Volume 107, Number 6, November–December 2002 Journal of Research of the National Institute of Standards and Technology [J. Res. Natl. Inst. Stand. Technol. 107, 681–685 (2002)] Smithsonian Microbeam Standards Volume 107 Number 6 November–December 2002 Eugene Jarosewich This is a short history of the Smithsonian Key words: electron microprobe standards; Microbeam Standards; their sources, se- microbeam standards; mineral standards; Department of Mineral Sciences, lection, preparation, and analyses. Fifty- reference materials. NMNH, eight minerals, natural glasses, and syn- Smithsonian Institution, thetic samples have been characterized in the past 25 years. During that time, over Accepted: August 22, 2002 Washington, DC 20560 750 requests were received for approxi- mately 11 000 individual samples. These reference samples are referred to as the Smithsonian Microbeam Standards. Available online: http://www.nist.gov/jres 1. Introduction The early 1960s was a period when various Federal minerals, mineral separates, and natural glasses, several agencies generously supported the development of new of which became electron-microprobe standards. These instrumentation and new techniques for the analysis of analyzed minerals and natural glasses were initially in- lunar samples whose return was anticipated. In the fall tended to serve our own needs; however, numerous re- of 1964, the Department of Mineral Sciences, National quests for these standards prompted the staff to publish Museum of Natural History, with funding from NASA, the data and make small quantities available to interested purchased an Applied Research Laboratory electron mi- researchers. croprobe and expanded its laboratory facilities for the As a rule, materials selected for standards should be study of meteorites. By then the electron microprobe analyzed by more than one laboratory and, if possible, had become an established instrument and most of the by two independent methods. However, in our setting basic analytical techniques were already developed. this ideal approach was constrained by both insufficient Since x-ray microanalysis is not based on first principles funds and the limited amount of many samples. For of physics or chemistry, but relies on comparison with these reasons most samples were analyzed only once by materials of known composition, a set of well-character- wet-chemical methods. Based on our experience, a care- ized standards is required for analysis of unknowns. At ful wet-chemical analysis will provide satisfactory ma- this early stage of electron microprobe analyses only a jor-element results. The only samples in our suite of few mineral standards were available. In the following standards analyzed by more than one laboratory or dif- years, the National Mineral Collection at the Smithso- ferent analysts were: Cr-bearing augite from Nevada; nian served as an invaluable source of minerals for refer- A-99 basaltic glass from Hawaii; and partial analyses of ence samples. Although the laboratory staff focused hornblende, pyrope, and augite from Kakanui, New primarily on meteorite research, they also analyzed Zealand, and VG-2 glass from the Juan de Fuca Ridge. 681 Volume 107, Number 6, November–December 2002 Journal of Research of the National Institute of Standards and Technology 2. Standards and the three standards were in good agreement with the wet-chemical analyses [4]. As the staff’s interest in electron-microprobe analyses As geochemical research at the Smithsonian broad- of different geological materials expanded, the demand ened, it became clear that a wide range of well charac- for additional standards increased. For example, in the terized materials were needed as primary and secondary late 1960s George Switzer, then chairman of our de- standards for electron-microprobe analyses, as well as partment, obtained questionable results with the avail- standards for special applications. Also, since a large able standards when analyzing garnets from diamond- number of requests were received for different stan- bearing kimberlites in South Africa. He addressed this dards, J. Nelen and J. Norberg of the laboratory staff and problem by preparing mineral separates of two garnets the author continued evaluation of the most requested and an omphacite, since no suitable garnet and omphac- minerals from our collections as potential standards. Al- ite specimens were available. These three separates were though the interest in standards was very broad, our then analyzed by wet-chemical methods and thereafter efforts were focused only on silicate materials. used as standards. When using these “like standards”, of Over a period of approximately 10 years (1968-1978), close composition to the unknowns, acceptable results 31 standards were characterized and made available for were obtained as judged by stochiometry and analytical distribution [5]. These standards have been widely used totals close to 100 %. by the geochemical community and their acceptance by About the same time Brian Mason, then curator of the the users gave us an additional impetus to continue with National Collection of Meteorites, considered re-analy- characterization of other standards. In the early 1980s, sis of hornblende and pyrope from Kakanui, New four carbonate standards were prepared for a study of Zealand. He felt that the values for these two minerals corals [6,7]. At the same time a large crystal of Cr-bear- in his original publication [1] were in error. After re- ing augite became available with approximately 0.8 % analysis, these two minerals plus the garnets and om- of Cr2O3, a useful standard for the routine analysis of phacite were routinely used as standards in subsequent low-concentration chromium in silicates [8]. An impor- studies [2]. tant addition to our reference material collection was the Further work in developing accurate standards were donation of fourteen synthetic single-element REE or- undertaken in the early 1970s, as Bill Melson, curator of thophosphates (plus Y and Sc) by Lynn Boatner of the the Petrology collection, began a major study of sea- Oak Ridge National Laboratory. These samples were floor volcanic glasses. One of the objectives of this work not chemically analyzed, but based on extensive crystal- was to determine compositions of glasses from different lographic data they were determined to be of stochio- localities around the globe. Basaltic glasses VG-2 and metric composition [9]. The Corning Glass Company A-99 were selected as standards for this project [3]. prepared three glasses, each containing 0.75 % of seven Several institutions used VG-2 as a standard for elec- elements commonly found in minor quantities in silicate tron-microprobe analyses, and in order to assure their minerals. Paul Carpenter, of the California Institute of quality, round-robin analyses were undertaken by three Technology, transferred these glasses to our collection in laboratories. A polished disk with VG-2, A-99, Kakanui 1997 for distribution. We also recently obtained from hornblende, and two glasses of unknown composition the Corning Museum four glass standards used in a was analyzed by the United States Geological Survey study of ancient glasses [10], with a useful range of (Reston, VA),the Massachusetts Institute of Technology, elements for special applications (Vicenzi et al., see p. and the Smithsonian Institution. In order to determine 719 of this Special Issue). the precision and accuracy of analyses, the samples were first analyzed with Kakanui hornblende as the standard for precision (uncorrected results), and then with the 3. Preparation of Standards preferred standards of each laboratory and finally com- pared with wet-chemical analyses (Smithsonian). The The preparation of standards is a time-consuming and round-robin test revealed excellent agreement in the pre- exacting process. Most of the standards used in geolog- cision of analyses from the three laboratories when the ical studies are natural minerals, although synthetic ma- samples were analyzed with Kakanui hornblende as the terials are also widely used. With both types of materi- standard. When the samples were analyzed with each als, the spatial homogeneity was determined by electron laboratory’s own standards, some with widely different microprobe before any wet-chemical analyses were un- compositions from those of the unknowns, there were dertaken (Jarosewich et al., 1980). Then, the stability considerable matrix corrections. Nevertheless, the re- under the electron beam was evaluated. This step sults among the three laboratories for the two unknowns requires care, as some samples initially appear to be 682 Volume 107, Number 6, November–December 2002 Journal of Research of the National Institute of Standards and Technology stable, yet count rates change when the sample is sub- lysts and other individuals in various institutions have jected to the electron beam for a prolonged period. For been actively engaged in the characterization of new example, dolomite is stable under standard operating geochemical standards. Much more needs to be done in conditions (15 kV, 15 ␮A, 5 ␮m beam diameter) for preparation of new standards and especially in the about 40 s to 50 s, but after that the count rate changes. timely characterization of these standards by the collab- For carbonates and high-sodium silicates, most of which orators. are not stable under the electron beam, but are essential Standards for trace-element analyses will be increas- standards, special techniques, such as wide beam di-