CERTIFICATION REPORT

The Certification of the Mass Fractions of Elements in Pig Kidney

Certified Reference Material ERM®-BB186

EU

R 25480 E

N -

2012

JRC-IRMM promotes a common and reliable European measurement system in support of EU policies. The primary task of JRC-IRMM is to build confidence in the comparability of measurement results by the production and dissemination of internationally accepted quality assurance tools. JRC-IRMM develops and validates testing methods, produces reference materials, organises measurement evaluation programmes, and provides reference measurements. European Commission Joint Research Centre Institute for Reference Materials and Measurements Contact information Reference materials sales Retieseweg 111 B-2440 Geel, Belgium E-mail: jrc-irmm-rm-sales@ec.europa.eu Tel.: +32 (0)14 571 705 Fax: +32 (0)14 590 406 http://irmm.jrc.ec.europa.eu/ http://www.jrc.ec.europa.eu/ Legal Notice Neither the European Commission nor any person acting on behalf of the Commission is responsible for the use which might be made of this publication.

Europe Direct is a service to help you find answers to your questions about the European Union

Freephone number (*):

00 800 6 7 8 9 10 11

(*) Certain mobile telephone operators do not allow access to 00 800 numbers or these calls may be billed.

A great deal of additional information on the European Union is available on the Internet. It can be accessed through the Europa server http://europa.eu/ JRC73895 EUR 25480 EN ISBN 978-92-79-26169-5 ISSN 1831-9424 doi:10.2787/64882 Luxembourg: Publications Office of the European Union © European Union, 2012 Reproduction is authorised provided the source is acknowledged. Printed in Belgium

CERTIFICATION REPORT

The Certification of the Mass Fractions of Elements in Pig Kidney

Certified Reference Material ERM®-BB186

J. Snell, K. Teipel, H. Schimmel

European Commission, Joint Research Centre Institute for Reference Materials and Measurements (IRMM)

Geel, Belgium

2

Summary

This report describes the preparation of a pig kidney matrix reference material ERM®-BB186, and the certification of the contents (mass fractions) of seven elements. All results are expressed as a mass fraction on a dry mass basis. The preparation and processing of the material, homogeneity studies, stability studies and characterisation are described hereafter and the results are discussed. Uncertainties were calculated in compliance with the Guide to the Expression of Uncertainty in Measurement (GUM) [1] and include uncertainties due to possible heterogeneity, instability and from characterisation. The certified values and their uncertainties are listed in Table 1: Table 1: Certified mass fractions of elements and their uncertainties in pig kidney (ERM®-BB186)

Element Certified value 1,2)

[mg/kg] Uncertainty 3)

[mg/kg]

Cd 1.09 0.05 Cu 36.5 1.8 Fe 255 13 Mn 7.26 0.25 Pb 0.040 0.005 Se 10.3 0.9 Zn 134 5

1) Unweighted mean value of the means of accepted sets of data, each set being obtained in a different laboratory and/or with a different method of determination.

2) The certified value and its uncertainty are traceable to the International System of Units (SI). 2) Certified mass fractions are corrected for the water content of the material (and expressed as dry mass), determined as described in the section "Instructions for use".

3) The certified uncertainty is the expanded uncertainty with a coverage factor k = 2 corresponding to a level of confidence of about 95 % estimated in accordance with ISO/IEC Guide 98-3, Guide to the Expression of Uncertainty in Measurement (GUM:1995), ISO, 2008.

The assigned values are based on a minimum sample intake of 0.2 g.

3

Table of contents

Summary ..................................................................................................................................2 Table of contents......................................................................................................................3 Glossary ...................................................................................................................................4 1 Abstract ...........................................................................................................................6 2 Introduction......................................................................................................................6 2.1 Background .................................................................................................. 6 3 Participants......................................................................................................................7 4 Processing of the material ...............................................................................................9 4.1 Material selection ......................................................................................... 9 4.2 Pig kidney processing .................................................................................. 9 4.3 Additional characterisation measurements................................................. 10 5 Homogeneity studies .....................................................................................................12 5.1 Design of the homogeneity study ............................................................... 12 5.2 Results of the homogeneity study .............................................................. 13 5.3 Minimum sample intake.............................................................................. 16 6 Stability studies..............................................................................................................17 6.1 Short-term stability study............................................................................ 17

6.1.1 Design of the short-term stability study ............................................... 17 6.1.2 Results of the short-term stability study .............................................. 17

6.2 Long-term stability study ............................................................................ 19 6.2.1 Design of the long-term stability study ................................................ 19 6.2.2 Results of the long-term stability study................................................ 20

7 Characterisation ............................................................................................................22 7.1 Design of the characterisation study .......................................................... 22 7.2 Results and technical evaluation – Principles ............................................ 22 7.3 Results and technical evaluation................................................................ 23 8 Certified values and uncertainties .................................................................................26 9 Metrological traceability .................................................................................................29 10 Instructions for use and intended use............................................................................30 10.1 Safety precautions...................................................................................... 30 10.2 Use of materials ......................................................................................... 30 10.3 Intended use .............................................................................................. 30 10.4 Storage conditions...................................................................................... 30 11 Acknowledgements .......................................................................................................31 12 References ....................................................................................................................31 Annex A ERM®-BB186 (Pig kidney) – Results of the homogeneity study .........................32 Annex B ERM®-BB186 (Pig kidney) – Results of the short-term stability study ................44 Annex C ERM®-BB186 (Pig kidney) – Results of the long-term stability study .................49 Annex D ERM®-BB186 (Pig kidney) – Characterisation data ............................................54 Annex E Methods used in the characterisation exercise...................................................64

4

Glossary

α.......................confidence level AAS..................atomic absorption spectrometry ANOVA ............analysis of variance b.......................slope of linear regression CRM.................certified reference material CV....................Cold vapour df ......................degree of freedom (regression) DM ...................dry mass ERM® ...............European Reference Material® GUM ................Guide to the Expression of Uncertainty in Measurement HG ...................Hydride generation ICP-MS ............ inductively coupled plasma mass spectrometry ICP-OES.......... inductively coupled plasma optical emission spectrometry ID ..................... Isotope dilution IRMM ............... Institute for Reference Materials and Measurements ISO................... International Organization for Standardisation JRC..................European Commission’s Joint Research Centre KFT..................Karl Fischer titration LOD ................. limit of detection LOQ ................. limit of quantification MSbetween ..........mean of squares between groups (ANOVA) MSI ..................minimum sample intake MSwithin .............mean of squares within groups (ANOVA) n.......................number of replicates NAA. ................Neutron activation analysis n/a....................not applicable n.c. ...................not calculable n.d....................not determined n.r.....................not reported p....................... level of significance PSA..................particle size analysis RSD ................. relative standard deviation RSDstab............. relative standard deviation of all results of stability study s .......................standard deviation sbb ....................between-bottle heterogeneity standard deviation swb ....................within-bottle heterogeneity standard deviation seb....................standard error of slope b of linear regression SI ..................... International System of Units SSAAS.............Solid-Sampling Atomic Absorption Spectrometry TIMS ................ thermal ionisation mass spectrometry ubb .................... relative standard uncertainty due to between-bottle heterogeneity u*

bb ................... relative standard uncertainty due to heterogeneity that can be hidden by method repeatability

uchar,rel ............... relative standard uncertainty of characterisation exercise uCRM .................combined standard uncertainty of certified value uCRM, rel..............combined relative standard uncertainty of certified value UCRM.................expanded uncertainty of certified value UCRM, rel .............expanded relative uncertainty of certified value ults .................... relative standard uncertainty of long-term stability umeas .................standard uncertainty of measurement result usts.................... relative standard uncertainty of short-term stability uΔ .....................combined standard uncertainty of certified value and measured value UΔ.....................expanded uncertainty of certified value and measured value tsl ......................pre-defined shelf life xi ...................... result at time point i in an isochronous stability study x........................average result of all time points in an isochronous stability study y........................average of all results of a homogeneity study

5

Δ ......................difference between two measurement results Δm ....................difference between measured and certified value νMSwithin .............degrees of freedom (ANOVA)

6

1 Abstract

2 Introduction

2.1 Background

This report describes the development of a reference material to replace the CRM BCR-186 (pig kidney). The target parameters for certification were the mass fractions of As, Cd, Cu, Cr, Fe, Hg, Mn, Pb, Se and Zn. The desired mass fractions were those at natural levels that are similar to BCR-186 and below regulatory limits for food contaminants (EC466/2001 (Cd < 0.05 µg/g, Pb < 0.1 µg/g "wet weight") with target uncertainties of 5 -10 % and values traceable to the SI, expressed as dry mass fraction. The purpose of the reference material is the quality control of measurement methods for trace element mass fractions in meat. Throughout this report, results are expressed as a mass fraction on a dry mass basis. For practical purposes, the dry mass is established by determining the "loss of mass on drying" under conditions defined in Section 10.2. It should be noted that determination of the dry mass correction factor under conditions other than specified in this report might lead to results that differ from the certified values.

7

3 Participants

Project management and evaluation European Commission (EC), Joint Research Centre (JRC) Institute for Reference Materials and Measurements (IRMM), Reference Materials Unit, Geel (BE) (Work performed under ISO Guide 34 accreditation; BELAC No. 268-RM)

Processing European Commission (EC), Joint Research Centre (JRC) Institute for Reference Materials and Measurements (IRMM), Reference Materials Unit, Geel (BE) (Work performed under ISO Guide 34 accreditation; BELAC No. 268-RM)

Homogeneity measurements ALS Scandinavia AB, Luleå (SE) (Measurements performed under ISO/IEC 17025 accreditation; SWEDAC 1087)

Short-term stability measurements Minton, Treharne & Davies Ltd., Llanelli (UK) (Measurements performed under ISO/IEC 17025 accreditation; UKAS 1946) ALS Scandinavia AB, Luleå (SE) (Measurements performed under ISO/IEC 17025 accreditation; SWEDAC 1087)

Long-term stability measurements ALS Scandinavia AB, Luleå (SE) (Measurements performed under ISO/IEC 17025 accreditation; SWEDAC 1087)

Characterisation measurements ALS Scandinavia AB, Luleå (SE) (Measurements performed under ISO/IEC 17025 accreditation; SWEDAC 1087) Ceinal, S.A. (Silliker), Área Análisis Físico-Químicos, Barcelona (ES) (Measurements performed under ISO/IEC 17025 accreditation; ENAC 257/LE413) Solvias AG- Elemental and Microanalytical Services, Basel (CH) (Measurements performed under ISO 9001 Certification; SQS 11237-04) ALS Czech Republic s.r.o., Praha (CZ) (Measurements performed under ISO/IEC 17025 accreditation; Czech Accreditation Institute 259/2006) Helmholtz Zentrum München - Deutsches Forschungszentrum für Gesundheit und Umwelt GmbH, München (DE) (Measurements performed under ISO/IEC 17025 accreditation; DACH DAC-PL-0141-01-10) The Food and Environment Research Agency, York (UK) (Measurements performed under ISO/IEC 17025 accreditation; UKAS 1642) LGC Ltd., Teddington (UK) (Measurements performed under ISO/IEC 17025 accreditation; UKAS 0003) Laboratoire National de Métrologie et d'Essais, Paris (FR)

8

(Measurements performed under ISO/IEC 17025 accreditation; COFRAC No 2-54) SCK-CEN, Mol (BE) (Measurements performed under ISO/IEC 17025 accreditation; BELAC 015-TEST)

9

4 Processing of the material

4.1 Material selection

Trace element mass fractions were measured in samples of pig kidney available on the Belgian market. The levels of elements in the kidneys regarded as contaminants (As, Cd, Hg and Pb) are lower than those of BCR-186, which may reflect environmental pollution reduction measures implemented since the production of BCR-186 in 1988. The levels of Cd and Pb were below the legal limits, (0.5 and 1.0 mg/kg wet weight, respectively, in Commission Regulation (EC) no 466/2001 and the subsequent amendment 1881/2006 "setting maximum levels for certain contaminants in foodstuffs"), and those of other elements are at natural levels. The source material for ERM BB186 was 90 kg of fresh pig kidneys from Belgium. The material was supplied in two batches of about 45 kg that were freeze-dried on two occasions. All raw material was certified as "Fit for human consumption". It consisted of fresh kidneys that were halved with major blood vessels removed, and were sealed in plastic bags of about 10 kg for transport to IRMM. The raw material was stored at 4 °C until processing, which took place less than 24 h after arrival at IRMM on both occasions. On arrival at IRMM, the material was weighed and sub-samples were taken and frozen at -20 ºC for later contamination checks on processing.

4.2 Pig kidney processing

The raw material was cut into cubes of 1 to 2 cm length using ceramic knives and plastic cutting boards. Care was taken to remove any remaining blood vessels and remnants of hepatic pelvis. The cutting took place in an area that was curtained off and supplied from above with HEPA-filtered air. Cubes were placed on Teflon-coated stainless steel trays, weighed, and placed in a freeze-dryer (Martin Christ Gefriertrocknungsanlagen GmbH, Osterode am Harz, DE). The material was dried under vacuum until a water content of about 3 % (m/m) was reached. The dried cubes of pig kidney were then placed in stainless-steel drums that were immersed in liquid nitrogen overnight. Drums were removed from the liquid nitrogen one-by-one and the contents were transferred to a cryogenic Palla vibrating mill with titanium inner surface and milling rods (KDH Humboldt Wedag, Köln, DE) that had been pre-cooled with liquid nitrogen, and was maintained below -100 ºC. The resulting powder was sieved three times through a 250 µm Nylon sieve (Russel Finex, London, UK). The powder that passed the sieve was retained and stored in acid washed 200 L HDPE drums, and the remaining fraction was discarded. About 18 kg of powder was produced, which was then homogenised by 3-dimensional mixing in a Dyna-MIX CM200 mixer (WAB, Basel, CH). The material was tested for suitability of water content and particle size distribution: Sub-samples indicated water contents of less than half the target maximum content of 7 % (m/m), while no particles were larger in size than 435 nm and the particle-size distribution approximated log-normal. The material was therefore considered suitable to be bottled.

10

A batch of 100 ml brown glass bottles were acid washed with 6 % HNO3, rinsed with Milli-Q purified water and dried in a drying cabinet. The powder was transferred to the bottles using an automatic filling machine (All Fill, Sandy, UK), fitted with PTFE coated hopper and auger. Lyo-inserts were automatically pressed into the bottles after filling, and the hopper and bottles were continually flushed with nitrogen gas. The bottles were then capped with aluminium lids and labelled. In total, 1846 units were produced, each containing about 10 g of material. Bottles for normal stock were stored at 18 °C.

4.3 Additional characterisation measurements

Five bottles were selected by random stratified sampling for measurement of water content by Karl Fischer titration (KFT) [2], particle size analysis, and for micrographs. All analyses were made in duplicate on each vial. KFT results showed a mean water content of 2.95 % (m/m) in the samples, with a standard deviation of 0.13 g/100 g (and estimated measurement uncertainty of 0.5 % (m/m), k=2). The material was therefore considered to be sufficiently dry to preserve the matrix over the desired shelf-life, and the water content was the same between bottles, within measurement uncertainty, at the time of their preparation. Particles size analysis (PSA) was performed with laser diffraction spectrometry using a Helos laser light scattering instrument (Sympatec GmbH System-Partikel-Technik, Clausthal-Zellerfeld, DE) that measured in the range 0.5 to 875 µm. Within the 10 measurements, the x90 bands (size beneath which 90 % of particles are measured) were found between 171 and 206 nm with uncertainty (k=2) of 9.5 µm. All distributions approximated log-normal. These results indicate that the majority of particles are below the 250 µm target, and that the distribution appears to be homogenous between bottles. Micrographs were made with a microscope (Stemi 2000-C, Zeiss, DE) with > 10 µm resolution. The particles were irregularly shaped, with no agglomerations or fibres apparent in any of the 10 images. The concentrations of major matrix components were estimated so that information values could be provided to assist in the optimisation of sample preparation methods. A selection of major elements were measured, in the samples which were used for certification measurement by SCK-CEN, using NAA. The element mass fractions provided as additional information are given in table 2. Table 2: Major elements measured by NAA in ERM-BB186 (pig kidney)

Element mg/kg Urel (k = 2) Ca* 288 5.1 %Cl 7104 1.4 %K 12034 1.2 %

Mg 787 0.9 %Na 6021 1.6 %

Where Urel = Relative uncertainty for 6 replicate measurements, with 2 replicates made on each of 3 bottles. *For Ca, 1 outlying measurement was excluded

11

In addition, as part of the homogeneity study, Co was included in addition to the target elements for certification. A total of 72 measurements by ICP-MS gave a mean mass fraction of 0.106 mg/kg, with a standard deviation of 0.004 mg/kg.

12

5 Homogeneity studies

5.1 Design of the homogeneity study

For the homogeneity study, 12 bottles of ERM®-BB186 (pig kidney) were selected using a random stratified sample picking scheme. In each vial, 6 replicate measurements of the mass fractions of each element for certification were made on sub-samples of about 0.4 g, together with measurement of the water content. The measurement methods used were ICP-MS for As, Cd, Co, Cr, Cu, Mn, Ni, Pb and Zn, ICP-OES for Fe and Se and CVAFS for Hg. Samples were measured in a sequence with replicates in ascending and descending order to allow distinction between an analytical trend and a trend in the filling sequence. All measurements were performed on one day, in a single analytical sequence. Water content was measured in each vial by drying a sub-sample to constant mass at 103 ºC, and measured element mass fractions were corrected to the dry mass. Grubbs tests at 95 and 99 % confidence levels were performed to detect potentially outlying individual results as well as outlying bottle averages. Regression analyses were performed to detect possible trends regarding analytical or filling sequence. The uncertainty contribution from possible heterogeneity was estimated by a one-way analysis of variance (ANOVA) [3]. Method repeatability (swb) expressed as a relative standard deviation is given in equation 1:

y

MSs within

wb (1)

MSwithin = mean square within a bottle from an ANOVA

y = average of all results of a homogeneity study

Between-unit variability (sbb) expressed as a relative standard deviation is given by equation 2:

yn

MSMS

s

withinbetween

bb

(2)

MSbetween = mean square among bottles from an ANOVA n = average number of replicates per bottle Heterogeneity that can be hidden by method repeatability is defined in equation 3:

4* 2

MSwithin

wbbb

n

su

(3)

νMSwithin = degrees of freedom of MSwithin The larger value of sbb or u*

bb was used as uncertainty contribution for (potentially hidden) heterogeneity, ubb (see Table 2 for a summary of results, values were converted into relative uncertainties).

13

5.2 Results of the homogeneity study

The results evaluation is presented in Table 3, with graphs of sample means against bottle number, and replicate measurements against sequence, in Annex A. Outliers were detected by the single Grubbs tests for all elements except As and Se. For 6 elements (Cd, Co, Cu, Fe, Mn and Zn) the measurement with position 2 in the sequence, was found to be outlying by the single test (at a = 0.01), and were the lowest results of the respective measurements sequences. A significant slope was found for Fe in the filling sequences at the 95 % confidence level, and for Co, Hg and Se in the analytical sequences, at 99 %, 99 % and 95 %, respectively. The consequences of these observations on the uncertainty estimations for between-bottle homogeneity are discussed below, by element: For As, no slopes in the data and no outliers were found. The results were in a range close to the method's limit of quantification, 0.001 mg/kg, and the estimation of ubb was therefore limited by the repeatability of the analysis, which was reflected in the high value of u*bb, relative to the other elements, of 3.24 %. For Cd, no slopes in the data were found. Measurement 2 in the sequence was found to be an outlier; however there was no technical reason to exclude the result, and it was included in calculation of ubb. For Cr, measurement RSD of over 60 % was found. Considering that measurement uncertainties by this method were typically 25 % (k=2, for elements with concentrations within the working range), this result indicates a level of heterogeneity of Cr in the material that renders it unsuitable for certification. For Cu, no slopes in the data were found. Measurement 2 in the sequence was found to be an outlier, but was retained as there were no technical reasons for its exclusion. For Fe, a slope was found at the 95 % confidence level for the sample filling sequence, of 0.0037 mg/kg/vial. On calculation it can be seen that over the range of 1846 bottles, the slope equates to a potential change of 2.7 %. If a user is equally likely to receive any vial from the sequence, the uncertainty is calculated as the potential difference divided by 3, to represent a rectangular distribution, which is 1.56 % in this case. As this value is higher than the measured u*bb of 0.58 %, the uncertainty introduced by the slope is taken as the ubb to cover any potential trend. Measurement 2 in the sequence was found to be an outlier; however there was no technical reason to exclude the result, and it was included in calculation of ubb. For Hg, a slope was found at the 99 % confidence level for the measurement sequence. As replicates of individual bottles were randomised in the measurement sequence, this could only lead to a conservative over-estimation of ubb. An outlier was detected at = 0.01, however, there was no technical reason to exclude the result from the calculation. For Mn, no slopes in the data were found. Measurement 2 in the sequence was found to be an outlier; however there was no technical reason to exclude the result, and it was included in calculation of ubb. For Ni, a measurement RSD of 33 % was found, and calculations returned sbb of the same magnitude. Considering that the measurement uncertainties were about 25 % (k=2, dependent on the concentration), this result indicates a level of heterogeneity of Ni in the material that renders it unsuitable for certification.

14

For Pb, no slopes in the data were found. One outlier was found by the single Grubbs test at level = 0.05, however there was no technical reason for exclusion and it was retained for the uncertainty calculation. For Se, no outliers were identified, however, a significant slope was found at the 95 % confidence level for the measurement sequence. As replicates of individual bottles were randomised in the measurement sequence, this could only lead to a slight over-estimation of ubb. For Zn, no slopes in the data were found. Measurement 2 in the sequence was found to be an outlier; however there was no technical reason to exclude the result, and it was included in calculation of ubb. All individual data and vial averages showed normal or unimodal distributions, for all analytes. In conclusion, the distribution of all elements except Cr and Ni can be considered as homogeneous, within acceptable uncertainties, in this material.

15

Table 3: Evaluation of homogeneity study for elements in ERM®-BB186 (pig kidney)

As Cd Co Cr Cu Fe Hg Mn Ni Pb Se Zn

Mean [mg/kg] 0.00715 1.0613 0.1061 0.1330 34.95 253.5 0.0198 7.015 0.2035 0.03628 9.408 131.1

RSD [%] ind pts 17.9 3.29 3.77 61 2.91 3.29 5.48 3.27 32.5 7.11 3.03 3.09

MSwithin 2 x 10-6 1.3 x 10-3 17 x 10-6 0.0004 1.1405 71.51 1.3 x 10-6 0.0561 0.0003 7 x 10-6 0.0869 17.53

MSbetween 1 x 10-6 0.8 x 10-3 9 x 10-6 0.0404 0.4626 58.11 0.3 x 10-6 0.0342 0.0266 5 x 10-6 0.0505 9.98

sbb [%] MSB<MSW MSB<MSW MSB<MSW 61 MSB<MSW MSB<MSW MSB<MSW MSB<MSW 32.5 MSB<MSW MSB<MSW MSB<MSW

swb [%] 18.6 3.40 3.93 15.4 3.06 3.34 5.82 3.38 8.70 7.29 3.13 4.19

u*bb [%] 3.24 0.59 0.69 2.68 0.53 0.58 1.02 0.59 1.52 1.27 0.55 0.56

ubb [%] 3.24 0.59 0.69 61 0.53 1.56 1.02 0.59 32.5 1.27 0.55 0.56 Significant slope in measured value/sample vial number, at confidence level;

95 % No No No No No Yes No No No No No No

99 % No No No No No No No No No No No No

Significant slope in measured value/measurement sequence number, at confidence level; 95 % No No Yes No No No Yes No No No Yes No

99 % No No Yes No No No Yes No No No No No

Outliers by Grubbs tests, individual measurements Single; a=0.05 None 1 1 1 1 1 1 1 1 1 None 1

Double; a=0.05 None n/a n/a n/a n/a n/a n/a n/a n/a n/a None n/a

Single; a=0.01 None 1 1 None 1 1 1 1 None None None 1

Double; a=0.01 None n/a n/a None n/a n/a n/a n/a None 2 None n/a

16

5.3 Minimum sample intake

The desired minimum sample intake for this material is 0.2 g. However, because the mass fractions of As and Hg in the material were close to the LOD of the chosen measurement method (0.001 mg/kg, by ICP-MS), the sample intake mass used for the homogeneity study for all elements was raised to 0.4 g. To verify the homogeneity of the material with smaller sample intake masses, a study was made at IRMM using solid sampling atomic absorption spectrometry, SSAAS, to measure the contents of selected elements in sub-milligram amounts of the material, without sample preparation. Initial tests showed that of all the target elements for certification, the mass fractions of Cd, Cu and Mn were in a suitable range for reliable measurement. The principle of the estimation of minimum sample intake by SSAAS is as follows: Between 3 and 50 absorbance peak area measurements are made on sub-samples weighed on a microbalance. These are blank subtracted by making an absorbance peak area measurement with an empty platform (or externally calculated mean of more than one measurement). A calibration slope (blank-subtracted absorbance peak area over sample mass) is calculated by the method of least-squares. For each measurement, the calibration predicted absorbance, residual absorbance (from the calibration) and relative residual are calculated, and thereby, the standard deviation of the relative residuals. This value is then used to calculate a minimum sample intake, based on the desired, or certified, standard uncertainty of the element mass fraction of the material, through the following equation:

mu

sk'=MSI

2m2

(4)

Where k'2 = the factor for two-sided tolerance limits for the number of measurements, sm = the standard deviation of the relative residuals, u = the relative standard uncertainty of the element mass fraction in the material, m = the mean sub-sample mass. The least-squares calibration and coefficient of determination are calculated according to common statistical convention. The estimation of minimum sample intake, based on this data, is made according to the calculations presented in reference [4]. The minimum sample intakes were estimated for the uCRM given for the respective element mass fractions in Table 13. This returned values of 49 mg for Cd, 4 mg for Cu and 91 mg for Mn, which lie well below the required minimum sample intake set for the material of 0.2 g.

17

6 Stability studies

6.1 Short-term stability study

6.1.1 Design of the short-term stability study

A four weeks isochronous study [5] was performed to evaluate the stability of the material during transport. For the short-term stability study, 10 bottles of ERM®-BB186 (pig kidney) were chosen using a random stratified sample picking scheme. Element concentrations were measured using ICP-MS and ICP-OES methods. Samples were stored at 18 °C and 60 °C as well as at a reference temperature of -20 °C. Two bottles were stored at each temperature for 0, 1, 2 and 4 weeks. After the indicated storage periods, the samples were transferred to storage at -20 °C until analysis. Samples were analysed under repeatability conditions in an order specified by IRMM (randomised sample order). In each of the 10 bottles a determination of dry mass content was performed. All results per vial were related to the mean of the respective duplicate dry mass determination. Grubbs tests at the 99 % confidence level were performed to detect potentially outlying results. Data points were plotted against time and the regression lines were calculated to check for significant trends (degradation, enrichment) due to shipping conditions (see Table 3). The observed slopes were tested for significance using a t-test, with t,df being the critical t-value (two-tailed) for a confidence level α = 0.05 (95 % confidence interval) and for a confidence level α = 0.01 (99 % confidence interval). The slope was considered as statistically significant when |b|/seb t,df. Graphs can be found in Annex B.

6.1.2 Results of the short-term stability study

Unfortunately, the analytical performance for the first study was not sufficient to draw conclusions from the data for As, Hg and Pb. Therefore, another set of bottles were held at 60 °C for 0, 2 and 4 weeks, and an additional study was made for As, Cd and Hg at a later date. For this second study, only the 60 °C temperature was tested as the long-term stability study described below concluded that all certified elements were stable in the material on storage at 18 ºC for 12 months. In addition, Se was re-measured in the second study, as a trend was noticed for samples held at 60 ºC in the first study. The results by element are discussed below: As: This element was assessed by the second study, at 60 ºC only, and no outlying data points or significant slope with time were found. Cd: Measurements showed a low RSD compared to those for the other elements, at 5.0 %, and no outliers or slope with time at elevated temperature were found. The dataset appeared reliable and uSTS could be estimated. Cr: Results for Cr were not assessed because evidence of heterogeneity was found by the homogeneity tests described in the previous section. Cu: For the 18 ºC data set, one outlier was found at the 99 % confidence level by the single Grubbs test. No technical reason could be found to exclude the outlier, it was retained. A significant negative slope, at 95 % confidence, was found with time for the 18 ºC test temperature. As no significant slope was found for the 60 ºC test temperature, this is most

18

likely to represent and analytical artefact rather than genuine instability of the material. However, when included as potential degradation in the study at 18 ºC, the uSTS rose from 0.6 % to 1.0 %, which remains insignificant in comparison to uCRM. Fe: No significant slopes with time were found at either test temperature, and no outliers were found in the dataset. Hg: This element was assessed by the second study, at 60 ºC only, and no outlying data points or significant slope with time were found. Mn: For both datasets, no outliers were found and no significant slopes with time were found at either test temperature. Pb: This element was assessed by the second study, at 60 ºC only. One outlying data point was found, however, this was retained as there was no technical reason to exclude it. No significant slope with time was found. Se: This element was assessed by the second study, at 60 ºC only, and no outlying data points or significant slope with time were found. Zn: No outlying measurements or slopes with time at elevated temperature were found. In summary, reliable estimations of uSTS were obtained for all elements to be certified. In each case, no slope was found and uSTS for 1 week varied between 0.1 and 1.5 %, which lie well below the respective uCRM and are thus negligible. The material is therefore considered to be sufficiently stable at elevated temperature to be transported without cooling elements. Table 4: Evaluation of short-term stability study for elements in ERM®-BB186 (pig kidney)

Element Test

temperaturesingle

Grubbs1 double

Grubbs1 slope2

uSTS, rel

%3 18ºC - - - - As 60 ºC none none no 1.5 18ºC - - - 0.6 Cd 60 ºC none none no 0.6 18ºC 1 n/a 95 % 1.0 Cu 60 ºC 1 n/a no 0.7 18ºC none none no 0.3 Fe 60 ºC none none no 0.2 18ºC - - - Hg 60 ºC none none no 0.3 18ºC none none no 0.5 Mn 60 ºC none none no 0.5 18ºC - - - Se 60 ºC none none no 0.1 18ºC none none no 0.4 Zn 60 ºC none none no 0.3

1 Number of outlying measurements by Grubbs tests at the 99 % confidence level 2 Significant slope in mean measured value (bottle) over time at raised temperature at 95 or 99 % confidence level 3 relative, for 1-week

19

6.2 Long-term stability study

6.2.1 Design of the long-term stability study

Data from two isochronous stability studies were combined to assess the stability of the CRM. For the first isochronous study, 8 bottles were chosen using a random stratified sample picking scheme. Sets of two bottles were stored at 18 °C for 0, 4, 8 and 12 months and were maintained before and after the study at the reference temperature of -20 °C. Samples were analysed in a sequence with randomised sample order. The measurements were made methods using ICP-MS for As, Cd, Cr, Cu, Mn, Ni, Pb and Zn, ICP-OES for Fe and Se and CVAFS for Hg. The second isochronous study compared vials stored at -20°C as reference samples, with those from the normal stock stored at 18 °C. Two vials per storage time were selected using a random stratified sampling scheme, and element mass fractions were measured 39 months after the start of the first isochronous study. From each vial, 6 samples were measured by ICP-MS. For both studies, water content was measured once in each bottle so that dry-mass correction of element concentrations could be performed Grubbs tests at the 99 % confidence level were performed to detect outlying results. Data points were plotted against time and the regression lines were calculated to check for significant trends of degradation or enrichment) due to storage conditions (see Table 5 for a summary). The observed slopes b were tested for significance using a t-test, with t,df being the critical t-value (two-tailed) for a confidence level α = 0.01 (99 % confidence interval). The slope was considered as statistically significant when |b|/seb t,df. Finally, the uncertainty of stability ults [6] was calculated for a pre-defined shelf life of 12 months applying equation 5:

sl

i

stablts t

xx

RSDu

2

(5)

with RSDstab being the relative standard deviation of all individual results of the stability study, xi being the time point for each replicate, x being the average of all time points and tsl being the pre-defined shelf-life. Graphs can be found in Annex C. A normalisation was applied to take into account differences between the two studies. This design allows separation of a potential analytical drift from a trend over storage time. Correction of biases, even if they are statistically not significant, was found to combine the smallest uncertainty with the highest probability to cover the true value [6]. t-tests were performed to evaluate potential significant difference between the measurements performed. In the case of a significant (95 % confidence level) difference between studies, results were normalised using the following equation:

12 / xxd (6)

1x mean measurement result of study 1

2x mean measurement result of study 2 The relative uncertainty of this correction ud, rel was calculated as:

22

2

21

1,

11RSD

nRSD

nu reld (7)

20

RSD1: relative standard deviation of all results in study 1 RSD2: relative standard deviation of all results in study 2 n1: number of data points is study 1 n2: number of data points in study 2 Whenever the bias correction is applied, the uncertainty on the correction is included in the calculation of RSDstab according to equation 8:

( )121

1

221

21

-+

-+

=

∑+

=

nn

yyun

RSD

nn

jjjrel,d

stab (8)

The corrected RSD is then used for the estimation of ults according to equation 5. The combined datasets were plotted against storage time and regression lines of concentration versus time were calculated. The slopes of the regression lines were tested for statistical significance.

6.2.2 Results of the long-term stability study

For As, Cd, Cu, Fe, Pb and Zn, significant differences in the means of all measurements were found between studies, and the results of the first studies were therefore normalised with the correction factors given in table 5. Table 5: Correction factors for the combination of the long-term stability tests Element d ud,rel

As 0.49 0.06

Cd 1.03 0.01

Cu 0.91 0.01 Fe 0.96 0.01 Hg 1 0 Mn 1 0

Pb 1.15 0.03

Se 1 0

Zn 0.90 0.01 The obtained data were evaluated individually for each study, and in combination. Results of the statistical tests on the data from the first, second and combined studies are given in table 5.

21

Table 5: Statistical results of the long-term stability tests

Number of individual outlying results Significance of the trend on a 99% confidence level

Element

first second combined first second combined As none none none yes no no Cd none none none no no no Cu none none none no no no Fe none none none no no no Hg 1 none 1 no no no Mn none none none no no no Pb none 1 1 no no no Se none none none no no no Zn none none none no no no

For As, analytical performance of the long-term stability studies was similar to that of the homogeneity study, in that the mass fraction in the samples was close to the method LOQ of 0.001 mg/kg. In the first study, a significant slope in concentration over time at elevated temperature was found with 99 % confidence. The slope indicates an increase in As content with time. This is most probably a statistical artefact, as any increase in element content could only be possible if the matrix degraded, and such a trend was not found for any other element. For the combined study, no significant trend was found. For Hg and Pb, single outlying measurements were detected by the single Grubbs test. In both cases there was no technical reason to exclude the results and they were retained for calculation. No significant slopes with time were found. For all other elements, no outliers or significant slopes with time were detected in either of the studies or the combined study. In summary, reliable estimates of uLTS were obtained for As, Cd, Cu, Fe, Hg, Mn, Pb, Se and Zn. The estimates are listed in table 6. No slopes were found in concentration over storage time at + 18 ºC. The material may therefore be stored at + 18 ºC. Table 6: Results of the 12 month stability study on element concentrations in ERM®-BB184 (bovine muscle)

Element uLTS, rel

%1

As 4.48 Cd 1.02 Cu 1.32 Fe 1.00 Hg 1.37 Mn 0.98 Pb 2.01 Se 1.04 Zn 0.87

1 relative, for 24-months

22

7 Characterisation

7.1 Design of the characterisation study

In providing measurements for the characterisation study, the laboratories applied validated methods. Measurements were either covered by the laboratory's scope of accreditation or measurement capabilities were proven through demonstration of previous experience and acceptable performance in the analysis of trace elements in comparable matrices. In commissioning the study, collaborators were asked if they were able to provide measurements for the same element(s) by more than one analytical technique. This allowed the provision of a greater number of independent results for elements, where possible. Each laboratory was provided with two bottles of ERM®-BB186 (pig kidney) and one vial of BCR-186 (pig kidney) that was to be measured with the samples, as a quality control. Because of the limited sample quantity per unit, laboratories that offered to measure one or more elements by more than one technique were supplied with extra bottles of the candidate reference material and quality control material, as required. With the exception of measurements by neutron activation analysis (NAA), measurements on each sample vial were spread over two days. NAA measurements were made under intermediate precision conditions. For each sample vial, three replicate measurements were made and in each vial of candidate CRM and QC material, the water content was determined to allow correction to dry mass for the element concentrations. A table of methods employed by the participants is given in Annex E.

7.2 Results and technical evaluation – Principles

After receipt of the data sets, the results were subjected to technical evaluation according to the following criteria. Certain data sets from laboratories were excluded when their measured values for the QC material did not agree with that certified, when compared according to equation 6:

222 yCRMCRMmean uucy +<- (6)

Where ymean = mean measured concentration in the QC, cCRM = certified concentration in the CRM, uCRM = standard uncertainty of the concentration in the CRM, uy = standard uncertainty reported for the QC measurements. As approaches to uncertainty estimation differ between laboratories, decisions to exclude datasets on the basis of failed QC criteria were taken considering the level of uncertainty reported by other participants for measurements of the same element by the same or similar techniques: The maximum reported measurement uncertainty, uy, of all labs reporting for the studied element was therefore applied in equation 6. Checks were made to see that results from the two different measurement days were consistent with the measurement uncertainty of the lab, and the level of between-bottle homogeneity established for the material. The reported values for each element were checked against the method LOQ reported by the laboratory and values below the LOQ were excluded from the calculation of certified values.

23

Finally, comments and reports of anomalies in the measurements reported by the labs were considered in deciding whether to include values in the certification. After passing the criteria above, the accepted sets of results were submitted to the following statistical tests:

Scheffe multiple t-test to check if the means of two labs are significantly different Dixon test to detect outlying laboratory means Nalimov t-test to detect outlying laboratory means Grubbs test to detect single and double outliers Cochran test to check for outlying laboratory variances Bartlett test to check for homogeneity of laboratory variances ANOVA to assess between laboratory and within laboratory variances and test their

significance employing the Snedecor F-test Skewness and kurtosis tests to assess the normality of the lab means distribution

7.3 Results and technical evaluation

The technical evaluation of results for inclusion in the accepted datasets is summarised in Table 6. The results of the accepted datasets, and statistical tests thereon are summarised in Table 7. Results from each participant and corresponding graphs can be found in Annex D. For all elements, variances between labs were significantly different (Snedecor F-test), therefore data could not be pooled and had to be grouped by labs. Unless a technical reason for exclusion was found, data sets that had outlying laboratory variances (Cochran test) were kept. One collaborator provided results using the two independent techniques of ICP-MS and ICP-OES, listed as labs 2 and 3. For all elements measured by either technique, results for QC measurements showed a positive bias of more than 50 %. The bias was reflected in that measurements on the candidate CRM were typically between 50 and 150 % higher than the mean of the other laboratories. The results of this collaborator using these techniques were therefore considered to be untrustworthy, and all were rejected. The same collaborator also provided results for Hg by an independent technique (solid-sampling pyrolysis AAS), and the performance of this was considered separately. Considering the remaining laboratories, the results of the statistical and other tests are discussed for each element, below: As: Lab 8 reported a mean of means below their LOQ of 0.1 mg/kg. The result was therefore rejected. For the remaining dataset, considering that only 4 independent results were obtained, and that the high standard deviation (relative to the other elements) resulted in a UCRM estimate of over 50 %, the value returned was not of sufficient quality for certification, and was given on the certificate as an indicative value only. Cd: The measurement of the QC sample of Lab 9 showed a bias towards low values of -13 %, which was not acceptable considering that the mass fraction of Cd in the material was at a level (about 1 mg/kg) easily measurable by the technique applied. The result was therefore rejected. On removing the result of lab 9, the result of Lab 6 was found to be an outlier by the Nalimov test at the 95 % confidence level, but not at the 99 % level. The result was therefore considered a straggler and was retained for calculation of the certified value. Cu: Lab 8 was found to have an outlying high lab variance by the Nalimov and test at the 95 % confidence interval, but not at the 99 % confidence interval. As such, it was considered to be a straggler and not significantly different from the remaining dataset. However, their QC measurements showed no bias, and as no technical reason could be found to exclude the result, it was retained.

24

Fe: After exclusion of results from Lab 3, discussed above, no suspicious data were found in the dataset, and all datasets were retained for calculation of the certified value. Hg: Lab 9 supplied a result for the QC with a bias of -24 %, which is greater than the combination of the maximum estimated measurement uncertainty and uncertainty of the certified value of 18 % (k=2). Their result was therefore excluded. Following this exclusion, the result of Lab 10 was found to be an outlier by the Nalimov and Grubbs tests at the 99 % confidence interval. As no technical problem was found with this result and it agreed with the remaining dataset within UCRM, it was retained for calculation. However, as only 5 independent results were obtained, and that the high standard deviation (relative to the other elements) resulted in a UCRM estimate of almost 50 %, the value returned was not of sufficient quality for certification, and was given on the certificate as an indicative value only. Mn: After exclusion of results from Labs 2 and 3, discussed above, no suspicious data were found in the dataset. Pb: After exclusion of the result of Lab 2, the result of Lab 8 was found to be an outlier by all tests at the 95 % confidence interval, and by the Nalimov test at the 99 % confidence interval. However, their measurement of the QC material showed no bias and no technical reason to exclude this lab was found. The result was therefore considered to be a straggler and not significantly different from the remaining dataset, and was retained. Se: Lab 9 supplied a result for the QC with a bias of -26 %, which is greater than the combination of the maximum estimated measurement uncertainty and uncertainty of the certified value of 19.7 % (k=2). Their result was therefore excluded. Following this exclusion, Lab 8 was found to have supplied an outlying mean by the Nalimov test (at the 95 % confidence interval). As their QC result was acceptable according to equation 6, the mean was not an outlier at 99 % confidence and as no other technical reason was found to exclude the result, it was retained. Zn: After exclusion of results from Labs 2 and 3, discussed above, no suspicious data were found in the dataset. Table 6: Summary of the technical evaluation Element As Cd Cu Fe Hg Mn Pb Se Zn

Lab 0 OK OK OK OK OK OK Lab 1 OK OK OK OK OK OK OK Lab 2 Bias Bias Bias Bias Bias Bias Bias Lab 3 Bias Bias Bias OK Bias Bias Bias Lab 4 OK OK OK OK OK OK OK OK Lab 5 OK OK OK OK Lab 6 OK OK OK OK OK OK OK OK Lab 7 OK OK OK OK Lab 8 <LOQ OK OK OK OK OK OK Lab 9 OK Bias OK OK Bias OK OK Bias OK Lab 10 OK OK OK OK

Summary of results that were accepted for the characterisation datasets, and those that were excluded due to being either below the reported method LOQ, showing evidence of bias on measurement of a QC material or with other indications of measurement unreliability reported by the Lab.

25

Table 7: Results of the characterisation exercise for the accepted datasets Element As Cd Cu Fe Hg Mn Pb Se Zn Number of accepted data sets 4 6 8 9 5 7 6 7 8 Number of replicate measurements 24 36 48 54 30 42 36 42 56 Mean of means [mg/kg] 0.0082 1.089 36.48 255.3 0.0234 7.263 0.0402 10.25 133.6 Relative standard deviation of mean of means [%]

40 2.93 5.4 4.4 26.7 3.23 13.1 10.9 3.5

Relative standard error of mean of means (uchar) [%]

19.9 1.20 1.92 1.47 12.0 1.22 5.3 4.1 1.25

All data sets compatible two by two? (Scheffe test)

no no no no no no no no no

Outlying means? (Dixon test; p = 0.05)

none none none none Lab 10 none Lab 8 none none

Outlying means? (Nalimov t-test; p = 0.05)

none Lab 6 Lab 8 none Lab 10 none Lab 8 Lab 8 none

Outlying means? (Grubbs test; p = 0.05)

none none none none Lab 10 none Lab 8 none none

Outlying lab variances? (Cochran test; p = 0.05, at consecutive passes)

no no no no Lab 10 no no no no

Lab variances homogeneous? (Bartlett test; p = 0.01)

yes yes yes yes yes yes yes yes error1

Variances between labs significantly different? (Snedecor F-test; p = 0.01)

yes yes yes yes yes yes yes yes yes

Distribution of means normal (p = 0.01)? (Skewness, kurtosis and normal probability plot)

insufficient data

insufficient data

yes yes insufficient data

yes insufficient data

yes yes

1It was not possible to run the Bartlett test, as for one lab the means of each vial were equal.

26

8 Certified values and uncertainties

The certified values for ERM®-BB186 (pig kidney) were calculated as the mean of means of the accepted data sets. The standard error of the mean of means was used as an estimation of the uncertainty contribution of the characterisation exercise. The standard error was calculated as the standard deviation divided by the square root of the number of accepted data sets. The combined standard uncertainty of the certified value includes contributions from the between-bottle heterogeneity, long-term storage and the characterisation study. The relative combined standard uncertainty is calculated according to equation 5:

222charltsbbCRM uuuu (5)

The combined expanded uncertainty, UCRM, is the expanded uncertainty with a coverage factor k = 2 corresponding to a level of confidence of about 95 % estimated in accordance with ISO/IEC Guide 98-3, Guide to the Expression of Uncertainty in Measurement (GUM:1995), ISO, 2008. Table 8 summarises the individual uncertainty contributions and the resulting expanded uncertainties as well as the certified values and their uncertainties after rounding for ERM®-BB186 (pig kidney). Table 9 provides the summary for As and Hg, to which indicative values were assigned.

27

Table 8: Certified values and uncertainties for ERM®-BB186 (pig kidney)

Element Cd Cu Fe Mn Pb Se Zn

ubb [%] 0.99 0.53 1.56 0.59 1.27 0.55 0.56

ults [%] 1) 1.02 1.32 1.00 0.98 2.01 1.04 0.87

uchar [%] 1.2 1.92 1.47 1.22 5.3 4.1 1.25

uCRM, rel [%] 1.86 2.39 2.37 1.67 5.9 4.3 1.62

n results 6 7 9 7 6 7 8

UCRM, rel (95 % ci.)2) [%] 3.7 4.8 4.7 3.34 11.7 8.6 3.2

Certified value [mg/kg]3) 1.09 36.5 255 7.26 0.040 10.3 134

UCRM (95 % ci.) [mg/kg] 3) 0.05 1.8 13 0.25 0.005 0.9 5 1) for a shelf life of 24 months

2) where the standard combined uncertainty is multiplied by a coverage factor k = 2 corresponding to a level of confidence of about 95 % estimated in accordance with ISO/IEC Guide 98-3, Guide to the Expression of Uncertainty in Measurement (GUM:1995), ISO, 2008

3) calculated from unrounded data

28

Table 9: Indicative values and uncertainties for ERM®-BB186 (pig kidney)

Element As Hg

ubb [%] 3.2 1.02

ults [%] 1) 4.5 1.37uchar [%] 19.9 12.0

uCRM, rel [%] 20.7 17.0n results 4 5

UCRM, rel (95 % ci.)2) [%] 66 47

Certified value [mg/kg] 0.008 0.023

UCRM (95 % ci.) [mg/kg] 0.006 0.0111) for a shelf life of 24 months

2) where the standard combined uncertainties are multiplied by coverage factors k = 3.18 for As and k = 2.78 for Hg corresponding to a level of confidence of about 95 % estimated in accordance with ISO/IEC Guide 98-3, Guide to the Expression of Uncertainty in Measurement (GUM:1995), ISO, 2008

29

9 Metrological traceability

The measurement results for assigning mass fractions of the elements were obtained by different digestion and extraction, sample concentration, matrix separation and quantification procedures. Measurements were calibrated with either external calibrants or by isotope dilution using reference materials of known purity and concentration. The certified mass fractions are traceable to the International System of Units (SI).

30

10 Instructions for use and intended use

10.1 Safety precautions

Usual laboratory safety precautions apply.

10.2 Use of materials

Allow the vial to reach ambient temperature before opening.

Shake the vial before taking aliquots.

Certified mass fractions are corrected for the water content of the material (dry-mass): To determine dry mass, accurately weigh an aliquot of approximately 1 g on an analytical balance and dry the sample in an oven at atmospheric pressure, at 103 °C ± 2 °C, until constant mass is attained. The weighing should be made at the same time as preparation of samples for element measurement.

10.3 Intended use

This material is intended to be used for method validation and performance control. To assess the method performance, the measured mass fractions are compared with the certified values following a procedure described by Linsinger [8]; described here in brief: Calculate the absolute difference between mean measured value and the certified value

(Δm).

Combine measurement uncertainty (umeas) with the uncertainty of the certified value (uCRM) according to equation 6:

22CRMmeas uuu (6)

Calculate the expanded uncertainty (UΔ) from the combined uncertainty (uΔ) using a coverage factor of two (k = 2), corresponding to a confidence interval of approximately 95 %.

If Δm ≤ UΔ then there is no significant difference between the measurement result and the certified value, at a confidence level of about 95 %.

10.4 Storage conditions

The material should be stored at a temperature of 18 °C ± 3 °C. However, the European Commission cannot be held responsible for changes that happen during storage of the material at the customer’s premises, especially after opening of the bottles.

31

11 Acknowledgements

The authors would like to thank Marta Dabrio, Ingrid Zegers and Heinz Schimmel for the IRMM internal review of this report and Steve Balsley, (International Atomic Energy Agency, IAEA, AT), Thomas Prohaska (University of Natural Resources and Life Sciences, AT) and Peter Vermaercke (Studiecentrum voor Kernenergie, SCK, BE) as members of the Certification Advisory Panel for reviewing the certification documents and for their constructive comments.

12 References

1. International Organisation for Standardization, ISO/IEC Guide 98-3: Uncertainty of measurement -- Part 3: Guide to the expression of uncertainty in measurement (GUM:1995) (2008), ISO, Geneva, CH.

2. Kestens, V., Conneely, P., Bernreuther, A. Vaporisation coulometric Karl Fischer titration: A perfect tool for water content determination of difficult matrix reference materials. Food Chem. 106 (2008) 1454.

3. Van der Veen, A.M.H., Linsinger, T.P., Pauwels, J. Uncertainty calculations in the certification of reference materials, 2. Homogeneity study. Accred. Qual. Assur. 6 (2001) 26.

4. Pauwels, J. and Vandecasteele, C., Fresenius J. Anal. Chem., 1993, 345, 121.

5. Lamberty, A., Schimmel, H., Pauwels, J. The study of the stability of reference materials by isochronous measurements, Fres. J. Anal. Chem. 360 (1998) 359.

6. T.P.J. Linsinger, A.M.H. van der Veen, B.M. Gawlik, J. Pauwels, A. Lamberty, Planning and combining of isochronous stabiliy studies of CRMs, Accred. Qual. Assur. 9 (2004) 464-472

7. Linsinger, T., Pauwels, J., Lamberty, A., Schimmel, H., van der Veen, A.M.H., Siekmann, L. Estimating the uncertainty of stability for matrix CRMs. Fres. J. Anal. Chem. 370 (2001) 183.

8. Linsinger, T.P.J. Comparison of measurement result with the certified value, ERM Application Note 1, July 2005, www.erm-crm.org.

32

Annex A ERM®-BB186 (Pig kidney) – Results of the homogeneity study

Graphs present mass fractions of bottle means relative to the grand mean, against bottle number, and individual measurement replicates, against sequence number. Vertical bars are a confidence interval of 95 % derived from swb of the homogeneity study.

Graph A1: As

50%60%70%80%90%

100%110%120%130%140%150%

0 500 1000 1500

bottle number

rela

tive

ma

ss fr

act

ion

0%20%40%60%80%

100%120%140%160%180%200%

0 20 40 60 80

sequence number

rela

tive

ma

ss fr

act

ion

33

Graph A2: Cd

80%

85%

90%

95%

100%

105%

110%

115%

120%

0 500 1000 1500

bottle number

rela

tive

ma

ss fr

act

ion

50%60%70%80%90%

100%110%120%130%140%150%

0 20 40 60 80

sequence number

rela

tive

ma

ss fr

act

ion

34

Graph A3: Co

80%

85%

90%

95%

100%

105%

110%

115%

120%

0 500 1000 1500

bottle number

rela

tive

ma

ss fr

act

ion

50%60%70%80%90%

100%110%120%130%140%150%

0 20 40 60 80

sequence number

rela

tive

ma

ss fr

act

ion

35

Graph A4: Cr

0%

50%

100%

150%

200%

250%

300%

0 500 1000 1500

bottle number

rela

tive

ma

ss fr

act

ion

0%

50%

100%

150%

200%

250%

300%

0 20 40 60 80

sequence number

rela

tive

ma

ss fr

act

ion

36

Graph A5: Cu

80%

85%

90%

95%

100%

105%

110%

115%

120%

0 500 1000 1500

bottle number

rela

tive

ma

ss fr

act

ion

50%60%70%80%90%

100%110%120%130%140%150%

0 20 40 60 80

sequence number

rela

tive

ma

ss fr

act

ion

37

Graph A6: Fe

80%

85%

90%

95%

100%

105%

110%

115%

120%

0 500 1000 1500

bottle number

rela

tive

ma

ss fr

act

ion

50%60%70%80%90%

100%110%120%130%140%150%

0 20 40 60 80

sequence number

rela

tive

ma

ss fr

act

ion

38

Graph A7: Hg

80%

85%

90%

95%

100%

105%

110%

115%

120%

0 500 1000 1500

bottle number

rela

tive

ma

ss fr

act

ion

50%60%70%80%90%

100%110%120%130%140%150%

0 20 40 60 80

sequence number

rela

tive

ma

ss fr

act

ion

39

Graph A8: Mn

80%

85%

90%

95%

100%

105%

110%

115%

120%

0 500 1000 1500

bottle number

rela

tive

ma

ss fr

act

ion

50%60%70%80%90%

100%110%120%130%140%150%

0 20 40 60 80

sequence number

rela

tive

ma

ss fr

act

ion

40

Graph A9: Ni

0%20%40%60%80%

100%120%140%160%180%200%

0 500 1000 1500

bottle number

rela

tive

ma

ss fr

act

ion

0%

50%

100%

150%

200%

250%

300%

0 20 40 60 80

sequence number

rela

tive

ma

ss fr

act

ion

41

Graph A10: Pb

80%

85%

90%

95%

100%

105%

110%

115%

120%

0 500 1000 1500

bottle number

rela

tive

ma

ss fr

act

ion

50%60%70%80%90%

100%110%120%130%140%150%

0 20 40 60 80

sequence number

rela

tive

ma

ss fr

act

ion

42

Graph A11: Se

80%

85%

90%

95%

100%

105%

110%

115%

120%

0 500 1000 1500

bottle number

rela

tive

ma

ss fr

act

ion

50%60%70%80%90%

100%110%120%130%140%150%

0 20 40 60 80

sequence number

rela

tive

ma

ss fr

act

ion

43

Graph A12: Zn

80%

85%

90%

95%

100%

105%

110%

115%

120%

0 500 1000 1500

bottle number

rela

tive

ma

ss fr

act

ion

50%60%70%80%90%

100%110%120%130%140%150%

0 20 40 60 80

sequence number

rela

tive

ma

ss fr

act

ion

44

Annex B ERM®-BB186 (Pig kidney) – Results of the short-term stability study

Figure B1: Short-term stability of As mass fraction at 60 °C with associated usts

80%

85%

90%

95%

100%

105%

110%

115%

120%

0 2 4

time / weeks

rela

tive

ma

ss fr

act

ion

Figure B2: Short-term stability of Cd mass fraction at 18 and 60 °C with associated usts

80%

85%

90%

95%

100%

105%

110%

115%

120%

0 2 4

time / weeks

rela

tive

ma

ss fr

act

ion

45

Figure B3: Short-term stability of Cu mass fraction at 18 and 60 °C with associated usts

80%

85%

90%

95%

100%

105%

110%

115%

120%

0 2 4

time / weeks

rela

tive

ma

ss fr

act

ion

Figure B4: Short-term stability of Fe mass fraction at 18 and 60 °C with associated usts

80%

85%

90%

95%

100%

105%

110%

115%

120%

0 2 4

time / weeks

rela

tive

ma

ss fr

act

ion

46

Figure B5: Short-term stability of Hg mass fraction at 60 °C with associated usts

80%

85%

90%

95%

100%

105%

110%

115%

120%

0 2 4

time / weeks

rela

tive

ma

ss fr

act

ion

Figure B6: Short-term stability of Mn mass fraction at 18 and 60 °C with associated usts

80%

85%

90%

95%

100%

105%

110%

115%

120%

0 2 4

time / weeks

rela

tive

ma

ss fr

act

ion

47

Figure B7: Short-term stability of Pb mass fraction at 60 °C with associated usts

80%

85%

90%

95%

100%

105%

110%

115%

120%

0 2 4

time / weeks

rela

tive

ma

ss fr

act

ion

Figure B8: Short-term stability of Se mass fraction at 60 °C with associated usts

80%

85%

90%

95%

100%

105%

110%

115%

120%

0 2 4

time / weeks

rela

tive

ma

ss fr

act

ion

48

Figure B9: Short-term stability of Zn mass fraction at 18 and 60 °C with associated usts

80%

85%

90%

95%

100%

105%

110%

115%

120%

0 2 4

time / weeks

rela

tive

ma

ss fr

act

ion

49

Annex C ERM®-BB186 (Pig kidney) – Results of the long-term stability study

Measured values are means at each time/temperature, relative to the mean of time-point zero. Vertical bars represent the 95 % confidence interval of the measurements, based on the variance of measurements for each time-point calculated by ANOVA.

Figure C1: Long-term stability of As at 18 °C with associated ults

50%

75%

100%

125%

150%

0 12 24 36

time / months

rela

tive

ma

ss fr

act

ion

50

Figure C2: Long-term stability of Cd at 18 °C with associated ults

90%

95%

100%

105%

110%

0 12 24 36

time / months

rela

tive

ma

ss fr

act

ion

Figure C3: Long-term stability of Cu at 18 °C with associated ults

90%

95%

100%

105%

110%

0 12 24 36

time / months

rela

tive

ma

ss fr

act

ion

51

Figure C4: Long-term stability of Fe at 18 °C with associated ults

90%

95%

100%

105%

110%

0 12 24 36

time / months

rela

tive

ma

ss fr

act

ion

Figure C5: Long-term stability of Hg at 18 °C with associated ults

90%

95%

100%

105%

110%

0 12 24 36

time / months

rela

tive

ma

ss fr

act

ion

52

Figure C6: Long-term stability of Mn at 18 °C with associated ults

90%

95%

100%

105%

110%

0 12 24 36

time / months

rela

tive

ma

ss fr

act

ion

Figure C7: Long-term stability of Pb at 18 °C with associated ults

90%

95%

100%

105%

110%

0 12 24 36

time / months

rela

tive

ma

ss fr

act

ion

53

Figure C9: Long-term stability of Se at 18 °C with associated ults

90%

95%

100%

105%

110%

0 12 24 36

time / months

rela

tive

ma

ss fr

act

ion

Figure C10: Long-term stability of Zn at 18 °C with associated ults

90%

95%

100%

105%

110%

0 12 24 36

time / months

rela

tive

ma

ss fr

act

ion

54

Annex D ERM®-BB186 (Pig kidney) – Characterisation data

Results of characterisation measurements (corrected to dry mass), as reported by the participating laboratories. Values struck-through were excluded from calculation of the certified values. Individual replicate values are given with the mean and measurement uncertainty.

As mass fraction for indicative value, mg/kg Lab

Code #1 #2 #3 #4 #5 #6 Mean U0 - - - - - - - -1 0.0081 0.0076 0.0079 0.0084 0.0095 0.0086 0.0084 0.00402 0.0375 0.0388 0.0464 0.0409 0.0417 0.035 0.0401 0.00603 - - - - - - - -4 0.0067 0.0058 0.0067 0.0107 0.004 0.0039 0.0063 0.00235 - - - - - - - -6 0.0137 0.0139 0.0129 0.0113 0.0121 0.0128 0.0128 0.00207 - - - - - - - -8 0.051 0.053 0.052 0.059 0.056 0.049 0.0533 0.00909 0.005779 0.00562 0.005253 0.004531 0.006984 0.004606 0.0055 0.0022

10 - - - - - - - -

0.000

0.002

0.004

0.006

0.008

0.010

0.012

0.014

0.016

9 4 1 6

Lab code

mg

/kg

55

Cd mass fraction, mg/kg

Lab Code #1 #2 #3 #4 #5 #6 Mean U

0 - - - - - - - -1 1.11 1.1 1.1 1.1 1.1 1.12 1.1050 0.12002 1.73 1.62 1.61 1.63 1.61 1.6 1.6333 0.24503 1.59 1.62 1.6 1.66 1.64 1.63 1.6233 0.16234 1.0639 1.0924 1.0854 1.086 1.1221 1.0706 1.0867 0.22445 1.060755 1.084405 1.09345 1.061791 1.065422 1.06375 1.0716 0.02586 1.15 1.14 1.14 1.16 1.12 1.16 1.1450 0.02007 - - - - - - - -8 1.03 1.06 1.08 1.01 1.11 1.08 1.0617 0.16009 0.892109 0.898274 0.900843 0.889528 0.910164 0.880185 0.8952 0.0448

10 1.04 1.06 1.05 1.1 1.07 1.06 1.0633 0.0200

0.8

0.9

1.0

1.1

1.2

1.3

1.4

8 10 5 4 1 6

Lab code

mg

/kg

56

Cr mass fraction not for certification, mg/kg

Lab Code #1 #2 #3 #4 #5 #6 Mean U

0 - - - - - - - -1 0.148 0.141 0.146 0.094 0.096 0.092 0.1195 0.03002 0.586 0.546 0.522 0.479 0.463 0.488 0.5140 0.03603 0.533 0.528 0.496 0.676 0.634 0.631 0.5830 0.05834 0.3939 0.4155 0.4273 0.1044 0.1118 0.1484 0.2669 0.0769

5, #0605 0.356 0.378 0.354 0.354 0.363 0.401 0.3677 0.01405, #1615 0.534 0.521 0.532 0.533 0.54 0.539 0.5332 0.0070

6 0.269 0.267 0.272 0.118 0.113 0.113 0.1920 0.01407 0.449 0.444 0.436 0.308 0.301 0.308 0.3743 0.01308 0.643 0.622 0.638 0.246 0.23 0.248 0.4378 0.08409 0.142931 0.128031 0.143547 0.159138 0.160164 0.191376 0.1542 0.009310 - - - - - - - -

0

0.1

0.2

0.3

0.4

0.5

0.6

1 9 6 4 7 8 5

Lab code

mg

/kg

57

Cu mass fraction, mg/kg

Lab Code #1 #2 #3 #4 #5 #6 Mean U

0 33.8 37.1 38.9 34.8 36.1 34.8 35.92 4.001 37.2 36.6 36.9 36.4 37.1 37.5 36.95 2.202 55.9 53 53.9 54.9 53.9 55.3 54.48 6.543 52.2 52.2 53.3 55 53.5 53.7 53.32 5.334 32.436 33.836 33.564 33.703 33.866 32.907 33.39 5.085 36.21355 36.70457 36.26847 36.14008 36.0343 35.80586 36.19 1.076 36.5 36.9 35.5 36.1 35.7 35.5 36.03 1.607 35.7 35.2 36.1 35.5 34.7 34.8 35.33 1.108 37.06 42.05 37.65 41.27 44.16 39.02 40.20 4.869 38.8358 37.8545 38.11139 37.10472 38.55236 36.71458 37.86 1.89

10 - - - - - - - -

25

30

35

40

45

50

4 7 0 6 5 1 9 8

Lab code

mg

/kg

58

Fe mass fraction, mg/kg Lab

Code #1 #2 #3 #4 #5 #6 Mean U0 260 259 254 257 263 250 257.2 15.01 252 263 268 264 267 269 263.8 20.02 - - - - - - - -3 371 378 373 389 383 385 379.8 22.84 243.2 252.65 253.49 248.94 252.83 240.48 248.6 22.85 261.6935 243.3133 255.2661 248.1179 250.3208 251.1069 251.6 2.76 245 248 238 235 243 240 241.5 9.57 241 243 241 242 240 242 241.5 2.18 279.4 277.2 292.8 276.2 260.7 257.3 273.9 31.09 267.7764 267.8792 277.538 258.5216 271.6632 261.191 267.4 13.4

10 249 242 252 251 260 259 252.2 7.0

200

220

240

260

280

300

320

6 5 4 5 10 0 1 9 8

Lab code

mg

/kg

59

Hg mass fraction for indicative value, mg/kg

Lab Code #1 #2 #3 #4 #5 #6 Mean U

0 - - - - - - - -1 0.0199 0.0195 0.0206 0.0193 0.0196 0.0191 0.0197 0.00502 - - - - - - - -3 0.0191 0.0204 0.0197 0.0197 0.019 0.0192 0.0195 0.00204 0.0204 0.021 0.0204 0.0218 0.0204 0.0202 0.0207 0.00375 - - - - - - - -6 0.0218 0.0223 0.0235 0.0218 0.0232 0.0238 0.0227 0.00217 - - - - - - - -8 - - - - - - - -9 0.01269 0.012197 0.012865 0.015411 0.016263 0.015062 0.0141 0.0025

10 0.037 0.037 0.033 0.034 0.033 0.032 0.0343 0.0030

0.000

0.005

0.010

0.015

0.020

0.025

0.030

0.035

0.040

3 1 4 6 10

Lab code

mg

/kg

60

Mn mass fraction, mg/kg

Lab Code #1 #2 #3 #4 #5 #6 Mean U

0 7 6.98 7.26 7.02 7.1 6.76 7.02 0.461 7.43 7.63 7.73 7.59 7.64 7.71 7.62 0.802 11.9 11.3 11.6 11.3 11.3 11.1 11.42 1.713 11.7 11.2 11.5 11.5 11.3 11.4 11.43 0.574 7.992 7.1655 7.2874 7.5614 7.7572 7.3485 7.52 1.125 - - - - - - - -6 6.93 7.24 6.99 7.09 7.25 7.11 7.10 0.267 7.33 7.25 7.31 7.26 7.31 7.25 7.29 0.108 6.61 6.54 7.63 7.31 7.24 6.91 7.04 0.959 7.471743 7.311961 7.404439 6.966119 7.330595 7.037988 7.25 0.36

10 - - - - - - - -

5.0

5.5

6.0

6.5

7.0

7.5

8.0

8.5

9.0

0 8 6 9 7 4 1

Lab code

mg

/kg

61

Pb mass fraction, mg/kg

Lab Code #1 #2 #3 #4 #5 #6 Mean U

0 - - - - - - - -1 0.0377 0.0364 0.0373 0.0366 0.0375 0.0374 0.0372 0.01302 0.0645 0.0674 0.0631 0.0618 0.065 0.0639 0.0643 0.00643 - - - - - - - -4 0.03 0.0367 0.0393 0.0317 0.0342 0.0485 0.0367 0.00635 - - - - - - - -6 0.0415 0.0403 0.0411 0.0408 0.0399 0.0423 0.0410 0.00177 - - - - - - - -8 0.052 0.042 0.047 0.053 0.054 0.053 0.0502 0.00709 0.035995 0.035673 0.034991 0.036068 0.03809 0.035626 0.0361 0.0029

10 0.038 0.041 0.046 0.036 0.041 0.038 0.0400 0.0050

0.01

0.02

0.03

0.04

0.05

0.06

9 4 1 10 6 8

Lab code

mg

/kg

62

Se mass fraction, mg/kg

Lab Code #1 #2 #3 #4 #5 #6 Mean U

0 10.5 10.2 10.5 10.7 10.4 9.9 10.37 0.751 10.4 10.3 10.1 10.2 10.6 10.2 10.30 1.002 24.7 23.6 24.4 24 24.4 23.3 24.07 3.853 24 24 22.1 24.5 24.3 24.7 23.93 2.394 11.0379 10.8333 11.0703 10.1457 11.0462 10.517 10.78 2.105 9.806451 9.887628 9.978184 9.837347 9.870215 10.07314 9.91 0.286 12.2 12.1 11.6 12.2 12 12.2 12.05 0.477 - - - - - - - -8 8.24 7.99 8.12 8.42 8.55 8.43 8.29 0.949 6.847513 6.993424 6.975956 7.093429 7.335729 6.99076 7.04 0.35

10 10.29 10.27 9.99 9.93 9.84 10.06 10.06 0.43

6

7

8

9

10

11

12

13

14

8 5 10 1 0 4 6

Lab code

mg

/kg

63

Zn mass fraction, mg/kg

Lab Code #1 #2 #3 #4 #5 #6 Mean U

0 136 132 135 132 135 128 133.0 8.51 136 136 134 134 136 136 135.3 9.02 217 216 220 214 220 223 218.3 21.83 209 208 208 215 213 215 211.3 21.14 135.4 139.6 138.2 141.1 142.7 136.4 138.9 22.85 129.1654 130.04 129.2131 130.4499 129.5827 130.6595 129.9 1.76 142 139 141 136 140 136 139.0 5.17 135 139 135 140 136 135 136.7 4.58 130.6 127 133.7 130.4 130 131.1 130.5 14.09 126.3872 125.3288 126.9729 123.152 128.2136 124.3532 125.7 6.3

10 - - - - - - - -

100

110

120

130

140

150

160

170

9 5 8 0 1 7 4 6

Lab code

mg

/kg

64

Annex E Methods used in the characterisation exercise

List of measurement methods used by participants in the characterisation exercise, as reported by participants.

Lab code Element(s) Sample preparation method

Measurement technique

Calibration method Quality Assurance Uncertainty estimation

0 Cu, Fe, Mn, Se, Zn

None k0-NAA k0 method - traceability in k0-NAA is guaranteed through the use of IRMM-530R Al-Au fluence rate monitors

BCR-186, BCR-278 Accoding to Robouch et al., Uncertainty budget for k0-NAA. J. Radioanal. Chem. 245, 195 (2000).

1 All MW-assisted digestion (0.5 g sample + 5ml HNO3+1mlH2O2), 600W, 1 h

ICP-SFMS in high resolution mode

External Preparation (method) blanks, DORM 2 CRM

In accordance with T.Ruth 'A model for the evaluation of uncertainty in routine multi-element analysis', Accred Qual Assur (2004) 9:349-354

2 As, Cd, Cr, Cu, Mn, Pb, Se, Zn

Acid digestion (2 g of sample, 5 ml HNO3, 1.5 ml H2O2, final volume - 25.0 ml)

ICP-MS, isotope 75As mathematically corrected for 40Ar35Cl interference

2 point calibration (0 and 100 µg/L)

Control of calibration for point 100 µg/L prepared from another standard material in each measurement series, measuring duplicate samples for each series of 20 samples, control of blank values, standard addition (50 µg/L) into each 20th sample.

Repeated measurement of standard materials and calculation of RSD values.

3 Cd, Cr, Cu, Fe, Mn, Pb, Se, Zn

Acid digestion (2 g of sample, 5 ml HNO3, 1.5 ml H2O2, final volume - 25.0 ml)

ICP-OES, axial view 4 point calibration in range 0 - 20 mg/L

Control of calibration for point 2 mg/L prepared from another standard material in each measurement series, measuring duplicate samples for each series of 20 samples, control of blank values, standard addition into each 20th sample.

Repeated measurement of standard materials and calculation of RSD values.

Hg

Direct analysis from solid sample, weight 0.1 g

AMA - Atomic absorption spectrometry

11 point calibration using standard solutions - 0.01 and 0.05 mg Hg/L, range 0 - 25 ng Hg

Control of calibration for point 1 ng Hg prepared from another standard material in each measurement series, measuring duplicate samples for each series of 20 samples

Repeated measurement of standard materials and calculation of RSD values.

4 All High pressure microwave assisted digestion using nitric and hydrochloric acid in quartz

ICP-MS with collision cell technology (Agilent 7500ce)

External, acid matched, plus internal standard correction using rhodium

Reagent blanks, spiked samples, and their following CRM: NIST1547, DOLT2, BCR185R. Applied in each batch.

Based on our performance in inter-laboratory comparisons.

65

vessels

5 Cd, Cr, Cu, Fe, Se, Zn

Microwave acid digestion (HNO3 and H2O2)

HR-ICP-MS Element 2, except Cd; Agilent 7500ce ICP-MS.

IDMS, with procedures detailed in report of analysis

Independent Std used as quality check, NIST SRM1577c used as QC for recoveries/matrix effects. Analysis performed following LGCs SOP (UKAS accredited, ISO 17025) for the provision of characterisation values by ID-ICP-MS

The uncertainty of the moisture corrected value was calculated in accordance with ISO Guide to the Expression of Uncertainty in Measurement (GUM) guidelines. Combined uncertainties were calculated following the Kragten spreadsheet method taking into account instrumental performance, IDMS equation, blend to blend variation and uncertainty of moisture determination.

6 As, Cd, Cr, Cu, Hg, Mn, Se, Zn

Microwave assisted digestion Multiwave 3.000, Anton Paar with 8-XF-100 pressure vessels. Sample intake: exactly weighed. Acid addition: 4 ml HNO3, 1 ml H2O2. When finished dilute the samples with H2O exactly to a final volume of 30 ml. (HNO3, subboiled distilled, H2O2 Suprapure, Merck)

HR-ICP-MS: Instrument: Element 1 Thermo Fisher Scientific, Sample introduction using a peristaltic pump, Spetec, at a flow rate of 1.2 ml/min, Meinhard nebulizer and cyclone spray chamber.

External calibration with 6 standards and internal standard Rh: Measurement series starts with blanks, calibration standards, blanks, control standard 1, followed by sample digests and control standard 2 + blanks (every tenth sample), ending with control-calibration standards and blanks. Samples not matching the calibration range were appropriately diluted and re-measured. Internal standard Ir and Rh.

Measurement of certified standards together with samples prepared at sample concentrations determined. Preparation and measurement of BCR-186 CRM together with the samples. Preparation and measurement of blank digestions together with the samples.

Uncertainty was estimated as twice the reproducibility standard deviation of BCR-186. Reproducibility standard deviation was derived from this sample, which was digested on 10 different days and analysed on 10 different days.

7 Cr, Cu, Fe, Mn, Zn

As above ICP-AES: Instrument: Optima 7300, Perkin-Elmer - LAS, Sample introduction using a peristaltic pump, Spetec, at a flow rate of 0.6 ml/min, low-flow Micromist nebulizer and heated cyclone spray chamber (APEX) for Cr and Seaspray nebulizer and cyclone spray chamber for the remaining elements.

External calibration with at least 7 standards; procedure and controls as above

As above As above

8 As, Cd, Cr, Cu, Fe, Pb

0.5 grams of sample. Microwave digestion with 6 ml HNO3 (67%), 1 ml H2O2 (33%). Dilution to volume with H2O

AAS-Graphite furnace with Zeeman background correction

External calibration curve with 4 standards

Quality control sample in each batch; calibration curve linearity; blank control; repeatability of measurements

We consider 2 contributions: uncertainty related to intermediate precision and uncertainty related to trueness/bias. These contributions are established based on the validation datas. For intermediate precision, 10 replicates obtained in 2-4 differents days were used. For trueness, MRC, materials from PT schemes or spiked samples were used, depending on the availability.

66

Hg 0.5 grams of sample. Microwave digestion with 6 ml HNO3 (67%), 1 ml H2O2 (33%). Dilution to volume with HCl (37%) and H2O

Cold vapour with Deuterium background correction

As above As above As above

Se 0.5 grams of sample. Microwave digestion with 6 ml HNO3 (67%), 1 ml H2O2 (33%). Ashing with Mg(NO3)2. Dilution of the ashes in HCl (37%). Reduction of Se at 90ºC, 10 min. Dilution to volume with H2O

Hydride generation, with Deuterium background correction

As above As above As above

Zn 0.5 grams of sample. Microwave digestion with 6 ml HNO3 (67%), 1 ml H2O2 (33%). Dilution to volume with Li (2%) in HNO3

AAS-Flame (Air-Acethylene)

As above As above As above

9 All Oxidative decomposition of the test material using a high pressure autoclave with microwave heating

ICP-MS external calibration (2 points) calibration check using two independent standards (2-3 dilutions per standard), analysis of 3 standard reference materials

sample preparation: weighing (0.05 %), filling of digestion solution to marker (1 %), dilution of digestion solutions (0.3 %), measurement itself (2.5 %) ---> in total 2.7 % rounding error in case of low analyte content [%]: (0.001 mg/kg / analyte content)*100

10 Cd, Fe, Se Around 0,2 g of sample with enriched spike of the element was digested with 7 ml HNO3 + 1 ml H2O2 in a microwave oven. Residues were diluted with high purity water.

ICP/MS (quadrupole) with no collision gas for Cd, in CCT mode (He - 6 ml/min + H2 – 1.2 ml/min) for Fe and in CCT mode (He - 2 ml/min + H2 - 3 ml/min) for Se

Double-IDMS with enriched spike of 111Cd (Ratio measured 112Cd/111Cd), and enriched spike of 57Fe (Ratio measured 56Fe/57Fe), and with enriched spike of 78Se (Ratio measured 80Se/78Se).

Validation of IDMS analysis during CCQM comparisons, for example CCQM K 49 bovine liver sample (nearest matrix for pig kidney). IDMS analysis accredited by COFRAC 2-54

An uncertainty budget had been established in compliance with the GUM comprising uncertainties for weighing, ratio measurements, fidelity, isotope abundances and corrections for blank . Our software calculate uncertainties propagation and the expanded final uncertainty (k=2)

Hg As above with the addition of BrCl for stabilisation (according to the EPA standard, 1631

ICP/MS sector field in low resolution

Double IDMS with enriched spike of 202Hg. Ratio measured 202Hg/200Hg

67

E)

Pb As for Cd, Fe except with 0,4 g sample intake and 10 ml HNO3 + 1 ml H2O2

ICP/MS (quadrupole) with no collision gas

Double IDMS with enriched spike of 206Pb. Ratio measured 208Pb/206Pb

European Commission EUR 25480 EN – Joint Research Centre – Institute for Reference Materials and Measurements Title: The Certification of the Mass Fractions of Elements in Pig Kidney Certified Reference Material ERM

®-BB186

Author(s): James Snell, Katharina Teipel, Heinz Schimmel European Commission, Joint Research Centre Institute for Reference Materials and Measurements (IRMM) Geel, Belgium

Luxembourg: Publications Office of the European Union 2012 – 65 pp. – 21.0 x 29.7 cm EUR – Scientific and Technical Research series – ISSN 1831-9424 ISBN 978-92-79-26169-5 doi:10.2787/64882 Abstract

This report describes the preparation of a pig kidney matrix reference material ERM

®-BB186, and the certification of

the contents (mass fractions) of seven elements. All results are expressed as a mass fraction on a dry mass basis. The preparation and processing of the material, homogeneity studies, stability studies and characterisation are described hereafter and the results are discussed. Uncertainties were calculated in compliance with the Guide to the Expression of Uncertainty in Measurement (GUM) [1] and include uncertainties due to possible heterogeneity, instability and from characterisation. The certified values and their uncertainties are listed in Table 1: Table 1: Certified mass fractions of elements and their uncertainties in pig kidney (ERM

®-BB186)

Element Certified value

1,2)

[mg/kg]

Uncertainty 3)

[mg/kg]

Cd 1.09 0.05

Cu 36.5 1.8

Fe 255 13

Mn 7.26 0.25

Pb 0.040 0.005

Se 10.3 0.9

Zn 134 5

1) Unweighted mean value of the means of accepted sets of data, each set being obtained in a different laboratory and/or with a different method of determination.

2) The certified value and its uncertainty are traceable to the International System of Units (SI). 2) Certified mass fractions are corrected for the water content of the material (and expressed as dry mass), determined as described in the section "Instructions for use".

3) The certified uncertainty is the expanded uncertainty with a coverage factor k = 2 corresponding to a level of confidence of about 95 % estimated in accordance with ISO/IEC Guide 98-3, Guide to the Expression of Uncertainty in Measurement (GUM:1995), ISO, 2008.

The assigned values are based on a minimum sample intake of 0.2 g.

How to obtain EU publications Our priced publications are available from EU Bookshop (http://bookshop.europa.eu), where you can place an order with the sales agent of your choice. The Publications Office has a worldwide network of sales agents. You can obtain their contact details by sending a fax to (352) 29 29-42758.

As the Commission's in-house science service, the Joint Research Centre's mission is to provide EU policies with independent, evidence-based scientific and technical support throughout the whole policy cycle. Working in close cooperation with policy Directorates-General, the JRC addresses key societal challenges while stimulating innovation through developing new methods, tools and standards, and sharing its know-how with the Member States, the scientific community and international partners. Key policy areas include: environment and climate change; energy and transport; agriculture and food security; health and consumer protection; information society and digital agenda; safety and security, including nuclear; all supported through a cross-cutting and multi-disciplinary approach.

L

A-N

A-254

80-EN

-N