JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY MARCH 1994 VOL. 9 385 Adaptation of a Glow Discharge Mass Spectrometer in a Glove-box for the Analysis of Nuclear Materials* Maria Betti Gert Rasmussen Tania Hiernaut and Lothar Koch Commission of the European Communities Joint Research Centre Institute for Transuranium Elements Postfach 2340 76125 Karlsruhe Germany Dafydd M.P. Milton and Robert C. Hutton Fisons Instruments VG Elemental Ion Path Road Three Winsford Cheshire UK C W7 3BX A VG9000 glow discharge mass spectrometer has been modified for the direct analysis of solid nuclear samples within a glove-box environment. Because containment is needed for the analysis of this kind of material the glove-box encloses all parts of the instrument that come into contact with the sample namely the ion source chamber sample interlock and associated pumping system. External modifications eliminate outside contamination by the fitting of absolute filters on all source supplies.Internally the design of the ion source has been altered to minimize the number of operations performed inside the glove-box thereby simplifing operation and routine maintenance. These modifications retain the ion extraction and focusing properties of the instrument. The data presented show that there is no compromise in the analytical performance of the instrument when placed in the glove-box. Data representative of nuclear materials is also shown. Keywords GIow discharge mass spectrometry; glove-box; nuclear material Glow discharge (GD) sources have a long history in analytical chemistry principally as sources for optical emission spec- tr~metry.’-~ Their advantages are related to direct solid sam- pling stable output and possibilities for depth profile and thin layer analysis.More recently GD sources have been coupled with mass spectrometry (GDMS) and the combination of these two established methods have made available to the analytical chemist a new exciting technique for the analysis of solid samples.46 The advantages of MS over conventional optical techniques lie in the increased sensitivity obtained by direct sampling of the ions and a much simpler spectrum which makes possible the determination of a wider range of elements. This has also been accompanied by much simpler quantification. Plasma sources coupled with mass spectrometers have thus dramati- cally extended both elemental coverage and detection limits in quantitative trace element analysis.This is true both of GDMS and the inductively coupled plasma (ICP) MS techniques. Glow discharge mass spectrometry has found widespread use in the determination of trace elements in a wide range of inorganic materials.”” Conducting samples such as metals alloys and semiconductors can be analysed directly by a d.c. GD. Insulating samples require indirect means to convert them into a conducting f0rm~9~ or the use of an r.f. powered GD”l’ for direct analysis. A priori the GD mass spectrometer is an analytical tool which would be ideally suited for the analysis of solid samples of nuclear origin requiring minimum chemical treatment. The elemental and isotopic capabilities of GDMS could be fully applied to materials having non-natural isotopes and/or non- natural isotopic abundances.Other advantages of GDMS are ( i ) virtually all elements can be determined; ( i i ) the wide dynamic range of the detector allows the determination of both major components and trace constituents within the same analytical cycle; and ( i i i ) decoupling of the atomization and ionization processes results in uniform sensitivities for many elements and also minimum matrix effects. For the analysis of nuclear materials difficulties arising from * Presented at the XXVIII Colloquium Spectroscopicurn Internationale (CSI) Post-Symposium on Analytical Applications of Glow Discharge in Optical and Mass Spectrometry York UK July 4-7 1993.the radioactive nature of the sample have to be overcome. Firstly the operator has to be protected from the material which means that the use of glove-boxes (alpha beta protec- tion) and/or hot cells (alpha beta and gamma protection) with master-slave manipulators is a necessity. Secondly in order to avoid contamination of the working area the analytical instruments have to be modified in order that containment is assured and no radioactive material leaks either into the laboratory or into the environment. Complete instruments cannot be introduced into a glove-box because electronics are very sensitive to radiation. In practice electronics and parts that might need special maintenance are kept outside the glove-box and only samples and the corresponding sampling stage are contained in the box.From the experience gained from other instrumentation,12 used for GDMS measurements the glove-box should enclose the ion source chamber sample interlock and associated pumping system. All supplies to the ion source (argon discharge support gas and liquid nitrogen for cryogenic cooling of the discharge cell) and pumping ports should be fitted with absolute filters to eliminate any external contamination. The ion source itself has been designed to minimize the number of operations and to simplify routine maintenance inside the glove-box area. This has been achieved through the use of simple plug-in components and by reducing the number of screws. In this paper a modified GDMS instrument in a glove-box for the analysis of radioactive materials of nuclear origin is described and some preliminary results are reported.Experimental Instrumentation Mass Spectrometer The VG 9000 glow discharge mass spectrometer has been described in detail previo~sly.’~ The instrument consists of a GD ion source coupled to a double-focusing mass spectrometer of reverse (Nier-Johnson) geometry. This provides high trans- mission and sensitivity whilst operating at high resolving power (typically 5000 10% valley definition > 75% trasmission). Ion detection is accomplished by means of a dual detection system comprised of a Faraday cup for the measurement of large386 JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY MARCH 1994 VOL. 9 Detectors Water cooling system Fig. 1 Schematic diagram of the installation of the discharge source housing in the glove-box (typically > Daly detector14 for the detection of lower signals.A) ion currents and a transverse mounted Source housing For the analysis of nuclear samples the GD ion-source housing and the associated pumps are placed within a glove-box. The glove-box is provided with an extraction system and filters for all gas lines. Absolute filters (> 99.99% efficiency Solfiltra Lagarenne France) are situated both inside and outside the glove-box which is kept at a lower pressure than the exterior (20mm water column). In Fig. 1 a schematic diagram of the installation of the GD source housing in a glove-box is given showing all connections to the glove-box. Stainless-steel filters ( 3 pm porosity) are installed in the argon compressed air and liquid nitrogen inlets and absolute filters (0.3 pm porosity) in the vacuum line to prevent any radioactive contamination.The cryogenic pump installed in the glove-box is shown in Fig. 2(a) and the absolute filters connected with the pump in Fig. 2(b). In Fig. 3(a) the instrument is shown in the Institute’s workshop after the first stage of its installation in the glove- box and in Fig. 3(b) in the hot lab after the closing of the glove-box for the handling of radioactive material. fb) Ion source The probe and sample interlock have been shortened to give easier access to the ion source within the glove-box. The source chamber door opens on a slider mechanism to give easier access to the ion source. The Wilson-seal can assembly provides a vacuum interlock region so that sample change over can be performed with the ion source maintained at high vacuum (Fig.4). The ion source itself has been designed to minimize the number of operations performed within the glove-box area and t o simplify routine maintenance. This has been achieved by utilizing a ‘universal’ cell for the analysis of both Pin and flat samples and a ‘plug-in’ focus stack. The source itself Fig.2 (a) Cryogenic pump installed in glove-box under the GD source housing. (b) Absolute filters installed for the containment of contamination right side inside glove-box; left side outside glove-boxJOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY MARCH 1994 VOL. 9 (4 387 Fig. 4 Modified probe and sample interlock Fig. 5 Modified source and its housing chamber Fig.3 (a) Complete view of the instrument in glove-box during the installation in the workshop.(b) Final view of the instrument installed in glove-box after closing the glove-box for nuclear material handling (Fig. 5) is split up into various components comprising a mounting plate with removable cell and focus stack assemblies. The source mounting position remains affixed to the back wall of the source housing chamber. The focus stack then plugs into a recess in the mounting plate and is held in place by four fixing rods. Electrical contact to the plates of the focus stack is made by a series of copper-beryllium contacts. This eliminates the need to disconnect any wires when removing the focus stack thus simplifying its removal. The contact assembly is connected to the high tension feed-through by kapton-coated wire (Dupont France).The focus stack assembly (Fig. 6) consists of a series of tantalum plates separated by PEEK spacers mounted onto a base containing the source-defining slit for the mass spec- trometer. The focus stack provides deflection and focusing of the ion beam in the y- and z-directions to give the best object on the source defining slit. The plates are shaped so that when Fig. 6 Modified focus stack assembly the focus stack is in position they make electrical contact with the appropriate connector on the contact assembly. The focus stack assembly also contains a mounting bracket for location of the cell and sample holder. The ‘universal’ cell has been designed to accommodate a range of pin and flat samples.The cell itself consists of a universal body that plugs into the focus stack. This cell body based around the existing flat geometry,” then remains located388 JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY MARCH 1994 VOL. 9 in position. The ‘universal pin holder’ can be used for the analysis of a wide range of pin samples. In Fig. 7(a) and (b) a sample pin assembled in the holder and the different parts used for the assembling respectively are shown. The use of a combination of cathode plates (A) anode chamber bodies (B) and sample chucks (C) allows analysis of pin and rod samples (D) up to 7 mm in diameter. With the appropriate sample holder combination the sample is held in position in the chuck with a screw. The chuck is then placed in the sample holder with the tapered end mating with the appropriate cathode plate.The spring ensures that a tight reproducible gas seal is maintained. Electrical insulation between the anode and cathode components is achieved by using a ceramic ring and nylon screws (F) are used to provide a gas-tight seal. The analysis of flat samples is performed using the flat sample holder as described previo~s1y.l~ In Fig. 7(c) the differ- ent parts used for assembling the flat sample in the holder are shown in detail. The sample (A) is held in the holder against an insulating disk (B) to provide electrical isolation from the front plate (C) of the sample holder which is at anode potential. Sample size can vary from approximately 10 to 38 mm in diameter and from wafer thin to a thickness of 20mm.To cope with different sample sizes a range of front plates and insulating discs are available. In Fig. 7(c) for sample (A) the points where the sputtering took place can also be clearly seen. Changing the sample geometry from pin to flat simply requires changing to the appropriate sample holder when loading the sample. It is no longer necessary to break vacuum 1.0 do this thus reducing the number of operations required inside the glove-box. Materials The argon discharge gas (BOC 99.9999%) enters the discharge cell via a heated getter inlet system (SAES GP50). The pressure is regulated using a leak valve (Fisons Cheshire UK) and monitored using an ion gauge situated above the cryogenic pump [Edwards Coolstar 1500 (Edwards Crawley Sussex UK)] serving the source housing. The discharge cell is cooled using a flow of liquid nitrogen to reduce background gases such as water vapour.A certified reference material CRM 115 depleted uranium metal (uranium and uranium-235 standard) obtained from New Brunswick Laboratory (US Department of Energy) was used. Once optimization of the discharge parameters had been per- formed a discharge voltage of 0.9 kV with a corresponding cur- rent of 0.6 mA was used throughout unless otherwise specified. Fig. 7 (a) Pin sample mounted in the proper holder; and (b) the different parts used for the assembling A cathode plate; B anode chamber body; C sample chuck; D pin or rod sample; E ceramic ring; F nylon screws; G spring; and H assembly holder (c) Flat samples together with the proper holder A samples; B insulating disk; C front plate of the sample holder (anode); and D loaded sample fixationJOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY MARCH 1994 VOL.9 389 Table 1 Comparison of GDMS instrumental specifications at various stages of its installation in a glove-box A before modifications and installation in glove-box; B after modifications and before installation in glove-box; and C after modifications and after installation in glove-box Specification Analyser vacuum/10-8 mbar Mass resolution A1 cu Au Matrix current/lO-" A Transmission (%) Stability (10 min) (RSD Yo) Stability (30 min) (RSD Yo) Peak position stability (ppm) Mass calibration (mmu)* Mass marking? (ppm) Ion counting efficiency (Yo) Abundance sensitivity (ppm) Reproducibility (RSD Yo) Major Minor Traces Major Minor Traces N C 0 Accuracy (RSD YO) Gas background (ppm) A 3.0 5384 4562 4139 70 4.9 2.94 4.22 9 7 234 76.0 0.43 0.8 1.2 3.5 1 .o 4.0 3.9 0.5 0.3 0.5 B 1.8 6940 6626 5466 75 6.2 2.52 3.87 13 8 100 80.7 0.38 1.6 2.6 1.8 2.5 2.7 2.1 4.3 3.6 0.7 C 2.4 6174 5624 6744 67 8.4 1.32 4.68 24 5 180 76 0.2 1 1 .o 3.7 5.2 1.5 4.4 2.8 0.4 0.9 2.4 * mmu = millimass unit.t Precision of repeat calibrations. The samples were pre-sputtered prior to analysis for 2 min with a 1 mA current. Results and Discussion GDMS Performance at Various Stages of Its Installation in the Glove-box As described above for the installation of the GDMS in a glove-box many modifications of the source housing and of the ion source were necessary. Particular attention was given to ensure that the instrumental specifications would not be affected and in the case that some of the specifications changed it was important to know to what extent the modifications influenced the instrumental performances.From this point of view the instrumental specifications of the GD mass spectrometer were checked throughout the technical modifications required for its installation in the glove- box and the performance of the spectrometer was found to be unaffected. In Table 1 the instrumental specifications tested at various stages are summarized. None of the parameters checked appeared to vary to a significant extent in the three stages tested particularly the specifications that might be expected to be directly affected by the modifications like source vacuum mass resolution matrix current transmission and gas background.Isotope Abundance Studies in Uranium Metal The isotopic analysis of uranium is important to establish the enrichment of the sample and thus its potential use as a nuclear fuel. Of primary interest is the 235U:238U ratio. In this investigation isotopic abundances were measured in uranium metal samples. A certified reference material (CRM 115) uranium metal depleted in 235U was measured. Three samples of this standard were analysed. The results of these measurements are shown in Table 2 where the signal intensities (expressed in amperes) measured for the three isotopes of uranium 234U 235U and 238U are reported. Each result represents the mean of ten measurements with each measurement lasting 5 s.The grand mean (in amperes and in concentration units) RSD CRM values and bias are shown at the base of each isotope column. The internal precision [% standard error (SE)] reported in parentheses for each isotope in each sample describes the reproducibility of each measurement of the individual isotopes tested for a single sample and is defined as %SE(l~)=lOO/xm{[ i = l n (xm-xi)2]1'2}/n(n-1) where x is the mean value and n is the number of measure- ments performed. The RSD values (external precision) reflect the reproduc- ibility of the isotopic concentration measurements from differ- ent specimens of the same original sample. This reproducibility is typically 0.2% or better even for 234U at a measured concentration of 8.26 ppm. This suggests that GDMS could be used to measure isotopic abundances at trace levels.In Table 3 results obtained for another uranium metal sample are shown. The abundance value measured for 235U in CRM 115 was used to produce a mass bias correction factor. This factor was used to correct the 235U abundance in a uranium sample with unknown isotopic composition. This correction gives an abundance which is accurate to within 0.12% of the natural abundance. No such correction was performed for the 234U isotope. From the data presented in Tables 2 and 3 it is clear however that CRM 115 is also depleted in 234U. Both CRM 115 depleted uranium and the natural uranium metal samples were analysed in the flat sample form. The RSDs obtained for both samples over 10 and 5 runs respectively indicate very good stability of the discharge and homogeneity of the sample surface.Comparison With Other Techniques The depleted uranium sample (CRM 115) was also analysed using isotope dilution analysis mass spectrometry (IDMS) and ICP-MS. Table 2 Analysis of CRM 115 depleted uranium reference sample. Signal intensities in amperes; values in parentheses are the SE (YO) Sample 1 2 Mean A Mean (concentration) CRM 115 value (%) Bias (YO) RSD (Yo) 234u 4.626 x (1.8) 4.638 x (1.6) 4.630 x (1.7) 4.6314 x 0.132 8.26 ppm Not certified Not applicable 235u 1.072 x (1.23) 1.075 x (0.88) 1.070 x (1.0) 1.0724 x 0.235 O.1973Yo 0.2008 (7) - 1.7 2 3 8 ~ 5.253 x lo-'' (0.7) 5.273 x lo-'' (0.4) 5.273 x lo-'' (0.6) 5.266 x lo-'' 0.219 99.770% 99.7762 -0.006214390 JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY MARCH 1994 VOL.9 - 2.1 ~ 0 1 c a 7 1.5 tA 1.2 2 0.9 - -. +-' .- . - - 0.6 0.3 - - Table 3 Analysis of uranium metal sample. Isotope values expressed in concentration units \ Measurement Mean SD RSD (%) Literature datai6 Bias (%) 234u (PPm) 57.950 59.090 59.150 59.330 59.33 1 59.9702 ppm 0.5803 ppm 0.98 9 55 PPm 235U (mass-%) 0.703 1 0.7064 0.7070 0.7145 0.7122 0.7086% 0.005 Yo 0.65 0.72% - 1.6 238U (mass-%) 99.2730 99.2745 99.2736 99.2729 99.2740 99.2736% 0.0007% 0.0007 99.2745% - 0.0009 When analysing CRM 115 by IDMS an accuracy of 0.1% was achieved. This represents an improvement in accuracy over GDMS mainly owing to the fact that an internal standard is used. The accuracy of GDMS is improved when the mass bias is accounted for.When this correction is made the accuracy obtained by GDMS for the natural sample is 0.12%. Without this correction the precision and accuracy of GDMS is comparable to that measured using ICP-MS. The major advantage of GDMS is the very quick and straightforward sample preparation. Both IDMS and ICP-MS require sample dissolution dilution and for IDMS further preparation of spiked aliquots. The GDMS method simply requires a flat surface of metal. As for handling of radioactive samples all techniques require the use of a glove-box. Consequently the ability of GDMS to perform direct analy- sis upon uranium provides an accurate and rapid isotopic screening process in preliminary investigations. Trace Element Determination in Uranium Metal The VG9000 is capable of analysis over the whole mass range and can thus be used to determine the concentrations of many other elements such as the transition metals and rare earth elements.Trace element analysis in uranium metal is important in the specification of the material. In Table 4 the elements observed in CRM 115 are listed. After the rapid complete survey the sample was analysed only for the trace elements observed. Among the elements listed in Table 4 some of particular importance are those elements with high neutron capture Table4 Trace element analysis of CRM 115 uranium metal. Concentrations expressed in ppm Element B Na Mg A1 Si Cr Mn Fe Ni c u Zr Mo Ru Ag I Pb Th Measured concentration 0.021 0.064 11.0 19.6 43.3 12.4 73.8 15.0 6.5 0.070 0.56 0.11 0.85 0.044 0.062 0.17 17.7 Standard deviation 0.001 0.003 0.3 0.5 1 .o 0.007 0.4 1.2 0.4 0.3 0.003 0.02 0.006 0.03 0.008 0.2 0.002 RSD (Yo) (n = 5 ) 5.9 4.0 3.1 2.7 2.2 3.7 3.1 1.6 2.6 4.7 4.3 3.5 5.5 3.8 17.4 1.3 3.5 2.4 (a) I -.10.008 10.010 10.012 10.014 10.016 10.018 1 4A v) c .- 11.004 11.006 11.008 11.010 11.012 11.014 11.016 m/z Fig. 8 GDMS spectra for (a) loB and (b) "B isotooes obtained in the analysis of an uianium metal sample.' honcentraiion of boron 20 PPb was 51.92 51.94 51.96 d Z Fig. 9 GDMS spectrum obtained when analysing 52Cr (170 ppb) in a uranium metal sample in the presence of carbon (540 ppm). The peak on the right is due to 40Ar'2C+; the mass resolution necessary to resolve "Cr from 40Ar'2C+ is 2400 cross-sections such as lithium boron cadmium etc. for which a precise and accurate determination is required.Boron for instance cannot be easily measured in diluted solutions at ppb levels (which correspond to ppm levels in the solid) by ICP-MS because of low ionization efficiency low sensitivity and blank contamination. In GDMS boron can easily be detected at low concentration levels directly in the solid sample. In Fig. 8 the peaks of loB [8(a)] and llB [8(b)] isotopes are shown measured on the same sample at a concentration level of 20 ppb in uranium matrix. Another example showing the power of GDMS for trace detection is reported in Fig. 9. Here the peak of "Cr which is often requested to be analysed as an impurity in uranium isJOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY MARCH 1994 VOL. 9 391 very well separated from the 40Ar'2C+ peak.The sample contained 540ppm of C and 52Cr has been determined at a level of 170 ppb. For the determination of trace elements in uranium metal five different runs were performed. In Table 4 the mean concen- tration values and the internal precision obtained over the five runs are reported for all elements found. The precision in the analysis is typically 5% RSD or less even at the ppb level. The exception is iodine which may be inhomogeneously distributed in the sample. Detection Limits In GDMS the detection limit depends on the number of points acquired the integration time and the resolution used. The data relevant to the elements of Table 4 were collected by acquiring 200 points for an integration time of 100ms for the Daly detector and using high resolution (5000) to overcome potential molecular and isobaric interferences.This resulted in typical detection limits of 3-10 ppb for most elements. Using low (1500) resolution acquiring the same number of points and for the same integration time as for high resolution acquisition the detection limits can be improved. The detection limit for thorium in uranium determined in this way was 0.2 ppb. Conclusions The modifications made to the GD source of a VG9000 GD mass spectrometer for its installation in a glove-box did not alter the instrumental specifications significantly. Elemental and isotopic capabilities of GDMS can be successfully applied to the characterization of samples of unknown isotopic com- position such as uranium metal. The GDMS technique has a precision and accuracy comparable to ICP-MS and for preliminary investigation the GDMS technique can give information on the isotopic abundance composition faster than IDMS.UP 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 The authors acknowledge warmly the workshop staff of the Transuranium Institute and Mr. Dockendorf Mr. Schrodt and Mr. Ougier for their fruitful technical support in setting _ - the instrumentation in the glove-box. - References Denoyer E. Van Grieken R. Adams F. and Natusch D. F. S. Anal. Chem. 1982 54 26A. Koch K. H. Spectrochim. Acta Part B 1984 39 1067. KO J. B. Spectrochim. Acta Part B 1984 39 1405. Fang D. and Seegopaul P. J. Anal. At. Spectrom. 1992 7 959. Jakubowski N. and Stuewer D. J. Anal. At. Spectrom. 1992 7 951. Shimamura T. Takahashi T. Honda M. and Nagai H. J . Anal. At. Spectrom. 1993 8 453. King F. L. and Harrison W. W. Mass Spectrom. Rev. 1990 9 285. Harrison W. W. in Inorganic Mass Spectrometry eds. Adams F. Gijbels R. and van Grieken R. Wiley New York 1988. Coburn J. W. and Kay E. Appl. Phys. Lett. 1971 18 435. Donohue D. L. and Harrison W. W. Anal. Chem. 1975,47,1528. Duckworth D. C. and Marcus R. K. Anal. Chem. 1989,61,1879. Garcia Alonso J. I. Thoby-Schultzendorff D. Giovannone B. and Koch L. J. Anal. At. Spectrom. 1993 8 673. Cantle J. E. Hall E. F. Shaw C. J. and Turner P. J. Int. J. Mass Spectrom. lon Process. 1983 46 11. Daly N. R. Reu. Sci. lnstrum. 1960 31 264. Milton D. M. P. Hutton R. C. and Ronan G. A. Fresenius' J. Anal. Chem. 1992 343 773. De Bievre P. and Barnes I. L. Int. J. Mass Spec. Ion Proc. 1985 65 211. Paper 3/05230C Received August 31 1993 Accepted November 2 1993