Determination of mercury, selenium, bismuth, arsenic ...

5 downloads 5840 Views 162KB Size Report
drip treatment for peptic ulcer complaints [35]. Patients treated with anti-ulcer ..... Results for C.R.M GBW0706 (Chinese human hair). Value obtained (ng g−1)a,b.
Talanta 52 (2000) 833 – 843 www.elsevier.com/locate/talanta

Determination of mercury, selenium, bismuth, arsenic and antimony in human hair by microwave digestion atomic fluorescence spectrometry L. Rahman a, W.T. Corns b,*, D.W. Bryce b, P.B. Stockwell b b

a Uni6ersity of Greenwich, Woolwich, London UK P S Analytical, Crayfields Industrial Estate, Main Road, St. Paul’s Cray, Orpington, Kent BR5 3HP, UK

Received 22 November 1999; received in revised form 19 April 2000; accepted 25 April 2000

Abstract A novel method for determination of Hg, Se, Bi, As and Sb based on microwave digestion followed by continuous flow vapour generation atomic fluorescence spectrometry was developed. The digestion for Hg was based on a two stage digestion involving HNO3 and H2O2, whilst for the hydride forming elements a common digestion using HCl and H2O2 was found to be the most effective. The instrumentation and chemistry were optimised in order to provide the best accuracy and precision. The method detection limit for hair samples was found to be 0.2 ng g − 1 for Hg and between 2 and 10 ng g − 1 for the hydride forming elements. The atomic fluorescence detector showed excellent linearity over the concentration ranges studied with linear correlation co-efficients between 0.99984 and 0.99997. To validate the accuracy of the method a human hair certified reference material (GBW 0706) was analysed and excellent agreement with the certified value was obtained for all elements. © 2000 Elsevier Science B.V. All rights reserved. Keywords: AFS; Hair; Elemental analysis

1. Introduction

1.1. Mercury Toxicological effects of mercury compounds on both plant and animal life have long been recognised, but it was not until the disaster at Minimata

* Corresponding author. Tel: + 44-1689-891211; fax: +441689-896009. E-mail address: [email protected] (W.T. Corns).

Bay in 1953 that the subject received world-wide attention [1,2]. Exposure of mercury to the general population is mainly through the diet and dental amalgam. In foodstuffs, mercury is usually in the inorganic forms and of very low concentration. The exceptions are fish and fish products, which are the main sources of methylmercury in the diet. Recent experimental studies [3] have shown that mercury is released from amalgam restoration in the mouth as a vapour and the rate of release may be increased by certain foods or by the action of chewing.

0039-9140/00/$ - see front matter © 2000 Elsevier Science B.V. All rights reserved. PII: S 0 0 3 9 - 9 1 4 0 ( 0 0 ) 0 0 4 3 6 - 7

834

L. Rahman et al. / Talanta 52 (2000) 833–843

There are several forms of metabolic transformation that can occur after exposure to mercury. These can be summarised as: oxidation of metallic mercury to divalent mercury, reduction of divalent mercury to metallic mercury, methylation of inorganic mercury and conversion of methylmercury to divalent inorganic mercury. Numerous papers have been published in this area [4–7] and these have been reviewed by WHO [8]. The faecal and urinary routes are main pathways for the elimination of inorganic mercury in humans, although some elemental mercury is exhaled in breath. The urinary route dominates when exposure is high. After exposure to metallic mercury vapour, a small fraction of the mercury in the urine may be present as elemental mercury [10]. Methylmercury is excreted in the urine to only a very limited extent. The mercury content in hair is a useful indicator of exposure to methylmercury via fish intake in non-occupationally exposed people [9]. When evaluating exposure to low concentrations of inorganic mercury, interference from methylmercury exposure can dominate blood analysis therefore an alternative biological matrix such as hair or urine is preferred [8]. There are numerous analytical techniques available for the determination of total mercury in biological materials. The most commonly used methods for measuring mercury are cold vapour– atomic absorption spectrometry (CVAAS) and neutron activation (NA) and inductively coupled plasma mass spectrometry (ICPMS). Although CVAAS has become the most widely used technique, there are several disadvantages associated with the use of AAS detection, such as limited linear calibration range, spectral interferences resulting from non-specific background absorption of volatile organics [11] and difficulties with measurements at lower levels. Neutron activation is generally regarded as being very accurate and sensitive, but extremely expensive and time consuming [12]. Recently, atomic fluorescence spectrometry (AFS) has been utilised for the detection of mercury in biological materials [13]. Atomic fluorescence is well suited for the determination of mercury as it absorbs and fluoresces at the same wavelength, (i.e. resonance fluorescence) in the

ultra violet (UV) region, intense mercury excitation sources are available and no optimisation is required when the technique is coupled to vapour generation. The reported limit of detection using this approach was 0.9 pg ml − 1, which equates to 2 pg absolute [13]. Various digestion procedures are available for total mercury in biological materials. A variety of combinations of strong acids (HCl, H2SO4, HNO3, HClO4) and oxidants (H2O2, KMnO4, K2Cr2O7, K2S2O8) have been used and recommended [14–17]. The main concerns of the digestion stages are related to the loss of analyte during elevated temperature digestion and interference from acid gases evolved during the final reduction step. Determinations at low concentrations are normally limited by blank contributions from reagents. Fentons reagent (Fe (II)+ H2O2) was utilised by Ping and Dasgupta [18] to minimise blank levels from reagent additives. Vermeir et al. [13] have used a microwave digestion procedure with nitric acid to mineralise a variety of biological materials including hair. Lind et al. [9] participated in a quality control program to determine mercury in hair for a study on ‘Mental effects of prenatal methylmercury exposure in New Zealand children’.

1.2. Hydride forming elements The essential trace element selenium has attracted increasing attention in recent years as evidence for its involvement in human health has become apparent. Selenium deficiency has been implicated in the development of severe and fatal cardiomyopathy (Keshen disease) [19] and of osteoarthropy (Kashin–Beck disease) [20,21]. Low selenium levels have also been found in many disease states, including various forms of cancer [22–24], acute myocardial infarction [25,26], severe rheumatoid arthritis [27], cirrhosis of the liver [28] and conditions exhibiting a compromised immune response [29]. Selenium is believed to exert its protective nature through a number of mechanisms, the best known of which is as an anti-oxidant and a constituent of the peroxidescavenging enzyme glutathione peroxidase. Alaejos and Romero [30] have published an extensive

L. Rahman et al. / Talanta 52 (2000) 833–843

review on the role of selenium in nutritional studies. With increasing recognition of the role of antioxidants in disease prevention, the need for accurate determination of selenium status has become more important. Very few papers have been published on the determination of selenium in hair, which allows long-term exposure to be monitored. Arsenic is a general protoplasmic poison. It is cumulative and all bodily systems are affected. Human exposure to inorganic arsenic occurs via inhalation of industrial dust and ingestion of contaminated drinking water and food. Exposure to organic arsenic occurs mainly through compounds biosynthesised across the food chain [31]. Inhalation may contribute 0.1 – 4 mg day − 1 of arsenic, but this amount can increase to high values in the event of atmospheric pollution. Drinking water may contribute up to 15 mg day − 1. Estimates of dietary intake range from 7 to 330 mg day − 1 [32]. Approximately 80 –100% of the inhaled and ingested arsenic is absorbed through the gastrointestinal tract and lungs but up to 50–70% of the absorbed arsenic is eliminated mainly through urine and to a lesser extent through hair, nails and faeces [32]. In high doses arsenic is toxic, with the toxicity depending on the oxidation state. Toxicity decreases in the following order: arsine, inorganic As(III), organic As(III), inorganic As(V), organic As(V), arsonium compounds and elemental arsenic [32]. Symptoms of arsenic poisoning include diarrhoea, vomiting, headaches, drowsiness, convulsions etc. [33]. Arsenic is also reported to be carcinogenic. Arsenic is distributed in all human tissues ranging from 0.01 to 0.09 mg g − 1 wet weight down to a few ng ml − 1 in biological fluids. The levels of arsenic found in hair, skin and nails are considerably higher compared to other tissues. Normal hair contains small quantities of As from 50 to 400 mg g − 1 but the level is greatly increased during excessive intake of arsenic. The profound accumulation of arsenic in hair during exposure is of value in the diagnosis of arsenic poisoning. Although antimony is non-essential for human life, it is found in biological specimens from persons who have been exposed to industrial sources

835

of antimony or who have been treated with drugs containing the element, whose main medicinal use is in the treatment of parasitic diseases. Pentavalent antimony is less toxic than trivalent antimony. In humans trivalent antimony is taken up by the red blood cells, whereas pentavalent antimony remains in the plasma and it is more easily excreted than the trivalent form. The most likely route in the body from industrial exposure is from inhalation. This can arise from dust and fumes from the grinding and high temperature processing of antimony and its compounds or from stibine (SbH3) which is particularly poisonous. This can be accidentally liberated from storage batteries when treated with nascent hydrogen under acidic conditions. Recently antimony has been controversially linked to sudden infant death syndrome (SIDS) more commonly known as ‘cot death’ because of the use of antimony flame retardants in cot mattresses. The fungus Scopulariopsis bre6icaulis is capable of generating stibine, which may cause anticholinesterase poisoning and cardiac failure in infants [34]. Typical levels of antimony in hair are between 0.09 and 3 mg g − 1. The bismuth content in most biological samples is very low, with biological fluids normally containing only a few ng ml − 1, while in biological tissues concentrations may range from 10 to 90 ng g − 1. Currently there is a greater interest in some clinics for the monitoring of patients on bismuth drip treatment for peptic ulcer complaints [35]. Patients treated with anti-ulcer drugs such as Zantac were found to contain much higher levels of bismuth in urine (e.g. 1460 ng ml − 1). No publications were found on bismuth in hair. There are numerous analytical techniques available for the determination of the hydride forming elements in biological matrices. The most common procedure used routinely by chemists in most laboratories is hydride generation coupled with atomic spectrometric techniques [36], such as atomic absorption spectrometry (AAS), plasma emission spectrometry (ICP, DCP, MIP), inductively coupled plasma mass spectrometry (ICPMS) and atomic fluorescence spectrometry (AFS). Of all the trace elemental analysis techniques AFS offers improved selectivity compared with optical emission spectrometry, because the resolu-

836

L. Rahman et al. / Talanta 52 (2000) 833–843

tion is determined by the line width rather than by the detection system. Atomic fluorescence techniques offer advantages in terms of linearity and detection levels. By its inherent nature the signal obtained can be increased by the intensity of the lamp or excitation source. The limitations are scatter and background levels of impurities. Atomic absorption, on the other hand, although it has been extensively used, suffers from the fact that it is non-linear and measurements at lower levels are extremely difficult. Whilst ICP MS offers similar limits of detection to HG AFS, it also suffers from some polyatomic interferences and systems are considerably more expensive to purchase and run. Corns et al. [37] have developed an atomic fluorescence spectrometer for the hydride forming elements. The analyte elements are introduced as their gaseous hydrides from a fully automated continuous hydride generator using a miniature argon–hydrogen diffusion flame as the atomiser. The hydrogen for the flame is chemically generated as a by-product of the sodium tetrahydroborate reduction. Excitation is achieved using a boosted-discharge hollow cathode lamp. Fluores-

cence wavelengths of interest were selected using an interference filter and a solar blind photomultiplier was used as the detector. This paper describes the development of a reliable method of determination for mercury and the hydride forming elements (Se, Bi, As and Sb) in human hair using hydride generation-AFS. Digestion procedures using HNO3/H2O2 for mercury and HCl/H2O2 for the hydrides were optimised and accelerated by the use of microwave digestion with closed vessels. The chemistry of vapour generation was studied for each element and with optimal conditions the analytical performance was evaluated. To validate the accuracy of the method, a human hair reference material (GBW 0706) was analysed for each element.

2. Experimental

2.1. Instrumentation and apparatus 2.1.1. Mercury determination A continuous flow vapour generation-atomic fluorescence spectrometer (PSA 10.025), was

Fig. 1. Schematic diagram of the continuous flow vapour generator.

L. Rahman et al. / Talanta 52 (2000) 833–843 Table 1 Instrumental and chemical conditions for Millennium Merlin Reductant flow rate (ml min−1) Blank flow rate (ml min−1) Sample flow rate (ml min−1) Carrier gas flow rate (ml min−1) Dryer gas flow rate (ml min−1) Blank concentration Reductant concentration Delay period (s) Analysis period (s) Memory period (s) Filter Wavelength (nm)

4.5 9.0 9.0 235 2500 5% v/v HCl 2% m/v SnCl2 in 10% v/v HCl 10 40 60 32 253.7

utilised to generate and detect gaseous mercury. A schematic diagram is shown in Fig. 1. Excitation was achieved using a low-pressure mercury vapour discharge lamp at a wavelength of 253.7 nm. The resonance fluorescence was detected at the same wavelength by a photomultiplier tube (PMT) located at right angles to the excitation source. The instrument has been described in detail elsewhere [38]. The operating and chemical conditions are summarised in Table 1.

2.1.2. Hydride determination For the determination of As, Se, Sb and Bi a continuous flow hydride generation-atomic fluorescence spectrometer (PSA 10.055) was utilised. A schematic diagram is shown in Fig. 1. Boosted-discharge hollow cathode lamps were used as the excitation source for each element. The gaseous hydride was atomised using a cool hydrogen diffusion flame, chemically generated from the reagents. Detection was achieved using a solar blind photomultiplier tube. The instrument is described in more detail elsewhere [37]. The operating and chemical conditions are summarised in Table 2. 2.1.3. Digestion apparatus Hair samples were digested using a 600-W closed bomb microwave system (Floyd Inc., RMS

837

150). Sample preparation for vapour generation was achieved using a block digestor (Lachat BD26).

2.1.4. Reagents All reagents were of analytical grade unless otherwise specified, and all solutions were prepared in doubly distilled water of high purity obtained from an Elga Option 3 system (Elga Ltd, High Wycombe, UK). A 2% m/v SnCl2 (Sigma– Aldrich Co. Ltd., Gillingham, Dorset) in 10% v/v HCl (QuadraChem Laboratories Ltd., Forest Row, East Sussex) solution was used as the reductant for mercury whereas for the hydride forming elements 0.73% m/v NaBH4 (Aldrich) in 0.1 M NaOH (Merck–BDH Laboratory Supplies, Poole) was used as reductant. HCl (Fisher), Table 2 Instrumental and chemical conditions for Millennium Excalibur Reductant flow rate (ml min−1)

4.5

Blank flow rate (ml min−1)

9.0

Sample flow rate (ml min−1)

9.0

Carrier gas flow rate (ml min−1)

235

Dryer gas flow rate (ml min−1)

2500

Blank concentration As and Sb

Se and Bi Reductant concentration

3 M HCl/1% m/v KI/0.2% m/v Ascorbic acid 6 M HCl 0.73% m/v NaBH4 in 0.1 M NaOH

Delay period (s)

10

Analysis period (s)

30

Memory period (s)

40

Filter

32

Wa6elengths (nm) As Se Sb Bi

189, 193.7, 197.2 196, 204, 206.3 206.8, 217.6 206.2

838

L. Rahman et al. / Talanta 52 (2000) 833–843

HNO3 (Merck, AristaR grade) and H2O2, 27.5%wt. in H2O (Aldrich) were used in microwave digestion procedures. A solution of 50% m/v KI (Aldrich) and 10% m/v L-ascorbic acid (Aldrich) was used for the pre-reduction of arsenic and antimony. Stock solutions (Merck, SpectrosoL) of 1000 ppm of mercury, selenium, bismuth, arsenic and antimony were used to prepare the calibration standards and spike solutions. The mercury was diluted to more appropriate levels in 10% v/v HNO3, while selenium and bismuth were diluted in 25% v/v HCl and arsenic and antimony in 25% v/v HCl, 1% m/v KI, 0.2% m/v ascorbic acid.

and antimony samples were treated with the appropriate pre-reductant to improve efficiency of hydride generation.

2.2. Sample preparation

2.2.4. Antimony and arsenic pre-reduction step The contents of each microwave vessel were allowed to cool before transferring carefully into 100-ml volumetric flasks, containing 20 ml conc. HCl and 2 ml KI-ascorbic acid solution. The contents were diluted up to the mark with DDW and left to stand for 30 min prior to analysis.

2.2.1. Mercury A portion of the hair sample (0.5 g) was accurately weighed using a four-decimal-place analytical balance and transferred to a closed PTFE bomb microwave digestion vessel, along with 5 ml conc. HNO3. The vessels were sealed and placed in the microwave chamber, and run at 30% power for 15 min. The vessels were allowed to cool for a few minutes before 1 ml H2O2 was added carefully and the digestion continued for a further 15 min at 30% power. Once the digestion was complete the vessels were allowed to cool before the contents were transferred to grade ‘A’ 100-ml volumetric flask, and made up to the mark with DDW. The sample was then ready for vapour generation.

2.2.3. Selenium pre-reduction step The vessels were allowed to cool before the contents were transferred carefully to the block digestor vessels, along with 45 ml conc. HCl and 20 ml DDW. The vessels were kept open and the digested samples were heated at 180°C for 1 h in a fume cupboard. The resulting solutions were cooled in a cold water bath prior to being transferred to volumetric flasks and made up to 100 ml with DDW.

3. Results and discussion The effect of reagent concentration on vapour generation was studied for both mercury and the hydride forming elements. This was achieved using univariant optimisation. In addition to this, the kinetic effect of the pre-reduction stage for arsenic and antimony was also studied

3.1. Optimisation 2.2.2. Antimony, arsenic, bismuth and selenium A portion of the hair sample (0.1 g) was accurately weighed and transferred to a PTFE microwave digestion vessel containing 5 ml conc. HCl. Once sealed, the vessels were placed in the microwave chamber and run at 40% power for 15 min. The contents were cooled for a few minutes before the addition of 1 ml H2O2. The digestion step was continued for a further 15 min at 40%, after which the vessels were allowed another few minutes for cooling prior to a second addition of 1 ml H2O2. One final microwave digestion step was performed again at 40% power for 15 min. Prior to hydride generation of arsenic, selenium

3.1.1. The effect of tin(II) chloride on the signal to background ratio for mercury The effect of tin(II) chloride (SnCl2) concentration on the signal to background ratio on the determination of mercury was investigated using univariant optimisation. This reagent is purified for mercury by purging with argon for 30 min. The influence of background is therefore negligible at increasing concentrations. The SnCl2 is present in excess at all concentrations and is shown to have little affect on the sensitivity of the method. The optimal concentration of 2% m/v SnCl2 was selected.

L. Rahman et al. / Talanta 52 (2000) 833–843

839

Fig. 2. SBR against NaBH4 concentration (% m/v) in 0.1 M NaOH.

3.1.2. The effect of sodium tetrahydroborate on signal to background ratio for arsenic, selenium, antimony and bismuth The effect of sodium tetrahydroborate (NaBH4) was investigated for arsenic, selenium, antimony and bismuth (Fig. 2). At concentrations above 1% m/v NaBH4 the background increased rapidly. This was attributed to the additional hydrogen produced from the reaction of NaBH4 with HCl, producing a larger flame. The background from the flame is predominately from OH emissions, which are detected by the solar blind photo-multiplier tube between 303 and 315 nm. In addition to this, the excess hydrogen produced dilutes the analyte of interest. An optimum concentration of 0.73% m/v was selected for all elements, which also minimises moisture carryover from the gas/ liquid separator. 3.1.3. Effecti6eness of nitric acid/hydrogen peroxide for the digestion of hair It is well known that the mercury found in hair samples is predominately methylmercury. Initially concentrated nitric acid (HNO3) was used without the addition of hydrogen peroxide (H2O2) and low, variable recoveries were observed. It is also well known that a strong oxidant is required to breakdown organomercury and therefore an addition of H2O2 was investigated. To avoid a high-

pressure build up in the microwave vessel it was decided to perform a two-stage digestion. The resulting sample digest was extremely clear with no undigested hair material observed. The HNO3/ H2O2 digests were also used for arsenic, selenium, antimony and bismuth. Residual nitrous fumes and the strong oxidising medium of nitric acid yielded low efficiency for the pre-reduction using potassium iodide (KI) for arsenic and antimony. The formation of iodine in these samples was clearly evident, rendering the solutions unsuitable for hydride generation. A procedure based on hydrochloric acid/hydrogen peroxide (HCl/H2O2) was therefore developed for the hydride forming elements.

3.1.4. Effecti6eness of hydrochloric acid/hydrogen peroxide for the digestion of hair It is generally accepted that HCl is the acid of choice for hydride generation. Digestions with HCl (3× 15 min) at 40% power were not sufficiently strong to obtain good recoveries of hydrides from hair. Furthermore, undigested material was observed at the end of the digestion. It was therefore decided to perform a three-stage digestion with HCl and take less sample (0.1 g) and to include the addition of H2O2 in the final two steps. This procedure produced a clean digest compatible with the pre-reduction step using potassium iodide and ascorbic acid.

840

L. Rahman et al. / Talanta 52 (2000) 833–843

3.1.5. The use of hydrochloric acid as an pre-reductant of selenate to selenite Selenate [Se(VI)] does not react with sodium tetrahydroborate to form selenium hydride (SeH2) and therefore it is necessary to pre-reduce all selenium to the Se(IV) state prior to hydride generation. Samples analysed directly after the digestion with HNO3/H2O2 and HCl/H2O2 yielded 0% recovery. The most probable explanation is that both digestion chemistries are sufficiently oxidising to convert all selenium to the Se(VI) state. One other possibility is that the selenium was lost during digestion. However this is unlikely with closed digestion apparatus. The most common procedure for pre-reduction involves heating the digested sample with HCl. This procedure is well documented and can be carried out using hot plate digestions or in microwave systems. Initial tests showed that although initial sample digestion was carried out in HCl in the closed microwave vessels, Se(VI) was not reduced to Se (IV). This was due to the fact that when Se(VI) is reduced by HCl to Se (IV), Cl2 is also formed. In closed systems this Cl2 can then oxidise Se(IV) to Se(VI). This was confirmed by adding Se(IV) to samples and digesting in the closed microwave system. The results showed that 0% of the expected Se(IV) was recovered. When the same samples were taken and heated in HCl in

an open block digester vessel after the microwave digestion the recoveries increased to 94.8% for the sample and 92.0% for the Se(IV) added (see Table 5).

3.1.6. The use of potassium iodide/ascorbic acid as reductant for arsenic and antimony After the digest with HCl/H2O2 it is reasonable to assume that arsenic and antimony are present as the pentavalent oxidation state. It is well known that As(V) and Sb(V) have much slower reaction kinetics for hydride formation. This is dependent on the hydride generation system used and the concentration of reagents. With a continuous flow manifold the hydride generation efficiency is around 30% compared to the trivalent oxidation state. The most common pre-reduction involves the addition of potassium iodide/ascorbic acid reagent. To test the efficiency of the pre-reductant stage, As(V) and Sb(V) standards were prepared in 3 M HCl. Each standard was analysed at 2-min intervals after the addition of the potassium iodide pre-reductant. The output was compared to a trivalent standard of the same concentration. The data for arsenic is presented in Fig. 3. For arsenic a pre-reduction time of 30 min was required in order to obtain 100% efficiency, whereas for antimony the pre-reduction was immediate. Bismuth does not exist in the pentavalent

Fig. 3. As(V) pre-reduction with KI in 3 M HCl against time.

L. Rahman et al. / Talanta 52 (2000) 833–843 Table 3 Summary of instrumental and method detection limits based on 3sn−1 of ten runs of the blank

841

marised in Table 3. All of the MDLs are well below the levels expected for hair samples.

Analyte

IDL (ng l−1)

MDL (ng g−1)a

4.2. Calibration data

Hg As Se Sb Bi

0.20 5.0 2.0 10.0 10.0

0.04 5.0 2.0 10.0 10.0

It is well known that atomic fluorescence is an extremely linear detection system. Linearity is typically achieved over seven orders of magnitude. In this study linearity was evaluated over the concentration range of interest for each analyte. The calibration data obtained for each element is summarised in Table 4.

a

The MDL could be improved with less dilution of the sample.

state and therefore the pre-reduction is unnecessary. It is worth noting that the pre-reduction stage could be accelerated by heating the samples if so desired.

4. Analytical performance characteristics The analytical performance characteristics were evaluated for each element. Figures of merit include detection limits, linearity, accuracy and precision of measurements.

4.1. Detection limits The instrumental detection limit (IDL) was evaluated by running the reagent blank solution ten times using the operating conditions outlined in Tables 1 and 2. The limit of detection was then calculated by multiplying the S.D. of ten runs of the blank by 3 (e.g. 3 sn − 1). Method detection limits (MDL) were then established by multiplying the IDL by the dilution factor employed for each sample preparation. The results are sum-

4.3. Accuracy and precision of measurements In order to validate the method for accuracy and precision, a certified reference material GBW 0706 (Chinese hair) was analysed for each element. Each sample was prepared in duplicate and analysed in triplicate (e.g. n=6). The results for each element are shown in Table 5. Excellent recoveries for all elements were obtained compared to the certified value. In addition to this, various hair samples (collected from laboratory staff) and spiked samples were also analysed (Table 6). In all cases the spike recoveries (92.2– 103.0%) were excellent, showing no matrix interferences.

5. Conclusions The determination of mercury and the hydride forming elements (As, Se, Sb and Bi) in hair is an important and useful tool to monitor long term occupational exposure. In most cases the hair pre-concentrates these elements thus enabling his-

Table 4 Calibration data for each analyte studied Analyte

Concentration range (mg l−1)

Equation of fita

Linear correlation co-efficient (r)

Hg Se Bi As Sb

0–5.0 0–1.5 0–1.0 0–0.5 0–1.0

y= 67.930x−0.767 y = 568.307x+0.110 y =161.952x+1.207 y =552.145x+1.472 y =166.252x+0.691

0.99996 0.99984 0.99997 0.99989 0.99991

a

A minimum of five standards was used for each calibration line.

L. Rahman et al. / Talanta 52 (2000) 833–843

842

Table 5 Results for C.R.M GBW0706 (Chinese human hair) Element

Certified value (ng g−1)

Value obtained (ng g−1)a,b

Hg Se Bi As Sb

360 9 50 6009 30 340920 280940 959 12

340 94 569 911 373 9 44 281 9 12 98 96

a b

Sample

Element

Value obtained (ng g−1)a,b

1 2 3 4

Sb Hg Hg Hg

226 95 733 930 24594 541 94

b

94 95 110 100 103

% Spike recovery – 92 – – 103

All results corrected for moisture content (i.e. 11.6%). Mean value 9S.D., n=3 (see text).

Table 6 Hair samples from laboratory staff

a

% Recovery

References % spike recovery

96 – – –

All results corrected for moisture content (i.e. 16.5%). Mean value 9S.D., n=3 (see text).

torical evidence of the excretion of these analytes in humans. Very few papers were found in the literature with regards to this subject, which is unsurprising for elements such as antimony and bismuth, which have only recently become of biological significance. Hydride generation coupled to atomic fluorescence spectrometry has successfully been applied for this application. Sample preparation for vapour generation is extremely critical and the importance for validating analytical measurements with certified reference materials is clearly indicated in this study. Speciation of elements such as mercury, arsenic and selenium is becoming very important and hydride generation atomic fluorescence can easily be coupled to separation techniques such as HPLC or GC, thus enabling the speciation of the important organometallics after suitable extraction procedures have been used. Future work would concentrate on this area with the view of providing analytical data to understand the excretion pathways of these toxic elements.

[1] W.E. Smith, A.M. Smith, Minimata, Holt, Reinhardt and Winston, New York, 1975. [2] K. Tsuchiya, Am. J. Ind. Med. 21 (1992) 275. [3] A.M. Aronsson, B. Lind, M. Nylander, M. Nordberg, Biol. Met. 2 (1989) 25. [4] J.B. Hursh, S.P. Sichak, T.W. Clarkson, Pharmacol. Toxicol. Engl. Transl. 63 (1988) 266. [5] M. Ogata, K. Kenmotsu, M. Hirota, T. Meguro, H. Aikoh, Arch. Environ. Health 42 (1987) 26. [6] U. Heintze, S. Edwardsson, T. Derand, D. Birkhead, Scand. J. Dent. Res. 91 (1983) 150. [7] B. Lind, L. Friberg, M. Nylander, J. Trace Elem. Exp. Med. 1 (1988) 49. [8] WHO, Inorganic Mercury, Environmental Health Criteria 118, World Health Organisation, Geneva, 1991. [9] B. Lind, E. Bigras, E. Cernichiari, T.W. Clarkson, L. Friberg, M. Hellman, P. Kennedy, J. Kirkbride, T. Kjellstrom, B. Ohlin, Fresenius Z. Anal. Chem. 332 (1988) 620. [10] M. Yoshida, Y. Yamamura, Int. Arch. Occup. Environ. Health 51 (1982) 99. [11] C.D. West, Anal. Chem. 46 (1974) 797. [12] J.M. Rottschafer, J.D. Jones, H.B. Mark Jr, Environ. Sci. Technol. 5 (1971) 336. [13] G. Vermeir, C. Vandecasteele, R. Dams, in: A. Aitio, A. Aro, J. Jarvisalo, H. Vainio (Eds.), Trace Elements in Health and Disease, Royal Society of Chemistry, Cambridge, 1991. [14] D.C. Baxter, W. Frech, Anal. Chim. Acta 236 (1990) 377. [15] M. Goto, T. Shibakawa, T. Arita, D. Ishii, Anal. Chim. Acta 140 (1982) 179. [16] A. Iverfeldt, Mar. Chem. 23 (1988) 441. [17] P. Freimann, D. Schmidt, Fresenius Z. Anal. Chem. 299 (1982) 97. [18] L. Ping, P.K. Dasgupta, Anal. Chem. 61 (1989) 1230. [19] X. Chen, G. Yang, X. Chen, K. Ge, Biol. Trace Elem. Res. 2 (1980) 91. [20] G. Yang, K. Ge, X. Chen, World Rev. Nutr. Diet 55 (1988) 98.

L. Rahman et al. / Talanta 52 (2000) 833–843 [21] L. Chong-Zheng, H. Jing-Rong, Cai-Xia, in: G.F. Combs, J.E. Spallholz, O.A. Levander, J.E. Oldfield (Eds.), Selenium in Biology and Medicine, Van Nostrand Reinhold, New York, 1986, p. 934. [22] W.C. Willet, J.S. Morris, S. Pressel, J.O. Taylor, B.F. Polk, M.J. Stampfer, B. Rosner, K. Schnieder, C.G. Hames, Lancet 2 (1983) 130. [23] K. Jaskiewicz, W.F. Marasas, J.E. Rossouw, F.E. Van Niekerk, E.W.P. Heine, Cancer 62 (1988) 26. [24] L.C. Clark, G.F. Graham, R.G. Crounse, R. Grimsone, B. Hulka, C.M. Shy, Nutr. Cancer 6 (1987) 13. [25] J.T. Salonen, G. Alfthan, J.K. Huttunen, J. Pikkarainen, P. Puska, Lancet 2 (1982) 175. [26] F.J. Kok, A. Hofman, J.C.M. Witteman, A.M. de Bruijin, D.H.C.M. Kruyssen, H.A. Valkenburg, J. Am. Med. Assoc. 261 (1989) 1161. [27] U. Tarp, K. Overvad, J.C. Hansen, E.B. Thorling, Scand. J. Rheumatol. 14 (1985) 97. [28] J. Aaseth, J. Alexander, Y. Thomassen, J.P. Blomhoff, S. Skrede, Clin. Biochem. 15 (1982) 281.

.

843

[29] B.M. Dworkin, W.S. Rosenthal, G.P. Wormer, L. Weiss, J. Parenter. Enteral Nutr. 10 (1986) 405. [30] M.S. Alaejos, C.D. Romero, Clin. Chem. 39 (1993) 2040. [31] L. Fishbein, Int. J. Environ. Anal. Chem. 17 (1984) 113. [32] D.L. Tsalev, Z.K. Zaparianov, Spectrometry in Occupational and Environmental Health Practice, vol. II, CRC Press, Boca Raton, FL, 1984. [33] M. Sitting, Handbook of Toxic and Hazardous Chemicals and Carcinogens, Naynes, NJ, 1985. [34] B.A. Richardson, J. Forensic Sci. Soc. 34 (1994) 199. [35] P.P. Chou, P.K. Jaynes, J.L. Bailey, J. Anal. Toxicol. 8 (1984) 158. [36] J. Dedina, D.L. Tsalev, Hydride Generation Atomic Absorption Spectrometry, Wiley, Chichester, 1995. [37] W.T. Corns, P.B. Stockwell, L. Ebdon, S.J. Hill, J. Anal. At. Spectrom 11 (1996) 61. [38] W.T. Corns, L. Ebdon, S.J. Hill, P.B. Stockwell, Analyst 117 (1992) 717.