A basic introduction to reference materials. POPs Strategy

Transcription

1 A basic introduction to reference materials POPs Strategy A tutorial 16 September 2009 Angelique Botha R&D metrologist

2 Contents Why do we need reference materials? comparability of results metrological traceability What are reference materials? Reference materials Certified reference materials Uses of reference materials How to select a reference material?

3 IRMM /IMEP proficiency test studies Traceability and Uncertainty Reality in measurements?

4 IMEP-16: Pb in Wine (Proficiency testing) Results submitted: December 2000 IMEP Report: September 2001 Published data: JAAS, (2001), 9: Metrologia Coordinator: IRMM, JRC European Commission Participants: 129 laboratories, 38 countries

5 Reliable measurement results Approach 1: Standardisation harmonised methods Approach 2: Metrology ultimate analysis Traceability Measurement uncertainty Prescribed methods/procedures Legislation prescribed methodperformance characteristics method-defined parameters SI traceable parameters required quality assurance?

6 ISO Scope 2. Normative references 3. Terms and definitions 4. Management requirements 5. Technical requirements 5.6 Measurement traceability General All equipment used for tests and/or calibrations shall be calibrated before being put into service. The laboratory shall have an established programme and procedure for the calibration of its equipment.

7 Metrological traceability VIM 2007 Property of a measurement result whereby the result can be related to a stated metrological reference through a documented unbroken chain of comparisons, all having stated measurement uncertainties. Why? Only measurement results where the metrological traceability is established can be compared, independent from when and where they were obtained. - IUPAC

8 Traceability key points Property of measurement result methods are not traceable Closely linked to measurement uncertainty Uncertainty + traceability = 2 main factors assuring measurement reliability Traceability to a common reference allows comparability of results Traceability chain = sequence of standards and comparisons that relate the result to a stated reference

9 Traceability final reference Important: the traceability chain shall be linked to a final reference: a measurement unit a measurement procedure a reference material ( artefact )

10 Traceability to a unit Preferably to the SI unit e.g. mass: 5 kg means: the measurement result of the property mass has been obtained by comparing the mass of the item with the mass of the kg in Paris Exists for most physical measurements independent of the method Not (yet) available for many chemical measurements: kg kg ; may depend on the mehod employed SI unit is not always relevant for the application (mass vs. catalytic concentration of an enzyme)

11 Traceability to a procedure A problem? No! These methods define the analytes dietary fibre has no real meaning a certified value without a statement of the method is meaningless the only useful reference is the procedure

12 Traceability to a reference material ( artefact ) Important for arbitrary units use the reference material directly as comparison can be method-dependent or independent e.g. WHO standard: by definition, this sample has 5000 IU

13 Traceability to institutes? The results are traceable to NIST Incorrect! what is meant is for example: The mass obtained is traceable to the kg in Paris, because it has been compared with a weight from NIST, which in turn has been compared to the kg in Paris. What matters is the end-point of the traceability chain

14 Metrological traceability in practice more complex? Example of the preparation of a Cd calibration standard for AAS: Weigh 10 g of cadmium metal, add nitric acid to dissolve the metal, transfer the dissolved metal to a volumetric flask, fill volumetric flask to the mark Analytical system: complete set of operations Weighing, heating, dissolving, quantitative transfer, fill, etc. Each operation has to be calibrated Mass for balances Not all operations are critical for a particular measurement result, i.e. required accuracy of the reference depends on the impact of the operation on the final analytical result analyst s (competent) judgement

15 Reference material (RM) - ISO Guide 30: 2008 Material Sufficiently homogeneous and stable Fit for its intended use in a measurement proces NOTES Generic term (family name) Quantitative and qualitative properties Uses Can only be used for a single purpose in a given measurement VIM3 definition

16 Certified reference material (CRM) - ISO Guide 30: 2008 Reference material Metrologically valid procedure Certificate Certified value has an uncertainty and statement of metrological traceability NOTES Quantitative and qualitative properties and their uncertainties Metrologically valid procedures can be found in ISO Guides 34 and 35 ISO Guide 32 gives guidance on the certificate VIM3 definition

17 Reference materials Other reference materials Homogeneous Appropriate stability Statements on homogeneity and stability performance controls method development material characteristics additional requirements accompanying information main applications Certified RMs Homogeneous Appropriate stability Metrologically valid establishment of property values property value(s) traceable to reference system measurement uncertainty stated homogeneity and stability intended use calibration trueness control full method validation all QA/QC issues

18 CRMs for short-cutting the traceability chain CRM = RM + certificate certified value with uncertainty CRMs to ensure traceability of stated traceability (to a final reference) 1. Measurement conditions, classical calibration 2. Results of final quantification (through calibration) 3. Results of overall method (through validation) 4. Results of daily work

19 Physical standards = classical calibration in the lab weights Volumetric flasks Thermometers Pipettes and burettes, etc.! Must be calibrated to ensure accuracy 1. Physical standards to ensure correctness of measurement conditions Correct weights, temperatures, volumes,. Basic requirements for all measurements Ensures that these operations are the same all the time, everywhere

20 2. CRMs for quantification sample Weiging, milling extraction? quantification Clean up Fill to volume

21 2. CRMs for traceability in final quantification Analysis result comparison Reference standard Standards with stated purity and uncertainty ISO section : shall, where possible, be traceable to SI units

22 2. Matrix CRMs for calibration? Usually not encouraged (uncertainty too large) For some methods necessary In the absence of pure substance materials

23 2. CRMs and traceability 1. Physical standards to ensure correctness of measurement conditions 2. Pure substance CRMs to ensure correctness and traceability of final quantification Substances must have stated identity, purity, concentration and uncertainty = must fulfil all requirements of CRMs Concentrations/purities must themselves be traceable: ISO Guide 31 demands traceability statement on the certificate Standards should be produced in conformance with ISO Guide 34. Third party assessment?

24 3. CRMs for validation sample Weighing, milling extraction?? quantification Clean up Fill to volume

25 3. CRMs for validation Analysis result Reference standard sample? comparison

26 Traceability in sample preparation Problem: every sample preparation step breaks the traceability chain One is not entirely sure that what went into the sample preparation step is the same that came out Method assumes correctness, absence of losses, etc. Any evidence? Therefore: ways to restore the traceability chain are needed: method validation Matrix CRMs can be used for method validation

27 Restoring the traceability chain sample Matrix CRMs weigh the sample X extraction/digestion clean up X dilution to a certain volume balance calibration impossible to link sample to final extract calibrated glassware quantification pure standards

28 3. CRMs and traceability 1. Physical standards to ensure correctness of measurement conditions 2. Pure standard CRMs to ensure correctness and traceability of final quantification 3. Matrix CRMs in method validation prove trueness, absence of losses and interferences and link the final quantification to the sample ISO Section 5.9: These demonstrations must be regularly repeated regular use of CRMs

29 4. CRMs in daily work Demonstration that conditions of validation study apply for this particular measurement: quality control charts In practice: CRMs often used Guaranteed homogeneity Better/documented stability compared to normal samples Ultimate test of accuracy and traceability of results in each measurement series

30 4. CRMS in daily work 1. Physical standards to ensure correctness of measurement conditions 2. Pure standard CRMs to ensure correctness and traceability of final quantification 3. Matrix CRMs in method validation prove correctness, absence of losses and interferences and link the final quantification to the sample 4. (C)RMs in control charts prove that the method has been applied correctly and that the conditions from the validation study apply

31 Conclusions ISO demands proof of traceability = comparability Traceability can be established by using several classes of RMs sample Ma atrix CRMs Weigh the sample Extraction/digestion Balance calibration Clean up Dilution to a certain volume Calibrated glassware quantification Pure standards

32 sample Conclusions ISO demands proof of traceability = comparability Traceability can be established by using several classes of CRMs weigh the sample physical standards Matrix CRM Ms QCMs extraction/digestion clean up dilution to a certain volume quantification pure standards

33 Reference materials Other reference materials Homogeneous Appropriate stability Statements on homogeneity and stability performance controls method development material characteristics additional requirements accompanying information main applications Certified RMs Homogeneous Appropriate stability Metrologically valid establishment of property values property value(s) traceable to reference system measurement uncertainty stated homogeneity and stability intended use calibration trueness control full method validation all QA/QC issues

34 Uses of reference materials Calibration Method development Method validation Evaluation of trueness Uncertainty estimation Proof of method performance Lab internal quality control ( charting ) Operator or equipment qualification Proficiency testing Training and verification of competence (external benchmarking)

35 Calibration RMs for calibration requires property values with stated uncertainty, and stated metrological traceability Many laboratories prepare their own RMs for calibration, often denoted as "calibrants" or "calibrators Different types of calibration: Single point Bracketing Multi-point

36 Single point calibration One calibrant (in this context the CRM) is used to calibrate the measurement equipment, which is then used to assign value(s) to the sample(s) measured x x sample CRM = y y sample CRM

37 Bracketing Two calibrants, one with a property value greater than the value(s) of the sample(s), and one with a property value smaller than those values By means of linear interpolation between the two calibrants, values are assigned to other samples x = + sample x x 2 1 y y x y y 2 1 ( ) sample 1 1

38 Multi-point calibration Ordinary least squares (OLS) EXCEL Generalised least squares (GLS) XGENLINE

39 Method validation Validation parameters

40 Concept of method validation Method validation is required to establish the fitness for purpose of a method for the specific requirements of customers when applied to a specific laboratory Method validation studies produce data on the Method validation studies produce data on the overall performance or individual influence quantities associated with the results of a method in normal use in the laboratory. This data can then be used for the estimation of the uncertainty associated with the results

41 Concept of method validation (cont.) Data on overall method performance parameters are obtained from method development (optimisation), interlaboratory studies, in-house validation protocols. Individual sources of error or uncertainty are typically only investigated when significant compared to the overall precision. Emphasis is on identifying and removing /reducing significant effects.

42 Validation parameters Precision Repeatability standard deviation s r Reproducibility standard deviation s R Intermediate precision s zi

43 Repeatability s r Short term precision Same sample Same analyst Same instrument Same method Short period of time

44 Reproducibility s R Long term precision Same sample Different analysts Different instruments Different methods Different times of the day Different days Different laboratories

45 Average Arithmetic mean n x = i= = 1 n x i

46 Standard deviation, s When repeated measurements give different results, we want to know how widely spread the readings are to be able to judge the quality of the measurements s n ( ) ( ) 2 x = x x i 1 n 1 i= 1 i

47 Validation parameters Bias ( Accuracy ) Studied through the use of reference materials or spiking studies. Analytical recovery (value observed divided by value expected). measuredvalue Re covery(%) = 100 expectedvalue Expected to be negligible or accounted for. Important uncertainty contributor.

48 t-test: Comparison of an experimental mean with a know value H : µ = 0 x and any difference is random H : x µ 1 and the difference is significant ( x µ ) n µ t calc = s t = t ;0,05 crit ν Level of confidence = 95% Reject if t calc > t crit

49 Comparison of the means of two samples Step 1: Compare the two standard deviations (F-test) H σ = : 1 σ 2 H1 : σ 2 2 : σ1 σ 2 F calc = s s the variances are not significantly different the variances are significantly different F = Fν ; ν 2;0,05 crit 1

50 Comparison of the means of two samples No significant difference ( ) ( ) = n n s n s n s pool n n s x x t pool calc + = = n n DF

51 Comparison of the means of two samples Significant difference n s n s x x s t calc + = 1 n 2 n + ( ) ( ) = n n s n n s n s n s DF

52 Validation parameters Linearity Not generally quantified; checked for by inspection or using significance tests for nonlinearity(correlation coefficient, F-test). Non-linearity corrected for by non-linear calibration or a more restricted operating range. Remaining deviations from linearity accounted for by overall precision estimates covering several concentrations, or within any uncertainties associated with calibration.

53 Validation parameters Detection limit Determined to establish the lower end of the practical operating range of a method. The detection limit is not of direct relevance to uncertainty estimation However, uncertainties near the detection limit may require careful consideration and special treatment

54 Limit of detection The concentration of analyte required to give a signal equal to the background (blank) plus three times the standard deviation of the blank y LOD = y + 3 s blank blank y x LOD LOD y = a + 3 s LOD x y

55 Limit of quantification For linear regression: y = a + 10 s LOQ y x

56 Validation parameters Robustness (ruggedness) Perform tests to determine the effect of one or more parameter changes on the results. If the effect is significant a more detailed study is carried out to measure the size of the effect. Significant effect: the precision deteriorated compared to the precision of the ruggedness test (ANOVA, F-test).

57 Validation parameters Selectivity/specificity The degree to which a method responds uniquely to the required analyte. Typical studies investigate the effects of likely interferents by adding the potential interferent to both blank and fortified samples and observing the response. Normally used to demonstrate insignificant effects. Can use the data to estimate the uncertainty associated with potential interferences.

58 Quality control: the RM An RM is needed with sufficient homogeneity and stability Sufficiently stable at least for the period of time that the precision checking measurements take If necessary, specific precautions should be taken to monitor the stability of the RM used Stability monitoring with the use of a CRM, or by using another measurement process of which the stability has been demonstrated Although the RM used for precision taking should not necessarily have known, metrological traceable property values for the properties of interest, precision measures may be dependent on the nominal value of the measurand, so knowledge about the nominal values of the parameters of interest is usually required to assess the appropriateness of the RM selected for the checking

59 Quality control: the measurement The user should perform independent replicate measurements. "Independent", in a practical sense, means that a replicate result is not influenced by previous measurements. To perform replicate measurements means to repeat the whole procedure. For example, in the chemical analyses of a solid material, the procedure should be repeated from the weighing of the test portion to the final reading or calculating of the result

60 Quality control Data treatment Check for outliers: Dixon s Q-test, Grubb s test Assessment of precision Calculate the average and standard deviation If you want an accurate nominal value Compare with a CRM

61 Quality control: control charts action limit warning limit 3s 2s nominal value average warning limit action limit 2s 3s

62 Selection of reference materials Checklist Intended use Measurand and measurement range Matrix match and potential interferences Measurement uncertainty and traceability Certification procedures Quality assurance issues Completeness and transparency of information Availability and cost

63 Intended use of the reference material Calibration or quality control? Pure material vs. matrix material Trueness or precision? Combined uncertainty vs. microhomogeneity Momentary or long-term control? Combined uncertainty vs. (long-term) stability

64 Measurand and measurement range Is the measurand for which the RM is certified identical to the measurand of the method to be checked/calibrated? Measurand definition unambiguous? Element vs. species specific? Method dependent of not? Measurement range compatible Concentration? Expected precision of your method?

65 Matrix match and potential interferences Matrix preferentially identical or similar? Cod fish vs. cod powder? Cod powder vs. fish powder? (fat content similar?) Fish muscles powder vs. pig meat powder? (matrix similar?) Presence of potential interferences?

66 Measurement uncertainty and traceability CRM uncertainty: Clear description? In accordance with the GUM? Proper consideration of homogeneity and stability? Uncertainty budget available? Does the information provided allow calculation of standard uncertainty? Does it satisfy your needs and requirements? (can this CRM e.g. be used for calibration?)

67 Measurement uncertainty and traceability Certified value Identity structurally defined? If yes: is the quantity value traceabe to the SI Identity operationally defined? E.g. according to a written standard, e.g. ISO If yes: is the quantity value Traceable to SI? Traceable to an artefact, e.g. expressed in IU?

68 Quality assurance issues Producer accredited as a reference material producer If not, quality system according to ISO Guide 34? Design and execution of certification process Selection of collaborators ISO accreditation Post certification activities (storage, monitoring, distribution, corrective actions) If not according to ISO Guide 34, does certification report provide sufficient evidence to guarantee equivalent quality, e.g. by third party evaluation (Certification committee, Certification advisory panel, ERM panel?)

69 Completeness and transparency of information Availability of Extended certificate? Full certification report? Sufficient details on Production process? Individual measurements? Data handling and treatment? Detailed instructions for use? Application notes/training

70 Availability and cost Where to find it? At what cost? www/catalogues Import/export limitations? Local legislative restrictions? Transport possible? Perishable CRMs to distant countries Local authorised distributor?