Tutorial ReviewX-ray Mapping in Electron-Beam InstrumentsJohn J. Friel1and Charles E. Lyman2,*1Princeton Gamma Tech, C/N 863, Princeton, NJ 08542, USA2Depart ment of Materials Science and Engineering, Lehigh University, 5 East Packer Avenue, Bethlehem, PA 18015, USAAbstract: This review traces the development of X-ray mapping from its beg inning 50 years ago throughcurrent analysis procedures that can reveal otherwise obscure elemental distributions and associations. X-raymapping or compositional imaging of elemental distributions is one of the major capabilities of electron beammicroanalysis because it frees the operator from the necessity of making decisions about which image featurescontain elements of interest. Elements in unexpected locations, or in unexpected association with otherelements, may be found easily without operator bias as to where to locate the electron probe for data collection.X-ray mapping in the SEM or EPMA may be applied to bulk specimens at a spatial resolution of a bout 1 mm.X-ray mapping of thin specimens in the TEM or STEM may be accomplished at a spatial resolution rangingfrom 2 to 100 nm, depending on specimen thickness and the microscope. Although mapping has traditionallybeen considered a qualitative technique, recent developments demonstrate the quantitative capabilities of X-raymapping techniques. Moreover, the long-desired ability to collect and store an entire spectrum at every pixel isnow a reality, and methods for mining these data are rapidly being developed.Key wor ds: X-ray mapping, compositional imaging, X-ray spectrometry , EDS, WDS, conc entration–concentrationhistograms, spectrum imaging, position-tagged spectrometry, principal component analysis, multivariate statis-tical analysisINTRODUCTIONProduction of images showing elemental distributions on afine scale is an important contribution of electron micros-copy to scientific investigations. X-ray maps are formed bycollecting characteristic X rays from elements in the speci-men as a focused electron beam is scanned in a raster acrossthe specimen. In the 50 years since the first compositionalimage was obtained in an electron-beam instrument ~Coss-lett & Duncumb, 1956!, there has been extraordinar yprogress. For the first 25 years, qualitative analog dot mapswere used to form X-ray maps of elemental distributions.Computer control of the electron beam and computer stor-age of digital images dramatically changed X-ray mappingto the point that digital methods are standard in all commer-cial systems.This article reviews X-ray mapping in the scanningelectron microscope ~SEM!, the electron probe microana-lyzer ~EPMA!, and the type of analytical transmission elec-tron microscope ~AEM! based on the scanning transmissionelectron microscope ~STEM!. Electrons are ideal for gener-ating X-ray compositional images because they can be fo-cused to a small probe, they can be deflected to form ascanned beam raster, and they can excite atoms in thespecimen to produce characteristic X-ray signals. All otherbeams that might be used to excite an element-specificsignal suffer from specimen preparation difficulties, poorspatial resolution, or quantification problems. Other com-positional imaging methods are compared with electronbeam methods in Table 1. Two important analytical param-eters are listed in the table for each method: the spatialresolution of analysis and the elemental detection limit, thesmallest amount of an element that can be detected in amatrix. For Table 1 these parameters have been estimatedfor the mapping mode of analysis where the values areabout an order of magnitude worse than the ultimate capa-bilities of each instrument. X-ray mapping remains themost convenient and popular method for producing com-positional images.ANALOG VERSUS DIGITAL MAPSEarly WorkCastaing ~1951! built the first practical EPMA in which anelectron beam excited characteristic X rays that were de-tected with an X-ray spectrometer; however, that instru-Received June 24, 2004; accepted November 7, 2005.*Corresponding author. E-mail: [email protected]. Microanal. 12, 2–25, 2006DOI: 10.1017/S1431927606060211MicroscopyANDMicroanalysis© MICROSCOPY SOCIETY OF AMERICA 2006ment could only analyze one specimen point at a time.Duncumb and Cosslett obtained the first X-ray “dot map”compositional image ~Cosslett & Duncumb, 1956; Dun-cumb & Cosslett, 1957! by modifying an EPMA such thatthe electron beam could be scanned across the specimensurface to generate characteristic X-ray signals as a functionof beam position ~see Fig. 1!. In this first X-ray map, theCu Kaand Ag Lasignals were separated by the energy-dispersive properties of a gas proportional counter ~about1000 eV energy resolution!. Because the detector was placedbeneath the specimen and X rays were collected in transmis-sion, dark lines appeared where X-ray absorption was thegreatest. Soon after this initial demonstration, thewavelength-dispersive spectrometer ~WDS! was employedto detect X rays for maps, extending the technique to moregeneral commercial applications ~Melford & Duncumb,1958!. A decade later the energy-dispersive X-ray spectrom-eter ~EDS! became available for X-ray mapping in electron-beam instruments ~Fitzgerald et al., 1968!.Analog Dot MapsBecause EDS and WDS systems can be attached to almostany SEM or EPMA, X-ray analog “dot maps” are possible onall such instruments, even those of old vintage. As the beamscans across the specimen in a continuous raster, a momen-tary bright flash is registered on the cathode ray tube ~CRT!screen when an X ray enters the spectrometer within apreselected X-ray energy range. These flashes ~dots! arecaptured directly on film by leaving the camera shutteropen. The resulting image is one of dots built up on thefilm, and the relative concentration of the element is in-ferred by observing the clustering ~areal density! of the dots.For guidelines on the setup of analog dot maps, the readeris referred to Goldstein et al. ~1981!. Because characteristicX-ray signals from the specimen are much weaker thanemitted electron signals, analog X-ray maps are often ac-quired over several scan rasters ~image frames!, a processthat may take up to an hour or longer for each element. Atotal of 250,000 counts ~dots on the film! is considered theminimum exposure for a high quality dot map ~Goldsteinet al., 1992!. An example of an analog dot map is shown inFigure 2. Although simple to acquire, dot maps have
View Full Document
Unlocking...