The M Emission Spectrum of68ErbiumJan Dellith and Michael Wendt*Institut für Physikalische Hochtechnologie, Albert-Einstein-Strasse 9, D-07745 Jena, GermanyAbstract: The M emission spectrum of68Er was reinvestigated using wavelength dispersive spectrometry, with aTAP diffracting crystal. By recording the spectra using the second-order reflection, an improved energyresolution was achieved, which is necessary to resolve the M5O3line from the neighboring a M5N7transition.In addition to the five lines/bands tabulated in the classical paper of Bearden, a number of further lines wereobserved. These are M1N3,M3O1,M2N1,M5O3,M3N1, and M4N3. For all the lines with an energy below the M5absorption structure ~M5O3,M3N1,M4N3, and z M5N3!, an increasing relative intensity with increasing energyof the exciting electrons, E0, was observed. This dependence has its origin in the fact that these lines arenormally absorbed whereas Ma~M5N7! and Mb~M4N6! are additionally affected by anomalous line-typeabsorption.Key words: X-ray spectrometry, electron microprobe analysis, wavelength dispersive spectrometry ~WDS!,erbium, M emission spectrum, relative intensitiesINTRODUCTIONRecently, when reinvestigating the M emission spectra ofthe elements 55 ⱕ Z ⱕ 58 w ith wave length dispersivespectrometry ~WDS!, a total of 34 M lines were observed~Dellith & Wendt, 2004!, of which only 13 are contained inBearden’s classical compilation ~Bearden, 1967!. In thatwork, for each rare earth ~RE! elementonly4or5Mlines/bands are tabulated. In contrast, 9 M lines are givenfor the element50Sn and 16 for73Ta. Therefore, whenstarting the present reinvestigation, we assumed that the Memission spectra of the RE elements were only incompletelyknown.Historically, the M emission and absorption spectra ofthe RE elements have been studied for a very long time. Theunderstanding of the complicated emission s tructure of Ma~M5N7! and Mb~M4N6! was the main interest in these earlyinvestigations. A milestone was achieved in 1956 ~Steward-son & Wilson, 1956! when these authors showed that thetrue emissions of Er Ma~M5N7! and Mb~M4N6! are broadsingle lines and that the observed structures are producedby self-absorption in the target material. A dozen years laterit was shown that this so-called anomalous line-type absorp-tion is responsible for all the structures observed in the Ma~M5N7! and Mb~M4N6! emissions of the elements57La to70Yb ~Fischer & Baun, 1967!.In practice, the analyst is only able to correctly identifyan observed peak if all the lines that are detectable with agiven type of X-ray spectrometer are known. Otherwise, thedanger of misinterpreting an unknown line as one causedby an impurity is high. For a correct identification, theknowledge of both the position of a line and its relativeintensity is needed. Table 1 summarizes what was knownwhen starting the present reinvestigation. The positionswere taken from Bearden’s above-cited paper and the rela-tive intensities from the popular ASTM tables ~White &Johnson, 1970!. From spectra taken by energ y dispersivespectrometry ~EDS!, it was shown that the relative intensitygiven in the ASTM tables for the line z M5N3of the REelements is too low by 3–4 orders of magnitude ~Wendt &Christ, 1985!, a finding that was corroborated by later WDSmeasurements ~Lábár & Salter, 1991!. However, even thisrelatively new paper does not contain any indication as towhether the M spectra of the RE elements consist of morelines than those given in Table 1.MATERIALS AND METHODSThe measurements were carried out using a JXA 8800 Lmicroprobe ~JEOL!. The take-off angle of this microprobeis 408. A TAP crystal with 2d ⫽ 25.757 Å was used as thedispersing element. The energy of the exciting electrons, E0,was varied in the range 2.5 ⱕ E0ⱕ 25 keV. All spectra weretaken in the pulse height analysis mode. The voltage of thegas flow proportional counter and the gain of the pulseamplifier were selected in such a manner that the meanpulse height was 4 V. For the spectra taken in first-orderreflection, the lower level of the discriminator was set to 2 VReceived January 2, 2007; accepted February 19, 2007.*Corresponding author. E-mail: [email protected]. Microanal. 13, 191–195, 2007DOI: 10.1017/S1431927607070389MicroscopyANDMicroanalysis© MICROSCOPY SOCIETY OF AMERICA 2007and the window width to 4 V. For spectr a taken in second-order reflection, the optimum discriminator setting is some-what different. To suppress the background, which has itsorigin in bremsstrahlung of low energy, the lower level wasset to 3 V and the window width to 2 V. A comparison ofthe central part of the spectr a taken at E0⫽ 20 keV infirst-and second-order reflection is shown in Figure 1. Thenumbers given in the figure refer to lines summarized inTable 2. The disadvantage of spect ra taken using second-order reflection is the loss of intensity, by approximately afactor of 40, compared to spectra taken in first order.Nevertheless, the detail in second-order reflection spectra isgreatly improved due to the higher energy resolution. Onlyin this manner were we able to identify line number 8without doubt as the M5O3line. In contrast, Lábár andSalter attributed this peak to be part of the Ma~M5N7!emission structure.A full spectrum, taken in first-order reflection at E0⫽12 keV, is shown in Figure 2a. In Figure 2b the samespectrum is shown with an ordinate scale magnified by afactor of 5.7. In addition, the linear background approxima-tion is shown. In our approximation, the relative net peakheights of the peak maxima are used as a measure of therelative intensities.The68Er standard used was nominally a 99.9% puremetal. Like the other RE elements, erbium is a material ofhigh reactivit y. Even a freshly polished sample, which wasinserted as quickly as possible into the microprobe, wascovered by an oxide layer the thickness of which was esti-mated to be of the order of 10 nm. Consequently, spectrataken from a carbon-coated pressed Er2O3powder werecompared with those taken from a freshly polished metalstandard. Even when using second order, no indication ofany line shift was observed. This result is in agreement withthe findings of Fischer and Baun that, except for63Eu and70Yb, neither the Ma, b emission ~M5N7,M4N6! nor theM4,5absorption spect ra appear to be significantly influ-enced by chemical combination.RESULTS AND DISCUSSIONThe results of our search for Er M peaks are summarized inTable 2. With the
View Full Document
Unlocking...