Trace Element Determinations in Biologicals Using Atomic Absorption Spectrometry

so far as possible, did not alter the chemical form of the element naturally present. The first point created problems, i.e., natural levels are close to the limits of reliable determination by atomic absorption spectrometry; dilution occurring on enzymolysis and separation of species produced nonmeasurable levels. One solution to this problem lies in the high sensitivity of ICP-MS [3] and the ability to use ICP-MS as a detector with LC columns. Particular attention has been paid to steric exclusion chromatography as one of the most gentle forms of separation achievable. It is also a relevant separation of soluble macro molecules (which are unlikely to be absorbed) from lower molecular weight species which are more likely candidates for absorption. Additionally, some work with reversed-phase HPLC linked to ICP-MS has also been carried out. With HPLC systems a guard column is used to protect the analytical column; post-column injections of standards are used for calibration checks. To minimize peak-broadening, aerosol transfer distance should be reduced by mounting the nebulizer system close to the torch. Flow rates above 0.5 to 0.7 mL/min can lead to signal instability unless precautions against condensation are taken. In an early piece of work, the cadmium content of crab meat with separation by HPLC and monitoring by UV and ICP-MS showed that the major protein peak has almost completely eluted before the cadmium peak corresponding to a somewhat lower molecular weight cadmium-containing protein. More recent studies of the HPLC separation of the proteins from pig kidney before and after cooking show that in the raw state there are three peaks-the higher molecular weight materials disappear during cooking, i.e., are rendered insoluble by heat denaturation, whereas the heat-resistant metallothionein remains intact. Another recent study concerns an organo-arsenical drug used in poultry production. In this, reversed-phase HPLC has been linked to the ICP-MS in much the same way for HPLC. This system allows good separation of "Roxarsone" from other low molecular weight substances and from interference from molecular ions formed from argon in the plasma in the presence of chloride ion. Arsenic in extracts from muscle of chickens fed on a base diet and one supplemented with "Roxarsone" have been compared. The "Roxarsone" peaks correspond to arsenic concentrations of 7 and 10 ng/g in the muscles from treated birds and demonstrates that in the untreated birds the concentration is less than 2 ng/g. The sensitivity foreseen with these systems is now a matter of reality rather than potential-and the ability to use it in speciation studies at tissue concentrations at the ng/g level is a matter of fact rather than of speculation.


Accuracy in Trace Analysis
so far as possible, did not alter the chemical form of the element naturally present. The first point created problems, i.e., natural levels are close to the limits of reliable determination by atomic absorption spectrometry; dilution occurring on enzymolysis and separation of species produced nonmeasurable levels. One solution to this problem lies in the high sensitivity of ICP-MS [3] and the ability to use ICP-MS as a detector with LC columns.
Particular attention has been paid to steric exclusion chromatography as one of the most gentle forms of separation achievable. It is also a relevant separation of soluble macro molecules (which are unlikely to be absorbed) from lower molecular weight species which are more likely candidates for absorption. Additionally, some work with reversed-phase HPLC linked to ICP-MS has also been carried out.
With HPLC systems a guard column is used to protect the analytical column; post-column injections of standards are used for calibration checks. To minimize peak-broadening, aerosol transfer distance should be reduced by mounting the nebulizer system close to the torch. Flow rates above 0.5 to 0.7 mL/min can lead to signal instability unless precautions against condensation are taken. In an early piece of work, the cadmium content of crab meat with separation by HPLC and monitoring by UV and ICP-MS showed that the major protein peak has almost completely eluted before the cadmium peak corresponding to a somewhat lower molecular weight cadmium-containing protein.
More recent studies of the HPLC separation of the proteins from pig kidney before and after cooking show that in the raw state there are three peaks-the higher molecular weight materials disappear during cooking, i.e., are rendered insoluble by heat denaturation, whereas the heat-resistant metallothionein remains intact.
Another recent study concerns an organo-arsenical drug used in poultry production. In this, reversed-phase HPLC has been linked to the ICP-MS in much the same way for HPLC. This system allows good separation of "Roxarsone" from other low molecular weight substances and from interference from molecular ions formed from argon in the plasma in the presence of chloride ion.
Arsenic in extracts from muscle of chickens fed on a base diet and one supplemented with "Roxarsone" have been compared. The "Roxarsone" peaks correspond to arsenic concentrations of 7 and 10 ng/g in the muscles from treated birds and demonstrates that in the untreated birds the concentration is less than 2 ng/g.
The sensitivity foreseen with these systems is now a matter of reality rather than potential-and the ability to use it in speciation studies at tissue concentrations at the ng/g level is a matter of fact rather than of speculation. Atomic absorption spectrometry (AAS) has been used for many years for trace element determinations in foods and other biological materials. The Nutrient Composition Laboratory of the U.S. Department of Agriculture has both conventional commercial single element AAS Spectrometers (Perkin-Elmer Models 603 (flame) and 3030 (Zeeman, furnace)) and a multielement AAS research spectrometer (SIMAAC) [1] which can provide multielement data using either flame or graphite furnace atomization [2][3]. The SIMAAC system consists of a 300 watt xenon arc lamp, an atomizer, an echelle polychromator modified for wavelength modulation, and a PDP 11/34 minicomputer responsible for high-speed data acquisition, data processing and report generation. The SIMAAC system can simultaneously determine up to 16 elements including: Al, Ca, Co, Cr, Cu, Fe, K, Mg, Mn, Mo, Na, Ni, Pb, Sn, V, and Zn. The extended range [4] provided by the system permits simultaneous measurement of elements whose concentrations may vary by as much as 4 orders of magni-Journal of Research of the National Bureau of Standards Accuracy in Trace Analysis tude of concentration. The SIMAAC system has been used for a wide range of applications but the maximum potential of multielement AAS can be seen using graphite furnace atomization (GFAAS) [7-9] where sub-ng/mL detection limits can be achieved using jtL-sized samples.

Trace Element Determinations in Biologicals Using Atomic Absorption Spectrometry
Successful single element or multielement AAS determinations cannot be achieved without suitable sample preparation procedures. As a result, research has been conducted to evaluate the following sample preparation procedures: 1) wet ashing; 2) dry ashing; 3) microwave dissolution; and 4) direct analysis of solids.

Wet Ashing
Wet ash sample preparation procedures are versatile and offer several advantages. First, wet ash procedures are applicable to a wide range of materials and suitable for all foodstuffs. Second, wet ash procedures are low temperature preparations and therefore provide reduced risk of loss of volatile elements. Third, high purity reagents are available for wet ash preparations and finally wet ash preparations are not labor intensive. Typically large batches of samples can be easily handled using a heating block or hot plates. A typical wet ash procedure appears in table 1. Basically, this is a HN0,/H 2 0 2 digestion done using ultrapure reagents (sub-boiling distilled HNO, National Bureau of Standards, Gaithersburg, MD; Perone Peroxide, DuPont, Wilmington, DE) and silanized quartz test tubes to prevent contamination. Typical wet ashing results for NBS SRM 1577a Bovine Liver and NBS SRM 1572 Citrus Leaves appear in tables 2 and 3. The average accuracy for the simultaneous multielement determination of these four elements was 100±9% and the average precision was 10% RSD. Table 1. Wet ash digestion procedure I. Weigh homogenized sample (0.5-1.0 g) into a silanized quartz test tube and add 2 mL 18 megohm water 2. Add 0.5-1.0 ml sub-boiling distilled nitric acid and heat samples (80 'C) overnight 3. Add I mL ultrapure hydrogen peroxide, repeating until samples are clear -Values= Trr, n=6.

Dry Ashing
Dry ashing also offers several advantages as a sample preparation procedure. Dry ashing requires very few reagents, utilizes simple apparatus, requires little sample manipulation and is excellent for biological fluids. Dry ashing also shares the benefit of requiring minimal operator attention so it lends itself well to large "batch" preparations of samples. Table 4 contains a typical dry ash procedure using serum as an example. This procedure provides an optional 4-fold concentration step. Data for NBS SRM urine prepared using this procedure without the 4-fold concentration step appear in

Microwave Dissolution
Microwave dissolution was evaluated because of its significant time savings and potential to provide lower blanks. The closed vessel procedure evaluated provides elevated temperatures and pressures leading to more complete digestions than can be attained using conventional preparation procedures. Microwave procedures also provide the ability to use reproducible conditions making sample dissolution less of an art and more of a science. The microwave digestion procedure utilized was based on the method of Kingston and Jassie [7] (see table 6). Equipment used consisted of an MDS 81 600 watt microwave oven (CEM Corp., Indian Trail, NC) equipped with a turntable and 60 mL teflon PFA vessels with relief valves (Savillex Corp., Minnetonka, MN). Analytical data for NBS SRM 1567 Wheat Flour appear in table 7 and data for an in-house tuna quality control material are summarized in table 8. Reference values for the tuna are based on analytical results obtained using conventional wet and dry ashing preparations and atomic absorption and emission methods. Both sets of data indicate that the microwave method provides suitable results (Wheat Flour: Average accuracy 100±7%, average precision 9% RSD).   bValues= Icr, n=6.

Direct Solids Analysis
The direct analysis of solids in the graphite furnace appears to be the ultimate solution to sample preparation problems related to contamination control and preparation time. Research has been conducted using slurry preparations of a wide range of materials [8]. Slurry preparations offer several advantages over conventional wet ashing and dry ashing procedures: I) reduced sample Journal of Research of the National Bureau of Standards Accuracy in Trace Analysis preparation time, 2) decreased possibility of analyte loss through premature volatilization, 3) reduced loss of analyte related to retention by insoluble residue, and 4) reduced opportunity for sample contamination. Slurry preparations also offer advantages over direct analysis of solids. First, slurries can be easily prepared and no special tools are required. In addition, any amount of the original powdered sample can be used and samples can be diluted if necessary. Finally, samples can be prepared in advance and handled easily. Slurries are typically prepared by weighing -10 mg of finely powdered homogeneous material into a clean polypropylene tube and then diluting with 5-10 mL of 5% HNO 3 containing 0.04% Triton X-100. Slurries are mixed well on a vortex mixer and while mixing, a I mL aliquot is removed and placed into an acid soaked teflon autosampler cup. Slurries are then mixed using an ultrasonic probe (RAI, Hauppauge, NY) until the autosampler withdraws an aliquot for injection into the graphite furnace. Table 9 contains analytical data from the analysis of NBS SRM Citrus Leaves and SRM 1572. Table 10 contains results for NBS SRM 1632a Coal. Average accuracies were: 100+10% (Citrus Leaves); 100±14% (Coal) and average precisions were: 11% RSD (Citrus Leaves); 6% RSD (Coal). These and similar data for more than 12 materials suggest that the slurry technique is well suited to the analysis of homogeneous powdered biological and botanical materials. bValues=1cr, n=5.

LValues=lor' n=5
Comparison of Metbods NBS SRM 1577a Bovine Liver was prepared in duplicate using each of the four methods discussed. Analytical results appear in table 11 with the amount of time each preparation procedure required under routine operating conditions in our laboratory. It is clear that all sample preparation methods provided accurate analytical data for the four elements determined. The microwave and slurry preparation procedures offered significant time savings over the wet ashing and dry ashing methods which were performed in the typical batch mode of operation routinely employed in the lab. Presumably, the times for wet ashing and dry ashing could be reduced somewhat but, it is clear that the microwave and slurry methods offer significant advantages with regard to sample contamination as well as speed. Additional work is needed to ascertain whether these two faster methods are well suited for as wide a range of sample matrices as conventional wet ashing and dry ashing preparation procedures are.

Introduction
A number of major, minor, and trace elements function as important nutrients in foodstuffs of crop and animal origin. These elements occur in a broad range of concentrations in an exceedingly vast array of foods available commercially and home-prepared. In disciplines such as regulatory compliance, product development, quality and safety, and research, reasonably reliable analytical information is mandatory for conclusions and decisions of impact. Although analytical scientists recognize the need for valid data, solid analytical information is at times elusive not only in the area of determination of chemical elements at low (trace) levels [1][2][3][4][5] but also when present as major constituents [6,7].
Of the many analytical techniques available for determination of inorganic nutrient constituents in biological matrices, those based on atomic spectrometry (AAS) are convenient and widely applied. The thrust of current research is to define impacts of a number of parameters on the performance of flame AAS (FAAS)-based methods of analysis of soil and biological materials to lead to reference methods for the measurement of major, minor and trace levels of elements of nutritional and toxicological pertinence. Some of these parameters as well as the excellent performance possible with well-applied versions of FAAS procedures are discussed.

High Reliability Flame Atomic Absorption Spectrometry
In spite of wide acceptance of the technique and the proliferation of published articles on its application, very few AAS-based methods have reached official method or reference method status [8]. Requirements for analytical data of the highest reliability (high precision and accuracy) in the author's research dealing with levels of elemental nutrients and toxicants in soils and biological tissues and with official and reference versions of procedures based on AAS have led to detailed investigations of a number of parameters bearing on method performance. Of the parameters listed in table 1, those Table 1. Some factors bearing on the reliable application of acid decomposition-flame atomic absorption spectrometry to the determination of major, minor and trace elements in biological materials

Sample decomposition and solution preparation
Volumetric ware verification, calibration, operational technique Sample drying and/or moisture determination Volatilization or retention losses in dry ashing or wet decomposition Incomplete destruction of matrix; recovery and analysis of insoluble residue Contamination from ashing acids, acids and vessels Procedural and standard reagent blanks, non-identical performance of method with pure reagents and samples Standard solutions, materials and preparative techniques Sample and standard solution dilution schemes Spectrometric measurement Instrument optimization, performance characteristics and operational techniques Physical and chemical interferences Non-atomic absorption Calibration solutions (single analyte, composite or matrixmatched) Calibration technique (calibration curve, bracketting)

Data handling and interpretation
Recording, data entry and calculation Calibration curve fitting and calculation techniques Interpretation and evaluation (controls, statistical treatment, data presentation basis) Overall Analyst, specialist training and experience Data quality control (accuracy verification by recovery testing and performance with appropriate reference materials)