ONG ET AL.: JOURNAL OF AOAC INTERNATIONAL VOL. 82, NO. 4, 1999

Transcription

1 ONG ET AL.: JOURNAL OF AOAC INTERNATIONAL VOL. 82, NO. 4, RESIDUES AND TRACE ELEMENTS Determination of Arsenic in Traditional Chinese Medicine by Microwave Digestion with Flow Injection-Inductively Coupled Plasma Mass Spectrometry (FI-ICP-MS) ONG ET AL.: JOURNAL OF AOAC INTERNATIONAL VOL. 82, NO. 4, 1999 ENG-SHI ONG, YUK-LIN YONG, and SOO-ON WOO Institute of Science and Forensic Medicine, Department of Scientific Services, 11 Outram Rd, Singapore A simple, rapid, and sensitive method with high sample throughput was developed for determining arsenic in traditional Chinese medicine (TCM) in the form of uncoated tablets, sugar-coated tablets, black pills, capsules, powders, and syrups. The method involves microwave digestion with flow injection-inductively coupled plasma mass spectrometry (FI-ICP-MS). Method precision was % (relative standard deviation, n = 6) for different concentrations of arsenic in different TCM samples analyzed by different analysts on different days. Method accuracy was checked with a certified reference material (sea lettuce, Ulva lactuca, BCR CRM 279) for external calibration and by spiking arsenic standard into different TCMs. Recoveries of 89 92% were obtained for the certified reference material and higher than 95% for spiked TCMs. Matrix interference was insignificant for samples analyzed by the method of standard addition. Hence, no correction equation was used in the analysis of arsenic in the samples studied. Sample preparation using microwave digestion gave results that were very similar to those obtained by conventional wet acid digestion using nitric acid. Analysis of arsenic in traditional Chinese medicine (TCM) is of importance because arsenic is a well-known toxic metal. A 1979 law in Singapore states that the content of arsenic in TCM should not be greater than 5 ppm (by weight; 1). TCMs usually contain herbal products and come in the form of tablets (coated or uncoated), concentrated pills, syrups, capsules, and soluble granules (2). Arsenic in TCM is usually determined by digesting samples with a single acid or a combination of mineral acids and analysis by hydride generation atomic absorption spectrometry (3). Other methods include use of a hyperpress digestor and analysis by Zeeman graphite furnace atomic absorption spectrometry (4). Received October 19, Accepted by JS February 23, Sample preparation is the most critical and most time-consuming step in chemical analysis. Wet acid digestion has been the method of choice because it gives good recoveries for most metals (5). With microwave digestion using a closed vessel, based on the coupling of electromagnetic radiation with mineral acid to produce high pressure and temperature, the reaction takes less time to complete than with conventional means (6, 7). Closed-vessel microwave digestion also minimizes loss of volatile analytes such as arsenic, and it has been used on botanical, vegetable, and plant reference materials (6, 8, 9) with good recovery and precision. Inductively coupled plasma mass spectrometry (ICP-MS) is an attractive system for determining trace metals in botanical materials because of its excellent detection limit, wide linear range, multielement capability, and ability to measure isotope ratios. ICP-MS had been used to analyze trace elements in plant reference materials and other vegetables (8 10). Use of flow injection (FI) with ICP-MS is an attractive option compared with direct aspiration of sample. The result is a transient signal from a well-defined volume of sample injected into a carrier stream that carries the sample to the detector. The advantages of combining FI analysis with ICP-MS include high sample throughput, low washout time, reduced sample volumes, and automation of sample preparation (11). Spectral interference of chloride in determination of arsenic by ICP-MS is well known (8, 10, 12). Different correction equations for determination of arsenic by ICP-MS in the presence of chloride have been used (8, 10, 13), the most common equation of which is as follows: 75 As = 75 Signal 3.065[ 77 Signal ( Signal)] (1) Equation 1 is based on the ratio of the 40 Ar 35 Cl signal to that of 40 Ar 37 Cl being equal to the ratio of the abundance of the 2 isotopes 35 Cl and 37 Cl; that is, 75.77: The signal at m/z 77 ( 77 Se) for 40 Ar 37 Cl indicates the amount of 40 Ar 35 Cl or the extent of chloride interference. Correction for the possible presence of selenium in the matrix is based on the 82 Se signal. To our best knowledge, there are no reports on determination of arsenic in TCM by closed-vessel microwave digestion and analysis by FI-ICP-MS. Our aim was to develop a simple, rapid, and accurate method to analyze arsenic in a large num-

2 964 ONG ET AL.: JOURNAL OF AOAC INTERNATIONAL VOL. 82, NO. 4, 1999 Table 1. ICP-MS Parameter Instrument operating conditions for the Setting Table 3. Counts obtained at m/z 77 signal in the presence or absence of chloride Counts at m/z 77 RF power 900W Nebulizer gas flow Argon at 0.75 L/min Nebulizer Cross flow Peristaltic pump M312, made in France, pumped at 1.0 ml/min Data acquisition Peak-hopping mode Dwell time 100 ms/amu Sweep/reading 1 Readings/replicate 100 No. of replicates 2 Mass analyzed: Analyte 75 As For correction 77 Se and 82 Se Internal standard 115 In ber of TCMs. The effect of spectral interference by chloride on 75 As determination by FI-ICP-MS was investigated, and matrix interference was checked by standard addition. Samples were digested by closed-vessel microwave digestion with nitric acid, and arsenic was determined by FI-ICP-MS. Method validation involving trace analysis of metals was based on various reports (14, 15). Finally, the efficiency of closed-vessel microwave digestion was compared with that of wet ashing with concentrated nitric acid. Experimental Reagents and Standard Reference Materials High-purity concentrated nitric acid (Aristar grade; BDH, Dorset, UK) was used for acid digestion. Deionized water from Millipore Alpha-Q system (Millipore Corporation, Bedford, MA) that had the capability to produce 18 m cm was used. One certified plant reference material (sea lettuce, BCR Analyte In the absence of Cl CRM No. 279; Commission of the European Communities, Brussels, Luxembourg) was used. Arsenic and indium reference standards (1000 mg/l) were obtained from Perkin-Elmer Corporation (Norwalk, CT). Working standards of arsenic containing 0, 10, 20, 50, and 100 µg/l with 5 µg indium/l as internal standard were prepared in 1% nitric acid by dilution from reference standards before use. Microwave Digestion System A CEM MDS-81D microwave digestion system (CEM Corporation, Matthews, NC) with a 12-position Teflon PFA vessel (from CEM user manual; registered trademark of E.I. Dupont Co.) and a capping station was used. The digestion vessels came with controlled-release valves and were capable of withstanding pressures of up to 1.7 kpa (250 psi). Before use, PFA vessels were soaked in 10% nitric acid. The system initially was programmed at 25% power for 5 min, 50% power for 5 min, and 65% power for 20 min. Microwave power range, 0 100% of full power (630 ± 70 W) in 1% increments. Sample Preparation With 100 mg Cl/L With 1000 mg Cl/L 10 µg As/L µg As/L µg As/L For pills and coated or uncoated tablets, 15 to 20 pills or tablets were taken randomly and ground with a mortar and pestle. For capsules, the contents of 15 to 20 capsules were emptied into a glass container and mixed. Powders and syrups were used without further treatment. A 0.5 g sample was weighed into the PFA vessel followed by addition of 5 ml Table 2. Chloride interference in determination of arsenic Analyte Concentration of As from equation 1, µg/l Concentration of As based on uncorrected signal, µg/l 10 µg As/L with 100 mg Cl/L µg/ As/L with 1000 mg Cl/L µg As/L with 100 mg Cl/L µg As/L with 1000 mg Cl/L µg As/L with 100 mg Cl/L µg As/L with 1000 mg Cl/L

3 ONG ET AL.: JOURNAL OF AOAC INTERNATIONAL VOL. 82, NO. 4, Table 4. Comparison of data obtained from external calibration and standard addition Sample type Ratio of slope for calibration plot using In as internal standard to slope for standard addition method As found by external calibration, µg/l As found by method of standard addition, µg/l TCM capsule TCM syrup TCM capsule TCM sugar-coated tablet TCM uncoated tablet TCM black pills concentrated nitric acid. For syrups, 2 ml was weighed into the PFA vessel. The resulting solution was filtered through Whatman No. 41 filter paper (Whatman International Ltd., Hillsboro, OR) and diluted to 25 or 50 ml with deionized water. This sample preparation also was used for addition of arsenic standard (0 50 µg/l) to the sample. For wet ashing, 0.5 g sample was weighed into a Nessler tube followed by 5 ml concentrated nitric acid. The sample was digested for 4 h until no brown fume was observed and made up to 25 or 50 ml. Blanks containing only concentrated nitric acid were prepared at the same time for microwave digestion and wet ashing. To check method recovery, 50 µl of a 10 mg/l arsenic standard was added to 0.5 g TCM followed by addition of 5 ml concentrated nitric acid. FI-ICP-MS Measurement A standard Perkin-Elmer SCIEX Elan 6000 ICP-MS (Thornhill, Ontario, Canada) was operated at the conditions listed in Table 1. The FI system was Perkin-Elmer FIAS 400, and the autosampler was Perkin-Elmer AS91. The program used to run the Perkin-Elmer FIAS 400 was that recommended in the Perkin-Elmer SCIEX Elan 6000 ICP-MS system software guide. The loop size for the Perkin-Elmer FIAS 400 was 500 µl, and water was used as the carrier solution. Results and Discussion Method Optimization The calibration plot for arsenic, obtained with indium as internal standard, was linear between 0 and 100 µg/l. Instrument limit of detection (LOD) and limit of quantitation (LOQ) based on 3 and 10 times, respectively, the standard deviation of the blank were 0.12 and 0.36 µg/l. Method LOD and LOQ based on 0.5 g sample were and mg/kg, respectively. Carryovers of arsenic and indium were observed after standards containing high levels of arsenic were run. The memory effect of arsenic and indium disappeared after applying to the FI system a wash cycle with 1% nitric acid. The effects of chloride in the determination of arsenic by FI-ICP-MS are shown in Table 2. The presence of chloride up to 100 mg/l did not significantly increase the 75 Signal. This conclusion was based on the observation that the concentrations of arsenic calculated on the basis of the corrected signal (equation 1) and on the basis of the uncorrected signal did not differ much at different levels of arsenic. The numbers of counts obtained at 77 Signal ( 40 Ar 37 Cl), which indicates the effect of interference, in the presence or absence of chloride at 100 mg/l did not differ significantly (Table 3). Table 5. Comparison of the number of counts at m/z 77 for the various sample types No. of counts at m/z 77 Sample type Sample blank Sample Figure 1. Calibration plot of external calibration (series 2) with standard addition (series 1) in black pills. TCM capsule TCM capsule TCM sugar-coated tablet TCM uncoated tablet TCM syrup TCM black pills

4 966 ONG ET AL.: JOURNAL OF AOAC INTERNATIONAL VOL. 82, NO. 4, 1999 Table 6. Sample Method precision Arsenic, mg/kg (n =6) Relative standard deviation, % TCM uncoated tablet 6.16 ± 0.29 a ± 0.47 b 7.87 TCM syrup ± b TCM sugar-coated tablet 0.31 ± 0.02 a 6.83 TCM capsule ± 0.05 c 2.73 TCM capsule ± 0.02 a 3.77 a, b, c Analysts 1, 2, and 3, respectively. Table 8. Comparison of microwave digestion and wet ashing digestion Samples microwave digestion, mg/kg wet ashing, mg/kg TCM capsule TCM sugar-coated tablet TCM syrup TCM black pills TCM uncoated tablet The method of standard addition commonly is used to determine analytes in a sample if analyte matrix interaction can occur that will produce an inaccurate result with external calibration. It is also used to check the accuracy of an analytical result when no reference material is available (16). The method of standard addition was used to validate the method (Table 4). Arsenic contents determined by standard addition were similar to those determined by external calibration. Figure 1 shows an external calibration plot with standard addition plot for black pills. Therefore, no significant analyte matrix effect in the TCM matrix was observed. The numbers of counts at m/z 77 for different samples (Table 5) did not differ significantly from those of the sample blank. The results of the interference studies and studies of the method of standard addition indicate that a correction equation does not need to be applied to the TCM samples analyzed in this work. However, a correction equation may be required for samples with high levels of chloride. Method Precision and Accuracy Table 7. Comparison of results obtained from microwave digestion by different analysts Samples analyst 1, mg/kg analyst 2, mg/kg Parameters that are of main concern in practical method validation include method precision, accuracy, specificity or any possible interference present in the sample matrix, and repeatability and reproducibility. When suitable certified reference materials are not available, checking method accuracy is more difficult. In this work, method accuracy was checked by use of a certified reference material (sea lettuce, CRM No. 279), comparison of wet acid digestion with microwave digestion, spiking of arsenic into TCM samples, and application of the standard addition method. TCM samples in the form of tablets (uncoated and sugar coated), capsules, powder, black pills, and syrup were used. Method precision was determined from data obtained by 3 different analysts on different days for different levels of arsenic in different TCM samples (Table 6). Method precision was % (relative standard deviation, n = 6). The results for CRM No. 279 from different analysts on different days (Table 7) were within the range certified. In this case, no correction equation was applied because the chloride content was certified to be 25 ppm. Amounts of arsenic in 2 different types of TCM determined by different analysts on different days (Table 7) were in close agreement. Comparison of the microwave digestion method with wet acid digestion (Table 8) showed that the content of arsenic determined by microwave digestion is in close agreement with that determined by wet acid digestion. Finally, recovery of arsenic standard spiked into TCM was good (Table 9). Conclusions Determination of arsenic in TCM using closed-vessel microwave digestion and analysis by FI-ICP-MS is simple, sensitive, and rapid, with good precision and accuracy. FI-ICP-MS allows high sample throughput with minimal memory effect. The method was demonstrated to be repeatable with different analysts in our laboratory. Low levels of chloride in the sample does not significantly enhance the 75 As signal or cause matrix interference. Closed-vessel microwave digestion and wet acid digestion do not give significantly different efficiencies. TCM uncoated tablet TCM sugar-coated tablet Sea lettuce (CRM No. 279) certified value of As = 3.09 mg/kg 2.75 (recovery = 89.0%) 2.85 (recovery = 92.2%) Table 9. Recovery of arsenic spiked into TCM Samples Arsenic recovered, % TCM uncoated tablet 97.6 TCM syrup 97.4

5 ONG ET AL.: JOURNAL OF AOAC INTERNATIONAL VOL. 82, NO. 4, Acknowledgments We thank Ng Tju Lik, director of the Department of Scientific Services, and Chua Teck Hock, director of the Health Science Division, for their interest and support. We also thank Chia Soon Lai and Divya Velaedam from Singapore Polytechnic for their assistance. References (1) The Sale of Drugs (Prohibited Substances) Regulations 1979, Government Printers, Singapore (2) Zhu, C.H. (1989) Clinical Handbook of Chinese Prepared Medicine, Paradigm Publications, Brookline, MA (3) Li, X.L., Zhou, P.S., Zheng, Y.M., Yang, C.L., & Wang, X.R. (1995) Fenxi Shiyanshi 14, (4) Liang, Q., Lu, H., & Li, X. (1995) Zhonggou Yaoke Daxue Xuebao 26, (5) Gorsuch, T.T. (1959) Analyst 84, (6) Kingston, H.M., & Jassie, L.B. (1986) Anal. Chem. 58, (7) Kingston, H.M., & Jassie, L.B. (1995) Introduction to Microwave Sample Preparation: Theory and Practice, American Chemical Society, Washington, DC (8) Wu, S., Feng, X., & Wittmeier, A. (1997) J. Anal. At. Spectrom. 12, (9) Bettinelli, M., Spezia, S., Baroni U., & Bizzarri, G. (1997) in Plasma Source Mass Spectrometry: Development and Applications, G. Holland & S.C. Tanner (Eds), The Royal Society of Chemistry, Cambridge, UK, pp (10) Broeck, K.V.D., Vandecasteele, C., & Genus, J.M.C. (1997) J. Anal. At. Spectrom. 12, (11) Anderson, S.T.G., Taylor, M.J.C., & Willians, S.J.S. (1993) in Applications of Plasma Source Mass Spectrometry II,G. Holland & N.A. Eaton (Eds), The Royal Society of Chemistry, Cambridge, UK, pp (12) Vanhoe, V., Goossens, J., Moens, L., & Dams, R. (1994) J. Anal. At. Spectrom. 9, (13) Amarasiriwardena, C.J., Lupoloi, N., Potula, V., Korrick S., & Hu, H. (1998) Analyst 123, (14) Bruce, P., Minkkinen, P., & Riekkola, M.L. (1998) Mikrochim. Acta 128, (15) Sargent, M., & Mackay, G. (1995) Guidelines for Achieving Quality in Trace Analysis, The Royal Society of Chemistry, Cambridge, UK (16) Rodriguez, L.C., Campana, A.M.G., Barrero, F.A., Linares, C.J., & Ceba, M.R. (1995) J. AOAC Int. 78,