Volume 107, Number 6, November–December 2002Journal of Research of the National Institute of Standards and Technology[J. Res. Natl. Inst. Stand. Technol. 107, 663–679 (2002)]Copper Oxide Precipitates in NBS StandardReference Material 482Volume 107 Number 6 November–December 2002Eric S. WindsorNational Institute of Standards andTechnology,Gaithersburg, MD 20899-8371Robert A. CarltonElan Drug Delivery, Inc.,King of Prussia, PA 19406Greg Gillen, Scott A. Wight, andDavid S. BrightNational Institute of Standards andTechnology,Gaithersburg, MD [email protected]@[email protected]@nist.govCopper oxide has been detected in the cop-per containing alloys of NBS StandardReference Material (SRM) 482. This occur-rence is significant because it representsheterogeneity within a standard referencematerial that was certified to be homoge-neous on a micrometer scale. Oxide occursas elliptically to spherically shaped pre-cipitates whose size differs with alloy com-position. The largest precipitates occur inthe Au20-Cu80 alloy and range in sizefrom submicrometer up to 2 mindi-ameter. Precipitates are observed usinglight microscopy, electron microscopy,and secondary ion mass spectrometry(SIMS). SIMS has demonstrated that theprecipitates are present within all the SRM482 wires that contain copper. Only thepure gold wire is precipitate free. Initial re-sults from the analysis of the Au20-Cu80alloy indicate that the percentage of precip-itates is less than 1 % by area. Elect ronprobe microanalysis (EPMA) of large(2 m) precipitates in this same alloy in-dicates that precipitates are detectable byEPMA and that their composition differssignificantly from the certif ied alloy com-position. The small size and low percent-age of these oxide precipitates minimizesthe impact that they have upon the in-tended use of this standard for electronprobe microanalysis. Heterogeneitycaused by these oxide precipitates mayhowever preclude the use of this standardfor automated EPMA analyses and othermicroanalysis techniques.Key words: copper-gold alloy; electronprobe microanalysis; metallography; NBSStandard Reference Material 482; oxideinclusions; sample preparation; secondaryion mass spectrometry.Accepted: August 22, 2002Available online: http://www.nist.gov/jres1. IntroductionStandard Reference Material (SRM) 482 was issuedin 1969 by the National Bureau of Standards (NBS)1.Ithas been continuously available to the public for 33years. The standard consists of a set of six wires (Fig. 1).Each wire is of a different composition within the cop-per-gold binary alloy system. Two uncoated wires repre-sent the pure end member compositions of pure copper(Cu) and pure gold (Au). The remaining four wires arealloys with nominal compositions varying in steps of 0.21NBS was renamed The National Institute of Standards and Technol-ogy (NIST) in 1987.mass fraction. For identification purposes, each alloywire was coated with a different colored paint. Theircomposition and color is as follows: Au20-Cu80 (red),Au40-Cu60 (blue), Au60-Cu40 (yellow), and Au80-Cu20 (gray). Each wire is approximately 5 cm long and0.5 mm in diameter. SRM 482 was issued specificallyas a standard for microanalysis. Therefore, each wirewas certified for both chemical composition and homo-geneity on a micrometer scale (Appendix A).At the time that SRM 482 was issued, Heinrich [1]reported that there was concern about the usefulness of663Volume 107, Number 6, November–December 2002Journal of Research of the National Institute of Standards and TechnologyFig. 1. SRM 482. Copper-Gold binary alloys for microanalysis. The SRM is a set of six wires. Wires are approxi-mately 5 cm long and 0.5 mm in diameter. Each wire has a different composition within the copper-gold binary system.the electron probe microanalyzer (EPMA) as a quantita-tive tool for chemical analysis. Often, analyses per-formed on the same material by different experiencedinvestigators varied in excess of 10 % relative. Factorscontributing to these high relative errors included sys-tematic errors in addition to errors resulting from theapplication of different matrix correction procedures.Additionally, Heinrich determined that a major con-tributing factor to these errors was the lack of standardmaterials of accurately known chemical compositionand microscopic homogeneity [2]. SRM 482 was thenissued in response to this need for chemically character-ized homogeneous standards.Since the time it was issued, SRM 482 has been wellaccepted and widely used throughout the microanalysiscommunity. Uses include the evaluation and modifica-tion of matrix correction procedures [3-7], evaluation ofEPMA instrument performance [8], and investigation ofsystematic errors associated with electron probe micro-analysis [9]. To this day, SRM 482 remains one of a verylimited number of standard reference materials availablefrom NIST that is certified to be homogeneous on amicroscopic scale [10].Recently, homogeneity of SRM 482 has been ques-tioned. Carlton reported the occurrence of spots onmetallographically prepared surfaces of the Au20-Cu80and the Au60-Cu40 wires [11]. The presence of thesespots was then verified by independent preparations per-formed at NIST [12].The occurrence of these spots raises the followingquestions:(1) Do the spots represent heterogeneity within thewires or are they artifacts that were created dur ingmetallographic sample preparation?(2) What is the chemical composition of these spots?(3) Does the occurrence of these spots affect the useof SRM 482 as a standard for electron probe microanal-ysis?(4) If these spots represent heterogeneity within thewires, then why were they not detected prior to certifi-cation and why have they not been reported in 33 yearssince SRM 482 was issued?The purpose of this manuscript is to report resultsfrom work that is in progress to answer these questionsand investigate the possibility of heterogeneity withinSRM 482. Our investigation centers on the Au20-Cu80wire because the spots are largest and appear most abun-dant in this composition. Therefore, if the spots do sig-nificantly affect the microanalysis of SRM 482, wewould expect this effect to be most noticeable in theAu20-Cu80 wire.2. Samples, Preparation, and InitialObservationsWhen the wires of SRM 482 were manufactured, theywere drawn as one continuous wire [1]. Wire length wasapproximately 150 m. Extensive homogeneity testing664Volume 107, Number 6,
View Full Document
Unlocking...