Determination of Physiological Platinum Levels in Human Urine Using Magnetic Sector Field Inductively Coupled Plasma Mass Spectrometry in Combination With Ultraviolet Photolysis Journal of Analytical Atomic Spectrometry JUTTA BEGEROW MARTINA TURFELD AND LOTHAR DUNEMANN Medizinisches Institut fur Umwelthygiene Department of Analytical Chemistry Auf 'm Hennekamp 50 0-40225 Dusseldorf Germany An extremely sensitive and reliable procedure for the determination of physiological (normal ) Pt levels in human urine is described based on UV photolysis of the sample followed by the determination of Pt with magnetic sector field ICP-MS. Owing to the low blank values which are a consequence of the minimal reagent addition required UV photolysis was used to decompose the organic matrix components.Magnetic sector field ICP-MS operated in the low resolution mode afforded detection limits that were 100 times lower than those obtained in the high resolution mode or with conventional quadrupole ICP-MS and was found to be advantageous for the ultratrace determination of non-interfered isotopes. The detection limit in urine samples was 0.24 ng I-' using standard nickel cones. The use of a high performance nickel skimmer cone did not result in an improvement in the detection limit because the analyte and background signals were enhanced to a similar extent. The urinary Pt levels in 16 non-exposed persons ranged between 0.48 and 7.65 ng I-' (arithmetic mean 1.72 ng I-'). Keywords Magnetic sector field inductively coupled plasrrra mass spectrometry; ultraviolet photolysis; platinum; urine With the introduction of Pt-containing automobile exhaust catalysts and anti-cancer drugs the widespread release of Pt into the environment has increased enormously.Belonging to the rarest elements in the Earth's crust Pt has long been denied as an environmental pollutant. Since investigations have shown that it has already accumulated in roadside dust,' airborne dust,2 soil3 and correlating with traffic density and distance from the road there is a growing concern about the possible effects of Pt on humans. The interest in analytical methods suitable for the determination of trace amounts of Pt in biological and environmental matrices has thus rapidly increased. An analytical procedure suitable for this application was first presented by Hoppstock et a/.' using high-pressure acid digestion followed by adsorptive voltammetry.By applying this procedure Messerschmidt et aL6 were able to determine baseline Pt levels in blood and urine. According to these workers,6 normal Pt values in urine ranged between 0.5 and 14.3 ng 1-'; the limit of detection was 0.2 ng 1-'. Because of their insufficient sensitivity other elemental detec- tion methods such as AAS ICP-AES and quadrupole (Q) ICP-MS do not allow the determination of environmental Pt levels in biological matrices without extensive enrichment. Recently a powerful analytical method has been presented offering the possibility of rapid multi-element analysis in com- bination with excellent detection These studies used a new type of magnetic sector field ICP-MS instrument (ELEMENT; Finnigan MAT Bremen Germany) permitting operation in three different resolution modes (m/Am z 300,3000 and 7500 10% valley definition). Irrespective of the possibility to eliminate interferences instrumental detection limits of non- interfered isotopes in the low resolution mode are improved by about two orders of magnitude compared with conventional Q-ICP-MS.7.8 This is mainly a result of the extremely low instrumental background level of < 1 count s-'.Thus the practical detection limits in real matrices are generally not affected by the instrumental background but are mainly limited by the blank concentrations of the reagents memory effects or spectral interferences which become more important with decreasing analyte concentration especially in complex matrices such as blood and urine.Conventional digestion methods such as heating with large amounts of acids in open or closed systems are inconvenient for this application because the background of the acids even at the highest commercial degree of purity and after additional sub-boiling distillation are unacceptably high. Additionally these procedures lead to high acid concentrations in the digestion solutions which are not tolerated by ICP-MS and thus have to be diluted prior to analysis. UV photolysis seems to be a promising alternative using hydroxyl radicals as the oxidizing agent. In principle the mineralization takes place without addition of oxidizing agent but in practice it is useful to add minimal amounts of acid and hydrogen peroxide (up to 1%) to enhance the efficiency and oxidation rate.Owing to the small amounts of reagent required UV digestion results in very low blank values. The major drawback of this digestion method is that its application is limited to fairly simple matrices with a relatively low content of organic matrix such as natural and waste water beverages urine and According to Kroder et al.,13 the carbon content of the sample solutions should not exceed 100 mg 1-'. This paper presents the results of the determination of Pt in human urine at physiological concentrations using UV pho- tolysis and magnetic sector field ICP-MS in the low resolution mode. The relevance and severity of blank values from reagents and spectral interferences in the determination of Pt in urine were investigated. EXPERIMENTAL Instrumentation Magnetic sector field ICP-MS was performed with the ELEMENT instrument (Finnigan MAT).Details of the instru- mental operating conditions are given in Table 1. The instru- ment was operated in an air-conditioned laboratory ( 18-22 "C) equipped with a filter to remove dust particles. Pt was measured at masses 194 195 and 196. Reagents and Standard Solutions In order to reduce the risk of contamination all work was performed on a clean bench. Before use all materials and chemicals were randomly checked for contamination. Journal of Analytical Atomic Spectrometry October 1996 Vol. 1 1 (91 3-91 6) 91 3Table 1 measurements Instrumental conditions for magnetic sector field ICP-MS Instrument Resolution Rf power Pump speed Scan type Integration time Torch Nebulizer Spray chamber Cones Intermediate gas Outer gas Nebulizer gas Wash time Masses Scan time Passes Runs Total scan time Total samples Sample time ELEMENT (Finnigan MAT) m/ Am = 3 00 1.30 kW 1.4 ml min-' Magnetic jump with electric scan 50 ms Fassel torch Meinhardt TR-30-A3 Scott type ('double-pass') at 20 "C Ni (standard or high-performance) Argon 0.76 1 min-' Argon 13.3 1 min-' 0.63 1 min-' 12 min 194 195 196 8 s 5 2 5.08 s 167 0.025 s Polyethylene vials were cleaned with HNO (1 +2 50"C 30 min) rinsed with ultrapure water and dried at 50 "C.Ultrapure HNO from two different manufacturers was checked for Pt background levels Ultrex I1 (70%; Baker GroB-Gerau Germany) was used without further purification; Suprapur ( 6 5 %; Merck Darmstadt Germany) was further purified by sub-boiling distillation in a quartz device.Formation of HfO' interferences (as shown in Fig. 1) was measured using a single-element standard solution in 0.5% HNO containing 10 mg I-' Hf (prepared from a single- element standard solution for AAS; Aldrich Heidenheim Germany). Sample Collection and Storage Morning urine samples from 16 healthy volunteers with no known exposure to Pt were collected in acid-washed poly (propylene) bottles. For stabilization purposes they were acidi- fied with concentrated HNO (Ultrex 11; 1 ml per 100 ml of urine) and stored at - 20 "C until analysis. Digestion UV photolysis was carried out with a UV-1000 digester (Hans Kurner Analysentechnik Rosenheim Germany) using a 1000 W medium-pressure Hg lamp.Before being divided into aliquots the sediment of the urine samples was distributed as homogeneously as possible by rigorous shaking to obtain a representative portion of the specimen. Filtration of the urine samples must be avoided because the sediment may include considerable amounts of analyte. Aliquots (20 ml) of acidified urine were mixed with 500 p1 of H202 (Suprapur; Merck) and digested for 10 min at a water cooling flow of 0.9 ml min-'. A second 500 p1 portion of H202 was then added and photolysis was continued for a further 35 min resulting in a clear colourless solution. Calibration Calibration was performed by the standard additions pro- cedure by adding dilute single-element standard solutions (for AAS; Aldrich) in 0.5% HNO (Ultrex 11) containing 5 10 and 50 ng I-' Pt which were prepared freshly every day.Aliquots (500 pl) of these standard solutions were added to 500 pl of digested sample solution and diluted with 2 ml of ultrapure 91 4 Journal of Analytical Atomic Spectrometry October 1996 water. The linearity of the calibration procedure was checked in the concentration range between 0.5 and 200ng 1-'. In routine use standards lower than 5 ng 1-' were not used for standard additions because the uncertainty of the calibra- tion process increases when the limit of determination is approached. No internal standard was used for calibration because in the lowerng 1-l range it is not possible to find an element that can reflect the behaviour of the analyte (e.g. its ionization energy and monoxide bond strength) and which is at the same time not present in the sample or interfered with by blanks and spectral interferences. lo3Rh for example is in this context unsuitable as an internal standard because its determination in the ng 1-' range is interfered with by spectral interferences caused by 206Pb2+ 63C~40Ar+ and s7Sr'60+ which have been described in detail in a previous paper.7 These spectral inter- ferences resulted in a Rh blank level of about 30 ng 1-' in blood samples.Detection Limit and Reliability Criteria The detection limit was calculated as three times the standard deviation of replicate measurements (n = 10) of blanks contain- ing the same concentrations of HNO and H,02 as the digested urine samples. All analyses were carried out under internal quality control conditions.The within-series imprecision was calculated from a urine sample (geometric mean =0.72 ng 1-' Pt) that was analysed ten times. For internal quality control a control sample with a known Pt concentration was analysed within each analytical series. As a control sample with a known or certified Pt concentration is not commercially available it was prepared in the laboratory by pooling acidified urine samples and stored in portions in a refrigerator at -18°C. External quality control was achieved by participation in an interlaboratory comparison programme (Deutsche Gesellschaft fur Arbeits- und Umweltmedizin Erlangen Germany). The designated Pt concentration of the circulating urine samples was 22.72 ng 1-'; our result was 23.05 ng 1-'.RESULTS AND DISCUSSION By using magnetic sector field ICP-MS in the low resolution mode in combination with UV photolysis as the sample preparation step Pt was detectable in the urine samples of all 16 volunteers. Urinary Pt levels ranged between 0.48 and 7.65 ng 1-' with an arithmetric mean of 1.72ng 1-'. The within-series imprecision of a urine sample (geometric mean 72.3 ng 1-' Pt) that was digested and analysed ten times was 17.7%. Our results are in complete accordance with those of Messerschmidt et ~ l . ~ who determined physiological Pt levels in the urine samples of 14 persons by adsorptive voltammetry. These workers reported Pt concentrations between 0.5 and 14.3 ng 1-' with a geometric mean of 1.1 ng 1-'. The detection limit of our procedure was 0.24ng 1-' using a standard skimmer cone made of nickel.A so-called high-performance (HP) nickel skimmer cone which was reported to enhance the sensitivity by a factor of 2-3,14 was also tested. However no improvement in the detection limit was achieved because the analyte and background signals were increased to a similar extent. With the HP skimmer cone the detection limit for Pt was 0.27 ng I-'. The detection limit obtained here with UV photolysis- magnetic sector field ICP-MS is comparable to that reported for the extremely sensitive voltammetric procedure described in ref. 6 viz. 0.2 ng 1-'. In comparison with magnetic sector field ICP-MS the voltammetric procedure is more susceptible to interferences and more labour-intensive. The voltammetric determination of Pt is severely interfered with by residual carbon surfactants and by HNO,.' In order to eliminate the VOl.11organic matrix a complete mineralization of all the organic sample components is required which was achieved6 by high- pressure ashing with a mixture of concentrated mineral acids including HNO as oxidizing agent. Prior to the determination step the excess of HNO has to be eliminated by heating in open vessels with concentrated H2S04 and HC1,6 a procedure which is prone to interferences caused by contamination or elemental losses. According to Messerschmidt et ul.," the detection limit of their procedure is also limited by blank values from reagents vessels etc. In contrast ICP-MS does not require the complete removal of the organic matrix; in fact the direct analysis of diluted urine samples without any sample pre-treatment has been described.15-17 Mulligan et ~ 1 .' ~ diluted the urine samples 10-fold with 0.28 mol 1-' HNO while Schramel et ~ 1 . ' ~ performed a 1+2 dilution. In our opinion UV photolysis prior to ICP-MS measurements is preferable to direct analysis if solution nebulization is applied for sample introduction. The removal of the organic matrix leads to a reduction of non-spectral interferences such as memory effects or blocking of the sampling cone. The use of UV photolysis for the digestion of urine samples was also employed by Ensslin et al.," who determined urinary Pt in occupationally exposed hospital personnel. They irradiated a mixture of 1 ml of urine 100 pI of concentrated H2S0 and 200 pl of H20z for 2 h.For creatinine contents of > 1.5 g 1-' they had to use smaller urine volumes to obtain complete mineralization. For ICP-MS measurements it is possible to irradiate a mixture of 20 ml of urine 0.7% HNO and 1 in1 of H202 because complete mineralization is not necessary. HNO was chosen because it is generally regarded as the best acidic medium for ICP-MS since the constituent elements (H N 0) are already present in the air entrained by the plasma. 4s a consequence the formation of polyatomic interferences is not increased by HNO addition." Additionally almost all elements form easily soluble nitrates preventing precipitation of salts of low solubility in the sample solution or the sample introduction system resulting in analyte losses memory etrects and blockage of the nebulizer and cones.The detection limit of our procedure is a factor of 4 lower than that reported by Schramel et who determined Pt directly in diluted urine samples (1 +2) by applying ETV-Q-ICP-MS. The proposed procedure is also about two orders of magnitude more sensitive than Q-ICP-MS with conventional pneumatic nebulization which has been used by other ~ o r k e r s . ~ ' ~ ' ~ The lower detection limits of ETV sample introduction over pneumatic nebulization are a result of the higher transport efficiency into the plasma and the possibility to remove interfering matrix component^.'^ According to Schramel et a1.,16 the application of ETV for sample introduc- tion instead of pneumatic nebulization resulted in a 20-fold improvement in the detection limit for Pt determination in urine.The disadvantages of ETV are the transient nature of the signals obtained and the fact that the analytes require different optimum furnace conditions both of which restrict the ability to perform multi-element determinations. Compared with the detection limit for ETV-ICP-AES given by Alimonti et our procedure is better by a factor i)f at least 1000. The detection limit of the procedure described here is lower by a factor of more than 100 than that obtained by ETXAS even after a 25-fold enrichment using solvent extraction8 21 A survey of the detection limits for Pt in urine obtained with different analytical procedures is given in Table 2. Our results indicate that the determination of physiological Pt levels in urine is not hampered by spectral interference$.As illustrated in Fig. 2 the isotope pattern of 194Pt I9'Pt and 196Pt obtained in urine samples is in conformity with the natural pattern. Although the formation of HfO+ interfering at rn/z=194 195 and 196 can be demonstrated by measuring 2.2e+008 1.8e+008 1.4e+008 1 e+008 6e+008 2e+007 0 - - 8000 - - 6000 - - 4000 - - 2000 - +n 176 177 178 179 I80 181 . " - 50 (6) 80HP60" .4e+006 - 40 30 20 10 0 193 194 195 196 197 Mass Fig. I LR mode) Spectral interferences by HfO' (10 mg 1-' Hf in 0.5% HNO,; Table 2 Determination of Pt in urine comparison of detection limits (DLs) obtained with different analytical procedures Method UV photolysis-HR-ICP-MS (LR mode standard Ni cones) UV photolysis-HR-ICP-MS (LR mode H P Ni cones) Dilution-Q-ICP-MS Dilution-Q-ICP-MS Dilution-ETV-Q-ICP-MS High-pressure digestion-adsorptive UV photolysis-adsorptive voltammetry Solvent extraction-ETAAS Dilution-ETV-ICP- AES vol tammetry DL/ ng l-' 0.24 0.27 60 20 1 .o 0.2 1.8 70 500 Ref.This work This work 15 16 16 6 11 8 20 50 Dl DIP1 195Pt I 30 s- 20 10 0 natural Pt abundance urine sample Fig. 2 Determination of physiological Pt levels in urine by magnetic sector field ICP-MS at LR mode-isotope ratios. The urine sample contained 28.7 ng 1-' Pt (calculated as lg5Pt) Jclurnul of Analyticul Atomic Spectrometry October 1996 Vol. 11 915I I 100 400 4501 Mass Fig. 3 Determination of physiological Pt levels in urine-typical mass spectrum. The urine sample contained 7.7 ng I-' Pt (calculated as lg5Pt) Hf standard solutions this interference is in practice of no relevance because of the very low Hf concentrations in urine producing a signal of up to 100 counts s-' at masses 178 179 and 180.The corresponding mass spectrum of a 10 mg 1-' Hf standard solution illustrating the formation of HfO + is given in Fig. 1; the average formation rate (HfO+/Hf+) under stan- dard instrumental multi-element conditions was 0.4% (range O.2-O.6%). Owing to the correct Pt isotope ratios obtained in all urine samples it is concluded that there are no significant spectral interferences at masses 194 195 and 196 (Fig. 3). The determination of Pt at these extremely low concen- trations is mainly affected by blank values originating from reagents such as HN03 and H202.The Pt blank values found in HNO [Ultrex I1 (Baker) and Suprapur (Merck)] and H202 [ p r o analysi and Suprapur (Merck)] related to a HNO concentration of 5% and a H202 concentration of 1.5% are given in Table 3. The given blank values consist of the blank values of the reagent and of the ultrapure water used for dilution and of the blank of the sample introduction system. The results show that Pt blank values originating from HN03 and H,02 are of the same order of magnitude as the lower physiological Pt levels in urine. After UV photolysis blanks containing 0.7% HN03 and 1.5% H202 had a mean total Pt concentration of 0.38 ng 1-' (standard deviation 0.075 ng 1-'). These results show that reagent addition should be restricted to the required minimum. CONCLUSIONS The results presented here clearly indicate the superiority of magnetic sector field ICP-MS for the determination of urinary Pt levels in both unexposed and exposed persons.For elements with non-interfered isotopes such as Pt the full sensitivity of the instrument can be used resulting in detection limits that are comparable to those of adsorptive voltammetry. In contrast to other precious metals such as Rh and Pd,7 the determination of Pt in urine is relatively unaffected by spectral interferences. The practical detection limits are for Pt not affected by the instrumental background but are mainly limited by the blank concentrations of the reagents and Table 3 Mean blank values (ng I-') in ultrapure HNO and H202 solution (LR mode) Aqueous solution 5% HNO [ Ultrex IT (Baker)] 5% HNO [ Suprapur (Merck)] 1.5% H20 [pro analysi (Merck)] 1.5% H202 [ Suprapur (Merck)] Pt/ng I-' (n= 10) 0.15 0.13 0.28 0.18 memory effects.It is expected that detection limits can be further improved if the potential of the ELEMENT instrument is fully utilized. In this context it is necessary to improve the analytical technique to reduce blanks and memory effects. The purity of commercially available reagents should also be improved. UV photolysis was found to be advantageous for the diges- tion of urine. With solution nebulization for sample introduc- tion UV photolysis is preferable to direct urine analysis because the removal of the organic matrix leads to a reduction of interferences such as memory effects and less blockage of the sample introduction system. In contrast to mineral acid digestion techniques or direct urine analysis an extensive dilution of the irradiated samples is not necessary.In order to correct for non-spectral interferences caused by high mat- rix concentrations the standard additions procedure is recommended. The described procedure is also applicable to the simul- taneous determination of other precious metals such as Pd Ag and Au and other environmentally relevant metals such as Cd Pb T1 and Hg." For the monoisotopic element Rh the application is so far restricted to elevated urine levels because its determination is interfered with by spectral interferences which cannot be separated at m/Am 2 7500.7*8 The authors thank Hans Kurner Analysentechnik (Rosenheim Germany) for temporary disposal of the UV digester C.Kranich for technical assistance and A. Landwehr foreign- language correspondent for linguistic revision of the manuscript. REFERENCES 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 Hodge V. F. and Stallard M. O. Environ. Sci. Technol. 1986 20 1058. Schierl R. and Fruhmann G. Sci. Total Environ. 1996 182 21. Zereini F. Zientek C. and Urban H. Z. Umweltchem. Okotox. 1993 5 130. Waber M. Laschka D. and Peichl L. Z. Umweltchem. Okotox. 1996 8 1. Hoppstock K. Alt F. Cammann K. and Weber G. Fresenius' Z. Anal. Chem. 1989 335 813. Messerschmidt J. Alt F. Tolg G. Angerer J. and Schaller K.-H. Fresenius' J. Anal. Chem. 1992 343 391. Begerow J. and Dunemann L. J. Anal. At. Specrrom. 1996 11 303. Begerow J. Dissertation Clausthal-Zellerfeld 1995. Moens L. Vanhaecke F. Riondata J. and Dams R. J. Anal. At. Spectrom. 1995 10 569. Moens L. Verrept P. Dams R. Greb U. Jung G. and Laser B. J. Anal. At. Spectrom. 1994 9 1075. Ensslin A. S. Pethran A. Schierl R. and Fruhmann G. Int. Arch. Occup. Environ. Health 1994 65 339. Saur D. and Spahn E. GITFachz. Lab. 1994 2 103. Kroder H. Bauer K.-H. and Saur D. Vom Wasser 1994,82,269. Hamester M. Greb U. and Rottmann L. poster presented at the European Winter Conference on Plasma Spectrochemistry Cambridge 1995. Mulligan K. J. Davidson T. M. and Caruso J. A. J. Anal. At. Spectrom. 1990 5 301. Schramel P. Wendler I. and Lustig S. Fresenius' J . Anal. Chem. 1995 353 115. Vanhoe H. J. Trace Elem. Electrolytes Heulth Dis. 1993 7 131. Jarvis I. in Handbook of Inductively Coupled Plasma Mass Spectrometry eds. Jarvis K.E. Gray A.L. and Houk R.S. Blackie Glasgow 1992 pp. 172-224. Carey J. M. Byrdy F. A. and Caruso J. A. J. Chromatogr. Sci. 1993 31 330. Alimonti A. Petrucci F. Dominici C. and Caroli S. J. Trace Elem. Electrolytes Health Dis. 1987 1 79. Begerow J. Turfeld M. and Dunemann L. Anal. Chim. Acta submitted for publication. Paper 6/03870K Received June 4 1996 Accepted July 10 1996 91 6 Journal of Analytical Atomic Spectrometry October 1996 Vol. 11