Albacore tuna Serum Urine Copepoda Fish flesh Bone Human hair Animal muscle Milk powder Mussel tissue Animal blood Hay powder Horse kidney Cotton cellulose Lobster Mussel Human hair Lyophilized serum Milk powder (spiked) Milk powder (spiked) Milk powder
Se concentration 0.134/zg/g 2.1 ~g/g 0.71 /zg/g 1.1 /xg/g 0.4/~g/g 0.11/zg/g 0.025/zg/g 30/xg/liter 460/zg/liter 105/xg/liter 16/zg/liter 0.242 gg/g 3.6/zg/g 90.7/zg/liter 49.4/xg/liter 3.0/zg/g 1.7/zg/g 0.0537 g.g/g 1.7/zg/g 0.28/xg/g 0.034/zg/g 2.27 ~g/g 0.24 p.g/g 0.022 gg/g 4.67/~g/g 0.015/xg/g 6.88 ~g/g 1.5/zg/g 1.4/xg/g 91 /zg/liter 0.127/xg/g 0.125/zg/g 0.088/zg/g
o Bowen, Reading, England; NBS, National Bureau of Standards, Gaithersburg, MD; NY, Nygaard & Company AS, Oslo, Norway; IAEA, International Atomic Energy Agency, Vienna, Austria; NRCC, National Research Council of Canada, Ottawa, Canada; NIES, National Institute for Environmental Studies, P.O. Yatabe, Tsukaba Ibaraki, Japan; IUPAC, International Union of Pure and Applied Chemistry; BCR, Community Bureau of Reference, Brussels, Belgium. b R. Alvarez (National Bureau of Standards, Gaithersburg, Maryland), personal communication. c Under development. d See Ref. 20. e US Distributors: Accurate Chemical & Scientific Corporation, Westbury, NY.
[31]
SELENIUM IN BIOLOGICAL MATRICES
395
primary reason for its development. 22 The samples are "spiked" with a known amount of an enriched stable isotope of selenium (82Se) and are digested to destroy the organic matter and acidified with hydrochloric acid to convert the selenium to the tetravalent oxidation state. Undigested lipids are extracted with chloroform. The selenite is reacted with 4-nitroo-phenylenediamine (NPD) to form the nitropiazselenol (Se-NPD) and this chelate is then extracted into chloroform. Aliquots of the chloroform extract are injected into the GC/MS and the individual ions are measured. The total selenium and specific isotopes, if required, are quantified from the isotope ratios, and the enriched stable isotope spike. The method presented herein contains the latest modifications to the original work.
Instrumentation In the development work, a Finnigan 4000 quadrupole GC/MS capable of monitoring multiple ions was used. However, this method is not limited to this equipment. 23
Chromatographic Conditions Injector, 175°; column, 160°; transfer line, 175°; carrier gas, He (20 ml/ min); column, 1.2 m x 2 mm i.d. silanized glass; packing, 3% OV-101 on 100/120 mesh Supelcoport (or equivalent). Sample volume, 5/~l in chloroform.
Reagents Enriched 82Se was obtained in elemental form from Oak Ridge National Laboratory, Oak Ridge, TN. A standard containing approximately 100/~g/g was prepared in hydrochloric acid. The exact concentration was obtained by a reverse-isotope dilution method. 21 The chelating agent, NPD, was obtained from Aldrich Chemical Company, Milwaukee, WI and was prepared by dissolving 1 g in 100 ml of 10% hydrochloric acid and extracting the excess NPD five times with chloroform. The reagent was then stored in the refrigerator. (Note: crystals may form.) Reagent grade nitric acid, phosphoric acid, formic acid, and 30-50% hydrogen peroxide were used. It was found necessary to use Ultrex grade hydrochloric acid in tracer studies because of unidentifiable background interference for the 74Se-NPD peak. Alternatively, extracting reagent grade HC1 several times with chloroform also removes the interference. However, monitoring of D. C. Reamer and C. Veillon, J. Nutr. 113, 786 (1983). z3 M. J. Christensen (Brigham Young University) and W. R. Wolf (U. S. Department of Agriculture), personal communications.
396
ANALYSISOF METALS
[31]
the 748e-NPD peak was only necessary in those studies utilizing 748e as a
tracer.
Digestion The appropriate amount of sample (0.5-10 g or 0.5-20 ml) was weighed or pipetted into 100-ml Kjeldahl flasks containing two glass boiling beads and spiked with a known quantity of 82Se. Five milliliters of concentrated nitric acid and 1 ml of phosphoric acid were added to the sample and allowed to stand at least 1 hr (or overnight) to partially digest the organic material and reduce foaming. The digests were then boiled. When the dark fumes of nitrogen dioxide subsided, hydrogen peroxide (30-50%) was added dropwise from a disposable plastic pipet. Hydrogen peroxide was continually added until the volume was reduced to about 1 ml and no visible nitrogen dioxide fumes were given off. The samples were removed from heat, cooled, 1 ml formic acid added, and then heated until the brown fumes of nitrogen dioxide were given off. The formic acid removed the residual nitric acid by reduction to nitrogen dioxide. Two milliliters of concentrated hydrochloric acid was then added and the samples were boiled gently for 10 min to convert the selenium to the tetravalent state. Ten milliliters of water was added and if lipids were present, the digests were extracted with one or two volumes of 3 - 5 ml of chloroform (10 min on a wrist action shaker). Derivatization with NPD was carried out by adding 0.2 ml of the NPD reagent and shaking for 10 min. The Se-NPD was extracted with chloroform, dried in a vacuum oven (60°, 1/3 atm), and reconstituted with chloroform immediately prior to injection into the GC/MS.
Tuning The individual channels of the multiple ion monitor on the GC/MS are adjusted to the specific Se isotopes using a standard preparation from SeNPD in the natural (unenriched) ratio. This is only needed to assess the tuning of the individual channels.
Discussion In the hands of the personnel at the Vitamin and Mineral Laboratory, USDA, Beltsville, Maryland, this method proved to be precise and extremely versatile for the analysis of total selenium in a wide variety of matrices. These included many foods, human breast milk, infant formulas, plasma and serum, red blood cells, feces, and urine. The sample
[31]
SELENIUM IN BIOLOGICAL MATRICES
397
preparation and digestion does not use perchloric acid, 24 and hence negates the need for a perchloric acid hood and other safety devices; however it is not as suitable for samples with a high fat content. The use of a stable selenium isotope as an internal standard added prior to digestion eliminates the need for quantitative sample recoveries--always a problem because of the volatility of selenium. There has been some criticism that this digestion procedure, as opposed to the perchloric acid procedure, does not completely digest some organic forms of selenium, but in the analysis of some 5000 samples of varying matrices this has not been found to be so. The availability and use of the many quality control materials shown in Table I permitted us to use GC/MS in a routine fashion. There is no good quality control material available for feces. Although we were involved in the development of several reference materials (bovine serum and mixed diet),25,26 we felt there was limited need for a fecal reference material. We did attempt to use BCR CRM 144 sewage sludge. However, because of a high silicon content the matrix proved to be dissimilar to fecal materials. We, therefore, used RM 8431 (mixed diet) on the premise that the matrices are similar. However, the analyte concentrations (especially selenium) for the diet are generally lower than the expected fecal content. We routinely used NBS RM 8419 bovine serum and NBS SRM 1577A bovine liver as a control for red blood cells. However, the IAEA A-13 animal blood could be used. This method had the accuracy one expects from a definitive method and over a 1.5-year period had an imprecision of about 2% at 100 ng/ml. The absolute detection limit is about 50 pg. As the required sample size ranges from 0.5 g (0.5 ml) to l0 g (10 ml), this requirement effectively removes this method as an effective technique for small samples such as encountered in the analysis of pediatric specimens. Fluorometry Introduction The reader is urged to consult the AOAC Official Methods of Analysis for the fluorometric determination of selenium in foods and plant material. The fluorometric method is well suited for many applications. The only D. C. Reamer and C. Veillon, Anal. Chem. 56, 605 (1983). C. VeiUon, S. A. Lewis, K. Y. Patterson, W. R. Wolf, J. M. Harnly, J. Versieck, L. Vanballenberghe, R. Cornelis, and T. C. O'Haver, Anal. Chem. 57, 2106 (1985). 26 N. J. MiUer-Ihli and W. R. Wolf, Anal. Chem. 58, 3225 (1986).
398
ANALYSIS OF METALS
[31]
equipment and resources needed are a fluorometer and a perchloric acid hood. The organic material is digested in a perchloric acid/nitric acid mixture. The selenium must be converted to the tetravalent state and then reacted with 2,3-diaminonaphthalene (DAN) to form the fluorescent piazselenol. This is extracted under subdued light into a suitable organic solvent. Selenium is then quantified by comparing the fluorescence of the unknowns to that of suitably prepared standards (excitation wavelength = 369 nm, emission wavelength = 525 nm). This method has been automated, 27 however the automated procedure does not automate the digestion which is the labor intensive part of the method. Instrumentation and Facilities
1. A fluorometer capable of excitation of 369 nm and emission at 525 nm. 2. A perchloric acid hood. Method
The reader can do no better than to consult the 1984 AOAC, Official Methods of Analysis for details of reagents, digestion, and precise instructions for apparatus, analytical conditions, and procedure. The methods are specifically designed for the analysis of selenium in foods and plants. However, the method has been adapted to determine selenium in a wide variety of matrices, namely, serum, urine, red blood cells, and/or whole blood, milk and infant formulas, feces, and hair. zS-3z The method has also been modified for the analysis of some inorganic matrices such as soils and sediments. 33,34 Discussion
The fluorometric method has the advantage that it is inexpensive to set up; however it is very labor intensive. It is mandatory that the digestion 27 M. W. Brown and J. H. Watkinson, Anal. Chem. Acta 89, 29 (1977). R. C. Ewan, C. A. Bauman, and A. L. Pope, J. Agric. Food Chem. 16, 212 (1968). 29 N. L. Zabel, J. Harland, A. T. Gormican, and H. E. Ganther, Am. J. Clin. Nutr. 31, 850
(1978). 3oA. Geahchan and P. Chambon, Clin. Chem. 26, 1272(1980). 3t j. A. Watkinson, Anal. Chem. 38, 92 (1966). 32AnalyticalMethods Committee,Analyst (London) 1114,778 (1979). 33"The FluorometricDeterminationof Selenium--A Literature Review." Turner Associates, 1972. Turner Associates, Palo Alto, CA. L. Lalonde, Y. Jean, K. D. Roberts, A. Chapdelaine, and G. Bleau, Clin. Chem. 28, 172 (1982).
[31]
SELENIUM IN BIOLOGICAL MATRICES
399
procedure be quantitative-- sometimes a difficult task due to the volatility of selenium. It requires the use of perchloric acid in the digestion mixture necessitating the use of a perchloric acid hood. Again, it is imperative to ensure that all of the selenium is in the tetravalent state prior to reaction with DAN. Reamer and co-workers developed a digestion procedure using a nitric acid/phosphoric acid mixture and this procedure can be used instead of the perchloric acid digestion for the fluorometric determination of selenium. However, this digestion procedure is not good for samples with high lipid content and there can be a positive fluorometric interference from the lipids. 35There are several disadvantages to the fluorometric method. The glassware, etc., used must be rigorously cleaned to avoid contamination with other fluorescing compounds. This method is not very good for very low concentrations of selenium. Detection limits of 10-100 ng have been reported 1°,28,32 and imprecisions range from 2 to 100%. 28,32 However, the consensus of opinion is that above 100 ng/g the percentage RSD should be better than 8%. Recoveries range from 75 to 110%, 28,32and are very dependent on the concentration and the complexity of the matrix. Again, many authors do not report detection limits, imprecisions, and recoveries, or the consistent use of reference materials. In the proper hands, with close attention to detail, and the enforcement of a good quality assurance program, the fluorometric method for selenium determinations is probably the simplest, least expensive, and most versatile method available. However the limitations of imprecision and detection limits must be realized.
Atomic Absorption Spectrometry (AAS) The development of AAS in 1955 by Walsh 36 brought about a revolution in the field of trace element analysis. Flame AAS can only be used for samples containing high concentrations of selenium as the detection limit varies from 0.05 to 3 mg/liter. Generation of volatile selenium hydride coupled to an atomic absorption spectrometer, and electrothermal atomic absorption spectrometry (especially with graphite furnace atomization) can both be used to determine selenium at a much lower concentration. Hydride generation atomic absorption spectrometry (HG-AAS) offers the advantage of good sensitivity and relatively simple instrumentation for the determination of selenium in a wide variety of matrices. In HGAAS, sodium tetraborohydride is added to an acidic solution of the sample containing selenium in the tetravalent state forming the gaseous selenium hydride, which is then stripped from solution with a gas and 35 D. C. Reamer and C. Veillon, Anal. Chem. 54, 1605 (1983). 36 A. Walsh, Spectrochim. Acta. 7, 108 (1955).
400
ANALYSISOF METALS
[31]
atomized--generally in a heated quartz cell. There are reports that the accuracy of this method is very dependent on the decomposition technique used. Welz and co-workers 37-39 have investigated the problems of sample digestion and have shown that provided adequate care is taken to ensure complete sample digestion and conversion of the selenium to selenite, this technique can be accurate. Siemer and Koteel 4° have examined different HG-AAS techniques and mention methods of optimizing systems to obtain maximum sensitivity. Reamer and co-workers 4~ used radiotracers to evaluate losses in different hydride generation systems. The HG-AAS approach has been used to determine selenium in waters, blood and blood products, urine, environmental samples, foods, and feeds. It is reported not to be a good technique for the analysis of all biological samples for selenium.I° The HG-AAS technique has been shown to have absolute detection limits ranging from 0.1 to 60 ng. 14Imprecision and recovery vary according to care and expertise of the user but have been reported to be 1-5% RSD and 70-110%, respectively.
Graphite Atomic Absorption Spectrometry (GFAAS) Graphite furnace atomic absorption spectrometry would seem to be the most appropriate technique to analyze samples for small amounts of selenium. As many semiautomated systems are now available, this technique could be used for the rapid direct determination of selenium in large batches of routine samples. However, GFAAS is not simple nor free from interferences--both spectral and matrix--and/or losses due to the volatility of selenium. The use of pyrolytically coated or total pyrolytic graphite tubes, a L'vov platform, peak area integration, matrix modification, and Zeeman background techniques (ZAAS) have enabled some advances to be made. Spectral interferences from iron and/or phosphate are compensated for by Zeeman background correction. The use of matrix modification--generally nitrates of Ni, Cu, Pd, Pt, Ag, Mo, and/or M g - help to thermally stabilize selenium and allow for ashing temperatures of up to 1200°. 1°A4'42-44 However, only a few sample types are suitable for 37 B. Welz, M. Melcher, and G. Schlemmer, in "Trace Element Analytical Chemistry in Medicine and Biology," Vol. 3. de Gruyter, Berlin, Federal Republic of Germany, 1984. 3s B. Welz, M. Melcher, and J. Neve, Anal. Chem. Acta 165, 131 (1984). 39 B. Welz and M. Verlinden, submitted for publication. 4o D. Siemer and P. Koteel, Anal. Chem. 49, 1096 (1977). 41 D. C. Reamer, C. Veillon, and P. T. Tokousbalides, Anal. Chem. 53, 245 (1981). 4z W. Slavin, G. R. Carnrick, D. C. Manning, and E, Pruszkowska, At. Spectrosc. 4, 69 (1983). 4~ R. D. Ediger, At. Absorpt. Newsl. 14, 127 (1975). 44 F. J. Fernandez, S. A. Myers, and W. Slavin, Anal. Chem. 52, 721 (1980).
[31]
SELENIUM IN BIOLOGICAL MATRICES
401
direct analyses even with the use of matrix modification. Blood, serum, semen, yeast, and food supplements and waters have been analyzed directly) °,14,42A ZAAS method was shown to correlate well with the definitive IDMS method (r = 0.987) for the analysis of 30 plasma samples for selenium, but was less precise. 45 Many commonly used digestion procedures and acids have been reported to cause matrix interferences for G F A A S . 46 Pretreatments involving the formation of selenium chelates and subsequent analysis by GFAAS again raise the question of quantitative sample recoveries. Contamination, generally a problem with many GFAAS techniques, is of less concern for selenium. GFAAS has good detection limits (20-50 pg) and reported imprecisions ranging from 2 to 15% RSD.
Gas Liquid Chromatography (GLC) The determination of selenium by GLC is usually based on the detection of a volatile piazselenol formed by the reaction of selenite, after destruction of the organic matrix, with an aromatic diamine, like 4-nitroo-phenylenediamine. This again presents the problem of quantitative sample recoveries and conversion to the tetravalent state. Various detectors l° have been used with this system, including flame ionization, thermal conductivity, atomic absorption spectrometry, mass spectrometry (see mass spectrometry method), and electron capture detectors (ECD). 47,48 When using GC/ECD, the absolute detection limit can be as low as 1 pg. Imprecisions are reported to be from 2 to 4% RSD. 48 This method has been adapted for the determination of selenium in the following matrices: water, blood, and blood products, grain, fish, milk, hair, urine, and soft tissues) °,47,48 The method has also been used to detect selenium species in biological materials and in waters. 9 Summary The definitive IDMS method has the advantage of good detection limits and precision. The advent of the new, smaller, easy-to-use GC/MS system should make this method more popular. The fluorometric method remains the method of choice for the determination of selenium in many matrices if the limitations of the method are understood. The newer methods include gas chromatography, HG-AAS, and GFAAS. The gas chromatography methods show promise because of the 45 W. Slavin and D. C. Manning, Prog. Anal. At. Spectrosc. 5, 243 (1982). S. A. Lewis, N. S. Hardison, and C. Veillon, Anal. Chem. 58, 1272 (1986). 47 T. P. McCarthy, B. Brodie, J. A. Milner, and R. F. Bevill, J. Chromatogr. 225, 9 (1981). C. J. Cappon and J. C. Smith, J. Anal. Toxicol. 2, 114 (1978).
402
ANALYSIS OF METALS
[32]
realtively simple instrumentation needed and the fact that the analyte is separated from the matrix. HG-AAS offers good sensitivity provided care is taken to ensure complete sample digestion and conversion of selenium to selenite. The advent of Zeeman background correction systems for GFAAS has greatly facilitated selenium determinations, particularly in biological matrices where iron and phosphorus are also present. The reference materials now available, used as part of a quality assurance program, should help to ensure accurate determinations, permit method validation, and allow performance evaluation in intedaboratory trials.
[32] V a n a d i u m By D O N N A
M. MARTIN
and N.
DENNIS CHASTEEN
Introduction Studies of the metabolism, biochemistry, clinical pathology, and environmental toxicology of vanadium require reliable analytical techniques for the determination of this element in biological materials. The ideal technique for vanadium analysis would permit trace amounts to be determined rapidly and economically but with sufficient sensitivity to detect normal and subnormal levels. Because vanadium concentrations in biological materials can vary widely, all these requirements cannot be met with a single analytical method. In the present chapter, we survey the most common techniques for determining vanadium and discuss some of the advantages and disadvantages of each. Often the instrumentation at hand dictates the analytical method to be employed. While most of this chapter is devoted to analysis of tissues and fluids, the methods are readily adapted to other types of samples. To our knowledge, methods for vanadium analysis of biomaterials have not been reviewed previously. Since much of the earlier work on vanadium analysis, particularly at the sub parts per million level, is of questionable reliability, this review focuses on the more recent literature. Atomic Absorption Spectrometry (AAS) Graphite furnace atomic absorption is the most conventional and economical technique for analyzing vanadium in biological samples in the pbb range. Analysis requires only a small amount of sample and matrix matchMETHODS IN ENZYMOLOGY, VOL. 158
Copyright © 1988by AcademicPress, Inc. All rightsof reproductionin any form reserved.
SELENIUM IN BIOLOGICAL MATRICES
391
nickel added in concentration of 8/zghiter to 16 serum specimens averaged 97 - 3%; recovery of nickel added in concentration of 200/zg/kg (dry weight) to homogenates of 7 rat kidneys averaged 101 -+ 8%. 11-13 No interferences in the described procedures were observed by additions of Fe, Cu, or Zn (0.5 mM) or As, Au, Ba, Bi, Cd, Co, Cr, Hg, Mn, Pb, or V (50/zM) as soluble salts to serum or urine specimens. 12,~3Standard additions curves, prepared by additions of nickel to specimens of serum, whole blood, urine, and tissue homogenates, were consistently parallel to the slopes of the calibration lines, so that computations by the method of standard additions are unnecessary.
[31] D e t e r m i n a t i o n of S e l e n i u m in Biological M a t r i c e s By S. A. LEwis Introduction
There has been an increasing need for methods to determine selenium in biological matrices because of its toxicological and physiological importance and its potential carcinopreventive and anticarcinogenic roles. There has also been a demand for the determination of selenium species, both organic and inorganic. Many reviews of selenium methodologiesT M J. H. Watkinson, in "Selenium in Biomedicine" (O. H. Muth, J. E. Oldfield, and P. H. Weswig, eds.), p. 97. AVI Publ., Westport, Connecticut, 1967. 2 0 . E. Olson, I. S. Palmer, and E. I. Whitehead, Methods Biochem. Anal. 21, 39 (1973). 3 j. E. Alcino and J. A. Kowald, in "Organic Selenium Compounds: Their Chemistry and Biology" (D. L. Kalyman and W. H. H. Gunther, eds.), p. 1049. Wiley, New York, 1973. 4 W. C. Cooper, in "Selenium" (R. A. Zingaro and W. C. Cooper, eds.), p. 615. Van Nostrand-Reinhold, New York, 1974. 5 A. D. Shrendrikar, Sci. Total Environ. 3, 155 (1974). 6 0 . E. Olson, Proc. Symp. Selenium-Tellurium Environ., Univ. Notre Dame, Indiana May 11-13 (1976). 7 N. T. Crosby, Analyst (London) 102, 225 (1977). 8 T. M. Florence, Talanta 29, 345 (1982). 9 H. Robberecht and R. Van Grieken, Talanta 29, 823 (1982). ~0 S. E. Raptis, G. Kaiser, and G. Trig, Anal. Chem. 316, 105 (1983). II H. J. Robberecht and H. A. Deelstra, Talanta 31, 497 (1984). 12 L. Fishbein, Int. J. Environ. Anal. Chem. 17, 113 (1984). " G. T61g, "Selenium Analysis in Biological Materials: Trace Element Analytical Chemistry in Medicine and Biology," Vol. 3. de Gruyter, Berlin, Federal Republic of Germany, 1984. 14 M. Verlinden, H. Deelstra, and E. Adriaenssens, Talanta 28, 637 (1981). 15 S. A. Lewis and C. Veilion, in "Selenium" (M. Ihnat, eds.). CRC Press, Boca Raton, Florida, in press.
METHODS IN ENZYMOLOGY,VOL. 158
Copyright© 1988by AcademicPress, Inc. All rightsof reproductionin any form reserved.
392
ANALYSIS OF METALS
[3 1]
have been published and the reader is encouraged to scrutinize these. However, it is not the purpose of this chapter to provide a comprehensive review, but rather to cover a few methods that are used to determine selenium in specific biological matrices, emphasizing applicability but at the same time pointing out the potential pitfalls and limitations of these methods. Isotope dilution gas chromatography/mass spectrometry (GC/MS) and fluorometry will be discussed in detail. Using the classification scheme established by the International Federation of Clinical Chemistry, ~6 the GC/MS method can be considered a definitive technique representing the ultimate in quality. Isotope dilution GC/MS employs an ideal internal standard--another isotope of the same element--negating the need for quantitative sample preparation and external standardization. 17Definitive methods are generally considered to be so sophisticated that they are outside the realm of most laboratories. However, with the advent of small, simple-to-use, benchtop mass spectrometers coupled to capillary gas chromatographs, this technique is now both feasible and affordable for routine selenium analyses. Reference methods are those that generally can be considered as representing the best method available for routine use under carefully controlled conditions. The Association of Official Analytical Chemists' (AOAC) fluorometric method for the determination of selenium in plants and foods ~8can be classified as a reference method. This method has been adapted for many other biological matrices; however, these adaptations do not necessarily confer reference method status to these other sample types. Some routine methods may or may not have analytical biases or systematic errors. The following methods fall into this category and will be discussed in less detail: atomic absorption spectrometry using either hydride generation or electrothermal atomization (preferably with Zeeman background correction) and gas chromatography coupled to detection systems other than mass spectrometry. However, emphasis will be placed on methods that have been developed for the analysis of specific biological matrices for selenium using these techniques. In the discussion of all of these methods, the limit of detection, imprecision, and suitability for the determination of selenium in specific matrices will be examined. 16j. Bittner, R. Bolton, J. A. Boutwell, and P. M. G. Broughton, Clin. Chem. 22, 532 (1976). z7 C. Veillon and R. Alvarez, in "Metal Ion in Biological Systems" (H. Sigel, ed.). Dekker, New York, 1983. ~s "Official Manual of Analysis." Association Official Analytical Chemists, Washington, D.C,, 1984.
[31]
SELENIUM IN BIOLOGICAL MATRICES
393
Quality Assurance All aspects of any analytical method, from sampling through method validation, must be incorporated if a quality assurance program is to be successful. Sample collection and subsequent sample manipulation must be documented. Standards and reagents must be chosen carefully and the judicious use of sample and reagent blanks is mandatory. Contamination is not generally a concern with selenium determinations; however, analyte loss because of the volatility of selenium can present a problem especially for biological matrices. Sample containers of borosilicate glass are reported to cause less problems through adsorption and desorption than plastics.19,2° The ideal way to establish the validity of a method is by the analysis of samples by two or more methods that utilize different physical characteristics. However, this procedure is generally beyond the scope of most laboratories. In the author's opinion, the single most effective way of assuring the accuracy and measuring the imprecision of a method is by consistent use of well characterized, matrix-suitable quality control materials with similar analyte concentrations that are treated exactly like samples. A list of available and potentially available quality assurance materials of varying biological matrices is shown in Table I. However, in spite of the availability of these reference materials, many authors do not report the use of suitable reference materials or the establishment of imprecision. Isotope Dilution Gas Chromatography/Mass Spectrometry (GC/MS)
Introduction
Combined GC/MS has the advantages of shorter analysis time and less expensive equipment over other definitive methods such as other mass spectrometry techniques, and neutron activation analysis. This is especially true since the advent of benchtop capillary GC/MS systems. Another limitation is the lack of suitable chelates for many elements of biological interest; however, a suitable chelate and GC/MS method for selenium has been developed. 21 This method has been logically extended to a double-isotope technique for use with nutritional tracer studies--the ,9 j. R. Moody and R. M. Lindstrom, Anal. Chem. 49, 2264 (1977). 20 M. Ihnat, Y. Thomassen, M. S. Wolynetz, and C. Veillon, Acta Pharmacol. Toxicol. submitted for publication (1985). 2t D. C. Reamer and C. VeiUon, Anal. Chem. 53, 2166 (1981).
TABLE I BIOLOGICAL REFERENCE MATERIALSFOR Se
Source a Bowen NBS NBS NBS NBS NBS NBS NBS
NBS NBS NBS NBS NBS NYE NYE IAEA
IAEA IAEA IAEA
IAEA IAEA
IAEA IAEA IAEA
IAEA IAEA NPCC NIES NIES IUPAC d BCR BCR BCR
Reference material Bowen's kale SRM 1566 SRM 1577A SRM 1567 SRM 1568 SRM 1549 SRM 1572 SRM 2670
SRM 909 RM 8419 RM 1589 RM 8431 RM 50 105 108 MA-A-I MA-A-2 H-5 HH-1 H-4 A-11 MA-M2 A-13 V-10 H-8 V-9 TORT- 1 CRM-6 CRM-5 Seronorm e (Batch 1 0 3 ) CRM- 150 CRM-151 CRM-063
Matrix Kale Oyster tissue Bovine liver Wheat flour Rice flour Nonfat milk Citrus leaves Urine Normal Elevated Human serum b Bovine serum Bovine serum c Mixed diet Albacore tuna Serum Urine Copepoda Fish flesh Bone Human hair Animal muscle Milk powder Mussel tissue Animal blood Hay powder Horse kidney Cotton cellulose Lobster Mussel Human hair Lyophilized serum Milk powder (spiked) Milk powder (spiked) Milk powder
Se concentration 0.134/zg/g 2.1 ~g/g 0.71 /zg/g 1.1 /xg/g 0.4/~g/g 0.11/zg/g 0.025/zg/g 30/xg/liter 460/zg/liter 105/xg/liter 16/zg/liter 0.242 gg/g 3.6/zg/g 90.7/zg/liter 49.4/xg/liter 3.0/zg/g 1.7/zg/g 0.0537 g.g/g 1.7/zg/g 0.28/xg/g 0.034/zg/g 2.27 ~g/g 0.24 p.g/g 0.022 gg/g 4.67/~g/g 0.015/xg/g 6.88 ~g/g 1.5/zg/g 1.4/xg/g 91 /zg/liter 0.127/xg/g 0.125/zg/g 0.088/zg/g
o Bowen, Reading, England; NBS, National Bureau of Standards, Gaithersburg, MD; NY, Nygaard & Company AS, Oslo, Norway; IAEA, International Atomic Energy Agency, Vienna, Austria; NRCC, National Research Council of Canada, Ottawa, Canada; NIES, National Institute for Environmental Studies, P.O. Yatabe, Tsukaba Ibaraki, Japan; IUPAC, International Union of Pure and Applied Chemistry; BCR, Community Bureau of Reference, Brussels, Belgium. b R. Alvarez (National Bureau of Standards, Gaithersburg, Maryland), personal communication. c Under development. d See Ref. 20. e US Distributors: Accurate Chemical & Scientific Corporation, Westbury, NY.
[31]
SELENIUM IN BIOLOGICAL MATRICES
395
primary reason for its development. 22 The samples are "spiked" with a known amount of an enriched stable isotope of selenium (82Se) and are digested to destroy the organic matter and acidified with hydrochloric acid to convert the selenium to the tetravalent oxidation state. Undigested lipids are extracted with chloroform. The selenite is reacted with 4-nitroo-phenylenediamine (NPD) to form the nitropiazselenol (Se-NPD) and this chelate is then extracted into chloroform. Aliquots of the chloroform extract are injected into the GC/MS and the individual ions are measured. The total selenium and specific isotopes, if required, are quantified from the isotope ratios, and the enriched stable isotope spike. The method presented herein contains the latest modifications to the original work.
Instrumentation In the development work, a Finnigan 4000 quadrupole GC/MS capable of monitoring multiple ions was used. However, this method is not limited to this equipment. 23
Chromatographic Conditions Injector, 175°; column, 160°; transfer line, 175°; carrier gas, He (20 ml/ min); column, 1.2 m x 2 mm i.d. silanized glass; packing, 3% OV-101 on 100/120 mesh Supelcoport (or equivalent). Sample volume, 5/~l in chloroform.
Reagents Enriched 82Se was obtained in elemental form from Oak Ridge National Laboratory, Oak Ridge, TN. A standard containing approximately 100/~g/g was prepared in hydrochloric acid. The exact concentration was obtained by a reverse-isotope dilution method. 21 The chelating agent, NPD, was obtained from Aldrich Chemical Company, Milwaukee, WI and was prepared by dissolving 1 g in 100 ml of 10% hydrochloric acid and extracting the excess NPD five times with chloroform. The reagent was then stored in the refrigerator. (Note: crystals may form.) Reagent grade nitric acid, phosphoric acid, formic acid, and 30-50% hydrogen peroxide were used. It was found necessary to use Ultrex grade hydrochloric acid in tracer studies because of unidentifiable background interference for the 74Se-NPD peak. Alternatively, extracting reagent grade HC1 several times with chloroform also removes the interference. However, monitoring of D. C. Reamer and C. Veillon, J. Nutr. 113, 786 (1983). z3 M. J. Christensen (Brigham Young University) and W. R. Wolf (U. S. Department of Agriculture), personal communications.
396
ANALYSISOF METALS
[31]
the 748e-NPD peak was only necessary in those studies utilizing 748e as a
tracer.
Digestion The appropriate amount of sample (0.5-10 g or 0.5-20 ml) was weighed or pipetted into 100-ml Kjeldahl flasks containing two glass boiling beads and spiked with a known quantity of 82Se. Five milliliters of concentrated nitric acid and 1 ml of phosphoric acid were added to the sample and allowed to stand at least 1 hr (or overnight) to partially digest the organic material and reduce foaming. The digests were then boiled. When the dark fumes of nitrogen dioxide subsided, hydrogen peroxide (30-50%) was added dropwise from a disposable plastic pipet. Hydrogen peroxide was continually added until the volume was reduced to about 1 ml and no visible nitrogen dioxide fumes were given off. The samples were removed from heat, cooled, 1 ml formic acid added, and then heated until the brown fumes of nitrogen dioxide were given off. The formic acid removed the residual nitric acid by reduction to nitrogen dioxide. Two milliliters of concentrated hydrochloric acid was then added and the samples were boiled gently for 10 min to convert the selenium to the tetravalent state. Ten milliliters of water was added and if lipids were present, the digests were extracted with one or two volumes of 3 - 5 ml of chloroform (10 min on a wrist action shaker). Derivatization with NPD was carried out by adding 0.2 ml of the NPD reagent and shaking for 10 min. The Se-NPD was extracted with chloroform, dried in a vacuum oven (60°, 1/3 atm), and reconstituted with chloroform immediately prior to injection into the GC/MS.
Tuning The individual channels of the multiple ion monitor on the GC/MS are adjusted to the specific Se isotopes using a standard preparation from SeNPD in the natural (unenriched) ratio. This is only needed to assess the tuning of the individual channels.
Discussion In the hands of the personnel at the Vitamin and Mineral Laboratory, USDA, Beltsville, Maryland, this method proved to be precise and extremely versatile for the analysis of total selenium in a wide variety of matrices. These included many foods, human breast milk, infant formulas, plasma and serum, red blood cells, feces, and urine. The sample
[31]
SELENIUM IN BIOLOGICAL MATRICES
397
preparation and digestion does not use perchloric acid, 24 and hence negates the need for a perchloric acid hood and other safety devices; however it is not as suitable for samples with a high fat content. The use of a stable selenium isotope as an internal standard added prior to digestion eliminates the need for quantitative sample recoveries--always a problem because of the volatility of selenium. There has been some criticism that this digestion procedure, as opposed to the perchloric acid procedure, does not completely digest some organic forms of selenium, but in the analysis of some 5000 samples of varying matrices this has not been found to be so. The availability and use of the many quality control materials shown in Table I permitted us to use GC/MS in a routine fashion. There is no good quality control material available for feces. Although we were involved in the development of several reference materials (bovine serum and mixed diet),25,26 we felt there was limited need for a fecal reference material. We did attempt to use BCR CRM 144 sewage sludge. However, because of a high silicon content the matrix proved to be dissimilar to fecal materials. We, therefore, used RM 8431 (mixed diet) on the premise that the matrices are similar. However, the analyte concentrations (especially selenium) for the diet are generally lower than the expected fecal content. We routinely used NBS RM 8419 bovine serum and NBS SRM 1577A bovine liver as a control for red blood cells. However, the IAEA A-13 animal blood could be used. This method had the accuracy one expects from a definitive method and over a 1.5-year period had an imprecision of about 2% at 100 ng/ml. The absolute detection limit is about 50 pg. As the required sample size ranges from 0.5 g (0.5 ml) to l0 g (10 ml), this requirement effectively removes this method as an effective technique for small samples such as encountered in the analysis of pediatric specimens. Fluorometry Introduction The reader is urged to consult the AOAC Official Methods of Analysis for the fluorometric determination of selenium in foods and plant material. The fluorometric method is well suited for many applications. The only D. C. Reamer and C. Veillon, Anal. Chem. 56, 605 (1983). C. VeiUon, S. A. Lewis, K. Y. Patterson, W. R. Wolf, J. M. Harnly, J. Versieck, L. Vanballenberghe, R. Cornelis, and T. C. O'Haver, Anal. Chem. 57, 2106 (1985). 26 N. J. MiUer-Ihli and W. R. Wolf, Anal. Chem. 58, 3225 (1986).
398
ANALYSIS OF METALS
[31]
equipment and resources needed are a fluorometer and a perchloric acid hood. The organic material is digested in a perchloric acid/nitric acid mixture. The selenium must be converted to the tetravalent state and then reacted with 2,3-diaminonaphthalene (DAN) to form the fluorescent piazselenol. This is extracted under subdued light into a suitable organic solvent. Selenium is then quantified by comparing the fluorescence of the unknowns to that of suitably prepared standards (excitation wavelength = 369 nm, emission wavelength = 525 nm). This method has been automated, 27 however the automated procedure does not automate the digestion which is the labor intensive part of the method. Instrumentation and Facilities
1. A fluorometer capable of excitation of 369 nm and emission at 525 nm. 2. A perchloric acid hood. Method
The reader can do no better than to consult the 1984 AOAC, Official Methods of Analysis for details of reagents, digestion, and precise instructions for apparatus, analytical conditions, and procedure. The methods are specifically designed for the analysis of selenium in foods and plants. However, the method has been adapted to determine selenium in a wide variety of matrices, namely, serum, urine, red blood cells, and/or whole blood, milk and infant formulas, feces, and hair. zS-3z The method has also been modified for the analysis of some inorganic matrices such as soils and sediments. 33,34 Discussion
The fluorometric method has the advantage that it is inexpensive to set up; however it is very labor intensive. It is mandatory that the digestion 27 M. W. Brown and J. H. Watkinson, Anal. Chem. Acta 89, 29 (1977). R. C. Ewan, C. A. Bauman, and A. L. Pope, J. Agric. Food Chem. 16, 212 (1968). 29 N. L. Zabel, J. Harland, A. T. Gormican, and H. E. Ganther, Am. J. Clin. Nutr. 31, 850
(1978). 3oA. Geahchan and P. Chambon, Clin. Chem. 26, 1272(1980). 3t j. A. Watkinson, Anal. Chem. 38, 92 (1966). 32AnalyticalMethods Committee,Analyst (London) 1114,778 (1979). 33"The FluorometricDeterminationof Selenium--A Literature Review." Turner Associates, 1972. Turner Associates, Palo Alto, CA. L. Lalonde, Y. Jean, K. D. Roberts, A. Chapdelaine, and G. Bleau, Clin. Chem. 28, 172 (1982).
[31]
SELENIUM IN BIOLOGICAL MATRICES
399
procedure be quantitative-- sometimes a difficult task due to the volatility of selenium. It requires the use of perchloric acid in the digestion mixture necessitating the use of a perchloric acid hood. Again, it is imperative to ensure that all of the selenium is in the tetravalent state prior to reaction with DAN. Reamer and co-workers developed a digestion procedure using a nitric acid/phosphoric acid mixture and this procedure can be used instead of the perchloric acid digestion for the fluorometric determination of selenium. However, this digestion procedure is not good for samples with high lipid content and there can be a positive fluorometric interference from the lipids. 35There are several disadvantages to the fluorometric method. The glassware, etc., used must be rigorously cleaned to avoid contamination with other fluorescing compounds. This method is not very good for very low concentrations of selenium. Detection limits of 10-100 ng have been reported 1°,28,32 and imprecisions range from 2 to 100%. 28,32 However, the consensus of opinion is that above 100 ng/g the percentage RSD should be better than 8%. Recoveries range from 75 to 110%, 28,32and are very dependent on the concentration and the complexity of the matrix. Again, many authors do not report detection limits, imprecisions, and recoveries, or the consistent use of reference materials. In the proper hands, with close attention to detail, and the enforcement of a good quality assurance program, the fluorometric method for selenium determinations is probably the simplest, least expensive, and most versatile method available. However the limitations of imprecision and detection limits must be realized.
Atomic Absorption Spectrometry (AAS) The development of AAS in 1955 by Walsh 36 brought about a revolution in the field of trace element analysis. Flame AAS can only be used for samples containing high concentrations of selenium as the detection limit varies from 0.05 to 3 mg/liter. Generation of volatile selenium hydride coupled to an atomic absorption spectrometer, and electrothermal atomic absorption spectrometry (especially with graphite furnace atomization) can both be used to determine selenium at a much lower concentration. Hydride generation atomic absorption spectrometry (HG-AAS) offers the advantage of good sensitivity and relatively simple instrumentation for the determination of selenium in a wide variety of matrices. In HGAAS, sodium tetraborohydride is added to an acidic solution of the sample containing selenium in the tetravalent state forming the gaseous selenium hydride, which is then stripped from solution with a gas and 35 D. C. Reamer and C. Veillon, Anal. Chem. 54, 1605 (1983). 36 A. Walsh, Spectrochim. Acta. 7, 108 (1955).
400
ANALYSISOF METALS
[31]
atomized--generally in a heated quartz cell. There are reports that the accuracy of this method is very dependent on the decomposition technique used. Welz and co-workers 37-39 have investigated the problems of sample digestion and have shown that provided adequate care is taken to ensure complete sample digestion and conversion of the selenium to selenite, this technique can be accurate. Siemer and Koteel 4° have examined different HG-AAS techniques and mention methods of optimizing systems to obtain maximum sensitivity. Reamer and co-workers 4~ used radiotracers to evaluate losses in different hydride generation systems. The HG-AAS approach has been used to determine selenium in waters, blood and blood products, urine, environmental samples, foods, and feeds. It is reported not to be a good technique for the analysis of all biological samples for selenium.I° The HG-AAS technique has been shown to have absolute detection limits ranging from 0.1 to 60 ng. 14Imprecision and recovery vary according to care and expertise of the user but have been reported to be 1-5% RSD and 70-110%, respectively.
Graphite Atomic Absorption Spectrometry (GFAAS) Graphite furnace atomic absorption spectrometry would seem to be the most appropriate technique to analyze samples for small amounts of selenium. As many semiautomated systems are now available, this technique could be used for the rapid direct determination of selenium in large batches of routine samples. However, GFAAS is not simple nor free from interferences--both spectral and matrix--and/or losses due to the volatility of selenium. The use of pyrolytically coated or total pyrolytic graphite tubes, a L'vov platform, peak area integration, matrix modification, and Zeeman background techniques (ZAAS) have enabled some advances to be made. Spectral interferences from iron and/or phosphate are compensated for by Zeeman background correction. The use of matrix modification--generally nitrates of Ni, Cu, Pd, Pt, Ag, Mo, and/or M g - help to thermally stabilize selenium and allow for ashing temperatures of up to 1200°. 1°A4'42-44 However, only a few sample types are suitable for 37 B. Welz, M. Melcher, and G. Schlemmer, in "Trace Element Analytical Chemistry in Medicine and Biology," Vol. 3. de Gruyter, Berlin, Federal Republic of Germany, 1984. 3s B. Welz, M. Melcher, and J. Neve, Anal. Chem. Acta 165, 131 (1984). 39 B. Welz and M. Verlinden, submitted for publication. 4o D. Siemer and P. Koteel, Anal. Chem. 49, 1096 (1977). 41 D. C. Reamer, C. Veillon, and P. T. Tokousbalides, Anal. Chem. 53, 245 (1981). 4z W. Slavin, G. R. Carnrick, D. C. Manning, and E, Pruszkowska, At. Spectrosc. 4, 69 (1983). 4~ R. D. Ediger, At. Absorpt. Newsl. 14, 127 (1975). 44 F. J. Fernandez, S. A. Myers, and W. Slavin, Anal. Chem. 52, 721 (1980).
[31]
SELENIUM IN BIOLOGICAL MATRICES
401
direct analyses even with the use of matrix modification. Blood, serum, semen, yeast, and food supplements and waters have been analyzed directly) °,14,42A ZAAS method was shown to correlate well with the definitive IDMS method (r = 0.987) for the analysis of 30 plasma samples for selenium, but was less precise. 45 Many commonly used digestion procedures and acids have been reported to cause matrix interferences for G F A A S . 46 Pretreatments involving the formation of selenium chelates and subsequent analysis by GFAAS again raise the question of quantitative sample recoveries. Contamination, generally a problem with many GFAAS techniques, is of less concern for selenium. GFAAS has good detection limits (20-50 pg) and reported imprecisions ranging from 2 to 15% RSD.
Gas Liquid Chromatography (GLC) The determination of selenium by GLC is usually based on the detection of a volatile piazselenol formed by the reaction of selenite, after destruction of the organic matrix, with an aromatic diamine, like 4-nitroo-phenylenediamine. This again presents the problem of quantitative sample recoveries and conversion to the tetravalent state. Various detectors l° have been used with this system, including flame ionization, thermal conductivity, atomic absorption spectrometry, mass spectrometry (see mass spectrometry method), and electron capture detectors (ECD). 47,48 When using GC/ECD, the absolute detection limit can be as low as 1 pg. Imprecisions are reported to be from 2 to 4% RSD. 48 This method has been adapted for the determination of selenium in the following matrices: water, blood, and blood products, grain, fish, milk, hair, urine, and soft tissues) °,47,48 The method has also been used to detect selenium species in biological materials and in waters. 9 Summary The definitive IDMS method has the advantage of good detection limits and precision. The advent of the new, smaller, easy-to-use GC/MS system should make this method more popular. The fluorometric method remains the method of choice for the determination of selenium in many matrices if the limitations of the method are understood. The newer methods include gas chromatography, HG-AAS, and GFAAS. The gas chromatography methods show promise because of the 45 W. Slavin and D. C. Manning, Prog. Anal. At. Spectrosc. 5, 243 (1982). S. A. Lewis, N. S. Hardison, and C. Veillon, Anal. Chem. 58, 1272 (1986). 47 T. P. McCarthy, B. Brodie, J. A. Milner, and R. F. Bevill, J. Chromatogr. 225, 9 (1981). C. J. Cappon and J. C. Smith, J. Anal. Toxicol. 2, 114 (1978).
402
ANALYSIS OF METALS
[32]
realtively simple instrumentation needed and the fact that the analyte is separated from the matrix. HG-AAS offers good sensitivity provided care is taken to ensure complete sample digestion and conversion of selenium to selenite. The advent of Zeeman background correction systems for GFAAS has greatly facilitated selenium determinations, particularly in biological matrices where iron and phosphorus are also present. The reference materials now available, used as part of a quality assurance program, should help to ensure accurate determinations, permit method validation, and allow performance evaluation in intedaboratory trials.
[32] V a n a d i u m By D O N N A
M. MARTIN
and N.
DENNIS CHASTEEN
Introduction Studies of the metabolism, biochemistry, clinical pathology, and environmental toxicology of vanadium require reliable analytical techniques for the determination of this element in biological materials. The ideal technique for vanadium analysis would permit trace amounts to be determined rapidly and economically but with sufficient sensitivity to detect normal and subnormal levels. Because vanadium concentrations in biological materials can vary widely, all these requirements cannot be met with a single analytical method. In the present chapter, we survey the most common techniques for determining vanadium and discuss some of the advantages and disadvantages of each. Often the instrumentation at hand dictates the analytical method to be employed. While most of this chapter is devoted to analysis of tissues and fluids, the methods are readily adapted to other types of samples. To our knowledge, methods for vanadium analysis of biomaterials have not been reviewed previously. Since much of the earlier work on vanadium analysis, particularly at the sub parts per million level, is of questionable reliability, this review focuses on the more recent literature. Atomic Absorption Spectrometry (AAS) Graphite furnace atomic absorption is the most conventional and economical technique for analyzing vanadium in biological samples in the pbb range. Analysis requires only a small amount of sample and matrix matchMETHODS IN ENZYMOLOGY, VOL. 158
Copyright © 1988by AcademicPress, Inc. All rightsof reproductionin any form reserved.