2 REFERENCE MATERIALS Anal. Proc. Reference Materials The following are summaries of six of the papers presented at a Meeting of the Analytical Division held on February 4th, 1981, at the Scientific Societies’ Lecture Theatre, Savile Row, London, W.1. Some Recent Work on Reference Materials at the National Physical Laboratory J. D. Cox Division of Chemical Standards, National Physical Laboratory, Teddington, Middlesex, T W11 OL W In this account of recent work on Certified Reference Materials (CRMs) at the NPL, the charac- terisation of chemical substances and materials by accurate measurement, both chemical and physical, is emphasised. Work on CRMs at Teddington began over 25 years ago as a result of a research programme on accurate measurement of the thermal properties of organic compounds.A prerequisite of that measurement programme was a stock of highly purified substances, which we prepared in-house and certified for purity.l It became our practice to offer for sale samples drawn from the same purified batches as were used for property measurement. The second phase of CRM activities began 12 years ago in response to a proposal from the Ministry of Agri- culture, Fisheries and Food that we should utilise our expertise in purification, for the preparation of pesticide chemicals as reference substances of certified high purity. The emphasis was initially on the analysis of commercial pesticides and pesticide formulations but interest was soon shown by environmental chemists. Pure samples of contaminants and metabolites were later included in the programme.By 1973 organic reference materials were being issued to a wide variety of customers, who made us aware of other unfilled needs for CRMs of various types. A decision was then made to launch a research programme into the provision of CRMs required by industry, with the Metrology and Standards Requirements Board of the Department of Industry as the major source of funds. This third phase of CRM activities has run from 1973 to the present but it will shortly give way to a fourth phase, when the resources available will be severely reduced as part of the: Government’s economy measures. From April 1981, most of the effort available will be used in sustaining NPL’s present range of CRMs, so that expansion of the range will be much slower than in the recent past.Some examples of groups of CRMs which have recently been developed by research at the NPL and are now, or will shortly be, available for distribution outside NPL, are now given. The tariff is given in a catalogue2 available free from NPL. In all of these examples CRMs are offered along with an NPL certificate, which thereby provides the user with traceability to national standards of measurement. In the field of chemical analysis the range of pesticide CRMs has been extended to 90 substances, and CRMs are offered for other areas of analysis where pure specimens are needed for reference purposes. For example, 16 metal - organic substances, which are oil-soluble, are offered for use in determining the metal content of lube oils by atomic-absorption or emission spectroscopy; these CRMs are certified for their metal contents.In the field of powder technology, the NPL offers eight particulate materials that are certified for specific surface area, by nitrogen adsorption measurements. Four of theseJanuary, 1982 REFERENCE MATERIALS 3 materials were certified by a collaborative campaign organised by the Society of Chemical I n d ~ s t r y , ~ and a set of four grades of a-alumina was characterised at NPL alone. Our general objective in certification is to make the measurements traceable, which in the context of specific surface area implies direct measurement in terms of square metres and grams. It can be argued that the existing certification in terms of nitrogen adsorption, involving a particular physical model of the adsorption process and an assumed value for the cross- sectional area of the nitrogen molecule, is not as directly traceable to national standards of measurement as would be desirable.Therefore, research into improved traceability is under way. In the case of one material, Sterling FT graphitised carbon black, the particles are sufficiently round, crack-free and homogeneous in size for a group of particles to be photographed in an electron microscope and subsequently counted in size-categories. The geometrically calculated mean surface area was found to agree with that derived from nitrogen adsorption measurements. It should be mentioned that because individual particles in a batch of powder are likely to differ from one another in their various physical properties, it is important to draw a representative sub-sample from a CRM, which in turn must constitute a representative sample of the whole CRM batch.For this reason NPL employs a rotating riffle machine to sub-divide a CRM batch into representative portions for sale. A recent development in the NPL programme concerns the provision of RMs for the calibration and checking of apparatus for measurement of low vapour pressures. It is becoming necessary for the physicochemical properties of new commercial chemicals to be measured as part of an assessment of their environmental impact. Vapour pressure is one of the parameters to be measured, which is not easy in the pressure range to lo2 Pa. Two solid CRMs certified for vapour pressure, naphthalene4 and hexamethylbenzene,5 are available, and two liquid substances, dibutyl phthalate and butyl 2,4-dichlorophenoxyacetate, have recently been evaluated for use at ambient temperatures (where the vapour pressure of the liquids is Pa).The method employed for measurements on the liquids was a novel one,6 and it has yet to be established whether the results show any method dependency. The final development area considered concerns CRMs certified for thermal properties ; here the nature and quality of the certification depend strongly on the likely end-use of the CRM. For example, for check measurements of the performance of melting-point apparatus, manual or automatic, a set of CRMs is offered, covering the temperature range 51-285 "C; these are certified for meniscus temperature and liquefaction temperature, for a stated rate of heating.For calibration of calorimeters, including differential-scanning calorimeters, we offer CRMs certified for enthalpy of fusion7 and final melting-point temperature at equi- librium, measured to 0.01 "C. For more precise thermometric measurements (to 0.001 "C) we offer triple-point cells containing water, benzene and phen0xybenzene.s A very recent development, not yet completed, concerns triple-point cells for accurate realisation of a temperature near 36.4 "C. Such cells are expected to have an important role in temperature measurement in enzymology and elsewhere in the biomedical field. The examples given above show that the NPL has been responsive to new, and often specialised, needs for calibrants and check substances.Suggestions regarding other unfilled needs for CRMs are always welcomed. 1. 2. 3. 4. 5. 6. 7. 8. References Cox, J. D., Proc. SOC. Anal. Chem., 1966, 3, 61. "Certified Reference Materials and Transfer Standards. A Catalogue of Items Available from the National Physical Laboratory, United Kingdom, and from the Community Bureau of Reference of the Commission of the. European Communities," Department of Industry, National Physical Laboratory, Teddington, 1979. Everett, D. H., Parfitt, G. D., Sing, K. S. W., and Wilson, R., J . APPZ. Chem. Biotechnol., 1974, 24, 199. Ambrose, D., Lawrenson, I. J., and Sprake, C. H. S., J . Chem. Thermodyn., 1975, 7 , 1173. Ambrose, D., Lawrenson, I. J., and Sprake, C. H. S . , J . Chem. Thermodyn., 1976, 8, 503.Hales, J. L., Cogman, R. C., and Frith, W. J., J . Chem. Thermodyn., 1981, in the press. Andon, R. J. L., and Connett, J. E., Thermochim. Acta, 1980, 42, 241. Cox, J. D., and Vaughan, M. F., MetroZogia, 1980, 16, 105.4 REFERENCE MATERIALS Anal. Proc. The Activities of the Community Bureau of Reference H. Marchandise Commissio.Pz of the Eurofiean Communities, Community Bureau of Reference, Rue de la Loi 200, B-1049 Brussels, Belgium The Community Bureau of Reference (BCR), which is a department of the General Directorate for Research, Science and Education of the Commission of the European Communities, is responsible for a programme on Applied Metrology and Reference Materials. The reason for this activity is that the Community has specific needs for measurements and in particular for measurements which should be made everywhere with the same level of accuracy.The need for the compatibility of the measurements is obvious from the point of view of trade. Whereas the uniformity of the basic units is ensured at the level of the metrology laboratories through Bureau International des Poids et Mesures (BIPM), there are many other measurements which present difficulties. Disagreements in the results of measure- ments could be the origin of trade barriers and it is precisely the objective of the Community to avoid trade barriers and to abolish them where they exist. The work on metrology started only at the end of 1979. It consists essentially in inter- comparisons of various kinds of measurements and in improvement of transfer standards.It has already proved very successful in establishing useful collaborations. The work on reference materials (RMs) started earlier (1973). The first reference materials were certified in 1977; at present 40 RMs are available and this number will increase to 70 by the end of 1981. To illustrate the importance for the Community, one could consider the case of a directive on pollution control that specifies maximum levels of carcinogenic compounds and of toxic elements (lead, cadmium, mercury). When the directive is adopted those who have to implement it would like to be sure that it is enforced in the same manner everywhere, that the measure- ments are made with equal care and that they are comparable. Otherwise, when industry is involved there could be distortions of competition.For use in environmental analyses we have prepared pure polycyclic aromatic hydrocarbons (carcinogenic) for the calibration of chromatography, and we are on the way to certifying toxic elements in plants, milk powder, blood, meat, liver, kidney, fly ash and sludges. We also hope to prepare reference materials for water analysis. We have in progress a project on reference gas mixtures, and we have RMs for flash point determination that can be useful in the context of the regulations on dangerous chemicals. On industrial products we have prepared or are preparing RMs for fertilisers, crystal glass, coke, etc. Of course, we welcome any request from industry that would point out specific needs to us. In this context it is perhaps of interest to mention that there is in our programme a group of highly skilled laboratories for analyses of non-metallic elements in metals, and as a result of the work done with their collaboration we have a series of metals (zirconium, titanium, aluminium, lead, copper, molybdenum and nickel) which are certified for oxygen content and sometimes for nitrogen content.We have also developed a programme on the properties of powders (particle size distribution, surface area, pore size). Finally, we have started working on reference materials for biomedical analyses. This field is in great need of reference materials on a large scale. All the work on analyses and measurements is carried out by laboratories of the member countries and, for some projects, with contributions from the laboratories of the Joint Research Centre of the Community.For the sake of brevity, details of the selection and preparation of the materials, the study of stability and homogeneity are omitted. The measurements are made by the best laboratories we can find in the Community. When the results are available, they are discussed in detail. Some of them are eventually rejected when reasons for inaccuracy have been identified as a result of the discussion of the methods. When the work is terminated, the draft report is submitted to the Certification Group, composed of experts from the member countries who have responsibilities in their respective countries for metrology or for reference materials. When all steps of the work have been found to be satisfactory the material is finally certified.Samples are available to anyone who requests them and are delivered with a certificate and a complete report. The objective is in principle the same as for metrology.January, 1982 REFERENCE MATERIALS 5 The final batch of reference materials is only one result of our work. The other important consequence is the benefit of the collaboration for all participants and the result of it is better compatibility and general improvement of the quality of the measurements. Lists of BCR Reference Materials certified before 1980 and in preparation are available from the author on request. NBS: Current Work and Future Plans in Reference Materials Stanley D. Rasberry Ofice of Standard Reference iwaterials, National Bureau of Standards, Washington, D.C. 20234, USA Standard Reference Materials (SRMs) have been produced, certified and issued by the NBS since 1906.These materials, with specific chemical or physical properties certified by the NBS, have found wide acceptance especially by industrial users needing to maintain or increase high levels of productivity. Today, 84 of the 100 largest manufacturers in the USA buy and use SRMs. Together with 10000 other users, they buy over 40000 SRM units per year at a total price of about $3.0 x lo6. The SRMs have a vital role in promoting industrial product- ivity by helping companies maintain : reliability and uniformity of materials in meeting specifications ; quality control in raw materials and produced goods or services ; and inter- changeability of materials and sub-components. In addition to the industrial customers, major SRM users also include federal and state governments, universities and non-profit research establishments, particularly for SRMs used in the areas of health, environmental protection, metrology and forensic science.Inter- national distribution of SRMs accounts for about 20% of the total. The last 14 years has seen the NBS-SRM programme greatly expand the nature and scope of its activities. In 1966, the NBS offered 559 different SRMs for sale, which were intended to serve the measurement needs of rather limited but important segments of the US economy. At that time, the majority of SRMs consisted of compositional standards (primarily certified for bulk constituents or purity), designed for use in quality control systems during mining and manufacturing of raw and processed industrial materials. The SRM inventory primarily consisted of metal bearing ores, metals and alloys, cement, glass and inorganic chemicals.By 1980, the SRM programme was considerably strengthened through the participation of 23 NBS technical divisions. This enabled the NBS to increase the number of SRMs to about 1000. Now included are SRMs in many new areas such as computer technology, fire safety, forensic science and radiopharmaceuticals, with particular recent emphasis on the two important areas of environmental and clinical measurements. The expansion of the SRM programme was accompanied by a corresponding expansion of the customer base as indicated by SRM sales.In the period 1969-80, SRM sales more than doubled to $3.0 x lo6. The number of units sold each year also increased steadily over this period, from 29600 in 1971 to approximately 40800 in 1980. Several con- siderations have helped shape these plans. From our perspective the four most important considerations or, as we call them, “external trends,’’ affecting the future of SRM production are outlined below. Table I summarises the types of SRMs now certified by the NBS. The scope of our future plans for SRM production is indicated in Table 11. 1. Increased Importance of Measurement Compatibility in a Technological World During the next decade, international SRM activities are expected to increase dramatically, including the joint development of multi-national certified reference materials and increased requirements for the use of reference materials to meet international standards.Great care must be taken to ensure that international standards, which require the use of reference materials, foster international trade rather than serve as a trade barrier. 2. Increased Cost of Raw/Processed Materials and Energy Recent increases in the cost of raw and processed materials, as well as the cost of energy, have resulted in the development of more stringent procurement specifications for materials6 REFERENCE MATERIALS TABLE I SRM TYPES CERTIFIED BY NBS Minerals Refractories Carbides Glasses Cements Trace elements Nuclear materials X-ray diffraction Isotopics Ion activity Mechanical and metrology Superconducting Freezing-points Melting-points Calorimetric Vapour pressure :a1 Radioactivity Anal.PYOC. Steels Steelmaking alloys Cast irons Cast steels Non-ferrous alloys Gases in metals High-purity metals Electron probe microanalytic Primary chemicals Clinicals Biologicals Botanicals Environmentals Industrial hygiene Metallo-organic compounds Fertilisers Ores Thermal conductivity Thermal expansion Thermal resistance Thermocouple materials Magnetic Optical Gas transmission Permittivity Reference fuels Resistivity Rubber materials Computer tapes Sizing standards Colour Photographic Surface flammability Smoke density sold in national and international commerce. SRMs are being increasingly used to provide a basis for the arbitration of disputes between producers and users of materials, as well as in the traditional role as part of production quality control systems.3. Increased Development of New Materials in High-performance Applications The development and use of materials for high-performance applications has increased considerably in recent years. This trend is leading to the development of a wide variety of new high-performance materials serving many diverse uses. Examples range from high- temperature, corrosion-resistant alloys used in aircraft turbine blades to plastic foams used in building construction. Many of these high-performance materials are used in critical applica- tions where their failure in service could result in serious safety hazards. This has led to the development of rigidly enforced trade specifications and to the need for reliable production quality control, which in turn has resulted in increased demand for SRMs.4. Increased Development and Implementation of Government Imposed Health and Safety Regulations The trend to greater federal, state and local government regulation of the environment, health care system and transportation systems is also resulting in intense demand for new SRMs. Effective development and enforcement of regulations depend on the availability of TABLE I1 PROJECTION OF DEMAND FOR SRMs FROM NBS IN 1982-86 SRM category Metals . . .. .. Chemicals/rubber/plas tic Nuclear . . .. .. Non-metalslglass . . .. Radioactivity . . .. Engineering . . .. Environmental gases . . Environmental liquidslsolids Health . . .. .. Science/metrology . . SRM Current renewals? inventory* 1982-86 ..316 53 .. 74 15 . . 105 23 .. 30 10 .. 156 71 .. 146 146 .. 54 145 .. 35 21 .. 42 33 . . 105 20 Current SRMs needed discontinued NewSRMs t o b e 1982-86 1982-86 80 108 30 10 24 32 20 0 20 38 46 58 25 10 30 6 25 6 10 48 Inventory needed 1986 288 94 97 50 138 134 69 59 61 67 Total . . .. .. .. 1063 537 310 316 1057 * As of April 30th, 1980. 7 Includes a number of multiple renewals in the Radioactivity, Engineering, Environmental gases and Health categories.January, I982 REFERENCE MATERIALS 7 accurate, reliable and compatible measurement systems. The NBS response to this trend has been the establishment of a number of programmes aimed in part at providing a reliable measurement base for regulatory agencies. These include the Environmental Measurements Programme, Nondestructive Evaluation Programme, Resource Recovery Programme, Recycled Oil Programme and the Nuclear Safeguards Programme.A programme for Measure- ments and Standards for Nuclear Waste Management has also been recently proposed. All of these programmes will require the development of new SRMs to serve as measurement tecli- nology transfer mechanisms. Geological Reference Materials with Particular Emphasis on Multi-element Trace Analysis Alan R. Date Institute of Geological Sciences, 64-78 Gray’s I n n Road, London, WClX 8NG Geochemical analysis, in common with many other branches of analytical chemistry, suffers from an absence of true standards, i.e., samples with accurately fixed element concentrations. The significance of this failing was not recognised for many years, probably as a result of the traditional lack of liaison between geologist and chemist and the latter’s faith in the absolute methods of analysis used in classical schemes.Following the pioneering work of Larsen in 1938,l which demonstrated the poor agreement achieved by two to four “better” analysts for six amphibole (rock-forming mineral) samples, a collaborative programme of analysis was initiated in the late 1940s by the Geology Depart- ment of the Massachusetts Institute of Technology, the Geophysical Laboratory and the United States Geological Survey. This project, intended initially to demonstrate that spectrographic analysis could match the precision and accuracy attained by wet chemical methods, involved major element determinations for the first two “standard” rocks, the granite G-1 and the diabase (dolerite) W-1. The results submitted by 34 chemists from 25 laboratories in 10 countries were published in 1951,2 and proved to have much more far reaching influence.The wide variation shown for most major elements was very disturbing, the data for Na20 and K20 being particularly instructive. Soda was found to vary from 2.61 to 3.95% in G-1 and from 1.61 to 2.62% in W-1, and potash was found to vary from 3.85 to 6.88% in G-1 and from 0.41 to 1.30% in W-1. Major element data are normally quoted as oxides to provide summation to 100% for the commonly reported components, and because one scheme of silicate rock classification depends on the relative proportions of normative minerals (a theoretical combination determined from chemical data), in which soda and potash play a crucial role. The ranges shown for G-1 encompassed almost all the world’s known granite variation, although several distinct groups may be identified under the microscope. Soda and potash may be particularly difficult major element determinations, but the position with trace elements is expected, by definition, to be far worse. The term trace in this context includes any element normally reported in the parts per million range or below.The period following publication of the first report on G-1 and W-1 may be considered in two parts. Although a few trace element results were included in this first report in 1951, during the succeeding 15-year period the number of determinations proliferated, and several additional reports were produced.In the process, a few papers drew attention to the possibility of between-bottle variation for some trace elements in G-l,3-5 and in 1967 Kleeman6 concluded that both rocks as originally prepared (G-1, -80 mesh, 180 pm; W-1, -100 mesh, 150 pm) were too coarse to serve as reference standards. In parallel with the work on G-1 and W-1, other reference materials became available, notably the syenite Sy-1 (in conjunction with a sulphide ore, Su-1) from Canada,’ and a series of silicate rocks from France.* By 1962 the supply of G-1 was almost exhausted and a suitable replacement was prepared. By the end of 1966, Flanagan and Gwyn could record more than 50 geological reference materials from 12 sources, although many were of very limited application. Their paper,g published in Geoclzimica et Cosmochirnica Acta in 1967, was followed by an Editorial Notice announcing The granite G-2 and five other silicate rocks were issued in 1964.Anal.Proc. future limits on the publication of reference material papers. In a very broad sense it may be said that the first 15-year period was characterised by unexpurgated data compilations, with statistical analysis taking account of all reported results. The initial impetus provided by the demands of igneous petrology, which resulted in a preponderance of silicate-based reference materials, was continued into the second 15-year period, with the issue of further reference materials from the USA, Canada and France, and new reference materials from South Africa, Japan, Scandinavia and the Eastern Bloc countries.Reference to the range of silicate rocks currently available may be obtained from a recent review paper by Abbey.lo It is interesting to note at this point that in the first compilation of datall for the six USGS rocks issued in 1964, the variation for K20 in G-2 was found to be worse than that for G-1 almost 20 years earlier. As a result of the difficulty of dealing with such data sets, the tendency during the second 15-year period has been towards the “select laboratories” concept as a means of limiting the initial spread of results subjected to statistical analysis. The increasing importance of geochemical prospecting during this period led to the intro- duction of six geochemical exploration reference materials from the United States Geo- logical Survey12 and four soils from the Canadian Certified Reference Materials Project .l3 Stream sediment and lake sediment reference materials are promised (CCRMP) .The range of sample types widened further to include reference materials of a more specialised nature, such as the USGS manganese n0du1es.l~ The number of ore standard reference materials increased dramatically, with contributions from several sources including the Institute of Geological Sciences.15 There is still very poor coverage for marine sediments and for sedi- mentary rocks in general, although the field is represented to a certain extent by standard reference materials from the National Bureau of Standards in the United States and the Bureau of Analysed Samples in the UK.The lead taken by Geochimica et Cosmochimica Acta in 1967 in limiting the number of reference material papers accepted for publication resulted in the formation of Geostandards Newsletter in 1977,16 and this journal should be consulted in any search for geological reference materials. 8 REFERENCE MATERIALS Multi-element Trace Analysis The first compilation of data for the USGS geochemical exploration reference materials (GXRs)12 serves as an illustration of the difficulty in standardising geological materials for trace elements. Although atomic-absorption spectrometry would be expected to provide reliable data for copper, the range of values reported for GXR-4 by this technique is 5- 8800 pg g-l. This also illustrates the danger in defining a method of analysis by the technique of measurement only; such anomalies owe more to sample attack.Although the user of such reference materials must be aware of the problems associated with data compila- tions, he may usually rely on the work of others in selecting a series of suitable reference materials for a particular project. In the first compilation of data for the six USGS rocks issued in 1964,11 results for the ten most frequently determined trace elements (Ba, Co, Cr, Cu, Ni, Pb, Rb, Sr, Zn, Zr) were obtained by six techniques, including spectrophotometry. Although atomic-absorption spectrometry and isotope dilution mass spectrometry may be developed for simultaneous determination of several elements, only atomic-emission spectrometry, X-ray fluorescence spectrometry and neutron-activation analysis can be considered true multi-element methods, offering simultaneous determination of more than ten elements.In a review paper in 1977, RubeSkal’ used this approach to identify a trend away from traditional methods of analysis (including atomic-emission spectrometry) towards X-ray fluorescence spectrometry, neutron- activation analysis and isotope dilution mass spectrometry. The growth of inductively coupled plasma techniques is expected to reverse this trend, and the advent of laser ablation allied with plasma emission spectrometry or plasma source mass spectrometry will create demand for a wider range of reliable solid reference materials for geochemical analysis. Reference Material Selection The author’s recent interest in geological reference materials is related to the multi-element analysis of Scottish stream sediment samples submitted under the Institute’s RegionalJanuary, 1982 REFERENCE MATERIALS 9 Geochemical Reconnaissance Programme.The samples cover a composition range similar to silicate rocks, with extremes for iron and manganese. Up to 30 elements are determined by d.c. arc direct-reading emission spectrometry. International geological reference materials are too valuable to be used for calibration purposes, but are used to monitor long-term precision and accuracy. In selecting a series of reference materials to cover a suitable range of concentration for all 30 elements, compilations of data such as the series published by Abbey have been consulted.The “1979” edition of Abbeylo may be used to illustrate potential problems in selection. Abbey lists reported values for major, minor and trace elements in all reference materials that may be applied to silicate analysis. The data he considers reliable (using various criteria, including the “select laboratories” concept) are shown unqualified. Other reported values are shown with question marks. The ratio of the first group to the second is the degree of acceptability. The selection of elements with acceptability in excess of 50% includes eight of the ten most frequently reported trace elements. Of these, zinc and chromium have similar determination ranges by direct reader, and matching reported ranges of concentration. A selection of the more readily available international reference materials provides good cover for zinc, while many have chromium below our detection limit (10 pg g--.l) or above the required top standard (2000 pg g-l).Reference materials with intermediate concentrations of chromium, the Canadian MRG-1 (450 pg g-l) and the French BR (38Opgg-l) are of great use. The second group of elements, with acceptability ratios below SO%, includes tin with a reported range of concentration from 1.4 to 1900 pgg-1. Twenty-five reference materials, from a total of thirty, have concentrations below our detection limit (10 pg g-l), and only two of the remaining five have reliable values at 11 and 1900 pg g-1. A similar exercise for rare earth data, and for environmentally important elements normally reported in the ng g-l range, shows that the recent upsurge in interest in these elements has not been matched by an increase in the number of reliable geological reference materials.Synthetic Reference Materials In the first report on G-1 and W-1 it was suggested that analytical error could be separately identified from sampling error by the use of a synthetic rock standard. The author is unaware that such a standard was ever prepared. The nearest approximation to such a scheme is the series of glasses prepared by Corning Glass for private use by the US Geological Survey, and reported in 1976.lS Glass reference materials are available from the US National Bureau of Standards and from ANRT in France (VS-N).IS Each contains a very wide range of trace elements in nominally identical concentrations, and in the case of VS-N the concentrations are unusually high.In an attempt to avoid calibration standards prepared by solid dilution, the author developed a method for the preparation of synthetic silicates in which the element concentra- tions could be changed at will. This method, a development of the technique used to manufacture starting materials for experimental petrology, was described in detail in The AnaZyst in 197tX2O It involves the preparation of two solutions, the first a solution of tetra- ethyl orthosilicate in ethanol, to render it miscible with the second solution, which is an aqueous phase containing the remaining major, minor and trace elements. The two solutions are mixed, ammonia is added and a “flash hydrolysis” occurs, producing a solid “gel” with the trace elements entrained in a silicate framework.The gel is dried, ignited a t a tempera- ture sufficiently high to cause slight sintering, which removes its hygroscopic properties, and ground lightly to reduce it to a convenient powder form ((300 mesh, 53 pm). The range of major element reference materials currently used in this work is shown in Table I. Only two major element reference materials from an earlier series have been independently analysed.20 The application of the current series as calibration standards in d.c. arc direct-reading emission spectrometry is illustrated in Fig. 1, a calibration graph showing magnesium channel count against concentration. There is good agreement between synthetic reference materials and international standard rocks.The trace element reference materials have been prepared in a limited number of matrices. The series currently used is shown in Table 11. With the exception of molybdenum, the They have very limited use in geochemical analysis.10 REFERENCE MATERIALS TABLE I MAJOR ELEMENT “GEL” SILICATE REFERENCE MATERIALS Major element data in yo. Anal. PYOC. Reference material Element 50, .. .. M203 . . .. TiO, .. . . Fe,O, . . .. JIgO . . .. CaO .. .. li,O . . .. Sa,O . . .. M n . . .. DR ES FR 70.0 40.0 30.0 18.0 14.4 5.72 2.0 1.0 0.5 2.0 1.0 5.0 2.0 20.0 45.0 2.0 20.0 10.0 2.0 - 0.2 2.0 2.0 2.0 - 1.0 1.0 GR 60.0 1.83 1 .o 20.0 1 .o 1 .o 10.0 2.0 2.0 HS IR 65.0 65.0 16.0 14.0 4.0 2.0 2.0 5.0 0.5 5.0 10.0 5.0 0.5 2.0 2.0 2.0 JR 45.0 29.7 0.5 0.5 1.0 20.0 0.5 2.0 0.5 KR 70.0 9.07 0.1 1 .o 0.5 0.5 1 .o 2.0 10.0 LR MR’ 50.0 70.0 32.8 3.38 0.2 0.2 10.0 0.5 2.0 10.0 2.0 1 .o 1.0 5.0 2.0 2.0 5.0 - trace elements appear to be present at the intended concentrations.Several trace element reference materials have been analysed by independent methods. Data for the current involatile element series (with high levels of zinc) are shown in Table 111. The author is confident that the variation shown is analytical rather than compositional in nature. The use of these reference materials as calibration standards is illustrated in Fig. 2, a graph showing lithium channel count against concentration. Here too there is good agreement between synthetic and natural reference materials. This method of reference material preparation is limited in several respects: (a) some elements, e.g., Be and B, are unstable under the silicate hydrolysis procedure, and have to be added by liquid - solid dilution; ( b ) some elements, e.g., Sn and V, are dissolved with the aid of standardised NaOH solution, which assumes the presence of sodium in the matrix; and (c) the technique is currently limited to small amounts ((5OOg).For silicate analysis, however, it allows one to prepare standard reference materials to order. TABLE I1 TRACE ELEMENT “GEL” SILICATE REFERENCE MATERIALS Trace element data in pg g-l. Element Li .. .. JSe . . .. I3 . . .. v . . .. Cr .. .. 3tn . . .. co . . .. Xi .. .. c u . . .. Zn .. .. Ga . . .. Ge . . .. Rb . . . - Sr . . .. Y .. .. Zr .... No* . . .. Sn . . .. Ha . . .. La . . .. Pb .. .. Bi .. .. +g .. .. f DRlO DR09 DR08 2 50 5 5 2 20 - 50 25 - 50 - 200 - 20 - 50 - 50 - - - - - - - 20 10 2 2 loo - 2: 5 20 - 50 100 - 20 - - 500 - 33 12 5 2 20 - 50 200 - 20 - 20 200 50 20 - 50 - - - - - - - DR07 100 - - 50 100 500 50 100 200 - - - 200 50 1000 - - 500 50 500 - Reference material DR06 DR05 10 200 10 - 100 - - 100 - 200 - 1000 - 100 - 200 100 - 50 500 10 10 A - - 100 I - 500 - 100 - 2 000 81 10 - 100 - - 1000 - 100 100 2000 100 - - DR04 20 20 200 - - 200 100 20 20 20@ - - 176 20 200 - 200 200 DR03 500 - - 250 500 2 000 200 500 1000 - - - 1000 200 5 000 - - 2 000 200 5 000 DR02 50 50 500 - - 500 200 50 50 500 - - 44 1 50 500 -- - 500 500 -7 DRO 1 1000 - - 500 1000 5 000 500 1000 2 000 - - - 2 000 500 10 000 - - 5 000 500 10000 - * Determined by solution absorptiometry (found to be in error).January, 1982 REFERENCE MATERIALS TABLE I11 ANALYSIS OF DR MATRIX TRACE ELEMENT REFERENCE MATERIALS BY THREE TECHNIQUES Techniques used : a, atomic-absorption spectrometry (P.T. S. Sandon, Geochemical Division, IGS) ; b, atomic-emission spectrography (B. A. R. Tait, Geochemical Division, IGS) ; and c, instrumental neutron-activation analysis (J. Herrington, AWRE, Aldermaston). 500 f 8 2 200 - C lu 11 - - DR09 DR07 DROB DR03 DROl (-'-, r-----. -7 & - Expected, Found, Expected, Found, Expected, Found, Expected, Found, Expected, Found, Element p.p.m. p.p.m. p.p.m. p.p.m. p.p.m. p.p.m. p.p.m. p.p.m. p.p.m. p.p.m. co .. .. 20 158 50 48a 100 1 0 s 200 200% 500 5408 2 i c 5SC ll0C 220c 510C 15b 44b lOOb 220b 433a Cr .... 50 54c 100 l0OC 200 180C 500 460C I000 9ooc .53oa 1000 1070a 2000 2050a 5000 5200" >In . . .. 200 -b 200a 500 190b 600b 95Ob 2150b 19oc 520c 980C 2 oooc 4 7OOC 540a 1000 1070a Ni .. .. 50 508 100 100a 200 2108 500 v .. .. 25 23b 50 45b 100 80b 250 200b 500 590b < 5OC 61C 120c 290C 590C Zn .. .. 100 1ooa 200 210a 500 5ooa 1000 1050a 2000 2000a 21oc 29oc 560C 1 oooc 190oc The author is grateful to Mrs. Dawn Hutchison and Miss E. Waine for critically reading the manuscript. "This paper is published with the permission of the Director, Iistitute Geological Sciences (NERC) . 999 I -1 100 50 20 .* 'I .* 0 * 0 $ * * * +i I I I I I 0.5 1.0 2.0 5.0 10 20 5p fl 999 7 8 ** I I I of 2.0 5.0 10 20 50 100 200 Magnesium concentration, % MgO Lithium concentration/vg g-' Fig. 1. Use of major element reference materials as calibration standards. 0, syn- Fig. 2. Use of trace element reference materials thetic reference (mean of three) ; and as calibration standards. Symbols as in Fig. 1. *, international standards. Data from Abbey.lo 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. References Larsen, E. S., Jr., A m . J . Sci., 1938, 35, 94. Fairbairn, H. W., Schlecht, W. G., Stevens, R. E., Dennen, W. H., Ahrens, L. H., and Chayes, F., Stevens, R. E., Niles, W. W., Chados, A. A., Philvy, R. H., Leininger, R. K., Ahrens, L. H., Fleischer, Butler, J. R., and Thompson, A. J., Geochim. Cosmochim. Acta, 1962, 26, 516. Butler, J. R., and Thompson, A. J., Geochim. Cosmochim. Acta, 1962, 26, 1349. Kleeman, A. W., J . Geol. SOC. Aust., 1967, 14, 43. Canadian Association for Applied Spectroscopy, Report of Nonmetallic Standards Committee, Roubault, M., de la Roche, H., and Govindaraju, K., Sci. Tewe, Nancy, 1966, 11, 105. Flanagan, F. J., and Gwyn, M. E., Geochim. Cosmochim. Acta, 1967, 31, 1211. Abbey, S.. Geostand. Newsl., 1980, 4, 163. Flanagan, F. J., Geochim. Cosmochim. Acta. 1969, 33, 81. Allcott. G. H., and Lakin, A. W., "Statistical Summary of GeochEmical Data Furnished by 85 US Geological Survey, Open U.S. Geol. Surv. Bull., No. 980, 1951. M., and Flanagan, F. J., U.S. Geol. Surv. Bull., No. 1113, 1960. A+@ Spectrosc., 1961, 15, 159. Laboratories for Six Geochemical Exploration Reference Samples, File Report. Denver. Colo., 1974.12 REFERENCE MATERIALS Anal. Proc. Bowman, W. S., Faye, G. H., Sutarno, R., McKeague, J. A., and Kodama, H.. Geostand. Newsl., Flanagan, F. J., and Gottfried, D., U.S. Geol. SUYV. Prof. Pap., No. 1155, 1980. Lister, B., Trans. Inst. Min. Metall., 1977, B86, 133. Geostand. Newsl., 1977, 1, No. 1. RubeSka, I., Geostand. Newsl., 1977, 1, 15. Myers, A. T., Havens, R. G., Connor, J. J., Conklin, N. M., and Rose, H. J., Jr., U.S. Geol. Surv. de la Roche, H., and Govindaraju, K., Analusis. 1973, 2, 59. Date, A. R., Analyst, 1978, 103, 84. 13. 14. 15. 16. 17. 18. 19. 20. 1979, 3, 109. Prof. Pap., No. 1013, 1976.