Introduction

International and national programmes require the monitoring of levels of contamination in environmental compartments to preserve the natural habitat we live in and to support life and the functioning of society now and in the future. The European Union adopted a comprehensive system of control to protect the environment from pollution and to improve the quality of Europe’s nature as a whole. Under this system monitoring requirements and quality objectives are stipulated, obliging Member States to monitor and assess the quality of air, water, soil, etc., and to identify and reverse trends in environmental pollution. The data produced by monitoring programmes constitute the foundations for decision making. Because such decisions might drastically affect individuals, communities, enterprises, or society as a whole in different ways, the quality of such data must be ensured by subjecting the measurement systems used to vigorous quality-assurance and quality control (QA/QC) procedures. Such procedures must ensure that the analytical techniques used in monitoring programmes remain temporally and spatially stable. An integrated and internationally accepted set of measures, which must be deployed as a concerted effort, has been developed over the years for QA/QC of analytical data. ISO/IEC 17025 [1] describes, in essence, the requirements necessary for setting up and operating such a system. Factors contributing to the reliability and comparability of measurement results are summarised in Fig. 1. Reference materials (RMs) and, in particular, certified reference materials (CRMs) play a key role in this ensemble of tools for ensuring confidence in measurement data.

Fig. 1
figure 1

Aspects of quality-control or quality-assurance systems for testing laboratories

Full size image

This article will review basic concepts of matrix-matched CRMs, their production, selection, and use, with special emphasis on environmental measurements. The topic has been much more thoroughly treated in other monographs [4–7] and in more specialised review articles [7–13]. Pure elements and compounds, elements and compounds in solution, and gases and gas mixtures, although the basis of instrumental analysis via calibration, are deemed beyond the scope of this review.

Reference material-formal definitions

A variety of terms have been used in the literature to describe the nature of these materials: reference materials, certified reference materials, reference standards, quality-control materials, etc. Correct usage, and common misconceptions in the terminology of RMs, have recently been critically discussed [14].

The current version of the International Vocabulary of Basic and General Terms in Metrology (VIM) [15] defines RM as:

Material or substance one or more of whose property values are sufficiently homogenous and well established to be used for the calibration of an apparatus, the assessment of a measurement method, and for assigning values to materials and CRMs as:

Reference material, accompanied by a certificate, which has one or more properties whose value is certified by a procedure that establishes traceability to the accurate realisation of the unit in which the values of the properties are expressed, and for which each certified value is accompanied by a stated uncertainty with a given level of confidence.

Thus, the three distinctive main features of CRMs are:

  • certified property value

  • uncertainty statement

  • traceability statement

The VIM is currently under revision; in the meantime ISO REMCO (the Committee on Reference Materials of the International Organization for Standardization) has suggested modification of these definitions to make them more meaningful for end-users and to avoid confusion about the underlying concepts [16]. One objective of these modifications is to make it clear that the term “reference material” is generic and should be understood as sort of a “family name”, and that this family comprises CRMs and non-certified RMs. The most notable clarifications relate to the nature of the certified property, which could be quantitative or qualitative (e.g. identity of substance or species), and to normative references related to RM preparation and certification (e.g. ISO Guides 34 and 35). Another salient feature of the ISO REMCO definitions is the specific reference to the homogeneity and stability of (C)RMs.

Types of certified reference material

CRMs are a key element of analytical data quality assurance and are used for four main purposes:

  • instrument calibration;

  • method validation, in particular for assessment of the trueness of a method;

  • ensuring the traceability of measurement results; and

  • statistical quality control (charting).

Other main uses include verification of the performance of a measurement instrument and training and/or qualification of operators.

Pure substances certified for their purity and stoichiometry, or pure substances dissolved in a suitable solvent, are pivotal, because in many instances they are the anchor point for building up a metrological traceability chain. They are used for calibration of a measurement instrument and this traceable calibration is a principal component in achieving traceability in chemistry. Some analytical techniques (X-ray fluorescence, spark-source emission spectrometry, solid-sampling atomic-absorption spectrometry, etc.) cannot be calibrated by use of pure substances or solutions, because it is essential the analytes of interest are contained in a matrix as similar as possible to the real samples.

Traceability is a relatively new concept in analytical chemistry and requires analytical chemists to link measurement results to stated references through an unbroken chain of comparisons, all having stated uncertainties. CRMs are such stated references with known uncertainties; linking measurement results to CRMs, or to the quantity values embodied in the CRMs [17], ties them to a reference, and if the CRMs themselves are traceable, e.g. to the International System of Units (SI), the measurement results obtained would be traceable to the same reference system.

Because chemical measurement processes for real-world materials include several sample pretreatment steps before the actual measurement, certified matrix materials are necessary for assessing the validity of the whole analytical process. In particular, a matrix CRM will enable straightforward assessment of the trueness of the applied method; if no suitable CRM is available this pivotal characteristic can only be approximated by using alternative approaches, for example spiking and recovery experiments. Matrix CRMs should be similar in composition (level of analyte, matrix type, interferences), format (liquid, semi-solid, solid), and physical processing history (fresh, frozen, sterilised, dried) to the samples regularly analysed.

Matrix CRMs are by far the QA/QC tools most often requested by practitioners. Certified calibration materials (pure substances and solutions) are regarded as less important, presumably because they can be obtained as non-certified chemicals with relative ease from specialised supply houses. Because of the prominent role of certified calibration materials in establishing traceability through calibration, however, the awareness of analysts should be directed toward use of these materials and the benefit gained when they are properly applied.

Preparation and value assignment of certified reference materials

Because preparing and value-assigning a reference material is a sophisticated and resource-intensive process in terms of knowledge, equipment, and man power, a rather limited number of institutions, either public or private, offer such services. A non-exhaustive list of producers and/or distributors of (C)RMs is given in Table 1; this compilation is intended to inform the reader and not to order the quality of the services offered. A very informative tool is the COMAR data base (http://www.comar.bam.de), which is free of charge and lists thousands of CRMs produced world-wide by approximately 200 producers in 27 countries. Another specialised web-site dealing with different aspects of CRMs, including a searchable catalogue, can be found at http://www.virm.org.

Table 1 Producers and distributors of (certified) reference materials for environmental studies
Full size table

Preparation of CRMs requires the certifying body to implement a highly integrated process-planning and quality-assurance scheme to realise the physical production and value assignment of CRMs (Fig. 2). To make the whole process transparent and to inform producers and users of CRMs about organisational and metrological requirements of CRM preparation and value-assignment, ISO has issued several guidelines, in particular ISO Guides 34 [18] and 35 [19]. Third-party accreditation of CRM production on the basis of the combined application of ISO Guide 34 and ISO/IEC 17025 is currently seen by many CRM producers as the best means of demonstrating competence and of complying with the stringent CRM quality requirements expressed by ISO/IEC17025-accredited calibration and testing laboratories.

Fig. 2
figure 2

ISO Guide 34-compliant preparation and value-assignment scheme (simplified) for CRMs

Full size image

From the principal layout of a certification procedure as given in Fig. 2 it follows that the three main contributors to the uncertainty of the certified value, i.e. the homogeneity and the stability of the material and the measurements leading to characterisation of the property value, must be taken into account when stating the expanded uncertainty (U) of the certified property value in a way compliant with the Guide to the Expression of Uncertainty in Measurements (GUM) and ISO Guide 35 [19]:

$$U = k \times {\sqrt {u_{{\hom o}} + u_{{{\text{stab}}}} + u_{{{\text{char}}}} } }$$

where U is the expanded uncertainty, k the coverage factor (usually set to k=2, which approximates a 95% confidence level), u homo the uncertainty due to the inhomogeneity of the material, u stab the uncertainty due to the instability of the material, and u char the uncertainty in the characterisation of the property value.

The underlying principles, the organisational aspects, and statistical evaluation of homogeneity and stability studies have been summarised in several review articles [20–23]; the reader is referred to these for in-depth information. In this context it is important to realise that, contrary to previous CRM guidelines, the inhomogeneity and instability of a CRM must be quantified and an allowance for these contributions must be made in the uncertainty budget for the certified property value.

Characterisation measurements may be performed by one or more laboratories using a primary method, whose measurement results are, by definition, traceable to the SI [24], a method for which all possible sources of bias are known, or several independent methods [25]. How the results of such certification measurements are combined for assignment of the certified property value would lead beyond the scope of this review, but can be found in ISO Guide 35 [19] and other related publications [26–28].

Many environmental and biological CRMs may degrade over time, even when stored under favourable conditions (at low temperature, in the absence of irradiation and oxygen), necessitating periodic monitoring to check whether the certified values are still valid. If not, the certified property values and the associated uncertainties must be revised or the material withdrawn from the market.

Selection of certified reference materials

Ideally, CRMs should be as similar as possible to the sample being routinely analysed. In reality, the physical form of any RM is always a compromise between the “as close to reality as possible” sample requested by the end user and the “as stable and homogeneous as possible” sample required for certification with the lowest possible uncertainty. Many environmental candidate (C)RMs are therefore transformed into dry, finely ground powders using techniques such as ball-milling, jet-milling, cryo-grinding, etc., in combination with oven-drying or freeze-drying. Other reasons for choosing such technology are the low cost of storage and transport of powder samples and their excellent stability under extreme transport conditions in comparison to frozen wet materials. With the increasing development of RMs for clinical analysis, however, producers became aware that insufficient matrix matching with real samples might make a CRM quite useless because of lack of “commutability”, i.e. the inter-assay properties of the CRM are not comparable with those of routine samples. Although exact CRM matrix format-matching is an issue in clinical chemistry, measurements on traditional environmental matrices, for example soil, sediment, and sludge, are not affected or are only slightly affected by this format mismatch. So far only the US National Institute of Standards and Technology (NIST) has offered, for some time, cryo-ground, frozen materials for biota (SRM 1974a, Organics in Mussel Tissue) [29] and a freeze-dried version of the same starting material (SRM 2974); SRM 2974 is out of stock and will be replaced by a freeze-dried material. The comparatively high costs and complicated logistics of distributing frozen materials using dry-ice transportation severely limit the widespread use of this alternative technology.

Use of certified reference materials

Calibration of a measurement system by using properly prepared calibration materials is the essential means of establishing traceability for chemical measurements. To determine an unknown quantity the signal produced by the measurement system is referenced to the calibration materials used. In most practical situations, testing laboratories prepare their calibration materials by dissolving a pure substance (the analytes of interest) in a suitable solvent using calibrated equipment (balance, volumetric glassware, thermometer). Performing such operations is regarded as rather trivial by many analysts. Calibration is, therefore, not usually regarded as a substantial source of bias affecting the trueness of measurement results. Consolidated experience from the European Commission’s CRM programme-previously known as the BCR programme-and other interlaboratory comparisons organised by the JRC-IRMM has proven that agreement of results improved when the same calibration material was made available to the participants. Certified calibration solutions, a special form of CRMs, free the end-user from the burden of manipulating small amounts of (hazardous) substances and provide an easy means of building a traceability chain. The uncertainty associated with the certified concentration or mass fraction value contributes directly to the overall uncertainty of the measurement results. A very comprehensive guideline on calibration and the use of CRMs in setting up calibration functions can be found elsewhere [6].

The other main field of application of CRMs is in assessment of trueness as part of the validation of an analytical method, and related processes (verification of the correct use of a standardised method, qualification of operators, etc). Matrix CRMs, preferably matching as closely as possible the samples investigated (matrix type, analyte level, matrix format, etc.) are needed for this crucial step. Although most certificates of analysis that accompany CRMs describe in detail the properties of the materials (certified values, uncertainties, traceability, recommended minimum sample intake, storage conditions, expiry dates, etc.) information about how to make best use of the CRM for, for example, estimating trueness is often not directly available; the difficulty users encounter when evaluating and reporting results from estimation of trueness using CRMs have been emphasized by Jorhem [30, 31], who has also made several proposals about how to use CRMs correctly for this purpose [32]. ISO published a guideline for the use of CRMs (ISO Guide 33) [33]; it is currently under revision.

A straightforward means of assessing whether the average analytical results obtained for a CRM are biased has been published by van der Veen [34]. In essence, the measurement is regarded as unbiased if the difference between the result obtained and the certified property value is less than the expanded uncertainty of this difference:

$${\left| {x_{{{\text{obs}}}} - x_{{{\text{CRM}}}} } \right|} < k \times {\sqrt {u^{2}_{{{\text{obs}}}} + u^{2}_{{{\text{CRM}}}} } }$$

where x obs is the average result obtained from analysis of the CRM, x CRM the certified value of the property for the CRM, u obs the uncertainty of the result, u CRM the uncertainty of the certified property value, and k the coverage factor (usually 2, which corresponds to approximately 95% confidence).

A similar approach to trueness estimation is recommended by the European Reference Materials (ERM) initiative [35].

Another important aspect of every quality-assurance programme is demonstration that the temporal performance of validated methods remains under statistical control. Various charting techniques have been proposed for this purpose [36]. All of these techniques rely on the availability of a homogenous and stable test material. CRMs are ideally suited to this, because of their demonstrated stability and homogeneity. Use of CRMs for charting not only enables checking of whether the process is under control over time but also concomitant assessment of the trueness of the method.

Certified reference materials for environmental studies

The number of analyte/matrix combinations potentially needed by the environmental chemist is gargantuan, even though, at first glance, the number of “classical” environmental matrices seems to be rather limited: soil, sediment, water, air. From each compartment, however, an enormously wide range of samples differing in composition and physicochemical properties, e.g. clay, sandy, calcareous, humic-rich soils, etc, may be presented to the chemist for analysis.

CRM producers therefore try to address these challenges by offering “typical” matrices for analytes of global interest or concern. Examples of currently available matrix CRMs are given in Tables 2, 3, 4, 5, 6, 7, 8, 9, 10, 11 and 12; they are complemented by a sufficiently wide range of pure element/compound and pure element/compound in solution CRMs, which are primarily used for instrument calibration (data not shown)

Table 2 CRMs for inorganic analytes in soil
Full size table
Table 3 CRMs for inorganic analytes in sediment
Full size table
Table 4 CRMs for inorganic analytes in sewage sludge
Full size table
Table 5 CRMs for inorganic analytes in water
Full size table
Table 6 CRMs for inorganic analytes in fly ash and particulate materials
Full size table
Table 7 CRMs for inorganic analytes in biota
Full size table
Table 8 CRMs for organic analytes in soil
Full size table
Table 9 CRMs for organic analytes in sediment
Full size table
Table 10 CRMs for organic analytes in sewage sludge
Full size table
Table 11 CRMs for organic analytes in fly ash, particulate materials, and transformer oil
Full size table
Table 12 CRMs for organic analytes in biota
Full size table

For the analysis of elements a rather wide array of analyte/matrix combinations and analyte levels, covering all relevant compartments, is currently available as exemplified by the entries given in Tables 2, 3, 4, 5, 6 and 7. CRMs certified for content of different elements were among the first to be produced by certifying bodies. These started with the pioneering work of H.J.M. Bowen (Bowen’s Kale) in the 1960s and this was followed by activity of the US National Bureau of Standards who produced NBS Orchard Leaves. As a result of these long-standing efforts the analyst can choose a CRM which best suits his needs, in terms of analyte levels and matrix properties, from a substantial variety of soil, sediment, sewage sludge, water, particulate matter, and biota CRMs certified for a wide range of elements.

Because the eco-toxicology and mobility of metals in the environment are very dependent on their chemical forms (species) and their binding to particulate matter, several soil and sediment CRMs are offered for which, in addition to the total content of various elements, the extractable (leachable) part is also certified. One example is BCR-483, a sewage sludge-amended soil certified for acetic acid and EDTA-extractable metal content. Because these are method-dependent properties, the user is required to follow the same procedure as used for certification analysis. Reporting total concentrations of trace elements is no longer sufficient when elucidating the toxicity of environmental samples, because the toxicity of some elements depends on whether they are present as inorganic or organic species [38]. Metal speciation, therefore, has recently become increasingly important. To support quality-assurance programmes for speciation analysis CRMs are available for which not only the total amounts of elements such as As, Cr, Pb, and Sn are certified, but also individual organometallic species of eco-toxicological concern [39]. Several materials are available which are certified for their content of organotin (BCR-462, BCR-646, ERM-CE477, NIES-11, NIES-12, HIPA-1, PACS-2, SOPH-1), organomercury (BCR-463, ERM-CC580, ERM-CE464, DORM-2, LUTS-1, TORT-2, SRM-1566b, SRM-1974b, SRM-2976, SRM-2977), organolead (BCR-605), and organoarsenic (BCR-627, DORM-2) compounds.

The range of environmental CRMs for organic trace analysis is more limited and, usually, restricted to compounds regarded as persistent organic pollutants (POPs), for example PCBs, PAHs, PCDD/Fs, and organochlorine pesticides (OCPs), e.g. hexachlorobenzene, α-HCH, γ-HCH, heptachlor epoxide, oxychlordane, mirex, cis-chlordane, cis-nonachlor, trans-nonachlor, 2,4′-DDE, 4,4′-DDE, 2,4′-DDD, 4,4′-DDD, 2,4′-DDT, and 4,4′-DDT, which are of global concern and for which worldwide, supra-national monitoring programmes exist (e.g. UN Environment Programme, Arctic Monitoring and Assessment Programme, OSPAR Commission, etc). Several suppliers offer CRMs for POPs in different matrices (soil, sediment, sewage sludge, particulate matter, biota), but also for method defined properties such as AOX in soil (ERM-CC010, ERM CC-011, ERM-CC012) and total petrol hydrocarbons in sediment (ERM-CC015) (Tables 8, 9, 10, 11 and 12).

CRMs for organic compounds in water do not exist, mainly because of stability problems and the large sample volumes (mostly 1–2 L) necessary, which complicates CRM distribution. Alternative concepts for production of water-based CRMs for organic trace analysis, based on “freeze-dried water”, stabilization of pesticides in a starch-rich matrix (pills which must be reconstituted), and spiking and matrix surrogate solutions, have been developed and are discussed elsewhere [40–43], although none of these alternatives has been commercialised.

CRMs for the vast group of so-called “emerging pollutants” are currently scarce, although the objective of several development programmes is to remedy this. For many of these emerging contaminants there are no national or international limits, because of incomplete or insufficient information about occurrence, risk assessment, or eco-toxicology. These man-made substances are not necessarily new-some have often been in use for a long time-but their presence and impact on the environment are only now being recognised. NIST has recently quantified some polybrominated diphenyl ether congeners (PBDEs) and toxaphenes, which are stable in the environment and readily bioaccumulate in wildlife and human tissue, in three marine SRMs (SRM 1588b Organics in Cod Liver Oil, SRM 1945 Organics in Whale Blubber, SRM 1946 Lake Superior Fish Tissue). Although certified values have not yet been assigned, the materials are certainly fit-for-purpose as quality-control materials. A European Commission-funded research project explored the feasibility of preparing CRMs for PBDEs in flounder and sediment and technology for providing sufficiently homogenous and stable materials containing these analytes has been developed (BROC project, J. de Boer, personal communication). The objective of another EU-funded project currently in progress (NORMAN, www.emerging-pollutants.net) is to establish a network of reference laboratories for monitoring emerging environmental pollutants; QA/QC tools are expected from the project.

The EU Water Framework Directive [44] requires monitoring of some of the emerging pollutants which have been grouped together on the list of Priority Substances [45]. Table 13 gives a non-exhaustive overview of relevant classes of emerging pollutants in the environment [46–48]. Because wastewater-treatment plants are major point sources for discharge of these substances, sewage sludge or sludge-amended soil might be a suitable matrix for CRM production.

Table 13 Emerging pollutants in the environment (modified after [46, 47])
Full size table

Conclusions

Chemical measurements play an ever increasing role in many economic, political, environmental, medical, and legal decisions; reliable and comparable data form an integral part of the decision-making process. CRMs are undoubtedly a key element in analytical data-quality assurance, because these tools enable building of a traceability chain to established and acknowledged reference systems and enable comprehensive assessment of the performance of a measurement method. The variety of matrices and substances encountered in environmental chemistry necessitates provision of a wide range of matrix (C)RMs by specialised RM producers. A perfect match (matrix composition and format, analyte level, etc.) between a particular CRM and a natural sample is difficult to achieve, however, and a compromise is usually necessary. A sufficiently large supply of CRMs is available for environmental studies, at least for most elements and “classical” organic pollutants, for example PCBs, PAHs, etc. Regular use of matrix CRMs by laboratories is seen as a key aspect of setting up internal quality-control and assurance programmes, because they enable the performance of methods applied by a particular operator or laboratory to be checked, irrespective of time constraints or the availability of collaborative study efforts. Consequently, regular use of CRMs has been made mandatory for laboratories accredited in accordance with ISO/IEC 17025, to ensure the comparability and reliability of chemical measurements.