VAM BULLETIN. Traceability for in-vitro diagnostic medical devices

Transcription

1 VAM BULLETIN An LGC publication in support of the National Measurement System Issue Nº 28 Spring 2003 Traceability for in-vitro diagnostic medical devices Process Analytical Technology a new initiative within pharmaceutical development and manufacture Assessing the performance of indoor and in-car air monitoring devices Isotope-Dilution Mass Spectrometry A quantum method for chemistry? VAM formulation update FUNDED BY THE DTI

2 C O N T E N T S Keith Marshall Editor General enquiries about VAM to: VAM Helpdesk vam@lgc.co.uk LGC s address: LGC, Queens Road TEDDINGTON Middlesex TW11 0LY. ISSN The DTI VAM programme: The DTI s programme on Valid Analytical Measurement (VAM) is an integral part of the UK National Measurement System. The VAM programme aims to help analytical laboratories demonstrate the validity of their data and to facilitate mutual recognition of the results of analytical measurements. The VAM programme sets out the following six principles of good analytical practice, backed up by technical support and management guidance, to enable laboratories to deliver reliable results consistently and thereby improve performance. 1. Analytical measurements should be made to satisfy an agreed requirement. 2. Analytical measurements should be made using methods and equipment, which have been tested to ensure they are fit for their purpose. 3. Staff making analytical measurements should be both qualified and competent to undertake the task. 4. There should be a regular independent assessment of the technical performance of a laboratory. 5. Analytical measurements made in one location should be consistent with those elsewhere. 6. Organisations making analytical measurements should have well defined quality control and quality assurance procedures. The VAM Bulletin is produced by LGC under contract with the UK Department of Trade and Industry as part of the National Measurement System Valid Analytical Measurement Programme. No liability is accepted for the accuracy of information published and the views expressed are not necessarily those of the Editor, LGC or DTI. Photography by Andrew Brookes Contents Guest column Traceability for in vitro diagnostic medical devices...3 Contributed articles VAM Programme Formulation Update...7 Assessment of the performance of low-cost indoor and in-car air monitoring devices...10 Process Analytical Technology a new initiative within pharmaceutical development and manufacture: Implications for the development and validation of methodology...13 Isotope-Dilution Mass Spectrometry a quantum method for chemistry?...16 Case study Meeting the needs of regulation and trade: determining low levels of sulfur in fuel...18 VAM in education Decline in skills...23 VAM news European Trial of Airborne Particle Measuring Instruments comes to NPL...24 UKAS appoints Lord Lindsay as Chairman...25 RSC to produce perfect cup of tea...25 Challenging the Limits Of Detection...26 ROMIL launches fully SI-traceable CRMs for trace element calibration...27 MCERTS Performance standard for laboratories undertaking chemical testing of soil...27 UK Analytical Partnership Science & Hazard Regulation claim or blame?...28 Product and process competitiveness...30 Chemical nomenclature Introduction to polymers...30 Proficiency testing update Accreditation of PT schemes puts UK in First Division...32 COEPT to examine the comparability of PT schemes...32 EEE Working Group discusses measurement uncertainty in PT...33 Reference materials update A network for users of reference materials...34 New products VAM publishes guides for measurement in the laboratory...35 New method validation software under development...36 Forthcoming events...37 Contact points V A M B U L L E T I N

3 G U E S T C O L U M N The quest for metrological traceability in laboratory medicine René Dybkaer H:S Frederiksberg Hospital, Denmark Introduction Effective globalisation of information and knowledge necessitates mutual trust in the reliability and comparability of data over space and time. Resources are saved if primary measurement results are accepted without need of confirmation. The credibility is obtained by reference to an international infrastructure for metrology 1. Its top elements are the Metre Convention, the General Conference on Weights and Measures (CGPM) defining the International System of Units (S1), the executive International Committee for Weights and Measures (CIPM), and the International Bureau of Weights and Measures (BIPM) at Sèvres, France. The CIPM is advised in various fields of metrology by consultative committees (CCs) of which a ten-year old newcomer is the Consultative Committee for Amount of Substance: Metrology in Chemistry (CCQM). An important task of the CCs is to identify primary measurement methods and provide the material underpinning of the CIPM Mutual Recognition Arrangement (MRA) in the form of Key Comparisons leading to the Calibration and Measurement Capabilities of the National Metrology Institutes (NMIs). The next level is the associated Regional Metrology Organisations (RMOs) with their key comparisons, followed by the Reference Measurement Laboratories. Resources are saved if primary measurement results are accepted without need of confirmation. Metrological Traceability This global complex structure provides dissemination of the SI 2 and the upper stretches of metrological traceability: currently defined by the International vocabulary of basic and general terms in metrology 3 as being the property of the result of a measurement or the value of a measurement standard whereby it can be related to stated references, usually national or international measurement standards, through an unbroken chain of comparisons all having stated uncertainties. The salient concepts here are stated references and uncertainties. Laboratory Medicine There is no long-standing tradition for stating global metrological traceability in laboratory medicine, including clinical chemistry, clinical immunology, and haematology. Each medical laboratory has had its own measurement procedures and biological reference intervals. For several reasons, this is now changing: Clinical specialisation Patients are increasingly being transferred from one service to another within a hospital, between hospitals, and even between countries. There is no long-standing tradition for stating global metrological traceability in laboratory medicine. Laboratory specialisation A referral laboratory may serve several hospitals and private practices, but its results should mesh with those from the originating sites. Accreditation according to the standards ISO or EN ISO is now being favoured as a competitive advantage, and the standards require metrological traceability of measurement results; Production of reliable, biological reference intervals often requires collaboration between several laboratories with common stated references to allow pooling of data. In vitro diagnostic medical devices (IVD MDs) A major impetus of the recent flurry of activities towards creating and organising the metrological wherewithal of traceability is undoubtedly the EU Directive 98/79/EC on IVD MDs 6, which requires that The traceability of values assigned to calibrators and/or control materials must be assured through available reference measurement procedures and/or available reference materials of a higher order. The directive has to be implemented by manufacturers in December of this year. The very general statements in a modern EU directive are explained and specified in harmonised European Standards (ENs) from the European Committee for Standardization (CEN). Thus, the above requirement has occasioned five such standards produced by Technical Committee (TC) 140 In vitro diagnostic medical devices over the last fifteen years and adopted by the International Organization for Standardization (ISO) through its 3 V A M B U L L E T I N

4 G U E S T C O L U M N TC 212 Clinical laboratory testing and in vitro diagnostic test systems. Presentation of reference measurement procedures; EN 12286: /A1:2000 = ISO 15193:2002 Description of reference materials; EN 12287:1999 = ISO 15194:2002 Laboratory medicine - Requirements for reference measurement laboratories; EN ISO 15195:2003 Metrological traceability of values assigned to calibrators and control materials; EN ISO 17511:2003 Metrological traceability of values for catalytic concentration of enzymes assigned to calibrators and control materials; EN ISO 18153:2003 Chemical calibration hierarchy The first three standards mentioned may be seen as detailing important elements in the chemical calibration hierarchies specified in the last two documents. In principle, a calibration hierarchy requires the use of a sequence of documents, measuring systems, calibrators, and events, in the form of calibrations and value assignments. A comprehensive chemical calibration hierarchy may look as follows: Definition of SI unit by CGPM, e.g. mole per litre (Level 1.) Primary reference measurement procedure specified by BIPM or an NMI, (e.g. isotope dilution mass spectrometry) with a completely described measurement equation in terms of SI units, governing a (Level 2.) mass spectrometer operated by an analyst assigning a quantity value and measurement uncertainty to a (Level 3.) primary calibrator issued with certificate by BIPM or an NMI and specified in a (Level 4.) secondary reference measurement procedure (e.g. based on atomic absorption), described by an NMI or accredited reference measurement laboratory (ARML) and governing an (Level 5.) atomic absorption spectrometer assigning a quantity value and larger measurement uncertainty to a secondary calibrator which is acquired for a (Level 6.) manufacturer s selected measurement procedure (perhaps based on flame emission), governing the A calibration hierarchy requires the use of a sequence of documents, measuring systems, calibrators, and events, in the form of calibrations and value assignments. (Level 7.).manufacturer s flame emission spectrometer assigning value and uncertainty to the (Level 8.) manufacturer s working calibrator used in the (Level 9.) manufacturer s standing measurement procedure, perhaps with a measurement principle close to that of the end-user s measurement, e.g. titration, governing the (Level 10.) manufacturer s standing measuring system assigning value and uncertainty to the (Level 11.) manufacturer s product calibrator utilised according to an (Level 12.) end-user s routine measurement procedure for his (Level 13.) routine measuring system assigning to a (Level 14.) routine sample a routine result, i.e. quantity value and final (largest) measurement uncertainty. 4 V A M B U L L E T I N

5 G U E S T C O L U M N Variation in hierarchical structure This quite extensive structure of a calibration hierarchy may be modified according to available elements and purpose. It is always possible to excise one or two segments each comprising proceduremeasuring system-calibrator or calibratormeasuring system-procedure. This simplification should reduce the measurement uncertainty, but may not be economical, such as applying primary measurement procedure with measuring system directly on routine samples. The EN ISO also outlines the less fortunate situations where no SI unit is applicable. One possibility is for an international scientific organisation or the World Health Organization to specify an (Level 1.) International conventional reference measurement procedure (ICRMP) governing a (Level 2.) measuring system, assigning quantity value and measurement uncertainty to an (Level 3.) international conventional calibrator, which is offered to the manufacturers. Sometimes, the ICRMP with measuring system operates directly on the manu-facturer s working calibrator. Or, with no ICRMP having been agreed, an international protocol for value assignment with varying measuring systems furnishes the value and uncertainty of the international conventional calibrator. Finally, any laboratory may specify a measurement procedure without preceding calibration hierarchy, but the ensuing results from the measuring system are not directly comparable with those obtained by other means. Prerequisites of a calibration hierarchy Prior to choosing a calibration hierarchy, that which is to be measured, i.e. the measurand, must first be defined with regard to system, component, and kind-of-quantity. Such requirements are described in publications from the International Union of Pure and Applied Chemistry (IUPAC) and the International Federation of Clinical Chemistry and Laboratory Medicine (IFCC). 7,8 The measurand must be of the same type throughout the hierarchy. Then, the allowed maximum measurement uncertainty should be specified in view of the clinical requirements and finally obtained by the principles presented in Guide to the expression of uncertainty in measurement 9, based on a measurement function. If several measurement procedures are involved in a hierarchy, as is usually the case, they must have the same analytical specificity and analytical selectivity. A calibrator must have sufficient commutability, i.e. the same behaviour towards preceding and following measurement procedure as routine samples. The calibration hierarchies described hitherto are all single-stranded between stated reference and measurement result. 5 V A M B U L L E T I N

6 G U E S T C O L U M N In principle, however, each input quantity of a (final) measurement function must have its own more or less complicated subsidiary calibration hierarchy, together yielding a pluri-stranded hierarchy. For example, if a measurement result for a mass concentration is obtained by separate weighing of component and determination of volume of system, their hierarchies start from the definitions of the kilogram and meter respectively. It should be added that it is possible to define a measurand with the specification that a certain measurement procedure be used. In this way the stated reference may be an SI unit, but the results cannot be expected to equate those of an otherwise look-alike measurand with another specified measurement procedure. Metrological traceability revisited Having established a suitable calibration hierarchy a priori, connecting stated reference to measurement result, the latter can now be claimed a posteriori to be metrologically traceable to the former. In physics, the retrograde series of elements from measurement result to stated reference is classically just seen as consecutive comparisons between measurement standards, where the quantity value of a lower standard is compared to that of the next higher standard. The unbroken series of comparisons is designated traceability chain. In chemistry, the comparisons are effected after calibration, measurement procedure, measuring system, and assignment of quantity value and measurement uncertainty from the higher calibrator to the next lower one. The challenge The end-user s task of establishing a calibration hierarchy for a given type of quantity from, e.g., the definition of an SI unit to measurements results, usually requires involvement of BIPM, NMI, ARML, manufacturer, and end-user. None of these stations can provide the hierarchy by itself. Any lower station has to plug in at a higher station, even if some level may be bypassed. The connection is usually achieved by acquiring a calibrator, often a reference material, preferably a certified reference material, embodying the higher hierarchical levels for the measurand. As laboratory medicine is concerned with thousands of types of quantities, the creation of the upper levels of all the individual hierarchies is an enormous challenge and beyond the capacity of a single institution or even a single country. Still, the manufacturers of IVD MDs in view of the EU Directive s requirement of metrological traceability are clamouring for assistance. The Joint Committee on Traceability in Laboratory Medicine (JCTLM) During the last few years, representatives of the responsible bodies have discussed the problems and finally, at a meeting in BIPM last year, the stakeholders agreed to create an organising Joint Committee on Traceability in Laboratory Medicine. Its general mission is to support world-wide comparability, reliability and equivalence of measurement results in Laboratory Medicine, for the purpose of improving health care. Specific means include promoting traceability, networks of NMIs and RMLs, and reference measurement systems. The four principal promoters and stakeholders are the CIPM/BIPM, IFCC, International Laboratory Accreditation Cooperation (ILAC), and the World Health Organization (WHO). Other important stakeholders include professional scientific organisations, CRM producers, IVD MD industry, written standards developers, external quality assessment organisers, networks of RMLs, and regulatory bodies. The second meeting of JCTLM is planned for this summer. Actual work will be done in Working Group 1 on Reference materials and reference procedures, formulating criteria for inclusion of such items in recommended lists, and Working Group 2 on Reference laboratories setting criteria for accreditation of reference laboratories and promoting networks. Progress reports should appear towards this year s end. A salient function of JCTLM is to identify and prioritise the major measurement problems requiring development of primary and secondary measurement procedures and calibrators. The CCQM should be advised on the selection of measurands in laboratory medicine needing CIPM key comparisons to complement the few already completed ones on amount-of-substance concentration in serum of cholesterol, creatinine, and glucose. The choice should be influenced by the expectation that the metrological knowledge gained from a given exercise may help in creating calibration hierarchies for similar measurands. As the saying goes, the question to ask is How far does the light shine? Conclusion With the description of various calibration hierarchies and the creation of JCTLM, the stakeholders in the provision of reliable and globally comparable medical laboratory results have fashioned an important infrastructure for adequate and economical laboratory examinations in healthcare. REFERENCES 1. Wielgosz, R., VAM Bulletin (2001) 25, pp Quinn, T.J., Metrologia, 1994/95, 31, pp BIPM, IEC, IFCC, ISO, IUPAC, and OIML, International vocabulary of basic and general terms in metrology, 2nd ed., Geneva, ISO, ISO/IEC 17025:1999 General requirements for the competence of testing and calibration laboratories 5. ISO 15189:2003 Medical laboratories. Particular requirements for quality and competence. 6. EU Directive 98/79/EC on in vitro diagnostic medical devices, Off J Eur Comm, L 332/ Rigg, J.C., Brown, S.S., Dybkaer, R., and Olesen, H., Compendium of terminology and nomenclature of properties in clinical laboratory sciences, IUPAC/IFCC Recommendations 1995, Oxford, Blackwell Science Ltd, BIPM, IEC, IFCC, ISO, IUPAC, and OIML, Guide to the expression of uncertainty in measurement, Geneva, ISO, 1993 and V A M B U L L E T I N

7 C O N T R I B U T E D A R T I C L E S VAM Programme Formulation Update Martin Milton NPL John Marriott LGC In the last issue of the VAM Bulletin 1 we announced plans for the formulation of the next VAM Programmes, which are due to start at the beginning of October 2003, subject to approval within the DTI. Since then, significant progress has been made and, following extensive consultation, the initial drafts of the new Physical and Chemical VAM Programmes have been drawn up. These have been published for public comment and can be viewed on the DTI s website The Physical and Chemical VAM Programmes (which are being formulated by NPL and LGC respectively) have been developed against a background of significant technological change and continued globalisation of trade. Here we review the main elements of the two programmes. The Physical VAM Programme NPL has carried out a substantial consultation exercise to establish the requirements for the Physical VAM Programme. In addition to ongoing contacts made through the knowledge transfer activities of the VAM Programme, the following special events have been held: A one-day focus group attended by invited representatives of industry and trade bodies with expertise in different industrial sectors and international standards bodies. The discussion was based on responses to a questionnaire to establish requirements for gas and particulate analysis circulated amongst a wider group. A one-day focus group of representatives from different industrial sectors, service providers and academics supported by a questionnaire to establish requirements for surface and nano-analysis measurements. A one-day focus group held with regulators from the Environment Agency to establish future regulatory drivers for VAM-Physical. A questionnaire circulated amongst participants at the RSC Electro-analytical Group meeting sponsored by VAM. A meeting of the Trace Gas Focus Group in order to develop requirements for trace gas measurements. The Physical and Chemical VAM Programmes have been developed against a background of significant technological change and continued globalisation of trade. The consultations have established four broad-based drivers for physical analytical measurements which support the objectives of the Physical Programme: Industrial competitiveness and trade: Industrial production and innovation depends on the capability to measure the specifications of outputs and to control efficiency in manufacture. This is only possible when access is available to a uniform and comparable basis for measurements. Innovative new technologies pose requirements that are at the cutting edge of measurement capability. The optimisation of process performance depends on the ability to monitor control parameters on a stable basis. Import and export of manufactured goods depends on the acceptance of a uniform basis for measurements. Regulation: Regulations developed at both EU and UK level act to ensure that emissions from industry are controlled within agreed limits. Such regulations are most effective when they define performance levels that are both achievable by industry with available technology and operate at levels that provide adequate protection for society. In cases where protection is required that goes beyond that which can be achieved with available technology, regulation can act to stimulate innovation in the development of improved technology. The imposition of such regulation on a basis that is fair to regulated industry and society requires an infrastructure for traceable measurements at a level of uncertainty appropriate for the application. Representing the UK Internationally: The effective achievement of the programme s objectives requires input to be made into representing the UK s interests in various international fora where international documentary standards are developed and where the validity of the UK s national measurement standards must be demonstrated. Activities of this type have become more intense as a result of the drive to provide a single set of standards across the EU and to provide a measurement infrastructure that can support global trade. Environmental Protection and Quality of Life: The protection of human health and the environment depends on the capability to monitor and reduce the concentrations of certain species in the environment and to measure whether sources of the most damaging species are below accepted levels. These objectives can only be achieved if both regulators and regulated industry have access to measurement standards that are valid at the very low levels appropriate for environmental protection. The four main themes of the Physical VAM Programme are as follows: 1. Gas and Particulate Analysis: Measurements of gases and particulates are fundamental to the control of emissions, the quality of ambient and indoor air, and to many aspects of industrial process control. These measurements are applicable across a wide range of industrial sectors including power generation, chemicals/petrochemicals, oil and other fuels, aerospace, vehicles, electronics, water, waste incineration and landfill, public health, analytical instrumentation (including sensors), metal and non-metal processing, and pharmaceuticals. The VAM Physical Theme on Gases and Particulates enables 7 V A M B U L L E T I N

8 C O N T R I B U T E D A R T I C L E S industry to demonstrate compliance with regulations and statutes in a fair and costeffective manner, it ensures the acceptability of results by regulators accreditation bodies and the public, and provides support to regulators to enable them to enforce legislation in a technically sound and impartial way. 2. Electrochemical Analysis: Electrochemistry represents an increasingly important area of quantitative analytical chemistry. In particular, the determination of ph is the most commonly made analytical measurement throughout the world. The majority of these measurements are made to fulfil QA/QC requirements where traceability is increasingly being required. The VAM Programme has supported the development of the UK s primary standard ph facility, which is used principally to provide experimental data to support the UK s position in international negotiations at CCQM and CEN. This has enabled the UK to make a major input into the redefinition of the IUPAC ph Scale which was published in November The next VAM Programme will meet the requirement to maintain standards and calibration facilities for ph and electrolytic conductivity, together with exploiting research established under the NMS Quantum Metrology Programme into the optimisation of electrochemical methods for use in quantitative sensing. 3. Surface and Nano-analysis: Surface analysis and nano-analysis are essential tools for today s high technology and innovative industries. Interactions of surfaces with the surrounding environment are key to durability, compatibility and enhanced product quality. This is of prime importance for aerospace, chemicals, pharmaceuticals, health, personal care, packaging, electronics and IT equipment, oil and petroleum products, polymers, sensors, and transport. For nanotechnology applications, the spatial resolution is now critical. Issues for measurement by both lateral and normal forces in the AFM (Atomic Force Microscopy) as well as chemical imaging to work close to the molecular level will be resolved and protocols and traceable calibration systems developed and promulgated. Methods will also be developed for improving the spatial resolution for valid analysis in Secondary and Gentle Secondary Ionisation Mass Spectrometry (SIMS). This will include methods to address the identification of different compounds where the phase size is less than the probe size. 4. Knowledge Transfer: The requirement for cross-programme knowledge transfer has increased in recent years as the key stakeholder communities have become more aware of the outputs and deliverables. The key knowledge transfer objectives for the Physical VAM Programme will address the requirement to maximise the impact and increase usage of the outputs from the Programme. This will involve promotion and awareness raising through a range of knowledge transfer activities, services and products tailored to target audience groups. The Chemical VAM Programme Chemical measurement is a complex and critical process. The results can inform decisions whose economic value are many orders of magnitude higher than the cost of the analysis, or that are critical to quality of life decisions, particularly in healthcare. These considerations were high in our minds as LGC formulated the Chemical VAM Programme, which has a basic aim to improve the quality and comparability of measurements made in the UK, in order to improve our competitiveness and support regulatory need. In formulating the Chemical VAM Programme, we have also taken into account the key trends and issues influencing chemical and biochemical measurement. These were highlighted in LGC s independent survey into the analytical sector, which targeted key decision-makers. They confirmed that the spending on analysis was continuing to grow and they indicated that the main qualities that they looked for from analytical suppliers were laboratory accreditation and the cost and speed of the analysis. Chemical measurement is a complex and critical process. Heading the major issues facing analysis over the next few years were: Difficulties in resourcing trained analysts Increasing legislation Cost of analysis Implementing New Technology The acute shortage of analytical skills and resources has been a consistent message coming from the consultation process. Clearly there is a danger that if staff carrying out analysis lack the basic understanding of how to make valid measurements, this could have serious implications for the reliability of the results and the consequential decisions. Other factors that have been taken into account in developing the new Programme include the continued globalisation of trade, which increases the need to have results that are valid and comparable on an international scale. The chemical measurement infrastructure being developed through the International Weights & Measures Organisation, the BIPM, is making a key contribution to this. We have now reached the point where the UK analyst can start to benefit directly from these efforts. Thus, the new Programme seeks to help the UK analyst consistently achieve valid results, which are globally acceptable and to embrace new analytical technologies. The main mechanisms through which the Programme seeks to achieve this are by: Providing UK analytical laboratories with a total analytical package to help them produce valid, globally acceptable results. Key to this is the provision of reference values and tools to help analysts implement the six VAM principles for good measurement practice Anchoring the UK measurement system to the emerging global system, through participation in the key international intercomparisons Developing and maintaining a high accuracy analytical capability to provide values for reference materials of importance to the UK, which are recognised globally Evaluating new technology, addressing the associated measurement issues and providing the tools that enable analysts to use these with confidence The proposals build on the considerable capabilities in chemical and biochemical measurements, which have been established in previous programmes, particularly those relating to high accuracy mass spectrometry measurements for the assignment of reference values and in DNA quantification and highly multiplexed analysis. In addition, we have introduced a new theme relating to new analytical technologies. The five main themes of the Chemical VAM Programme are: 8 V A M B U L L E T I N

9 C O N T R I B U T E D A R T I C L E S 1. Chemical Metrology: This is an integrated work programme at the heart of the Chemical VAM Programme, which will deliver reference materials, reference values and validated high accuracy methods. It is aimed at improving measurement reliability within a framework which relates the accuracy of the measurements to national and international standards. The work will be focused on working with UK industry, regulators and laboratories in key areas of regulation, trade, health and the environment. Following consultations with UK experts, UK and European bodies and in a specially convened workshop on the use of reference materials we have identified a wide range of measurement issues to be addressed. These have been highlighted in the individual projects. Also included is work on selected international comparisons of the UK s chemical measurement capabilities with those of other countries. This is vital for demonstrating the accuracy and international uniformity of our methods and the global acceptance of our standards and reference materials. 2. New Analytical Technologies: A new theme for VAM, introduced in response to the significant impact of new technology. During the consultation, input and feedback was sought from over 50 UK experts in a variety of sectors, which indicated that the UK was lagging in the uptake of new analytical technology. The work is aimed at correcting this by addressing the need for validation and awareness of the new technologies. In particular it focuses on the development of an infrastructure to help UK industry implement new technology with confidence. It covers support for the adoption of new technology for characterising complex systems and the quality infrastructure needed to apply new technology to outof-laboratory measurements. 3. Nucleic Acid Measurements: The rapid advances in the technology and the introduction of highly multiplexed measurements raise some critical issues related to the validity and comparability of results. The proposals have been designed to strike a balance between building on the substantial outputs from the current programme and addressing new areas of technology. Included are the provision of techniques and standards for quantitative DNA analysis, quality assurance for array-based measurements and work to address the measurement issues associated with key emerging technologies. We have introduced a new theme relating to new analytical technologies. 4. Tools for Analytical Quality: The proposed work is focused on delivering practical help to analysts in the form of software and on-line tools for key statistical applications, including sampling, method validation and measurement uncertainty. It will provide improved access to method validation using web-based tools and address the need for ready access to information on sampling plans and for software to evaluate the results. 5. Knowledge Transfer: The training work package is aimed at addressing one of the key issues highlighted in the consultation process, namely the shortage of skilled analysts, by focusing on support for the professional analyst through sector-based training networks and training the trainer. The work was proposed following direct contacts with stakeholders from the public and private sectors and academe, to determine where the focus of the VAM work should be. Also included are activities to support the effective dissemination of the outputs of the Programme through the VAM website, the VAM Bulletin and technology transfer events. International Comparability through Traceability A common continuing objective for both the Physical and Chemical VAM Programmes is to provide UK users with direct access to the reference methods and measurement standards that are needed for traceable quantitative measurements of chemical and biochemical species and the determination of the composition of materials and compounds. Much of the demand for traceable measurements is to support fair international trade and the free movement of products and services. Measurement traceability ensures that the results of measurements can be accepted around the world without the need for additional testing of products and commodities. It is this driver that provides the principal justification for the UK s involvement in the international measurement system being developed under the auspices of the International Bureau of Weights and Measures, the BIPM. As a signatory to the Mutual Recognition Arrangement 2 (which involves 41 nations from around the world) the UK is committed to the development of measurement traceability to the international system of units (SI) demonstrated through international comparisons, termed Key Comparisons. For many analytical laboratories, the concept of measurement traceability is relatively new. However, the concept, requirement and benefit of measurement traceability is gaining increasing recognition, particularly following the adoption of ISO 17025, the internationally-accepted standard for test and calibration laboratories, which requires that laboratories make use of traceable measurement standards where available. This can only be met by ensuring that measurements made in the Key Comparisons are made on a basis that is transparent and recognised internationally, and by ensuring that the results of routine measurements and tests can be linked to the internationallyagreed reference values established through Key Comparisons. Whilst it is not possible to provide traceable measurement standards for every analytical quantity that is measured, the Physical and Chemical VAM Programmes will continue to support the development of standards and reference materials to underpin key measurements in trade, health and consumer and environmental protection. Further information Further information on proposals for the new Physical and Chemical VAM Programmes is available in the programme drafts which have been published for public review on the DTI s website REFERENCES 1. VAM Bulletin (2002) 28: Mutual recognition of national measurement standards and of calibration and measurement certificates issued by national metrology institutes, Paris, October V A M B U L L E T I N

10 C O N T R I B U T E D A R T I C L E S Assessment of the performance of low-cost indoor and in-car air monitoring devices Lesley Hanna Sira Ltd Introduction Indoor and in-car air quality have already been recognised as significant measurement topics in recent years. It has been estimated that people in developed countries can spend as much as 90% of their time indoors 1. In addition, the most vulnerable people in society, the very young, the infirm and the very old, are the most likely to spend a higher proportion of their time inside. Problems generated by indoor pollution have already been identified and have been labelled Sick Building Syndrome or Building Related Illness by the public. Poor quality air has even been linked to reduced productivity and lost time in industry. Many of these problems arise in the context of a wide variety of pollutants being present in new buildings from construction materials and new furnishings. Other potentially hazardous materials may be present due to combustion, air exchange with the outdoor environment and other human activities. The air quality inside cars also raises questions since new cars can contain a high proportion of materials which can emit volatile organic compounds (VOCs) such as plastics, fabric, carpet, paint and leather. Concerns about the effects of poor quality in-car air has raised the possibility of regulations to limit the import of vehicles that do not meet air quality testing limits. Project objectives As a result of the consequences of exposure to poor air quality, guidelines have been formulated to control indoor air and for this strategy to be effective it must be possible to measure air quality definitively and accurately. A summary of CEN and ISO standards applying to measurement of indoor pollutants is shown in Table 1. In recognition of the importance of indoor air chemistry this VAM project was formulated to look at the influence of reactive air chemistry on the accuracy and stability of low-cost devices used to monitor indoor air. A consortium of organisations was assembled to carry out the work. The lead organisation was Sira Ltd, the leading independent RTO specialising in instrumentation and intelligent systems. The National Physical Laboratory (NPL), with great expertise in the development and application of measurement techniques, and BRE, a world-class centre of capability in building research, joined Sira in the test programme. The University of York Environment Department provided a modelling facility and Optimat Ltd, an independent strategy consultancy, carried out a survey of the issues and requirements underlying the topics. Species of interest There are a large number of materials that can contribute to indoor pollution, from inorganic gases, such as nitrogen oxides, CEN standards ENV :1999 ENV :1999 ENV :1999 Building products. Determination of the emission of volatile organic compounds. Part 1: Emission test chamber method Building products. Determination of the emission of volatile organic compounds. Part 2: Emission test cell method Building products. Determination of the emission of volatile organic compounds. Part 3: Procedure for sampling storage of samples and preparation of test specimens pr EN Indoor air quality Diffusive samplers for the determination of gases and vapours Requirements and test methods Part 4: Guide for selection use and maintenance BS EN ISO :2001 Air quality sampling and analysis of volatile organic compounds in ambient air, indoor air and workplace air by sorbent tube/thermal desorption/capillary gas chromatography Part 1: Pumped sampling pr EN ISO :1999 Air quality sampling and analysis of volatile organic compounds in ambient air, indoor air and workplace air by sorbent tube/thermal desorption/capillary gas chromatography Part 2: Diffusive sampling ISO standards ISO/DIS ISO/DIS ISO/DIS ISO/DIS ISO/DIS Indoor air Part 1: General aspects of sampling strategy Indoor air Part 2: Sampling strategy for formaldehyde Indoor air Part 3: Determination of formaldehyde and other carbonyls Active sampling method Indoor air Part 4: Determination of formaldehyde Diffusive sampling method Indoor air Part 6: Determination of volatile organic compounds in indoor and chamber air by active sampling on Tenax TA sorbent, thermal desorption and gas chromatography using MS/FID. Table 1: Standards on measurement of indoor pollutants V A M B U L L E T I N

11 C O N T R I B U T E D A R T I C L E S through volatile organic compounds, such as solvents, to particles, bacteria and fungal spores. The major species of interest in this work were VOCs. VOCs can arise from many sources: building materials, soft furnishings, paint, solvents, personal care products and cleaning materials to name but a few. Measurement of air quality Current methods of sampling air quality are primarily centred around off-line sampling methods such as diffusion or adsorption tubes. These devices contain a packing material that adsorbs VOCs from the air. Subsequent analysis of the tubes will enable identification of the substances present. Two sampling strategies exist: active and passive. Active sampling involves drawing air through the tube and is suited to applications where time is limited or the concentrations are low. Passive sampling requires a long sampling time and is best when long-term exposure is to be assessed. Once the sampling is complete, the VOCs must be transferred to some kind of analytical instrument such as a gas chromatograph or mass spectrometer. This analysis will give details of the compounds present. Although commonly used and very straightforward, use of sampling tubes requires access to expensive equipment and trained operators for analysis, and this must also occur off-line. As interest in air quality increases there will be a greater requirement for lower cost and real-time analysis. Moving to a real-time measurement may also necessitate moving away from the ability to distinguish between individual compounds to a measurement giving an indication of the total amount of VOCs present in the air. As interest in air quality increases there will be a greater requirement for lower cost and real-time analysis. This approach has the disadvantage that the measurement does not reflect the fact that some VOCs present a greater hazard than others, and in addition does not reflect the absolute VOC concentration, being expressed in terms of equivalence to a calibration compound. However a real-time total VOC (TVOC) measurement is a significant advantage in identifying the optimal sampling location and period, where high VOC concentrations occur intermittently, where ventilation or the effect of outdoor concentrations are being investigated or where the rate of decay of emissions is to be investigated 3. Instruments such as electronic noses, photoacoustic spectrometers, flame ionisation detectors (FIDs) and photoionisation detectors (PIDs) can all be used to make real-time measurements. Each technique has strengths and weaknesses, however, and PIDs were selected as the most appropriate devices, along with the more conventional sampling tubes for evaluation within this study. Test facilities Controlled and characterised environments are required for the validity of measurements to be assessed within them. The needs of the project required the ability to provide both a constant environment and to simulate indoor air under controlled conditions but permitting chemical reactions to occur. For a constant environment, NPL made available their Controlled Atmosphere Test Facility (CATFAC). The CATFAC, shown below, allows a multi-component mixture to be maintained at a constant concentration with temperature and humidity also controlled and measured to a high degree of accuracy. NPL CATFAC. (Photo courtesy of NPL.) 1 1 V A M B U L L E T I N

12 C O N T R I B U T E D A R T I C L E S Sira test chamber. The atmosphere generated is then circulated at a predetermined velocity. The CATFAC is therefore ideal for exposure of measuring devices over a long period, in this case passive diffusive sampling. To simulate reactive indoor air chemistry a test facility was designed and built at Sira as part of the project (see photo above). This facility allowed a known environment to be created, after which reactions could occur and be measured both with the instruments under test and equipment designed to monitor the environment. The test atmosphere was contained within a chamber made from fluoropolymer and stainless steel. Fluoropolymer was chosen since it is not only chemically inert but also has good transmission of ultraviolet and visible light, which is required to initiate photochemical reactions. The stainless steel plates were provided with ports for gas inlets, instrumentation mounting and analytical measurements. This facility was ideal for shorter-term measurements using the PID devices described earlier. A mixture of typical VOCs were selected for the test programme, with seven compounds selected, representing the major functional groups commonly found in indoor air (aromatic and aliphatic hydrocarbons and oxygenated compounds). The accuracy of measurement of sampling tubes and PIDs were evaluated using these compounds in the presence and absence of ozone gas. Ozone is highly reactive and the project team wanted to find out what effect the presence of ozone had on both the species present in indoor air and the measurements carried out by the instruments themselves. Modelling of reactions In the presence of reactive species such as hydroxyl radicals, nitrogen oxides and ozone, oxidation of VOCs will occur, resulting ultimately in the formation of ozone and other secondary pollutants such as carbonyl compounds and particulates. Some change is therefore seen in the compounds present. In order to give an approximation of the course of the reactions arising, models have been designed which will calculate the expected reaction paths and products formed. The model used by the University of York could calculate the atmospheric degradation of 125 VOCs and was specially modified for use in this project. Use of the model increased the understanding of the reactions occurring during the reaction chamber test programme and complemented the work carried out using reactive systems. Results and conclusions The field of indoor air chemistry is a huge subject and there are many methods of measuring the species present in indoor air. It is also clear that accurate measurements of indoor air quality are of great importance. The course of the project has shown that significant reactions can be expected in a typical indoor environment where VOCs are present in a reactive atmosphere, and that this will change the composition of the atmosphere. Limitations of the measurement techniques used have been encountered during the project: choice of the sorbant for analysis is important and ozone was shown to have some effect on the results of analysis. The PID instruments used showed unexpectedly large variations even in the absence of ozone and further work is required to understand the source of these errors. Further information is available from: Lesley Hanna Sira Ltd Tel: lesley.hanna@sira.co.uk REFERENCES 1. Dimitropolou et. al., Atmos. Env., 35, , D. Crump, BRE, personal communication 3. L.E.Ekberg, Real Time Monitoring of Organic Compounds, Organic Indoor Air Pollutants, ed. T.Salthammer, Wiley- VCH, V A M B U L L E T I N

13 C O N T R I B U T E D A R T I C L E S Process Analytical Technology a new initiative within pharmaceutical development and manufacture Implications for the development and validation of methodology David Rudd GlaxoSmithKline Introduction There is a potential revolution in the air in the field of pharmaceutical development and manufacture. After many years of threatening to follow the lead taken by other manufacturing sectors, the pharmaceutical industry now seems closer than ever to adopting Parametric Release and Quality by Design concepts in order to improve its manufacturing capability and enhance profitability in an increasingly competitive commercial world. This article describes what is meant by the terms Parametric Release and Quality by Design and discusses why these concepts are now receiving so much attention within the pharmaceutical sector despite having been around for several decades. In turn, the impact of these approaches on analytical method development is considered, suggesting that a step-change in method validation philosophy is also required. The need for improved development and manufacturing concepts What do we mean by Parametric Release and Quality by Design and what has brought about the need for change? First, it should be said that the term Parametric Release is a dangerous one to use. It has been around for so long within the industry, without ever really being employed in earnest, that it has come to mean all things to all men. However, for the purposes of this article (and without choosing to show preference for any of the multiplicity of definitions which already exists), the author takes the term to mean a: system of quality assurance based on wellunderstood and well-characterised manufacturing processes which allows product quality to be established and demonstrated during the manufacturing process itself. A natural corollary of this definition is that the demonstration of product quality by end-product testing becomes irrelevant. And it is this aspect, rather than the objective of guaranteed product quality based on robust manufacturing processes, well-characterised raw materials etc, which many people have come to associate with the term Parametric Release. Nevertheless, whether the commercial driver is to reduce the volume and expense of end-product testing or simply to improve manufacturing efficiency, the concepts embraced by Parametric Release remain laudable and highly-prized. In many ways, the principles of Quality by Design are no different to those described above, although the emphasis of this manufacturing strategy is much more clearly associated with ensuring finished product quality rather than the removal of end-product testing. A Quality by Design approach clearly sets out to provide finished product of a guaranteed standard, similarly based on well-understood and well-characterised manufacturing processes, but perhaps with more emphasis on control and regulation of such processes within pre-determined operating limits. 1 3 V A M B U L L E T I N

14 C O N T R I B U T E D A R T I C L E S Thus, if finished product quality is likely to be affected by variation in input raw material properties, for example, then a Quality by Design approach might allow slight, but controlled variations in manufacturing conditions to achieve finished product of the appropriate quality. Both Parametric Release and Quality by Design concepts are well-established in other manufacturing sectors, but are still fairly peripheral within pharmaceutical development and manufacture. Of course, there are examples of certain unit processes or partial manufacturing steps which incorporate such philosophies (some antibiotic production, for example, and most sterilisation procedures) but, in general, the implementation of such approaches as part of an integrated development and manufacturing strategy is yet to be widely achieved. In the author s opinion, this is largely attributable to two main reasons: Regulatory authorities have commented that the quality of pharmaceutical development is often determined by time to market. Key dates generally need to be met during the product development phase (for example, when developing products for seasonal markets such as rhinitis or influenza) and this will inevitably affect the extent to which full process understanding and characterisation can be achieved. Thus, at the point of registration, the level of product and process understanding will not necessarily be complete. While this is hardly surprising (and while attempts at achieving perfection in this respect are not necessarily to be encouraged, due to their inevitable delaying influence in product availability), the consequence is that there may be aspects of routine product manufacture which eventually require modification or improvement in terms of robustness and/or overall efficiency of operation. seem to be operating to similar principles and to similar levels of capability, the overall impression is that the industry remains competitive and reasonably efficient. It is only when comparison is made across other manufacturing sectors (motor, aircraft, food and beverage etc) that the relative inefficiency of pharmaceutical manufacture becomes apparent. The recent infatuation of the pharmaceutical industry with Lean Manufacturing and Six Sigma concepts is evidence enough that there are significant lessons to be learned from the experiences of other manufacturing sectors. However, the relative conservatism which has existed within the pharmaceutical sector is now being subjected to considerable scrutiny, both from within (as manufacturing groups try to maximise their productivity and efficiency in an increasingly competitive commercial climate) and externally (as regulatory agencies recognise the opportunities for improved pharmaceutical development and manufacture coupled with a more science-based agency inspection and review process 1 ). And, while these commercial and regulatory agents for change are arguably both of equal significance, it is the regulatory perspective which is presently having the greater impact on the way in which pharmaceutical development and manufacture need to evolve. The Food and Drug Administration (FDA) Process Analytical Technology (PAT) initiative In response to the perceived opportunity for improved pharmaceutical development and manufacture, FDA have announced their science and risk-based approach to product quality regulation incorporating an integrated quality systems approach 1. This reflects an increasing concern regarding instances of product recalls, batch failures and manufacturing inefficiencies, coupled with a need to balance an already overstretched inspection and review programme with limited agency resources. In this context, it is interesting to note that, in 2002, FDA approved a mere sixteen drugs for use in the US 2 the lowest for more than a decade and this despite the investment of more than 30 billion in pharmaceutical research and development worldwide. Regulatory authorities have commented that the quality of pharmaceutical development is often determined by time to market. Generally pharmaceutical manufacturing efficiency has been measured and benchmarked within the industry sector. And, as most major pharmaceutical manufacturers 1 4 V A M B U L L E T I N

15 C O N T R I B U T E D A R T I C L E S As part of this new risk-based initiative, it is recognised that improved product development and greater manufacturing efficiency can be achieved using, for example, the Quality by Design concepts described previously. To this end, it is appreciated that enhanced approaches to process monitoring and control as opposed to the traditional reliance on end-product testing are necessary. As a result, the FDA Process Analytical Technology (PAT) initiative has been established and developed during the last 12 to 18 months. Broadly PAT describes the type of analytical monitoring technology which allows measurements to be made on the pharmaceutical manufacturing process itself, rather than on the material (the finished product or end-product) which results from such a process. In this respect, PAT may be considered as in-line or at-line monitoring (that is, at point of manufacture) rather than off-line or laboratory-based measurement. This has the huge advantage that, with measurements being made in real-time (rather than after the event), decisions can be made based on the real-time monitoring data and actions taken themselves in realtime. This allows process modifications to be made via feedback control systems; thus ensuring that process operating conditions remain in control at all times. As part of this new risk-based initiative, it is recognised that improved product development and greater manufacturing efficiency can be achieved. PAT is seen as part of the toolbox which will allow real-time monitoring (and, hence, real-time control) of pharmaceutical manufacturing processes both during development and routine application. The use of a wide range of diverse measurement techniques (chemical, physical, spectroscopic, acoustic etc) during product and process development, allows true process understanding to be achieved and enables the relationship between critical process parameters (for example, those which affect powder blend uniformity during mixing) and end-product quality attributes (in this case, tablet content uniformity) to be established. In addition, PAT monitoring techniques can be used during routine manufacture to ensure that the process operating conditions (or process specification as established during process and product development) remain in control. As a consequence of this, finished product of the desired standard is routinely produced despite any minor variations in input material quality or processing conditions: the Quality by Design concept at its most effective level. Implications for analytical method validation Although there are many sources of guidance regarding analytical method validation (some of which have gained a high level of regulatory acceptance 3,4 ), the re-location of analytical assessment from the laboratory to the manufacturing process results in a number of significant implications when considering the development and validation of such technologies. Many process analytical technologies are sensor-based. As a non-sampling technique, it becomes important to consider how many sensors need to be used, where such sensors are to be positioned and the reproducibility of sensitivity from one sensor to another. Technologies such as acoustic monitoring, which are well-established in other industry sectors, but which are relatively new in terms of pharmaceutical application, require sophisticated data interpretation methodologies. The reliability, principles of operation and general effectiveness of such algorithms need to be established and demonstrated. Spectroscopic techniques (such as near infra-red) depend on the chemometric treatment of data and the development of calibration models from well-understood training sets of data. The authenticity and reliability of such models need to be established and demonstrated. Suitable reference standards need to be developed for some process-based monitoring technologies (for example, acoustic monitoring). The importance of existing validation parameters (for example, linearity or repeatability) may need to be questioned when considering process-based applications. For example, most processes are dynamic and it may therefore be impossible to demonstrate repeatability of measurement when the sample itself is changing with time. Indeed, many process-based measure-ment techniques are themselves simply concerned with change, rather than with any absolute quantitative measurement: for example, near infra-red spectroscopy for powder blend monitoring during mixing processes. So these may simply require demonstration of discrimination at an appropriate level of sensitivity, rather than assessment of any truly quantitative performance. Conclusions A significant opportunity presents itself in terms of further development of existing guidance for analytical method validation and performance assessment when applied to process-based applications. The FDA science and risk-based approach is undeniably with us already and is likely to have significant impact on product and process development in the pharmaceutical industry within a very short time-frame. As a result, it is incumbent upon the analytical community within the pharmaceutical sector to recognise the impact of this initiative and to address the issues (and others) raised in this article regarding development and validation of process analytical methodology. REFERENCES 1. Pharmaceutical cgmps for 21st century: A risk-based approach, FDA, August 2002, gmp.html 2. The Financial Mail, 5/1/03, Simon Watkins. 3. ICH Q2A: Text on validation of analytical procedures, guidance/ichq2a.pdf 4. ICH Q2B: Validation of analytical procedures: methodology, guidance/1320fnl.pdf All views and opinions expressed in this article are solely those of the author, Dr David Rudd, and are not necessarily endorsed by GSK. 1 5 V A M B U L L E T I N

16 C O N T R I B U T E D A R T I C L E S Isotope-Dilution Mass Spectrometry a quantum method for chemistry? Martin Milton & Jian Wang National Physical Laboratory Introduction One of the features of most highaccuracy analytical methods is their dependence on Certified Reference Materials (CRMs) to give them well-defined traceability to stated references. In many cases, suitable CRMs are not available to cover the required range of analytes, and when they are available, they are often very expensive. The limitations on the availability of such CRMs is often considered to be the principal difficulty in bringing traceability to a wider range of chemical measurements. Consequently, there is substantial interest in the development of methods that operate without the need for external references. In the area of physical measurements such methods are known as quantum methods since they are ultimately anchored to realisations of fundamental atomic properties and therefore do not require access to reference standards from external laboratories. A project at NPL has the objective of developing an analytical method for use in chemical measurement that has analogous properties to the quantum methods used in physical measurement. This can be achieved by making measurements using a method that provides a link to the most basic components of chemistry atoms and molecules. Isotope Dilution One of the most fundamental principles used in chemical analysis is isotope dilution. It relies on the fact that different isotopes of the same element have extremely similar chemical behaviour. When combined with analysis by mass spectrometry, isotope dilution forms the basis of a powerful analytical method called Isotope Dilution Mass Spectrometry (IDMS). Despite its potential to give results that are independent of both the matrix and the analyte, most IDMS methods require the use of isotopic reference materials to overcome inherent limitations in their accuracy. Box 1 explains the operation of the two most widely used IDMS methods. It shows how the direct method requires an enriched spike material with a certified isotope ratio and known purity. Additionally, reference materials with certified isotope ratios are required for the highest accuracy applications in order to calibrate the isotope ratio scale of the mass spectrometer. The two-step method does not require any independent certification of the purity of the spike, but still requires independent certification of its isotope ratio. Box 1: The basic equation governing IDMS for a single blend is where x is the ratio of the mass of the unknown to the mass of an enriched spike. R sp, R s and R are the isotope ratios of the spike, unknown and blend respectively. Q is a parameter that involves a summation over all other isotopes in each sample. The use of this equation without further modification is usually referred to as direct or one-step IDMS. This method has the advantage of simplicity. The disadvantage, however, is that independent measurements of the purity of the spike and of its isotope ratio are required to determine Q. A development of the direct method involves the use of a reverse step in which the spike is blended with a pure sample of the same material as the analyte. This twostep or reverse IDMS method has the advantage that it is not necessary to have independent information about the purity of the spike or Q, but it is still necessary to have an independent value for R sp. NMS Quantum Metrology Programme A research project being carried out under the National Measurement System s Quantum Metrology Programme ( aims to show how some of these limitations of the IDMS method can be overcome. The research uses well-characterised gases as model systems for investigating the operation of IDMS and takes advantage of highly accurate mass spectrometers available for gas measurements. Additionally, the results can be validated against state of the art gas measurements based on gravimetrically prepared primary gas standards that have been demonstrated to be traceable to the SI. The principles being developed in the project have the potential for application to other areas where IDMS is used, including organic and inorganic analysis as well as the analysis of trace amounts of radioactive contamination. Isotope-Dilution Curve Method The major innovation resulting from the project is a new IDMS method known as the isotope-dilution curve method (Box 2) 1. The principal of the method is that the isotope ratio of any blend made from a given spike and a given unknown must lie on a mathematical curve known as the isotope dilution curve. It is then clear that when two blends are prepared gravimetrically, with known ratios of the spike to unknown, the measurements of the corresponding isotope ratios will define the parameters of the isotope-dilution curve. When the unknown is blended with the spike, a measurement of its isotope ratio enables the corresponding ratio of material blended to be calculated using the equations given in Box 3. The absolute mass of analyte in the unknown can then be calculated by multiplying by the mass of blend added. 1 6 V A M B U L L E T I N

17 C O N T R I B U T E D A R T I C L E S Box 2: The isotope dilution curve method has been validated using gravimetrically prepared blends (mixtures of natural abundance and spike carbon dioxide). The mass ratio of the natural carbon dioxide to the spike is determined by gravimetry during its preparation and is then used as the reference to be compared with the IDMS result. Figure 1 shows the quadruple input mass spectrometer used at NPL and Figure 2 shows a typical set of results from five repeat analyses. Box 3: The basic equation for isotope dilution is given in Box 1. The principle of the isotope-dilution curve method is that this equation is re-written in a form that can be plotted as a single curve on a graph of the isotope ratio of the blend versus the amount ratio of blend to unknown. This curve is shown in Figure 3 where the isotope ratio of the blend is plotted relative to the isotope ratio of the pure reference material. Each point on this curve corresponds to a particular blend prepared from the given spike and the unknown sample. The curve can be uniquely defined through the measurement of any two blends that are known to lie on it. When the curve has been defined, the measurement of the isotope ratio for any blend enables the corresponding value x 1 to be read from the curve, which can then be used to calculate the unknown amount of analyte in the blend using: Figure 1: The four-channel stable gas mass spectrometer used at NPL for isotope dilution studies. where D is a function of the isotope ratios for the measured blends, and x 2 and x 3 are the amount fractions in the two reference blends. Since it is only the relative position of points on the isotope dilution curve that is important, the method is relatively insensitive to systematic errors arising from either the mass spectrometer or the procedure used to implement the method. Figure 2: Results of five analyses using the IDMS curve method. The solid line indicates the reference value from gravimetry and the dotted lines indicate its uncertainty (k=2). Validation of the Isotope- Dilution Curve Method Laboratory work is underway at NPL to validate the isotope-dilution curve method and to demonstrate its application to the measurement of primary standard gas mixtures of carbon dioxide. This uses a custom-built mass spectrometer with an array of detectors configured to measure the isotopes of carbon dioxide. The most important feature of the instrument is that it has four independent gas sample inputs, each of which can be precisely controlled using a bellows valve driven by a stepper-motor. This enables the samples to be changed without any change in the inlet conditions. Initial experiments using isotopically depleted carbon dioxide diluted in natural Figure 3: The isotope dilution curve for a blend of a natural isotope blended with a spike that is depleted in the natural isotope. carbon dioxide have demonstrated the principal of operation of the method. These have shown agreement to better than 100 parts-per-million (relative to the stated value). The validity of this result is further assured by observations that the results of 1 7 V A M B U L L E T I N

18 C O N T R I B U T E D A R T I C L E S the measurement are largely independent of the drift of the instrument. This property of the isotope-dilution curve method will enable its application to instruments with substantially poorer performance than the one used at NPL. Recent experiments at NPL have substantially increased the scope of application of the method by including a GC pre-separation stage which enables the measurements to be made in the presence of a matrix which is separated before the sample is introduced into the mass spectrometer. In the experiments carried out at NPL, the matrix has been nitrogen and the results of the method have been validated against the value of NPL s internationally recognised primary standard gas mixtures. The values have been shown to be comparable to 0.3% (relative to value) and are now largely limited by the repeatability of the GC separation process. Further work is aimed at improving this performance and a number of publications have been produced to give analysts working on the Physical and Chemical themes of the VAM Programme the opportunity to exploit the technique. Although it seemed unlikely before the inception of this project, it now appears possible that there are some primary methods used for chemical analysis, that have some of the benefits of the quantum methods exploited so successfully for high-accuracy measurements of time, length and electricity. In particular, the isotope-dilution curve method developed at NPL can enable IDMS measurements to be carried out without reference to externally certified reference materials. REFERENCES 1. Milton, M. J. T., Wang, J., International Journal of Mass Spectrometry, 218 (2002), BIBLIOGRAPHY 1. Milton, M. J. T., Wielgosz, R. I., Rapid Communications in Mass Spectrometry, 16 (2002), Milton, M. J. T., Wang, J., Harris, P. M., Implementation of isotope dilution mass spectrometry with one, two and three reverse steps, NPL Report COAM 7 (2002). 3. Milton, M. J. T., Wielgosz, R. I., Metrologia 37 (2000) C A S E S T U D Y Meeting the needs of regulation and trade: determining low levels of sulfur in fuel Jim Crighton & Harry Read BP plc Mike Sargent LGC Introduction The European commitment to a cleaner environment has resulted in fuel regulations that require a steadily decreasing sulfur content for road transport fuels (see Table 1). In addition, sulfur free fuels (< 10 mg kg -1 sulfur) are becoming increasingly available in EU member states and should be available in all member states by It is anticipated that the 10 mg kg -1 maximum sulfur concentration for road transport fuels will become mandatory in the EU from With such a significant reduction in sulfur concentrations, many existing methods of analysis have had to be adapted to deal with these new requirements and need to be validated at the lower levels of sulfur. An important issue for the oil industry has been the lack of CRMs or reference methods able to provide reference values of sufficiently small uncertainty for reliable verification of methods at the new, low levels. Year Sulfur Content (mg kg -1 ) EC Directive /12/EEC (diesel) 98/70/EC 150 (petrol) /70/EC Table 1: European regulatory requirements. LGC have recently 1 developed a novel reference method for S in fuel based on isotope dilution-mass spectrometry (IDMS) using microwave digestion of samples followed by inductively-coupled plasma mass spectrometry (ICP-MS) for the isotope ratio measurement. This methodology has itself been verified by comparison of results with overseas national measurement institutes which provide reference values based on IDMS using well-established thermal ionisation mass spectrometry (TIMS) techniques. The latter methods are, however, less well-suited to very low levels of sulfur due to problems of reagent contamination. LGC reference values obtained with the new ICP-MS method have contributed (in association with BP plc), to the development of CEN/ISO methods appropriate to the new regulatory requirements and to validation of these methods at the new limits required by the EN228 and EN590 fuel specifications. The methodology has also been used for certifying new reference materials for diesel fuel at levels embracing the new lower limits required for the EN fuel specifications. Existing methodology Following many national workshops considering the implications of the proposed mandatory limits on sulfur concentrations in 1 8 V A M B U L L E T I N

19 C A S E S T U D Y road transport fuels, it became clear that existing analytical methods quoted in EN fuel specifications were not suitable for sulfur measurements at the proposed new lower levels. As a result, at the plenary meeting of CEN Technical Committee 19 held in Madrid in June 1997, a resolution was passed creating a European Working Group (WG27) which would be concerned with the comparison of sulfur content determination methods for sulfur levels not greater than 0.1 % m/m. The working group consisted of national experts in sulfur analysis from a number of European community states and was charged (in collaboration with national standardisation bodies), with developing a consensus view on which methods should be referenced in the revised EN228 and EN590 fuel specifications. The working group considered and ranked methods on the basis of a number of criteria, which included: Sensitivity/limit of quantitation Precision Availability Cost and maintenance Robustness/ease of use (including use by shift/less skilled staff) Existing methods quoted in EN228 and EN590 Based on these criteria, the following analytical methods were short-listed: Wavelength Dispersive X-ray fluorescence (WDXRF) Energy Dispersive X-ray Fluorescence (EDXRF) Wickbold Combustion Microcoulometry Combustion/Ultra-violet Fluorescence (UVF) A European wide round robin was carried out in involving 69 laboratories from 9 countries in which the precision (repeatability and reproducibility) Φ of the above analytical methods was tested using 15 fuel samples (8 petrol and 7 diesel) covering sulfur concentrations in the range 5 to 500 mg kg -1. According to EN ISO 4295, for a method to be acceptable for inclusion in specifications, the reproducibility R should be less than half of the maximum concentration specified. On this basis, it was found that the EDXRF method specified in EN228 and EN 590 (EN ISO 8754 : 1995) was not suitable for use at the sulfur concentrations proposed in the 2000 specifications. However, a new EDXRF method (ISO/DIS 20847) developed by the working group was found to be suitable for measurement of sulfur concentrations down to 50 mg kg -1 (i.e specifications) and it was proposed that this method replaced EN ISO 4295 in EN 228 and EN590 specifications. The working group consisted of national experts in sulfur analysis from a number of European community states. WDXRF (EN ISO 14596) and a new UVF method (ISO/DIS 20846), developed by the working group were found to be suitable for inclusion in both the 2000 and the proposed 2005 specifications. However, only the UVF method was found to have adequate reproducibility for measurement of sulfur at the 10 mg kg -1 level (i.e. sulfur free fuel). The performance (reproducibility) of the methods at levels below 50 mg kg -1 was found to be very disappointing and did not reflect the inherent limits of detection which most of the techniques are capable of achieving. In particular, the reproducibilities for the Wickbold and microcoulometry methods (EN and ISO/CD 16591) calculated from the round-robin data were found to be considerably worse than those quoted in the methods. In the case of the Wickbold method, this was considered to be due to the loss of skilled operators, as more and more laboratories are replacing this technique with alternatives (particularly on safety grounds!). For this reason, it was proposed that the Wickbold technique was dropped from the EN229 and EN590 specifications (and replaced with UVF). The deterioration in performance of the microcoulometry method however, was more worrying. This, together with the relatively poor performance of the other analytical techniques at low sulfur concentrations, was thought to be due to problems associated with the inexperience of many laboratories in working at these new, lower levels and possibly due to problems associated with standards and instrument calibration. It was therefore recommended that a further round-robin study be carried out, concentrating on sulfur concentrations lower than 50 mg kg -1. It was also felt that inclusion of some reference materials of appropriate matrix and sulfur concentration could perhaps minimise some of the problems experienced in the first round-robin and allow the reproducibilities to better reflect the inherent sensitivities of the techniques. The availability of these reference materials would also allow any biases between the methods to be evaluated. Provision of reference values using high accuracy IDMS The IDMS method used to obtain reference values is based on a closed vessel microwave digestion followed by measurement with a sector-field ICP-MS in medium resolution mode (R=4000). Medium resolution is required in order to resolve the sulfur isotope peaks from oxygen polyatomic interferences. The advantages of this method over existing IDMS methods based on thermal ionisation mass spectrometry (TIMS) are that it is quicker, there is less risk of contamination of low level samples and it has the potential to achieve lower detection limits. The method is applicable to the quantitative determination of sulfur in fuel samples at concentrations up to ca 1000 mg kg -1 in all fuel types but particular care must be taken when handling volatile sample types such as petrol. Sample aliquots are weighed into the microwave digestion Φ repeatability, r the difference between two test results obtained by the same operator with the same apparatus under constant conditions on identical test material would in the long run, in the normal and correct operation of the test method, exceed this value in only one case in twenty. Reproducibility, R the difference between two single and independent test results obtained by different operators working in different laboratories on identical test material would in the long run, in the normal and correct operation of the test method, exceed this value in only one case in twenty. 1 9 V A M B U L L E T I N

20 CASE STUDY vessel and weighed aliquots of a solution of isotopically enriched sulfur ( 34 S= 94.2%) are added to each such that the ratio of 32 S/ 34 S is 1. The blended sample/spike mixture is digested with HNO 3 and H 2 O 2 in the microwave for 40 minutes then diluted by weight to the optimum sulfur concentration for ICP-MS measurement of the isotope ratio. A spiked solution of a natural sulfur calibration standard is also prepared and measured with the samples; i.e. the measurements are traceable to a natural sulfur standard and do not rely on the accurate characterisation of the isotopically enriched spike material. The methodology incorporates two key developments, which simplify the IDMS measurements and reduce the likelihood of errors. First, the isotope dilution technique used for this work is an exact matching procedure in which the isotopically enriched material is added to the sample and the natural calibration standard such that the ratio and intensity of the isotopes in each measurement solution is the same. In the case of this sulfur analysis, the solutions were blended to give a 32 S/ 34 S ratio close to unity. This method, which has been described in detail elsewhere 4, has the advantage that it eliminates the effects of instrumental mass bias, detector deadtime and linearity since all samples and calibration standard solutions are affected to the same degree. The second development addresses the problem that isotope ratios of sulfur are known to vary in nature due to biogenic and thermal fractionation and it cannot be assumed that the isotope ratios in samples and standards are identical. Therefore, all test samples must be measured for natural isotope ratios before IDMS analysis. Accurate measurement of this ratio by ICP-MS has been achieved through use of silicon to correct for instrumental mass bias. Silicon has been found to be suitable for this purpose since the 30 Si/ 28 Si ratio (29.88) is similar to that of the 32 S/ 34 S ratio (22.0) with the same mass/charge difference in the isotope pairs. Silicon has been found to have minimal natural variations in isotope ratio. The methodology incorporates two key developments, which simplify the IDMS measurements and reduce the likelihood of errors. The full method for this procedure has also been described previously 1. The precision and accuracy of the method at higher sulfur concentrations is demonstrated in Figure 1 which shows results for six replicate analyses of a NIST reference material. Development and validation of revised CEN/ISO methodology Following on from experiences gained in the first round-robin, CEN TC19 Working Group 27 decided in 2000 to carry out a second round-robin 3 concentrating on the range 1 to 60 mg kg -1 sulfur. This concentration range was selected in order to obtain robust precision statements for the analytical methods at levels consistent with current and future sulfur specification limits. Five analytical methods were included in the study, including three new methods (ISO/DIS 20846, ISO/DIS and ISO/DIS 20884) developed by the working group. The methods included were: EN ISO WDXRF (with internal standard) ISO/CD UVF ISO/CD EDXRF ISO/CD WDXRF (without internal standard) ISO/DIS Oxidative microcoulometry Figure 1: Results of analysis of NIST SRM2724b by IDMS. Error bars represent ± 95% expanded uncertainty (k=2) for each aliquot. Horizontal lines represent certified value and expanded uncertainty. 2 0 V A M B U L L E T I N

21 CASE STUDY Sulfur Concentration (mg kg -1 ) Method Technique EN ISO * WDXRF (internal standard) ISO/CD Petrol Combustion/UVF ISO/CD Diesel Combustion/UVF ISO/CD Petrol EDXRF ISO/CD Diesel EDXRF ISO/CD * WDXRF (no internal standard) ISO/DIS Petrol Oxidative microcoulometry ISO/DIS Diesel Oxidative microcoulometry * All products, X-ray tube power > 1 kw Table 2: Comparison of the reproducibilities obtained in the round robin. Petrol Diesel Sulfur, (mg kg -1 ) r Sulfur, (mg kg -1 ) r IDMS ± ± 0.65 EN ISO ISO/CD ISO/CD ISO/CD ISO/DIS r = repeatability; = absolute difference between mean value and IDMS reference value Table 3: Comparison of Method Mean Concentrations with Reference Values provided by IDMS. All of these methods were essentially the same as those included in the roundrobin, with the exception of ISO/CD 20884, which differs from EN ISO in that it does not require use of an internal standard (and is therefore much faster and more convenient to use). Seven petrol samples and seven diesel samples covering the range 1 to 60 mg kg -1 sulfur were included in the round robin. In addition, one of the diesel samples was blended with 5% FAME (fatty acid methyl ester) from two different sources to produce two additional samples for inclusion in the round robin, in order to test the applicability of the methods for these products (allowed under EN590 specifications). In order to try to minimise any potential degradation in reproducibility caused by calibration issues, two quality control samples (one petrol and one diesel) were included in the round-robin. These two materials (labelled QCP and QCG) were certified by LGC using high accuracy IDMS as described above. The reference values were provided to the laboratories participating in the round-robin study so that they could check performance/ calibration of their instruments prior to analysing the round-robin samples. A total of 92 laboratories from 10 European countries participated in the round-robin. Table 2 shows the reproducibilities of the various methods calculated from the roundrobin data for three critical sulfur concentrations (corresponding to low sulfur (< 50 mg kg -1 ), ultra low sulfur (< 30 mg kg -1 ) and sulfur free fuel (< 10 mg kg -1 )). In order to try to minimise any potential degradation in reproducibility caused by calibration issues, two quality control samples were included in the round-robin. In order to check potential methods for inclusion in the EN228 and EN590 specifications for any potential bias, the mean results obtained from the round-robin were compared with results obtained by LGC using high accuracy IDMS. The results are shown in Table 3. In all cases, the differences in concentrations between the method mean values and the IDMS reference values are small relative to the repeatabilities of the methods, indicating that any biases are insignificant from a practical point of view. Comparison of the results obtained for the diesel fuel with and without the addition of 5% FAME also showed no significant differences indicating that the methods are also applicable to these products. Based on the ISO 2R rule, it can be seen that all of the analytical methods included in the round-robin study are suitable for measuring sulfur in road transport fuels down to and including 30 mg kg -1 S (i.e. ultra low sulfur fuel). Both WDXRF and UVF were found to be suitable for measuring sulfur in sulfur free fuels (i.e. < 10 mg kg -1 S), although it was found that some WDXRF instruments utilising lower power X-ray tubes (1 kw and less) could not achieve adequate precision (reproducibility) at this level. EDXRF (ISO/CD 20847) was found not to be suitable for measuring sulfur in fuels at concentrations below 30 mg kg -1. Although, it was found that a new generation of EDXRF instruments utilising polarised X- ray sources gave significantly better precision figures for low sulfur concentrations than those for conventional EDXRF instruments and may be suitable for measuring sulfur in fuels at concentrations below 10 mg kg -1 S. From the above, it is clear the all three methods developed by CEN TC19 Working Group 27 are suitable for use at the 350, 2 1 V A M B U L L E T I N

22 CASE STUDY BP plc (2002) 150, 50 and 30 mg kg -1 sulfur specification levels. ISO/CD (UVF) and ISO/CD (WDXRF, no internal standard) are suitable for use at sulfur concentrations of 10 mg kg -1. EN ISO (WDXRF, with internal standard) was also found to be suitable for measuring sulfur at this level, but since ISO/CD is faster and easier to use, the working group recommended that the former technique was replaced by the latter in EN228 and EN590. In general, the precision obtained from the 2001 round-robin was significantly better than that achieved in the roundrobin Whilst this may be partly explained by the greater experience which laboratories had gained in measuring low sulfur products, the availability of the two quality control samples with reference sulfur concentrations certified using high accuracy IDMS, was undoubtedly a major contributing factor. The availability of these reference samples also permitted the methods to be validated in terms of freedom from bias. Production of low sulfur CRMs Six diesel samples have been collected and bottled for certification as reference materials for sulfur in fuel. These materials represent the range of analyses currently required and the lowest legislative limits that may expected in Europe within the next 10 years. Preliminary analysis of the six materials indicate the concentration levels detailed in Table 4. Experimental work for the certification of the 50 mg kg -1 material has been Reference Estimated S amount Material content (mg kg -1 ) LGC LGC LGC LGC LGC LGC Table 4: Preliminary Analyses of 6 LGC sulfur in fuel reference materials. completed and indicates a sulfur content of 52.4 mg kg -1 with an expanded (95% confidence) uncertainty of 1.3 mg kg -1 (2.4%). Certification of the other materials is expected to follow in due course, with priority being given to the 30 and 10 mg kg -1 levels. Conclusions International agreement on methodology for the determination of S in fuel is of major commercial importance as well as being an essential requirement to facilitate trade in the face of ever more demanding regulatory requirements. Development of a high accuracy method of analysis within VAM, together with its verification through international collaboration between LGC and other national measurement institutes, has allowed determination of reference values with small uncertainties at low sulfur levels. The application of these reference values has been demonstrated for the development of CEN/ISO methods, for validation of methods used by UK industry, and for certification of new reference materials. 2 2 V A M B U L L E T I N

23 CASE STUDY Acknowledgements The round-robin studies referred to in this article were carried out under the direction of CEN TC19 Working Group 27, comprising the following national representatives: Paolo Tittarelli, Convenor Stazione Sperimentale Combustibili San Donato Italy Harry Read, Former Convenor BP plc Sunbury UK Jim Crighton, Secretary BP plc Sunbury UK Franck Baco IFP CEDI Vernaison France Sophie Collet TotalFinaElf CReS Solaize France Paul Crezee Nerefco Rozenburg The Netherlands Claudia Do Marco TotalFinaElf CReG Harfleur France Francoise Douce TotalFinaElf CReS Solaize France Bo Edroth Preem Raffinaderi Goteborg Sweden Ralph Hensel Shell Global Solutions Deutschland Hamburg Germany Charles-Philippe Lienemann IFP CEDI Vernaison France Jose Gomez-Martinench CEPSA Research Centre Madrid Spain Patrizia Ruggieri AgipPetroli CREA San Donato Italy Mario Van Driessche ChevronTexaco Technology Ghent Gent Belgium Antoni Marchut, Observer Instytut Technologii Nafty Kraków Poland Certification data for the LGC reference materials has been provided by Peter Evans and Ruth Hearn, with the assistance of Celine Wolff Briche for estimation of uncertainty. The authors are also very grateful to all the laboratories who participated in the round-robin studies through the various national standardisation bodies. The work on development of the IDMS method described in this paper and the preparation of the CRMs is supported by the Department of Trade and Industry (UK) as part of the National Measurement System Valid Analytical Measurement Programme. REFERENCES 1. Evans, P., Wolff-Briche, C., Fairman, B.E., J. Anal. At. Spectrom., 2001, 16, Sulfur Methods for EN 228 and EN 590 Fuel Specifications, CEN TC19 WG27 Round Robin report Test Methods for the Determination of Sulfur Content, CEN TC19 WG27 Round Robin report Guidelines for achieving high accuracy in isotope dilution mass spectrometry (IDMS), Edited by M. Sargent, C. Harrrington, R. Harte, The Royal Society of Chemistry, Cambridge, UK, ISBN Decline in skills V A M I N E D U C A T I O N During the last few months we have been busy formulating the next VAM programme. Part of this process is to solicit the views of instrument manufacturers, analytical staff in a range of companies, regulators and those involved in education. Most sectors of industry are concerned about the lack of basic laboratory skills of those graduating from UK universities and those leaving school. The reason for this decline is frequently given as reduced budgets for consumables and increased safety regulations in schools, colleges and universities. The effect of safety regulations is probably a perceived problem rather than a real one. We are aware that the whole of the education sector is under extreme pressure, but if UK industry is unable to recruit from within the UK it will go elsewhere and there are signs that is already happening. What has this to do with VAM? For the last ten years the VAM programme has worked with schools, universities and industry to both assess the level of competence of laboratory staff and to provide training tools and courses to supplement materials provided by instrument manufacturers and academia. Just this last month, five training guides have been published by the Royal Society of Chemistry (RSC) as part of the VAM product portfolio. The guides cover mass, volume, ph, HPLC and GC. These were developed with SANG (the Chemical Industries Association Specialised Organic Chemicals Sector Analytical Networking Group) and have been in use by their organisations for the last few years. As part of the pack there are 2 CD-ROMs that deal with the basic practical skills required by staff who work in analytical laboratories. 2 3 V A M B U L L E T I N

24 V A M I N E D U C A T I O N We continue to work with schools, colleges, universities and those professionals who are required to make measurements. One of the hot topics at the Association for Science Edcation (ASE) conference in Birmingham in January was the introduction of applied courses, but there is a concern that schools do not have the equipment to deliver the practical work required and teachers lack confidence to teach the practical aspects. Most sectors of industry are concerned about the lack of basic laboratory skills of those graduating from UK universities and those leaving school. We are working with the RSC, 4 Science and Dstl to develop INSET training for teachers to support these courses. The concept of reliable measurements and words such as precision, accuracy and bias are sprinkled through course specifications. At the ASE we had a workshop session, on measurement, for teachers and technicians. It was clear that there was confusion about the exact meaning of the terms and they were appreciative of some of the simple tasks we used to illustrate what they believed to be complicated issues. The Qualification and Curriculum Authority (QCA) are interested to talk with us about help with terminology. Centres taking part in the PT competitions appear to have a better understanding of measurement issues after participation in a few rounds. We are awaiting the results from the centres that have taken part in Universities have not been neglected. We are still able to supply lectures on VAM issues and implementation of best practice to undergraduates and postgraduates. Workshops for postgraduates are also planned. The three self-help sector-based training networks supported under VAM are progressing well. Each of the networks has had a hot topic that has implications on measurement performance. The environmental analysis group (EMTN) has been discussing the Environment Agency s monitoring certification scheme MCERTS, Performance standard for laboratories undertaking chemical testing of soils. Clinical Governance is having an impact on point of care testing and this has been on the agenda of POCT meetings, along with the possible changes to the accreditation standard. SANG, the specialised organic chemicals network, is working on ways to improve efficiency through better measurement procedures. If UK industry is unable to recruit from within the UK it will go elsewhere. Further information can be obtained from the VAM Helpdesk. V A M N E W S European Trial of Airborne Particle Measuring Instruments comes to NPL Airborne particulate matter (PM) is strongly implicated as being the worst air pollutant in the UK. The effects of such material on human health can be worse than airborne gases such as ozone or nitrogen dioxide. Measuring particulate matter is complicated. Firstly, larger particles are less likely to enter human lungs and cause medical problems. There is therefore a desire to exclude these particles from measurement results. These larger particles can also have a disproportional effect on results. Secondly, measurement is further complicated by there being no self-evident attribute of the particles, i.e. size or composition, which is most relevant for monitoring purposes. Current air quality standards define the amount of particulate material in terms of the mass below a certain aerodynamic diameter, in a given volume of air. In principle, such measurements are simply a matter of weighing suitable filters before and after a known volume of air is pumped through them via a size-selective inlet. In practice, inconsistencies can arise due to a number of factors such as semi-volatile matter leaving the filter after sampling, and the absorption of water by the filter or particles. NPL is taking part in a major European trial to improve the measurements of airborne particulate matter less than 2.5 microns in diameter (known as PM2.5). The trial will involve running 20 separate instruments in parallel in order to compare results and ultimately help choose a reference method. The work will involve comparing different filter materials and quantifying sources of measurement uncertainty. NPL is one of the eight participating sites chosen across Europe. The trial is jointly funded by the VAM programme and the EC and is expected to run until May Further information about the trial is available from: Paul Quincey National Physical Laboratory Tel: paul.quincey@npl.co.uk Web: V A M B U L L E T I N

25 UKAS appoints Lord Lindsay as chairman V A M N E W S Former Scottish Office Minister, Lord Lindsay has been appointed as the new chairman of the United Kingdom Accreditation Service (UKAS). His appointment was approved at the organisation s Board Meeting, following the AGM on 8 October Commenting on his appointment, Lord Lindsay said: As the sole body recognised by government to accredit organisations providing testing, certification, inspection and calibration services, UKAS plays a vital role in maintaining confidence in standards throughout the British public and private sectors, enabling markets to function more efficiently. I look forward to helping UKAS to build on its achievements since it was formed in Welcoming UKAS new chairman, Chief Executive, Linda Campbell, said: Jamie brings broad experience of industry, environmental issues and government policy. His work in the food sector and the environment means he has first hand knowledge of the importance of technical standards in minimising risk. He will play a valuable part in providing UKAS with strategic guidance while acting as a first class ambassador for the company. Lord Lindsay replaces UKAS former chairman, Dr Bryan Smith, who has retired after seven years in the post. RSC to produce the perfect cup of tea To honour George Orwell on the centenary of his birth, the Royal Society of Chemistry is to decree the definitive method by which the perfect cup of tea should be made and served. The RSC will also try to persuade cookery book publishers to include the recipe in future publications, thereby fulfilling Orwell s published wish that the step-by-step preparation of the national drink be properly defined. Orwell, best known for his books Animal Farm and Nineteen Eighty Four, wrote the celebrated essay A Nice Cup of Tea three years before his death. In the essay, which stands alongside masterpieces such as A Hanging and Shooting an Elephant, Orwell expresses surprise and concern that the extremely serious matter of tea brewing is not featured in cookery books, a situation that has not changed. Orwell says: This is curious, not only because tea is one of the mainstays of civilisation in this country but because the best manner of making it is the subject of violent disputes. Now, to celebrate his 100th birthday, the RSC is to remedy the oversight. The effort to define what exactly makes the best cup of tea is also the RSC s way of saying that it does not bear ill will for Orwell s negative opinions of scientists, and of chemists in particular, as portrayed in his essay, Science. The RSC is to consult its own drinks and flavour specialists in order to identify ways to prepare and to serve the unbeatable cup of tea. The RSC will also be seeking the teamaking views of the general public through its local branches around the UK. Orwell defines in his essay the way to make a perfect cup of tea, saying: I find no fewer than eleven outstanding points. On perhaps two or three of them there would be pretty general agreement but at least four others are acutely controversial. He asserts that Indian or Ceylonese teas must be used, dismissing China tea as having no stimulation. One does not feel wiser, braver or more optimistic after drinking it. The novelists insists that the tea should be strong, claiming that six heaped teaspoons are required for a quart pot. No strainers or tea bags (muslin at that time) should be used and it is perfectly right that leaves should reach the cup and for them to be eaten in considerable quantities. He also insists that a cylindrical breakfast cup must be used, not a flat-based cup. He is adamant that the tea should be poured before the milk, claiming that it is the only way by which to regulate the amount of milk. His last point is that sugar should never be taken. How can you call yourself a true tealover if you destroy the flavour of your tea by putting sugar in it? It would be equally reasonable to put in pepper or salt. Tea is meant to be bitter just as beer is meant to be bitter. 2 5 V A M B U L L E T I N

26 V A M N E W S Challenging the Limits Of Detection Current trends in bio-analytical methods show an increased interest in detection of trace levels of DNA. Such sensitive detection methods are central to regulatory, public health, medical, environmental, and quality control sectors. One such technique, which is used for a wide range of applications, is fluorogenic quantitative realtime PCR, as it provides a route to semiquantitative DNA analysis, and sensitivity is increased compared to conventional amplification-based methods. Introduction of robust quantitative PCR analysis for applications such as the detection of genetically modified organisms in foodstuffs, or pathogens in food and environmental samples requires that the performance characteristics of the method be accurately assessed in method validation. The sensitivity of a method is associated with the lower limit of applicability of that method. Derived from this, the performance characteristic of the limit of detection (LOD) can be defined as the smallest concentration of a target analyte that can be detected and distinguished from a zero result with a given probability. The evaluation of the LOD of an assay is critical for trace detection methods, especially where the result will be used for regulatory or public health applications. Introduction of robust quantitative PCR analysis requires that the performance characteristics of the method be accurately assessed in method validation. Formal derivations concerning LOD do not take into account atypical data sets that are generated from real-time PCR techniques. Fluorogenic real-time PCR measures the accumulation of relative fluorescence during the course of the reaction, which at each cycle is proportional to the amount of PCR product formed (Figure 1). Results are not always normally distributed, and analyses using parametric statistics are thus not valid. Furthermore, negative controls do not give a zero measurand result, as a true negative reaction would give a sample value of infinity. An approach is being developed to determine this LOD. It uses a data set characteristic of typical non-normally distributed results produced from real-time PCR, using an experimental design representative of usual laboratory conditions. A computer simulation calculates the probabilities of detecting PCR products from the data set, using bootstrapping techniques. Both experimentally and theoretically determined LODs are produced, and provide an estimate for the sensitivity of the method, based on the real laboratory performance of the technique. By using the data analysis approach, the probability of detecting an analyte at a specific concentration can be determined, based on the known sensitivity of the method. Such information facilitates the design of robust analytical strategies, and thus enables reliable practical application of real-time PCR analysis. The wide applicability of this bootstrapping and data modelling approach should be of general interest to laboratories involved in trace-level detection. For further information, please contact: Malcolm Burns LGC malcolm.burns@lgc.co.uk Relative fluorescence Copy Number Data processing 100% Probability of detecting target analyte Threshold 0% Sample values PCR Cycle number Log (copy number) Figure 1: Generation, processing, and interpretation of real-time PCR data for limits of detection. Figure 1 illustrates a typical real-time PCR amplification plot where sample values are generated according to the cycle number at which point a sample copy number crosses a predetermined noise threshold. These sample values are then used in a simulation to generate a plot of the probability of detecting a PCR product dependent upon initial copy number. 2 6 V A M B U L L E T I N

27 ROMIL launches fully SI-traceable CRMs for trace element calibration V A M N E W S ROMIL Limited (Cambridge, UK) has been accredited to ISO 17025, as a Calibration Laboratory, for the production of elemental Certified Reference Materials (CRMs), enabling ROMIL to certify reference materials and to market ROMIL PrimAg CRMs bearing the UKAS (United Kingdom Accreditation Service) Logo. The ROMIL PrimAg CRMs provide analysts with a ready supply of fully SItraceable CRMs for trace element calibration. ROMIL PrimAg CRMs require no further analytical validation to demonstrate traceability to the SI and meet all the requirements of laboratories working to ISO ROMIL is extending and developing the concept to offer fully traceable analytical reagents for use in other applications. Further information: ROMIL Ltd The Source, Convent Drive Waterbeach, Cambridge CB5 9QT Tel: +44 (0) Fax: +44 (0) pure.chemistry@romil.com Web: MCERTS Performance standard for laboratories undertaking chemical testing of soil The Environment Agency has re-launched its MCERTS scheme for the chemical testing of soils with publication of Version 2 of the MCERTS performance standard. MCERTS provides assurance to all stakeholders (e.g. laboratories, Local Authorities, consultants, members of the public) of the reliability of data from such tests. Chemical test data on soils is used by the Agency to support its regulatory activities under a number of regimes, such as Part IIA of EPA 1990, Pollution, Prevention and Control (England and Wales Regulations) 2000 and Waste Management Licensing Regulations For the chemical testing of soil where results are to be submitted to the Agency for regulatory purposes, the Agency requires a laboratory to be accredited to the European and international standard, BS EN ISO/IEC 17025:2000. Accreditation is undertaken by an appropriate national organisation, which in the United Kingdom is the United Kingdom Accreditation Service (UKAS). The MCERTS performance standard for laboratories undertaking the chemical testing of soil provides an application of BS EN ISO/IEC 17025:2000 specifically for the chemical testing of soil and covers: the selection and validation of methods; sampling pre-treatment and preparation; the estimation of measurement uncertainty; participation in proficiency testing schemes; and the reporting of results and information. To allow laboratories to bring their soil testing methods up to the MCERTS standard, the Agency will be implementing the scheme over an 18-month period, starting from March The Agency is aware that it will take time for laboratories to gain approval through the appropriate accreditation process. To allow laboratories to bring their soil testing methods up to the MCERTS standard, the Agency will be implementing the scheme over an 18-month period, starting from March In this period laboratories reporting data to the Agency have as a minimum to be accredited to BS EN ISO/IEC 17025:2000 for the test methods, and test reports should include a brief method description and bias and precision estimates. From September 2004, only data from laboratories that have been accredited to BS EN ISO/IEC 17025:2000 for MCERTS will be accepted. For further information contact: Mike Healy Environment Agency Tel: mike.healy@environmentagency.gov.uk Copies of the standard can be obtained from the Agency s website: V A M B U L L E T I N

28 U K A N A L Y T I C A L P A R T N E R S H I P Science & Hazard Regulation claim or blame? Peter Frier & Kevin Thurlow LGC There is a special relationship between science, technology and regulation. Recent years have seen an increase in public awareness and concern for health and environmental issues. New technology and improvements in analytical techniques have increased our capability to detect potential risks, but technical uncertainties leave considerable leeway for interpretation of results. Scientific disputes often arise because different interest groups exploit these uncertainties to promote their point of view. Sometimes it is ill-informed scaremongering. Some years ago, it was announced that drinking instant coffee doubled the consumer s chance of getting a particular form of cancer. This dramatic headline was undermined somewhat when the text actually reported that the chance of getting that form of cancer increased from one chance in a million, to two chances in a million. Instant coffee drinkers, who read that far, breathed a sigh of relief. Of course, some people only read the headlines. Thus a climate of claim and counter claim can arise which muddies the interface between science and policy, with the effect that judgements made about regulatory policy (e.g. commands and controls) and science policy (e.g. prioritising research and surveillance) may be misdirected. The extent to which this occurs depends on other variables, such as public awareness/concern, media attention, legal and political pressure and scientific consensus. Scientific disputes often arise because different interest groups exploit uncertainties to promote their point of view. These factors are clearly evident when perceived hazards or risks are associated with a particular chemical or process. Science is then a tool to assess the potential harm, 2 8 V A M B U L L E T I N