Determination of Long-lived Radioisotopes Using Electrothermal Vaporization-Inductively Coupled Plasma Mass Spectrometry* JORGE S. ALVARADO AND MITCHELL D. EKICKSON Environmental Research Division Argonne National Laboratory 9700 South Cass Avenue Argonne IL60439 USA A general method for the determination of long-lived radioisotopes by integrating electrothermal vaporization and inductively coupled plasma-mass spectrometry (ETV- ICP-MS) to vaporize environmental samples with complex inorganic matrices is described. The method required no sample pretreatment and minimized sample size. The rationale was to use chemical modifiers such as CHF to form metal fluorides with much lower boiling-points than other metal compounds (such as oxides and carbides). Given sufficiently high temperatures and long reaction times samples in other chemical forms are converted into elemental halides and vaporized.The characterization and application of ETV- ICP-MS for the determination of radioisotopes is described. The detection limits for 99Tc 238U u6U U2Th 230Th and 226Ra were similar to those obtained with ultrasonic nebulization ( USN-ICP-MS). Absolute detection limits ranged from 0.6 fg for 226Ra to 5 fg for 238U. Analytical calibration plots were linear over a range of 2-3 orders of magnitude. Matrix effects caused by Group IA and IIA elements were minimized by changing the nature of the sample and by using temporal-thermal programming without affecting analytical performance. Comparison studies between ETV- ICP-MS and classical radiometric techniques were performed for various environmental samples.Keywords Electrothermal vaporization; sample introduction; inductively coupled plasma mass spectrometry; long-lived radioisotope; chemical modifier ICP-MS has proved to be a cost-effective trace analytical technique for many transition elements but has only recently been applied to the radiochemical field. This technique offers advantages such as low detection limits (typically n,g 1-I) mass-selective detection and multicomponent detection. In determining long-lived radionuclides ICP-MS is rapidly sur- passing other techniques such as differential-pulse polarogra- phy,' radiochemical neutron activation,2 ion chromatography3 and classical ph~tometry.~ The main advantage of ICP-MS is its ability to determine (1) long-lived radionuclides with low- intensity radiation and (2) alpha-emitting radionuclides that require tedious radiochemical separations.Despite the fact that ICP-MS itself was a major improve- ment techniques of sample introduction are critical for the performance of the plasma source. Browner and Boorn5v6 define the goals of sample introduction as 'the reproducible transfer of a representative portion of sample material to the atomizer cell with high efficiency and with no adverse inter- ference effects.' Owing to its simplicity and high reproducibility * The submitted manuscript has been authorised by a contractor of the US Government under contract No. W-31-109-ENG-38. Accordingly the US Government retains a non-exclusive royalty-free license to publish or reproduce the published form of this contribution or allow others to do so for US Government purposes.Journal of Analytical Atomic Spectrometry pneumatic nebulization (PN) is the most common method of introducing aqueous samples; however the efficiency is poor with reported typical efficiencies for PN of 1% with ICP spe~trometry.~ Ultrasonic nebulization (USN) has been used as a replacement for PN. Although USN has a more efficient production of droplets ranging up to 30% the instrument has longer clean-out times and tends to enhance matrix effects. Since its introduction electrothermal vaporization (ETV) has also been used for the introduction of samples into the plasma. Some of the advantages over traditional methods of sample introduction are high analyte transport efficiency producing higher sensitivity; reduction of polyatomic ion inter- ferences; reduction of oxide formation of the analyte plasma gas and matrix components; small volumes of sample required; ability to analyse organic liquids strong acids liquid high in solids and slurries; reduction of non-spectroscopic interferences through selective volatilization; and improved performance using chemical modification techniques.However the poor reproducibility and the formation of refractory molecular species involving the furnace (or cup) material are considered major disadvantages. Physical carriers such as the halogens and halocarbons have been used as aids to volatilization in order to reduce losses and improve sample transport. Ng and Caruso7 showed that pyrolytic carbon cups typically used during the ETV process exhibit memory effects multiple peaks and broad bands that can be minimized by using a tantalum coating procedure.8 Studies by Barnes and Fodorg and by Gun et a1.I' reported that when nickel was added elements such as selenium and arsenic owing to their volatility were retained during the ashing process increasing memory effects." Ediger and Beres12 performed an interesting experiment with sodium chloride magnesium nitrate palladium nitrate and tellurium nitrate as chemical modifiers.This study was performed on 20 elements and the amount of modifier present was more important than the physical characteristic of the modifier itself. The major exception here was uranium which seemed to show large suppressions for all of the solution-based modifiers evaluated except for palladium in moderate amounts.Fonseca and Miller- Ihli13 used palladium as a physical carrier to reduce the difference in response between samples and aqueous standards in slurry analysis. Nickel and Zadg~rska,'~ using an ETV device showed that the total evaporation of impurities in ceramic powders was achievable by mixing the sample with BaO and CoF,. Although many of these experiments showed promising results an increase in memory effects and the matrix compo- nents make the efficacy of chemical modification techniques very specific to the determination of specific analytes in specific matrices. In order to create a general procedure for the determination of any element in any type of sample and to be able to compare with standards a method that eliminates the matrix matrix interferences and memory effects should be Journal of Analytical Atomic Spectrometry October 1996 Vol.11 (923-928) 923studied. Ren and Salin" described the determination of trans- ition elements by using Freon 12 for the analysis of sediments and coal fly ash. Experimental results showed that by using a gas halogenation reagent nearly 100% vaporization was achieved for samples with a high content of aluminium silica zirconium and tantalum. Matrix effects were reduced and the formation of oxides and carbides was minimized. The objective of this work was to develop a general method that can be used to introduce any sample for the determination of radionuclides in environmental samples by ETV-ICP-MS.Our approach builds on that described previously by Ren and Salin," where samples are vaporized in the presence of halogen- ated gas by using high temperatures to decompose the halogen- ated compounds and create fluoride compounds that have lower vaporization temperatures and to eliminate interferences from the formation of oxides and carbides. For these experi- ments CHF3 was used because of its high purity and longer reaction times. Minimal modification makes the instrument easy to use and the process convenient for automation. EXPERIMENTAL Instrumentation The ICP mass spectrometer used in these experiments was a PlasmaQuad I1 (VG Elemental Winsford Cheshire UK). Instrument operating conditions are listed in Table 1. Two types of sample introduction technique were employed a U-5OOOAT ultrasonic nebulizer (Cetac Technologies Omaha NE USA) and a VG Microtherm Mark I11 electrothermal vaporizer (also manufactured by Fisons Instruments).The electrothermal vaporizer consists of a graphite tube open on both sides to permit the passage of argon used as the carrier gas. Pyrolytic coated carbon tubes from Fisons Instruments were used. Some of the carbon tubes were soaked in a tantalum solution as described by Zatka16 and by Ng and Caruso' to form the tantalum carbide layer. The development of a gold colour on the surface of the cup indicated the formation of the tantalum carbide.8 In ETV the sample is introduced through a small inlet on top of the graphite tube which is closed by a graphite rod when vaporization is taking place.On either side of the graphite cup are two graphite electrodes that heat the graphite tube when high voltages are passed through them. The sample undergoes three processes drying ashing and atomization. Operating conditions are described in Fig. 1. Reagents The standard solutions of the radionuclides of interest were quantified by isotope dilution employing radionuclides trace- Table 1 Instrument operating conditions Instrument and parameter Operating value ICP-MS Fisons PlasmaQuad I1 ( VG Elementa1)- Incident power 1350 W Reflected power ow Coolant gas flow rate Auxiliary gas flow rate Carrier gas flow rate Ultrasonic nebulizer U-5000AT (Cetac Technologies)- Condenser temperature 0 "C Heater temperature 150 "C Sample uptake rate Argon carrier gas flow rate Trifluoromethane gas flow rate Sample injection volume 25 p1 13.8 1 min-' 0.94 1 min - ' 0.8-1.0 1 min-' 1.2 ml min-' 0.8-0.9 1 min-' 1.0 ml min-' Electrothermal vaporizer Microtherm Mark III ( VG Elemental)- "Flash" Program "Ramp" Program Timels TemperaturePC Timds TemperaturePC Drying 0-60 0-80 60- 120 80-110 Ashin 120-190 130-160 Va orizarion 169-160 160-168 ~~ ~~.168- 168 168-180 110-400 400-600 600-2600 2600-2600 2600-25 25 -25 Drying 0-40 0-100 40-50 100-100 Ashing 50-70 70-100 100-100 Va orizarion 115-118 106-1 15 118-1 18 118-128 100-750 750-750 750- 1 100 1100-2400 2400-2400 2400-25 25-25 Fig. 1 Electrothermal vaporization operating conditions able to NIST. Standard solutions of radium-226 (SRM 4958) technetium-99 uranium-238 uranium-236 and thorium-232 were used.Standard solutions were prepared by volumetric dilution of each radionuclide with 5% v/v nitric acid (Ultrex 11-Ultra Pure Reagent J. T. Baker Phillipsburg NJ USA) and distilled de-ionized water (18 mR cm Nanopure System Barnstead Thermolyte Dubuque IA USA). Matrix effects were studied by using nitrate salts of sodium calcium potassium and magnesium from J. T. Baker. Solutions containing 1-1000 mg 1-' of each element were prepared. Tantalum metal powder hydrofluoric acid (1 + 1) and oxalic acid dihydrate from Aldrich (Milwaukee WI USA) were used to prepare the soaking solution necessary to produce the tantalum carbide coatings. This soaking solution contained 6% m/v tantalum. The gases used in ICP-MS and ETV were liquid argon and 99.995 electronic-grade Halogen-23 (CHF3) from AGA Gas (Cleveland OH USA).Samples The following samples were used tap water from Chicago Illinois and Lemont Illinois; river water from the Fox River and Kankakee River in Illinois; well water from Lemont Illinois and Borden Indiana; lake water from Herrick Lake in Illinois; spring water; and groundwater from the Gaseous Diffusion Plant in Paducah Kentucky were used. The samples were collected in plastic bottles and acidified to a pH of about 2 with nitric acid. Samples were divided into two fractions and one of the fractions was analysed for uranium,I7 thorium,I7 radium,18 and technetiumlg by using radiochemical techniques for confirmation. RESULTS AND DISCUSSION Instrument Optimization ICP mass spectrometers are used more often for the analysis of liquids via continuous solution nebulization which produces a stable ion signal.In a traditional procedure for optimization instrument parameters such as gas flow rates and ion lens stack settings are tuned to obtain a maximum signal. Owing to the transient gaseous and 'dry' nature of the ETV signal and the changes in the physical and chemical conditions during each step in the vaporization process the optimization param- 924 Journal of Analytical Atomic Spectrometry October 1996 Vol. 11eters were shown to be different from the typical conditions for optimization. Different tuning parameters have been studied to obtain optimum conditions during the ETV process. Gray et aL20 used the 38Ar2 dimer to optimize ion lens voltages torch position and nebulizer flow rate.Further optimization was performed by using single-ion monitoring of 115i[n to compensate for the vaporization pulse effect. For these studies USN was used to optimize the instrument using a 0.5 pg 1-' solution of beryllium magnesium cobalt bismuth indium lead and uranium. The signal was maximized for '151n. The standard deviation was less than 2% with ten consecutive 60 s integrations. USN was used for optimization because the extra desolvation provided by this instrument makes conditions that resemble (to a first approximation) the dry plasma conditions in ETV. After the gross calibration the ETV device was calibrated for each individual element. The typical response is shown in Fig. 2. When the auxiliary gas flow rate was increased from 0.80 to 1.06 1 min-l a plateau was reached with no changes in the signal until high flow rates were used.A different response was obtained when the nebulizer gas flow rate was changed from 0.86 to 1.06 1 niin-' as shown in Fig.2. The signal proved to be optimum for a 0.05 ppb solution of 238U when a carrier gas flow r'rte of 0.96 1 min-' was used. This signal showed a maximum that is very sensitive to any changes in flow rate. A small change can produce a decrease in the signal of up to 25% of the optimum signal. Chemical Modifiers ETV has been shown to exhibit memory effects multiple peaks and broad bands when the pyrolytically coated tubes interact with the element under study. Initial observations showed a response similar to that shown in Fig.3(u) where a reproduc- ible broad band was obtained for 232Th. An explanation for these effects is the possible formation of carbide or oxide compounds (with much higher boiling-points) that compete in the vaporization-atomization process. The formation of tantalum carbide (mp = 3983 "C) is neces- sary to restrict carbide formation with the element under study. Atomization and carbide formation are competing processes. More specifically the tantalum carbide formed is more stable than the carbides of elements such as thorium arsenic uranium or radium. When atomization occurs the tantalum carbide layer prevents the formation of any other carbide species and as a consequence single and sharp peaks of the element under study are obtained. Unfortunately no noticeable change in the 2m i Argon flow/l min" Fig.2 Variation of uranium-238 (0.05 pg I-') signal as a function of nebulizer and auxiliary gas flow rates 40 36 32 28 24 20 16 12 8.0 4.0 $2 = o z Fig. 3 Vaporization behaviour of thorium-232 (a) argon as carrier gas; and (b) trifluoromethane as carrier gas signal occurred when the tantalum-coated graphite tube was used in place of the uncoated tube. The formation of carbide was not the source of the broad signal. Chemical modifiers are known to increase sample transport and to minimize the formation of other interfering compounds. Elements such as thorium and uranium have a tendency to form stable oxides with boiling-points as high as 2350 and 3390 "C for UOz and Tho2 respectively. Fluoride compounds are usually characterized by having lower boiling-points than most oxides ( UF6 = 64.8 "C; UF = 1400 "C).When chemical modifiers such as CHF3 were used a single sharp signal [Fig. 3(b)] was observed. When CHF3 is heated it decomposes to form free fluorine radicals that react with the element under study producing fluoride complexes with lower boiling-points. These complexes have a much higher volatility than the oxide compounds. In addition the CHFJ helps in the elimination of matrix components the reduction of memory effects and the reduction of the background levels. The fluorine radicals in addition to reacting with the analyte react with other parts of the matrix such as silicates carbides and oxides; for example the Si02 (bp = 2230 "C) present as particles in liquid samples and present as a major constituent in soils can be eliminated Journal of Analytical Atomic Spectrometry October 1996 Vol.11 925by the formation of SiF (bp = - 86 "C). In addition to improv- ing the signal for uranium and thorium a reduction in the background levels and in memory effects was observed. No signal improvement was observed for radium and technetium. Linearity and Detection Limits Detection limits were determined and the results are listed in Table 2. Detection limits were defined as the concentration that yields a signal-to-noise ratio of three. The noise was defined as the standard deviation of ten non-consecutive blank determi- nations. These detection limits were calculated three times on three non-consecutive days to test for reproducibility.Table 2 shows the detection limits in femtograms when no chemical modifier was used and when CHF was used as a chemical modifier. An improvement of two or three orders of magnitude was observed. Trifluoromethane did not improve the signal of radium or technetium but improvement in the background caused a slight improvement in detection limits. Differences were observed in the detection limits of similar isotopes. Uranium-238 showed detection limits five times higher than 236U. Because 236U is not a natural isotope in the environment lower back- grounds were observed and thus better detection limits. This behaviour and the detection limits obtained are comparable to the results obtained by USN-ICP-MS2' and with the exception of radium meet the minimum requirements for the detection of radioisotopes in the environment.Ebdon and Gooda1122 described the thermochemical effects in hexafluoroethane-modified argon and concluded that the introduction of halocarbons resulted in an apparently unusual plasma spectrochemistry which yields non-linear calibration graphs. This effect was attributed to classical mass action buffering involving highly stable metal fluoride compounds. In contrast to this earlier work,22 we found that when 25 pl aliquots of 226Ra 236U 238U "Tc 230Th and 232Th were introduced into the ETV device the system showed a linear response over two orders of magnitude for each element. The log-log slopes were 0.99 for radium 1.00 for thorium 0.99 for uranium and 1.07 for technetium. No 'roll-over' was observed at higher concentrations.Fig. 4 shows the calibration plot for each element. Matrix Effects ETV offers distinct advantages over conventional methods of sample introduction. The gaseous nature of the sample reduces the effects of solvent and sample matrix interferences. In addition the ability to use temporal-thermal programming allows selective removal of the sample matrix constituents. Matrix effects can increase signal intensities in ICP-MS by facilitating ionization processes; conversely matrix effects can quench signals. These competing results as well as secondary effects produced by changing plasma temperature such as analyte desolvation solute volatilization compound dis- Table 2 Detection limits for long-lived radioisotopes. *Detection limit is defined as three times the pooled standard deviation of ten non- consecutive blank analyses ETV unmodified j Isotope fg Uranium-238 180 Uranium-236 48 Thorium-232 1600 Thorium-230 -$ Radium-226 1 Technetium-99 -1 ETV trifluorome t hane/ fg 5 0.9 2 1.4 0.6 1.5 USNt/ fg 18 25 0.5 0.3 0.3 0.8 * Sample size 25 pl.t From ref. 21. -$ Not determined. 5.61 5.22 - 4.83- 7 I w 2 4.44 - $ 4.06- c *g 3.67 - cn ,O 3-28 - 2.89 - 2.50 ' -3150 -2193 -2136 -159 -dl -0k4 -0107 log concentration added (ppm) Fig. 4 Calibration plots for technetium (H) radium (+) thorium (0) and uranium (+ ) sociation ionization and excitation can lead to complex relationships between the analyte and various matrix constitu- ents. Tanner et aL2 described space-charge effects that take place in ICP-MS due to a perturbation of the electrostatic field in an ion flow system.If the space-charge field is sufficiently strong a self-repulsion develops which acts to spread the ion radial distribution and thereby reduces the transmission of other ions. Matrix effects have been observed not only in the ICP-MS system but also in dc plasma24 and microwave-induced plasma atomic emission ~pectrornetry.~'.~~ Following the classical Le Chatelier's principle when easily ionizable elements (EIEs) are added to the sample in high concentrations the classical description states that the ioniz- ation of EIEs produces electrons that shift the ion-atom analyte equilibrium in the direction of neutral analyte atoms decreasing the signal observed on the mass spectrometer. Although analytical intensities-to-background ratios can be improved two or three times by deliberate EIE doping enhancement is generally regarded as an interference because of the different EIE content of samples and the unpredictable behaviour of the plasma.Matrix effects on uranium and thorium were studied by monitoring the mass of 236U and 232Th respectively. Matrix effects for uranium are shown in Fig. 5. For uranium the typically described EIE effect was observed when elements such as potassium sodium and magnesium were introduced. A comparison of the ionization potentials of sodium (5.138 eV) and potassium (4.339 eV) can be used to explain why potassium 120 I I 0 ' 0.b 0.50 1.b 1.k 2.b 230 3. log concentration added (ppm) Fig. 5 Effect of sodium (+) potassium (0 ) calcium (m) and mag- nesium (+) on the uranium-236 signal using flash vaporization 926 Journal of Analytical Atomic Spectrometry October 1996 Vol.1 1produces a larger quenching than sodium on the uranium signal. When calcium was introduced a maximum signal was observed at concentrations close to 10 ppm. Depressions at high concentrations of added EIEs were probably due to overloading of the plasma. Effects of EIEs on the thorium signal are shown in Fig.6. Sodium and potassium produced a decrease in the signal; however in this instance sodium produced a decrease in the signal of about 50% when a concentration of approxirnately 5 ppm was introduced and potassium produced a 5 -10% decrease in the signal. This behaviour is not consisteni with the EIE effect and is a good example of combination cffects taking place in the plasma and at the plasma interface.'When calcium and magnesium were introduced as matrix inter- ferences an increase in the thorium signal was observed. Magnesium and calcium produced an increase in the th+)rium signal of 100 and 8O% respectively when concentrations of 5 ppm of each element were introduced. In general when elements from Groups IA and IIA were introduced into the plasma the effects on analyte ion and atom ratios could be described by classical EIE effect3 in a few instances; however much of the behaviour was not consist- ent and depended on both the EIE and the analyte. The behaviour is not yet fully understood but may be the re>ult of a combination of classical equilibrium and space-charge effects.For this reason the presence of EIEs and other elements in the sample matrix can be considered more as an interference than as a signal improvement. In an attempt to minimize the matrix effects we used the unique characteristics of the ETV of temporal-thermal pro- gramming to vaporize and remove matrix interferences selec- tively. Figs. 7 and 8 show the effects of sodium and calcium respectively on the uranium signal when flash vaporization and ramp vaporization as described in Fig. 2 were used. When sodium was introduced no effects were observed for concen- trations up to 100 ppm. Calcium showed no effects at concen- trations close to 80 ppm. By performing thermal programming it was possible to minimize matrix effects when samples exhibited concentrations of approximately two orders of mag- nitude of each interfering element.In addition temporal- -ther- mal programming can be used to vaporize compounds with similar chemical and different physical compositions (e.g. U308 or UO,) selectively. A decrease in the signal at concentrations higher than these concentrations was observed when both elements calcium and sodium were introduced. This effect can be explained in the same way as before due to overloading of the plasma. Thermal programming allows the matrix effects in environmental samples to be minimized usually with concen- trations of EIEs lower than those used in this experiment. 120.00 m.00 $ 3 80.00 u 60.00 h v c ul v) .- .- A - a 0 z 40.00 20.00 0.00 log concentration added (ppm) Fig. 7 Effect of sodium on the uranium-236 signal using flash vaporiz- ation and thermal programming (ramp) 120.00 ' 100.00 h 8 Y 80.00 0) v) .- 60.00 .- 8 - a 0 z 40.00 20.00 0.00 ' 0.h 0.h tbo 1.h 2.b 230 3.log concentration added (ppm) Fig. 8 Effect of calcium on the uranium-236 signal using flash vaporiz- ation and thermal programming (ramp) Precision Precision measurements are illustrated in Fig. 9 for 99Tc. The data represent the worst-case scenario during this analysis. For ten different samples the blank and the standard with a concentration of 99Tc of 1.21 ng l-' were introduced into the ETV device; the average signal for the sample was 232 counts s - l with a relative standard deviation of 6.6%. For the blank the average signal was 24 counts s-' with a relative standard deviation of 9.2%.Other elements such as radium uranium and thorium presented relative standard deviations lower than 7%. 2Ml 1 500 Acid blank Te-sample 2.2 15.4 Mean 24 232.2 v) RSD I 9 .2% 6.6% . log concentration added (ppm) Fig.6 nesium (+ ) on the thorium-232 signal using flash vaporization Effect of sodium (+) potassium (0) calcium (m) and mag- mm 0 Sample number 1 Fig. 9 Precision measurement of technetium-99 and blank samples Journal of Analytical Atomic Spectrometry October 1996 Vol. 1 1 927Table3 Comparison of the results for the determination of uranium-238 in water Sample ETV-ICP-MS/pg 1- a-Spectrometry/pg 1-' Tap water- Chicago IL Lemont IL River water- Fox River IL Kankakee River IL Well water- Lemont IL Borden IN Herrick Lake Spring water Others- 0.001 5 f 0.0004 ND* 0.006 f 0.002 0.0043 f 0.0007 ND* 0.012 f0.003 0.003 f 0.001 0.001 5 f 0.0004 < 0.09 < 0.09 3.27 0.12 < 0.06 < 0.09 < 0.06 < 0.09 < 0.06 * Not detected Table4 Comparison of the results for the determination of thorium-232 in water ~~ Sample ETV-ICP-MS/pg 1-' a-Spectrometry/pg 1- ' Tap water- Chicago IL 0.17 & 0.04 0.27 & 0.05 Lemont IL 0.27 & 0.08 0.28 f 0.06 Fox River IL 0.8 f 0.2 0.92 & 0.09 Kankakee River IL 1.4 f 0.3 1.5f0.1 Lemont IL 0.24 f 0.07 0.30 f 0.07 Borden IN 0.07 f 0.03 0.06 f 0.02 Herrick Lake 0.19 f 0.06 0.16 & 0.05 Spring water 0.15 f 0.05 0.13 f 0.04 River water- Well water- Others- Table5 Comparison of the results for the determination of technetium-99 in water ETV-ICP-MS/ Membranelfi-counter*/ Sample ng 1-' ng I-' Paducah-5920 1.4+0.2 Paducah-6275 26f2 1.2 _+ 0.1 25f3 * Low-background proportional counter Results of Environmental Sample Analysis Samples of tap water river water well water lake water and spring water from the Illinois and Indiana areas were collected as described previously and analysed for 226Ra 238U 235U 232Th 230Th and "Tc.Owing to the very low concentrations in natural samples and the detection limits of the system no results were obtained for 226Ra 235U and 23@Th. Results obtained by using external calibrations for 238U 232Th and "Tc are reported in Tables 3-5 which also compare each isotope measurement obtained in this work with the results obtained by using isotope dilution alpha-spectrometry and low-background proportional counters.There was good agree- ment between the values obtained in this work and the values obtained by the radioanalytical measurements. CONCLUSIONS ETV-ICP-MS shows several advantages the system is easy to use is capable of handling microlitre sample volumes and is highly sensitive. Also plasma excitation processes are more efficient with ETV owing to the removal of solvent before the sample is introduced. The detection limits for long-lived isotopes such as "Tc 238U 236U 232Th and 226Ra were sim- ilar to those obtained with USN-ICP-MS. No sample pre-treatment was required. ETV can be used for the analysis of real samples owing to the elimination of matrix interferences. The use of a halogen- ated gas such as CHF3 as a carrier/chemical reactor gas proved to be very effective as a sample digester in the ETV- ICP-MS analysis.This gas minimized the effect of carbides and oxides by creating compounds with lower boiling-points. In addition the gas can be used to eliminate matrices such as Si02 particles or as a major component in a soil sample by forming fluorinated compounds such as SiF4 (gas). The use of temporal-thermal programming in addition to halogenation selectively eliminates matrix constituents without adversely affecting analytical performance. The authors thank Lesa L. Smith and Kent A. Orlandini for their work in preparing and characterizing the water samples used in this study. This work was funded by the Laboratory Management Division of the Office for Environmental Restoration and Waste Management US Department of Energy. Argonne National Laboratory is operated by the University of Chicago for the US Department of Energy under contract number W-3 1-109-ENG-38.REFERENCES 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 Daes S. K. Kulkarni A. V. and Dhaneshwar R. G. Analyst 1993 118 1153. Franek M. and Krivan V. Anal. Chim. Acta 1993 274 317. Jackson P. E. Carnevale J. Fuping H. and Haddad P. R. J. Chromatogr. A 1994 67 181. Rohr U. Meckel L. and Ortner H. M. Fresenius' J. Anal. Chem. 1994 348 356. Browner R. F. and Boorn A. W. Anal. Chem. 1984 56 787A. Browner R. F. and Boorn A. W. Anal. Chem. 1984 56 875A. Ng K. C. and Caruso J. A. Anal. Chim. Acta 1982 143 209. Toth L E. Transition Metal Carbides and Nitrates (Refractory Material Vol.7) Academic Press New York 1971. Barnes R. M. and Fodor P. Spectrochim. Acta Part B 1983 38 1191. Gunn A. M. Millard D. L. and Kirkbright G. F. Analyst 1978 103 1066. Alvarado J. S. PhD Thesis Northern Illinois University 1990. Ediger R. D. and Beres S. A. Spectrochim. Acta Part B 1992 47 907. Fonseca R. W. and Miller-Ihli N. J. Appl Spectrosc 1995 49 1403. Nickel H. and Zadgorska Z. Fresenius' J. Anal. Chem. 1995 351 158. Ren J. M. and Salin E. D. Spectrochim. Acta Part B 1994 49 555. Zatka J. V. Anal. Chem. 1978 50 538. Smith L. S. Crain J. S. Yaeger J. S. Horwitz E. P. Diamond H. and Chiarizia R. J Radioanal. Nucl. Chem. 1995 194 151. Alvarado J. S. Orlandini K. A. and Erickson M. D. J. Radioanal. Nucl. Chem. 1995 194 163. Orlandini K. A. King J. and Erickson M. D. Methods for Evaluating Environmental and Waste Management Samples DOE/EM-O089T Rev-2 US Department of Energy Richland Washington September 1994. Gray D. J. Wang S. and Brown R. Appl. Spectrosc. 1994 48 1316 Crain J. S. Smith L. L. Yaeger J. S. and Alvarado J. S. J. Radioanal. Nucl. Chem. 1995 194 133. Ebdon L. and Goodall P. Spectrochim. Acta Part B 1992 47 1247. Tanner S. D. Cousins L. M. and Douglas D. J. Appl. Spectrosc. 1994,48 1367. Miller M. Keating E. Eastwood D. and Hendrick M. S. Spectrochim. Acta Part B 1985 40 593. Alvarado J. S. and Carnahan J. W. Appl. Spectrosc 1993 47 2036. Wu M. and Carnahan J. W. Appl. Spectrosc. 1992 46 163. Paper 6/02394K Received April 9 1996 Accepted July 10 1996 928 Journal of Analytical Atomic Spectrometry October 1996 Vol. 11