673 JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY AUGUST 1993 VOL. 8 Performance Characteristics of a Glove Box Inductively Coupled Plasma Mass Spectrometer for the Analysis of Nuclear Materials Jose lgnacio Garcia Alonso Dominique Thoby-Schultrendorff Bruno Giovanonne and Lothar Koch Commission of the European Communities Joint Research Centre Institute for Transuranium Elements Postfach 2340 76125 Karlsruhe Germany Helmut Wiesmann Spectrotec GmbH Mainzerstrasse 46 64572 Buttelborn Germany A commercial inductively coupled plasma mass spectrometer (Elan 250) was modified in order to analyse nuclear materials in a glove box. The nebulizer plasma torch and sliding interface are situated inside the glove box while the mass spectrometer and associated electronics are outside. The sensitivity of the modified instrument is slightly reduced compared with the original owing to a flange that separates the mass spectrometer from the vacuum interface.This has modified the original distance between the skimmer cone and the ion lens system. The plasma torch is mounted in a fixed position and the load coil is now separated 25 mm from the tip of the sampling cone. Optimum plasma operating conditions stability of the signal and isotopic ratios levels of oxide and hydroxide polyatomic ions were evaluated in the modified instrument for selected fission products and actinides. The effect of the ion lens settings on sensitivity and mass discrimination were studied in detail. Interference effects due to heavy matrix elements (U and Pu) were also studied.Keywords Inductively coupled plasma mass spectrometry; glove box; nuclear material; performance characteris tics The inductively coupled plasma mass spectrometer is an analytical tool which a priori would be ideally suited for the analysis of materials of nuclear origin. The elemental and isotopic capabilities of inductively coupled plasma mass spectrometry (ICP-MS) could be fully applied to those materials having non-natural isotopes and/or non-natural isotopic abundances. The low detection limits provided by the technique allow the use of very dilute samples with reduced radiation risks for the operator compared with other atomic spectrometric techniques (e.g. ICP atomic emission spectrometry and atomic absorption spectrome- try) used currently in the nuclear field.' The ICP-MS technique can be used to determine both stable and radioactive nuclides simultaneously and can discriminate between natural and artificial sources by the use of isotopic abundances which is an added advantage over other optical spectroscopic techniques.Also the method of isotope dilution can be applied and in many cases natural elements can be used as spikes (e.g. for the analysis of fission products) because of the different isotope abun- dances in the sample. In spite of all these potential advantages very few reports have been published on the use of ICP-MS in the nuclear field.2-13 Only the analysis of natural impurities in non-irradiated uranium 0xide~9~ and zirconium alloy^^-^ by ICP-MS have been reported. Non-natural isotopes have been analysed by ICP-MS in environmental ~amples,~-~O which shows the suitability of the technique for the analysis of long-lived radionuclides compared with radioanalytical techniquesg The use of an ICP mass spectrometer installed in a glove box has also recently been described by Crain and GallimorelI for the analysis of Tc in nuclear materials.Smith et a1.12 compared the intrinsic sensitivity of ICP-MS and radiochemical methods for the determination of radionuclides. It was demonstrated that for half-lives longer than a few hun- dred years ICP-MS can provide lower detection limits than the corresponding counting technique. The ICP-MS technique has been applied recently for the determination of traces of Np in enriched U solutions.13 The tailing of the 238U peak at the low mass side requires separation of U and Np for the determination of ppb levels of Np in U solutions. Nuclear materials to be analysed for trace elements and isotopes include mainly spent nuclear fuel (fission products and actinides) fresh fuels (natural element impurities) input and output solutions from reprocessing plants (minor actinides) high level waste from reprocessing and leachates of spent fuel and high level waste glasses.The presence of a heavy actinide matrix element (U Pu) has to be taken into account for the development of analytical methodologies for ICP-MS because of possible matrix interferences (space- charge effects). For the analysis of these materials other practical difficulties which arise from 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 protection) and/or hot cells (alpha beta 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 is leaked either to the laboratory or to the environment. Whole 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 of the glove box and only the samples and the corresponding sampling stage are contained in the box. For atomic spectrometry instruments only the atomization source (plasma graphite furnace etc.) is installed inside the glove box and the rest of the instrument remains outside.For optical instruments con- tainment is easily achieved by the use of optical lenses or fibres built into the walls of the glove box. However when MS measurements are made an opening has to be provided for the ions and absolute containment is no longer achieved. For ICP-MS measurements modification of the commercial instruments is necessary to minimize contami- nation risks. In this paper a modified ICP-MS instrument working since 1988 in a glove box and its performance characteristics for the analysis of radioactive materials of nuclear origin is described.674 JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY AUGUST 1993 VOL.8 Experimental Instrumentation The ICP-MS instrument used is an Elan 250 from SCIEX (Canada). Fig. 1 shows a schematic diagram of the glove box in which the instrument is installed. The glove box is provided with an extraction system cooling water and filters for all argon streams. Absolute filters (>99.99% efficiency) are situated both inside and outside of the glove box which is kept at a lower pressure than the exterior (20 mm water column). The peristaltic pump nebulizer assembly plasma torch and sliding interface are situated inside the glove box. The interface with the sampling and skimmer cones is mounted on the stainless- steel left wall of the glove box. The interface is evacuated by an additional rotary pump of the same type as the original one of the instrument which is now used only for pumping down and cryo-cleaning.Absolute filters are installed in the vacuum line inside the glove box to prevent any radioactive material from the plasma con- taminating the primary pump. Owing to security require- ments in the box the load coil had to be mounted in a fixed position at a distance of 25 mm from the sampling cone which is suitable for long ICP torches (the usual distance for the Elan 250 is 15 mm). The plasma off-gases (exiting the plasma) are cooled in a radiator-type water cooling system above the plasma torch before being evacuated through filters to the ventilation system of the laboratory. All vacuum pump outlets are also connected to the ventilation system. Pressure water and tempera- ture detectors are installed in the box.The ICP torch is protected by a moving stainless-steel cover. In order to facilitate maintenance the nebulizer and spray chamber are separated from the torch by a 20 cm long glass tube (1 0 mm i.d.) provided with an additional tangential argon flow (about 0.1 1 min-l) in order to minimize deposition of spray droplets on the wall (Fig. 1). Fig. 2 shows a detail of the connection of the glove box to the quadrupole mass spectrometer. The interface had to be separated from the mass spectrometer by a flange which connects the glove box to the MS housing with a V-type retainer coupling. In case of maintenance the mass spectro- meter could be easily separated from the glove box which with the interface closed remains gas tight.The use of this flange increased by 15 mm the original distance between the skimmer cone and the ion lens assembly in the mass A nn n n n Fig. 1 Schematic diagram of the ICP-MS installed in the glove box A channel electron multiplier; B ion deflector; C quadrupole; D cryoshells; E ion lens; F sliding interface; G r.f. induction coil; H plasma torch; I off-gas cooler; J absolute filters off-gas; K absolute filters sliding interface; L cooling water in-out; M glove box manometer; N nebulizer and spray chamber; 0 peristaltic pump; P liquid sample; Q absolute filters inlet; and R glove box frame A Fig. 2 Detail of the connection between the mass spectrometer and the glove box (scale not accurate) A mass spcctrometer housing; B two-piece flange; C V-retainer coupling; D stainless- steel wall of glove box; € photon stop; F original connection piece to mass spectrometer housing showing O-rings; G sliding interface (horizontal movement); H skimmer cone; and I sampling cone spectrometer. The analytical effect of this flange is a reduction of the ion transmission compared with the original instrument.In the interior of the glove box the interface housing is installed as in the original instrument but in a horizontal configuration and is moved by pneuma- tic activation. The series of O-rings installed in the interface and flange assure that no radioactive contamination can leak outside of the instrument. Maintenance An opening for the ion beam must exist during the analysis by ICP-MS. This means that the ions that enter the interface and mass spectrometer do cause radioactive contamination.Most of the contamination is confined to the interface itself which is situated inside the glove box but some reaches the mass spectrometer. Catastrophic failure of the channel electron multiplier (CEM) in December 1990 required the MS housing to be opened in order to check for contamination in the vacuum chamber. Firstly argon was connected to the spectrometer in order to bleed the chamber and passed through a filter in the outlet of the vacuum chamber. The filter was measured for alpha and beta contamination and no measurable contamination was found. With the sliding interface closed the V-retainer coupling was removed from the flange (Fig. 2) the glove box was separated from the mass spectrometer and the housing could be removed.Plastic surgical gloves and gas masks were used during all the maintenance work. Critical sections like the interface outlet cryo-shells quadrupole and CEM housing were checked for contamination by wiping tests. Light contam- ination was only found at the interface outlet where the photon stop is located (Fig. 2). Finally the CEM was replaced satisfactorily. In March 1992 contamination on the helium line was observed which created malfunctioning in the cold head. The cold head was decontaminated the adsorber changedJOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY AUGUST 1993 VOL. 8 675 and new high purity helium introduced into the system. No radioactive contamination was found in any part of the cryo-pump which shows the good performance of this type of pump for radioactive applications.No contam- ination was found in the vacuum oil for the primary pumps. Operating Conditions The plasma and quadrupole operating conditions used are summarized in Table 1. Different radiofrequency (r.f.) power and nebulizer flow settings were used to test M+ MO+ and MOH+ signals. The optimum plasma com- promise conditions are those listed in Table I and are similar to those proposed by the manufacturer. Different ion lens settings were studied as shown in Table 1. The optimization procedure proposed by the manufacturer (optimum signal for lo3Rh and similar signals for 'Li and *OSPb) (a) was compared with single element optimizations for lo3Rh (b) and *W (c). Sensitivity curves and mass discrimination errors were evaluated for various ion lens settings.The measurements were made in scanning mode low resolution 1 point per peak and 1 s integration time. Generally five scans were run for each experiment and the data transferred to a PC for evaluation using an electronic spreadsheet. Reagents and Materials Used Stock solutions of natural elements were obtained from Spex (Grasbrunn Germany) as 1000 ppm standards and diluted as necessary with 1% nitric acid. Neptunium-237 and 239Pu standards were obtained from Los Alamos National Laboratory (NM USA) as the metal or oxide and dissolved in nitric acid to prepare stock solutions. These solutions were standardized by titrationI4 and handled in glove boxes. Pure Am and Cm oxides were obtained from the Lenin Research Institute for Atomic Reactors (St Petersburg Russia) and dissolved also in nitric acid.Nitric acid was Merck Suprapur (Darmstadt Germany) and Milli- Q water (Millipore Eschborn Germany) was used through- out this work. Inactive standards were prepared by simple dilution in acid-washed calibrated flasks. Radioactive samples and standards were diluted by mass in the glove box or in the Table I Operating conditions R.f. power/W Reflected power/W Argon outer gas flow ratell min-' Argon intermediate gas flow Argon nebulizer gas flow ratell m h - ' Sample uptake rate/ml min-I Nebulizer type Spray chamber Load coil/sampler cone distancelmm Interface pressure/Pa Quadrupole working pressure/Pa Sampler and skimmer cones rate/l min-I 1200 t 5 12 1.4 0.86 (30 psi) 1 Meinhard Scott type double pass 25 (fixed) 266.64 266.64~ Nickel Settings (O/o range) Lens Range/V (a) (b) (4 Bessel box B 0 +10 39 70 72 Bessel box P 0 -60 14 15 15 Einzel lens E l 0 -20 62 25 14 Photon stop S2 0 -20 36 10 25 hot cell facility.Polyethylene bottles (10 or 20 ml) were used for all radioactive material. Results and Discussion In order to study the performance of the modified instru- ment for the analysis of nuclear samples a series of tests were carried out including sensitivity stability of signals and isotopic ratios suitable internal standards optimum plasma operating conditions oxide and hydroxide ion levels mass discrimination errors and matrix interferences due to heavy elements. Sensitivity and Stability For the sensitivity and stability studies a solution contain- ing 100 ng ml-' of Li Cu Rh Cd Pb and U was nebulized for 2 h and the intensity values for each isotope measured every 5-10 min approximately (n= 19).The results ob- tained for the raw intensities can be compared with similar unmodified Elan 250 instruments using the same rec- ommended plasma operating conditions and similarly optimized ion lenses [settings (a) in Table 13. The results obtained which appear in Table 2 show that the intensity counts are lower by a factor of 2-10 compared with other Elan 250 instruments installed in Germany during 1986- 1988 depending on the particular element and instrument considered. This reduction in sensitivity must be due to the change in the original distance between the skimmer cone and the ion lens assembly by the use of the flange depicted in Fig. 2.However the background counts were not affected so an increase in the detection limits is observed in the modified instrument. The relative standard deviations (RSDs) using each element in solution as internal standard (IS) for the other elements were between 2 and 3%. As expected lower values were found when the mass of the IS was close to that of the other element. Rhodium-103 as IS seemed to give better results than the other elements as far as RSD is concerned. The RSD of the Cu (65:63) Cd (1 12 1 14) and Pb (206:208) isotopic ratios ranged between 1.7 and 1.9% for the 100 ppb standard after 2 h of measurement owing to the poor counting statistics. Significant differences between the expected and observed isotopic ratios were also found which demonstrates the existence of mass discrimination errors.Plasma Operating Conditions Plasma operating conditions were studied for the elements discussed earlier also using configuration (a) (Table 1) of the Table 2 Comparison of sensitivity for different Elan 250 instru- ments Raw intensity*/ions s-' Other instrument* Isotope This instrument? 'Li 2 5 32( 5.3) 12500 8500 15380 63CU 472 5(4.3) 9500 17000 31890 Io3Rh 1258 l(4.8) - - 125400 '14Cd 2193(4.9) 7000 9000 - 208Pb 39 15(4.5) 12000 12000 18820 238U - - - 8 2 7 5( 4.7) * Obtained for 100 ppb solutions of the corresponding elements. 7 Values in parentheses are RSD(%) for 2 h of measurement. $ Elan 250 installed in Germany during 1986-1988 by H. Wiesmann. The data are from performance tests after installation.676 JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY AUGUST 1993 VOL.8 3000 I 1 2500 2000 1500 1000 500 0 A 12000 fn fn 0 3 8000 c .- fn C c 4000 0 4000 3000 2000 1000 18 22 26 30 34 6000 5000 4000 3000 2000 1000 0 3000 2500 2000 1500 1000 500 0 8000 6000 4000 2000 0 I (4 I I ~ B A 18 22 26 30 34 Nebulizer pressure/psi Fig. 3 Influence of plasma operating conditions on the response for (a) 'Li; (6) 63Cu; (c) lo3Rh; (d) lI4Cd; (e) zOsPb; and cf) 238U. A nebulizer pressure of 30 psi ( 1 psi=6895 Pa) corresponds to 0.86 1 min-' Ar. R.f. power A 900; B 1000; C 1100; D 1200; E 1300; and F 1400 W ion lenses. The results obtained for the influence of r.f. power on the optimum nebulizer pressure are shown in Fig. 3. These results are comparable to previously published data15J6 showing the interdependence of applied power and nebulizer flow rate.However there is a clear mass depen- dent difference going from Li to U. In the case of Li [Fig. 3(a)] the maximum intensity in the nebulizer pressure curve for each r.f. power setting increased drastically with increasing r.f. power so no optimum conditions for Li could be found. In the case of U [Fig. 3 0 1 the maximum intensity is independent of the r.f. power setting above 1 100 W. For the other elements intermediate values were found betwen these two extremes. Similar results have been obtained p r e v i o u ~ l y ~ ~ ~ ~ ~ for the Elan 250 using both the old and improved ion lens system. Two effects seem to play a role in the optimization of the plasma parameters.Firstly the r.f. applied power has to be balanced with the nebulizer flow in order to obtain maximum ionization at the tip of the sampler cone (interdependent effect) and secondly the change in the plasma applied power could modify the kinetic energy of the ions as they enter the interface. Fulford and DouglasI7 reported that the kinetic energy of the ions depends on their mass on plasma space potential and possibly on ion beam space charge. They suggested a value of 2 eV for the plasma potential which should be added to the kinetic energies as determined by the supersonic expansion. This kinetic energy appeared to be unaffected by changes in plasma operating ~0nditions.l~ Based on the results obtained in the present study for the instrument used here the kinetic energy of the ions seems to be affected by increasing applied power but only to a limited extent.However the relative change in kinetic energy would be greater for lighter than for heavier elements and that would affect ion transmission of lighter elements preferentially; increasing the kinetic energy of lighter ions will increase their transmission through the ion lens system which is optimized for heavier ions (higher kinetic energy). This effect will be less noticeable for heavier ions as their kinetic energy is governed mainly by supersonic expansion. Fur- ther experiments on ion kinetic energy measurements such as those of Chambers and Hieftje,l8 for different plasma applied powers are needed to prove this assumption. The influence of operating conditions on the levels of oxide hydroxide and doubly charged ions was studied using Zr La Nd Th U Np Pu Am and Cm as test elements.The first three elements were selected because of their high fission yield which can produce isobaric inter- ferences on other MO+ or MOH+ lower yield isotopes. The levels of oxide and hydroxide ions for transuranium elements are reported here for the first time. Peaks for hydrides (MH+) were not detected for the actinides. The ion lenses used were those of configuration (b) (fission products) and (c) (actinides) in Table 1. Fig. 4 shows the results obtained for La Nd and Zr at 1200 W forwardJOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY AUGUST 1993 VOL. 8 677 ( b ) 0.05 1 0*04 t 0.03 2 0 E 0.02 0.01 Nd 0 ' I 1 I I I I I 0.05 ( C) 0- - I I - A 20 22 24 26 28 30 32 34 36 Nebulizer pressu re/psi Fig.4 Levels of M+ MO+ and MOH+ for Zr La and Nd for different nebulizer pressures at 1200 W forward power. (a) M+ raw intensities for the ions shown at the 2 ppm levels. (6) Intensity ratios MO+:M+ 106:90 155 139 and 158 142 for Zr La and Nd respectively. (c) Intensity ratios MOH+:M+ 109:92 156 139 and 163 146 for Zr La and Nd respectively power and different nebulizer pressures. As can be ob- served levels of oxides and hydroxides are below 1 O/o under the standard operating conditions (Table l) which for La and Nd are much lower than previously published results for the Elan 250.19320 The results obtained for the actinides are summarized in Table 3. As can be seen the levels of oxide are reduced drastically from Th to Cm.The levels of hydroxides are lower than 0.05% in all cases. The use of a 25 mm distance between the load coil and the sampler might be responsible for the low oxide levels observed. The determination of 239Pu in the presence of high amounts of 238U can be hindered by the presence of UH+ isobaric interferences. In the instrument used here the 238UH+ peak is less than 0.005°/6 of the 238U+ peak. Effect of Ion Lens Settings on Mass Discrimination The determination of the isotopic abundances of the elements produced by fission in nuclear samples is of great importance in nuclear research in monitoring reactor performance. Traditionally this is carried out by thermal ionization mass spectrometry (TIMS) after chemical separ- ation of the elements.The ability of ICP-MS to provide Table 3 Levels of oxides and hydroxides for actinides under standard operating conditions Element MO+:M+ MOH+:M+ Th 0.026 - U 0.016 - NP 0.0 12 t 0.000 5 Pu 0.005 t0.0005 Am 0.00 1 t0.0005 Cm 0.002 <0.0005 quick and reliable isotopic information for most of the fission products and actinides could provide a new method for this type of study. In order to investigate the precision and accuracy achievable by ICP-MS in the determination of isotope abundances a solution containing natural Li B Mg Ti Cu Rb Mo Sn Ba Nd Eu Gd Yb W Pt Ir T1 and Pb (all 1250 ng ml-l in order to improve counting statistics) along with solutions of certified U isotopic abundances [National Bureau of Standards (NBS) (now National Institute of Standards and Technology (NIST) U350 U500 and U9001 and Pu (in-house standard ana- lysed by TIMS) were tested for mass discrimination errors.Three different ion lens settings were used in the measure- ments [configurations (a)-(c) in Table I] under the standard plasma operating conditions. A series of five scans with 1 point per peak and 1 s integration time per point were measured. Systematic differences between the experimental and theoretical isotope ratios that depended on the ion lens settings were observed. Longerich et alto showed that mass discrimination errors were proportional to the mass differ- ence between the two isotopes used to calculate the ratio. The same results were observed by us for polyisotopic elements but the slope of the error curve was different for different mass ranges and seemed to be independent of the element considered for elements having similar mass. The relative error of the isotopic ratios could be expressed as where IRexg and IRthm are the observed and theoreticalt1 isotope ratios between the isotope of mass M and the isotope of mass M + M respectively and K is the mass discrimination factor.If the expression is transformed to and the K value is calculated for all elements and isotopes present in the solution tested the variation in K for the entire mass range is obtained. The mass discrimination factor K indicates the relative error produced in the measurement of isotopic ratios per mass unit of difference between the isotopes considered. For our calculations the isotopic ratios were always referred to the main isotope for each element and the K values plotted versus the mass of the less abundant isotope.The results obtained for the three ion lens configuration studied are illustrated in Fig. 5. As can be observed in all cases there is a trend of negative mass discrimination errors for the light elements and positive errors for the heavy elements depending on the ion lens configuration used. When the ion lens optimization recommended by the manufacturer (maximum signal for Rh and similar signals for 'Li and 208Pb) is followed mass discrimination errors are important for the actinides [Fig. 5(a)]. When optimizing only for Rh optimum settings for the Besel box B lens and the Einzel lens El change dramatically as can be seen in Table 1. Under those conditions mass discrimination errors increase for elements below mass 100 but decrease for those above that mass.678 4 - JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY AUGUST 1993 VOL.8 (a) 0.08 -0.08 1 I 1 - ( a ) rn 0.04 1 3 140 100 - e 0 0 60- 20 0 . I - .rr. .. La Nd E~ -. -'+-= *.- Sr Mo Rh Pd - I I I I I I 1 0.08 1 0 50 100 150 200 250 m/z Fig. 5 Variation of the mass discrimination factor K between B and Pu for the different ion lens settings (a) (b) and (c) from Table 1 Finally when optimizing only for U mass discrimination errors are eliminated for the actinides but aggravated for the middle of the Periodic Table. The determination of correct isotopic abundances for fission products and actin- ides is strongly influenced by the ion lens settings employed.In order to clarify the role of the ion lenses in mass discrimination the raw intensity data have been converted into ions s-' per pmol of each isotope taking into account the elemental concentration in the solution the atomic mass of the considered isotope and its known isotopic abundance. The so-called molar sensitivity curve for the different ion lens configurations can be seen in Fig. 6. Note that no ionization corrections were taken into account and that affects the response for W Ir and Pt principally. As can be observed there is not a uniform response for all isotopes and elements. Light elements (Li B Mg etc.) show in all cases a much lower response than heavier elements. The maximum sensitivity and its location on the response curve depends on the ion lens used.Also for isotopes of the same element increasing or decreasing response can be obtained which explains the sign and extent of the mass discrimination factors observed. In Fig. 6(a) the slope of the sensitivity curve is lower than in Fig. 6(b) and (c) and hence mass discrimination errors are smaller. For Pu a decrease in sensitivity is observed in Fig. 6(a) and (b) while constant values are obtained in Fig. 6(c) where the slope of the sensitivity curve is steep for the middle of the Periodic Table. The molar sensitivity curves of Fig. 6 show that ion transmission and/or detection for the modified Elan 250 is strongly mass dependent which is also a characteristic of most ICP-MS instruments where the ion kinetic energy is obtained only by supersonic expansion of the plasma gases (i.e. no accelerating electrodes are present). For the analysis of fission products in spent fuel the ion Yb Pu Sn M~ # Eu Ir 5 t Ti Cu I 0 50 100 150 200 250 mlz Fig. 6 Molar sensitivity curves for the different ion lens settings (a) (b) and (c) from Table 1 180 t Fig. 7 Influence of 500 ppm of U on the response for selected fission products. All isotopes of the elements were considered lenses used were those of configuration (b) (Table I) which gave adequate sensitivity [Fig. 6(b)] and constant mass discrimination errors [Fig. 5(b)] for the range to be explored [note that the sensitivity curve in Fig. 6(b) can be assumed linear between 80 and 160 u]. Actinides at low concentra- tions were analysed using the lens optimized for U which gave best sensitivity and minimum mass discrimination.Future reports will show examples of application to various nuclear samples. Matrix Interferences Interferences from heavy elements (U and Pu) were evaluated for the analysis of natural element impurities fission products and other actinides. The effect of levels of U up to 500 ppm on the signal was studied for selected fission products covering the mass range 80- 160 u. Higher concentrations of U were not evaluated as the radiation679 JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY AUGUST 1993 VOL. 8 I 180.00 140.00 1' % 2 20.00 B ~ 0 40 80 120 160 200 240 mlz Fig. 8 Recovery for 18 natural elements spiked into Pu solutions 1 1700 ppm of Pu; and 2,2000 ppm of Pu Several isotopes of the elements were considered for the analysis using aqueous calibra- tion solutions (1% nitric acid) 600 I A 0 0 100 200 300 400 500 600 Certified concentration/pg g-'of U Fig.9 Comparison between certified and found concentrations for U30s reference materials (NBL 95.1 and 95.2). Final U concentration 2000 ppm. Line of slope 1 is shown. A calibration with aqueous standards (slopex0.7); and B calibration using Rh as internal standard dose of the corresponding spent fuel solution would be too high for glove box analysis. As can be observed in Fig. 7 no effect of the presence of 500 ppm of U can be observed on the signal for 1 ppm of the corresponding fission products between Sr and Eu. In the analysis of natural element impurities in fresh nuclear fuel higher concentrations of U or Pu can be tested.Fig. 8 shows the recovery for natural elements which can be present as traces in actinides in the presence of (1) 1700 or (2) 2000 ppm of Pu. The recovery increases as the mass of the test element increases and quantitative recoveries were observed for the elements above Gd regardless of the concentration of Pu. Fig. 9 shows the results for the analysis of two U30s reference materials [New Brunswick Laboratory (NBL) NJ USA samples 95-1 and 95-21 containing different levels of impurities and prepared to contain 2000 ppm of U. As can be observed the recovery is about 70% for most elements using aqueous calibration solutions (based on the slope of the line A). However when Rh was used as IS satisfactory results were obtained. In the case of the actinides no matrix effect of U on Pu or Np was observed up to 2000 ppm of U in the sample.However low concentrations of 237Np in the presence of a high U matrix required manual adjustment of the mass resolution due to the tailing of the 238U peak at the low mass side. For the determination of ultratrace levels of Np in U chemical separation of the elements would be necessary.I3 Conclusions The performance characteristics of a glove box ICP mass spectrometer have been found to be similar in most respects to standard instruments. The reduction in sensitivity due to the glove box installation is perhaps the most striking effect but other parameters like the low oxide levels and the low interferences observed make up for this loss. For the samples to be measured the sensitivity can be considered adequate. Mass discrimination errors were studied in detail as the isotopic abundances of fission products and actinides have to be measured.Correct isotopic abundances can be obtained by the use of the mass discrimination factor K when the instrument is calibrated for the appropriate mass range. The interference effect arising from high concentration of heavy elements in the sample is perhaps the main drawback when applying ICP-MS to the analysis of real samples. In the present case owing to the lower sensitivity space- charge effects in the mass spectrometer seemed to be drastically reduced and no serious interference effects were observed in the presence of high concentrations of U and Pu which a priori would show the most pronounced effects.The limited interference effects would facilitate the use of semiquantitative approaches when high precision and accuracy is not necessary. 1 2 3 4 5 6 7 8 9 10 1 1 12 13 14 20 21 References Nickel H. Spectrochim. Acta Part B 1992 47 27. Cantle J. E. in Analytical Chemistry Instrumentation ed. Laing W. R. Lewis Publishers Chelsea MI 1986 p. 139. Allenby P. Clarkson A. S. and Makinson P. in Analytical Chemistry Instrumentation ed. Laing W. R. Lewis Publish- ers Chelsea MI 1986. Long S. E. and Brown R. M. in Analytical Chemistry Instrumentation ed. Laing W. R. Lewis Publishers Chelsea MI 1986. Luo S. K. and Chang F. C. Spectrochim. Acta Part B 1990 45 527. Beck G. L. and Farmer 0. T. 111 J. Anal. At. Spectrom. 1988 3 771. Kim C. Takaku A. Yamamoto M.Kawamura K. Shi- raishi K. Igarashi Y. Igarashi S. Takayama H. and Ikeda N. J. Radioanal. Nucl. Chem. Art. 1989 132 131. Scott R. D. Baxter M. S. Hursthouse A. S. MacKay K. Sampson K. and Toole J. Anal. Proc. 1991 28 382. Kim C. K. Morita S. Seki R. Takaku Y. Ikeda N. and Assinder D. J. J. Radioanal. Nucl. Chem. Art. 1992,156,201. Hursthouse A. S. Baxter M. S. McKay K. and Livens F. R. J. Radioanal. Nucl. Chem. Art. 1992 157 281. Crain J. S. and Gallimore D. L. Appl. Spectrosc. 1992 46 547. Smith M. R. Wyse E. J. and Koppenaal D. W. J Radioanal. Nucl. Chem. Art. 1992 160 341. Riglet C. Provitinia O. Dautheribes J.-L. and Revy D. J. Anal. At. Spectrom. 1992 7 923. Cromboom O. Garcia Alonso J. I. Koch L. Goerten J. Roesgen E. Wagner H. G. Ottmar H. and Eberle H. Proceedings of the 4th International Conference on Facility Operations-Safeguards Interface ed. American Nuclear Society Illinois 1992 p. 431. Horlick G. Tan S. H. Vaughan M. A. and Rose C. A. Spectrochim. Acta Part B 1985 40 1555. Vaughan M. A. Horlick G. and Tan S. H. J. Anal. At. Spectrom. 1987 2 765. Fulford J. E. and Douglas D. J. Appl. Spectrosc. 1986 40 971. Chambers D. M. and Hieftje G. M. Spectrochim. Acta Part B 1991 46 761. Longerich H. P. Fryer B. J. Strong D. F. and Kantipuly C. J. Spectrochim. Acta Part B 1987 42 75. Longerich H. P. Strong D. F. and Kantipuly C. J. Can. J. Spectrosc. 1986 31 1 1 1. White F. A. and Wood G. M. Mass Spectrometry Applica- tions in Science and Engineering Wiley New York 1986 pp. Paper 3/0 I0 I OD Received February 19 1993 Accepted April 19 1993 759-762.