Microsoft word - full_paper13.doc

International Conference on Metrology of Environmental, Food and Nutritional Measurements
9 – 12 September 2008, Budapest, Hungary DEVELOPMENT OF REFERENCE MATERIAL FOR ORGANOCHLORINE
PESTICIDES IN HERBAL SAMPLE
Yiu-chung Wong1, Tat-ting Kam2, Serena Chan3
1 Analytical & Advisory Services Division, Government Laboratory, Hong Kong, [email protected] 2 Analytical & Advisory Services Division, Government Laboratory, Hong Kong, [email protected] 3 Analytical Laboratories in Anaheim, Inc., Los Angeles, USA, [email protected]
Abstract: Development of reference material for four
Analytical methods for OCP are widely available IN organochlorine pesticides, namely hexachlorobenzene and the literature. Among all the commonly used screening and three isomers (α-, β- and γ-) of hexachlorocyclohexane, in quantification methodologies, GC-ECD and GC-MSD are ginseng root sample (Panax ginseng) to ascertain the quality the most preferred methods because the instrumentations control and validation processes was presented. A total of offer good degree of sensitivity and selectivity for OCP in a more than 300 bottles each containing 25g of Panax ginseng variety of matrices. Despite the majority of OCP test samples was prepared and the reference values were methods that claimed having high recovery and reasonably characterised using a primary measurement method, isotope good precision was mainly relied upon the results of spiked dilution gas chromatography-mass spectrometry. The or fortified samples, it was found that the actual bias might concentrations (± expanded uncertainty) of always be underestimated. As reviewed by a recent survey hexachlorobenzene, α-, β- and γ-hexachlorocyclohexane in on the measurements of residual OCP in food matrices [2], the reference material are 0.198 ± 0.015 mg/kg, 0.450 ± the expectable minimum of the combined relative standard 0.025 mg/kg, 0.213 ± 0.014 mg/kg and 0.370 ± 0.031 mg/kg uncertainty was significantly in the range of 33–49%. To respectively. A portion of the samples was also used in a ensure the reliability of analytical data and to fully proficiency testing (PT) scheme for assessing the testing comprehend the deficiencies of OCP analysis, quality capabilities of field laboratories and the consensus mean assurance through the use of reference materials (RM) and values from the 70 participants in the programme were the participation of proficiency testing schemes are regarded deviated from -2.7% to -14.1% as the assigned reference as the crucial prerequisites in ISO17025 [3]. The usefulness and applications of RM, either non-certified or certified RM (CRM), in quality controls of chemical analysis have been Keywords: reference material, organochlorine pesticides,
clearly explicated [4]. The certification processes for CRM isotope dilution mass spectrometry, proficiency testing are tedious and the production, in terms of quantity and variety, is often too small to satisfy laboratories’ demands. Commercially available matrix CRM for OCP are even 1. INTRODUCTION
severely limited [5] and such CRM in herbal matrices is nowhere identified. The circumstances have recently been Organochlorine pesticides (OCP) have been used discussed by the Reference Materials Committee of extensively to control harmful pests and prevent vegetation ISO/REMCO [6] and the uses of non-certified RM for infections during the last century. Owing to the persistent quality control work of analytical methods were properties, large amount of these chemicals that released recommended as the best alternative whenever suitable through all kinds of anthropogenic activities led to CRM is not present. As a universal rule for all good practice ubiquitous long-term contaminations in the environment and laboratories, analytical techniques that are used to monitor food chains. As a consequence, determinations of residual the levels and fate of contaminants in the samples must be OCP in agricultural products and food are of great calibrated using appropriate calibration materials, and the importance for comprehensive understandings of methods must be properly validated using fit-for-purpose contamination patterns, physiological absorption, matrix-matched RM, to ensure validity of data being distribution, transport pathways as well as the estimation of average daily intake in risk assessments. Since the recognition of the adverse effects of OCP to our ecosystem, We had recently developed a highly precise isotope a huge quantity of accessible scientific data concerning the dilution gas chromatography mass spectrometry (ID-GCMS) analysis of residual OCP in various types of environmental, method for analysing OCP in ginseng root [7] and reported agricultural and animal samples has been reported in the it as a method of choice for the production of RM. In view literature. In particular, with the signing of the Stockholm of rapid growth of herbal medicine trade and its subsequent convention on persistent organic pollutants and frequent of testing volume and lacking relevant CRM in laboratory global corporation programs [1], there is an ever-increasing testing for OCP in herbal matrices, we present the trend for laboratories worldwide to determine OCP. development RM for HCB, α-, β- and γ-HCH in ginseng root sample for method validation and quality control. The assigned reference values of the RM were determined by an 2.4. Moisture Content
accurate GC-IDMS method and the associated expanded uncertainties were estimated from the homogeneity and The bottles were kept in electronic desiccators at stability testing, and the precision of the GC-IDMS method. 25°C and 50% RH. The moisture content of the sub-samples (n = 10) was estimated by taking five bottles randomly and found to be 5.1% (RSD = 0.85%) at the time of bottling.
2. EXPERIMENTAL
2.1. Preparation of bulk sample
2.5. Preparation of Calibration Blend and Sample Blend
About 20 kg of raw radix ginseng (Panax ginseng) Calibration blend (by spiking appropriate quantity of that contained certain levels of targeted residual OCP were 12C6-OCP and 13C6-OCPs) and sample blend samples (by purchased from local market. The samples were rinsed with spiking appropriate quantity of 13C6) were respectively distilled water to remove dirt and other foreign matters, then prepared in accordance with the reported procedure [7]. The were freeze-dried at about -50 °C and < 70 mT for 7 days. concentration ratios of the natural to labelled isotopes in The dried ginseng root was ground to coarse powder by a both blends should be close to unity (0.9 – 1.1) in order to domestic blender, then to fine powder by a high-speed achieve a high degree of accuracy in the ID-GCMS centrifugal mill (ZM200, Retsch, Germany). The powder obtained was subject to passing through 100 µm calibrated sieves. About 8 kg of fine powder were collected and 2.6. Instrumental conditions for ID-GCMS
transferred to a 40-litre commercial blender for thorough mixing within a humidity- and temperature-controlled environment (relative humidity and temperature were HP5973, Palo Alto, USA) with a 30 m x 0.25 mm, 0.25 µm respectively maintained at 50% and 20ºC). Homogeneity of film HP-1707MS column (J & W Scientific, Folsom, USA) the bulk was checked regularly by taking portions of sample was used. Helium carrier gas was set at a flow-rate of 1.0 during mixing and analysing the concentrations of the OCP mL/min, and separation of the four OCP was carried out using a validated GC-ECD method. Satisfactory and under a temperature programme as follows: injector complete mixing was confirmed at day seventh after temperature at 200°C, column temperature at 90°C for 1 min, operation. Aliquots of about 25 g were packaged into clean ramped to 200°C at 50°C/min and held for 6 min, then to and nitrogen-purge amber glass bottles in a clean room 280°C at 20°C/min and held for 5 min. The transfer line and (Class 1000) and 300 bottles were finally prepared. All the ion source were set at 280°C, and the ionisation energy bottled samples were disinfected by γ-ray (Gammacell-1000 under electron ionisation mode at 70 eV. Aliquots of 1 µL Elite, MDS Nordion, Canada) at a dose of about 0.8 kGy to were injected into the GC-MS system under splitless mode, prevent microbial growth, then vacuum-sealed in and the analytes were respectively monitored under the polypropylene bags. single ion monitoring mode at multiple mass channels at m/z 181, 183, 284 for 12C-OCP and 187, 189, 290 for 13C 2.2. Homogeneity Testing
3. RESULTS and DISCUSSIONS
A total of 16 randomly selected samples were analysed in duplicate by a validated GC-ECD method and 3.1. Homogeneity and Stability Testing
the sequence of analysis (32 injections) was arranged in a randomised order. The analysis was performed using the Within variance (CVw) and between variance (CVb) same instrument and completed within the shortest time as of samples were derived from the duplicate analysis of the possible for minimizing errors arising from instrumental 16 samples. At the 95% confidence level, the critical value bias. Statistical approaches were applied to evaluate: A one- is 2.352 for n = 16. Since Ftest < Fcritical, statistical significant way ANOVA was applied to evaluate within-bottle and difference between samples was not detected. The between-bottle difference, hence the inhomogeneity status uncertainty of the sample inhomogeneity (U ) was of the samples. Samples were considered homogeneous if the F-test values were smaller than those of the critical 2.3. Stability Testing
When CVw was equal to or greater than CVb, U was estimated as the upper detection limit by the following Stability testing aiming at assessing the stability of equation, where n is the number of replicate and df is the residual OCPs of randomly selected samples was performed at 25ºC for twelve months (Jan 2006 to Jan 2007). Samples were considered statistically stable if the mean in each standard deviation of the mean value of the OCP obtained As shown in Table 1, the uncertainty arising from sample inhomogeneity was ranging from 0.46% to 2.8%.
Table 1. Inhomogeneity of 16 randomly selected samples
are in good agreements to our previous study. The measurement uncertainty of OCP obtained by IDMS were significantly better than those by conventional GC-ECD and HCB 2.252 3.8 2.5 2.8 GC-MS methods which could be as high as 50% in the concentration range of 0.01 – 10 mg/kg. As shown in Fig. 1, 0.800 1.6 1.8 1.1 all the error components were below 3%; and similar to 1.042 3.1 3.0 0.46 other IDMS studies, RB and RBC were amongst the major 0.492 2.1 3.0 1.8 contribution. Furthermore, the overall bias of the method was assessed with an animal feed reference material (CRM Stability testing indicated that all the four OCP were BCR 115) with less than 2% deviation from the certified stable at both 25°C and 37ºC over the 12-month study period. Deviations of OCP from the mean values obtained Figure 1. Distribution of error components in the IDMS method by homogeneity testing at were ranging from 0.1 to 7.5% at 25°C and from 0.1 to 13.5% at 37°C respectively. The RSDs (U ) of OCP at 25°C were the uncertainty component of stability testing for the reference material.
Table 2. Stability of OCP at 25°C and 37°C

3.2. Results of ID-GCMS
IDMS is recognised as a primary method for chemical analysis and hence an ideal method used for 3.3. Performance of PT
assigning reference values by RM producers [8]. Apart from
A portion of the samples was used in operating a PT being a high accurate technique, IDMS also offers a programme for assessing the competence of routine OCP provision of well definable uncertainty values, in which the testing methods amongst participating laboratories [9]. The overall uncertainty could be conveniently derived. As shown consensus means obtained from the 70 participants using in equation 1, the uncertainty of the reference values (Cx) robust statistics were found to be 0.170 ± 0.095 mg/kg for HCB; 0.393 ± 0.188 mg/kg for α-HCH; 0.219 ± 0.103 mg/kg β-HCH and 0.339 ± 0.149 mg/kg for γ-HCH z) and the bias of isotope ratios in the respectively. The consensus values were in the same order but deviated from -2.7% to -14.1% when compared with the IDMS values. Generally, the performance of most pesticide PT programmes was found to be not impressive [10] and it In this experiment, three independent determinations also happened in the present programme. The wide each comprising of duplicate analysis of five samples were distribution of the OCP results, with RSD ranging from 44% performed from February to May 2007. Using bracketing to 56%, indicated the difficulties and significant uncertainty technique, analyses were arranged by interspersing a in the determination of incurred OCP in herbal matrices calibration blend between each sample blend and the from participants. Hence, the RM prepared could serve as an respective ratios, ie. R essential source for identifying the bias and improving the quality control of OCP determinations. Due to the (ranging from 0.96 to 1.05). Mean concentration (± SD) of unsatisfactory results from testing laboratories in the PT 15 samples for HCB was 0.198 ± 0.004 mg/kg; α-HCH was programme, the results were considered not suitable to be 0.450 ± 0.008 mg/kg; β-HCH was 0.213 ± 0.004 mg/kg and included in the calculation of reference values. γ-HCH was 0.370 ± 0.012 mg/kg respectively. The inter-day variations of the OCP were ranging from 1.7 to 3.2% indicating that the characterisation using the described 3.4. Uncertainty of Assigned Reference
Three main contributors, ie. homogeneity and stability of the materials and the measurements using ID- The relative expanded uncertainties at a coverage GCMS technique, to the uncertainty of the reference values factor (k) of 2 were estimated as the square root of the sum were combined to estimate the relative expanded uncertainty of square of all error components that stated in equation (3) (U) at a coverage factor (k) of 2 as follows: and found to be 5.0% for HCB, 4.4% for α-HCH, 4.9% for β-HCH and 7.8% for γ-HCH respectively. The estimations S. Chan, M.F. Kong, Y.C. Wong, S.K. Wong and D.W.M. Sin, “Isotope Dilution Gas Chromatography-Mass Accordingly, the relative expanded uncertainty of the Spectrometry Analysis of Organochlorine Pesticide Residues in Ginseng Root,” J. Agric. Food Chem., Vol. 55, four OCP in the reference material was ranging from 5.6 % T. Yarita, A. Takatsu, K. Inagaki, M. Numata, K. Chiba and K. Okamoto, “Matrix certified reference materials for Table 3. Expanded Uncertainty of the Reference Material
environmental monitoring from the National Metrology Institute of Japan (NMIJ),” Accred. Qual. Assur., Vol. 12, M.K. Kong, S. Chan, Y.C. Wong, S.K. Wong and W.M. Sin, “Inter-laboratory comparison for the determination of five residual organochlorine pesticides in ginseng root sample,” J. AOAC. Int., Vol. 90, pp. 1133-1141, 2007. 4. CONCLUSION
[10] H.Z. Şenyuva and J. Gilbert, “Assessment of the An RM for four OCP in ginseng root sample was performance of pesticide-testing laboratories world-wide developed. The characterisation for the reference values was through proficiency testing, ” Trends Anal. Chem., Vol. 25, determined by an accurate ID-GCMS technique while the homogeneity and stability of the samples were confirmed satisfactory. The relative expanded uncertainty of the
reference values was estimated from all three components
and found to be within 9%. The product was considered as a
good alternative to use for method validation processes for
residual OCP analysis in herbal matrices in the absence of
any available CRM today.
ACKNOWLEDMENTS

The authors wish to express sincere thanks to Dr. T.L. Ting, Government Chemist of the Government Laboratory
of HKSAR for his support and encouragement during the
course of study.
REFERENCES
[1]
J. de Boer, H. Leslie, S.P.J. van Leeuwen, J.-W. Wegener, B. van Bavel, G. Lindström, N. Lahoutifard and H. Fiedler, “United Nations Environment Programme Capacity Building Pilot Project—Training and interlaboratory study on persistent organic pollutant analysis under the Stockholm Convention,” Anal. Chim. Acta, Vol. 617, pp. 208-215, 2008. A. Ambrus, “Reliability of measurements of pesticide residues in food,” Accred. Qual. Assur., Vol. 9, pp. 288-304, 2004. ISO 17025, “General requirements for the competence of testing and calibration laboratories,” International Standards Organization, Geneva, Switzerland, 2005. H. Emons, T.P.J. Linsinger and B. Gawlik, “Reference materials: terminology and use. Can't one see the forest for the trees?” Trends Anal. Chem., Vol. 23, pp. 442-449, 2003. J. de Boer and E. McGovern, “Certified reference materials for organic contaminants for use in monitoring of the aquatic environment,” Trends Anal. Chem., Vol. 20, pp. 140-159, 2001. A.M.H. van der Veen, A. Fajgelj, H. Emons and S. Sauvage, “Report of the 30th meeting of ISO/REMCO,” Accred. Qual. Assur., Vol. 13, pp. 53-55, 2008.

Source: http://www.imeko.org/publications/tc24-2008/IMEKO-TC24-2008-006.pdf