JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY SEPTEMBER 1993 VOL. 8 833 Application of Ultrasonic Nebulization for the Determination of Rare Earth Elements in Phosphates and Related Sedimentary Rocks Using Inductively Coupled Plasma Atomic Emission Spectrometry with Comments on Dissolution Procedures* 1. B. Brenner and E. Dorfman Geological Survey of Israel 30 Malkhe Israel Street Jerusalem Israel 95507 Low concentrations of all the rare earth elements (REE) can be determined in geological materials using ultrasonic nebulization and inductively coupled plasma atomic emission spectrometry (USN-ICP-AES). Using a matrix matched calibration procedure to compensate for an acid effect on REE spectral line intensities caused by high concentrations of nitric acid in-house and international geological reference materials were analysed.Provided that a total decomposition method was used the data obtained compared well with those from the literature. Yttrium La Ce Eu and Yb were also determined directly because of their superior limits of detection relatively high abundance in geological samples and freedom from spectral line interference as a result of the high resolution of the spectrometers. Various sample decomposition procedures were evaluated for the determination of REE in sedimentary rocks. The solutions obtained by decomposition with various acids (HCI HNO HC104 and HF) and sodium peroxide sintering and leaching were preconcentrated using a DOWEX 50 W x 8 ion-exchange column. Significant differences in Y La and Ce concentrations occurred when the sediment samples were submitted to different methods of sample decomposition. Indeed REE distributions between acid soluble and total samples differed significantly having different Y:Ce Y:La and La:Ce ratios.Keywords Inductively coupled plasma atomic emission spectrometry; ultrasonic nebulization; rare earth elements; phosphates; sedimentary rocks The growing interest in the distribution of the rare earth elements (REE) Ce in particular in sedimentary rocks has increased owing to their application in the unravelling of sedimentary processes such as the role of redox conditions in ancient and modern basins of deposi- tion,'- contribution of terrestrial provenance^,^-^ and rock-water interactiow6 Redox conditions can lead to fractionation of Ce relative to the other REE resulting in significant deviations from the smooth REE chondrite distributions. The distribution of REE in the individual fractions of sediments (phosphorites siliceous ferru- ginous and calcareous phosphates bituminous shales carbonaceous clays and their acid insoluble residues etc.) can also vary significantly. The partition of REE in littoral sediments is controlled by the individual fractions of the sediments and might be fractionated during estuarine mixing.Thus the distribution of REE can be used as a diagnostic tool for determining the role of terrestrial detritus in deposition basins and for characterizing redox conditions in ancient and modern sedimentary basins of d e p o ~ i t i o n . ~ ~ ~ Owing to the fact that the REE concentrations in some of these fractions in particular those that contain high amounts of siliceous and organic material may be low it is necessary to preconcentrate the REE fraction and exclude the interfering matrix.Indeed the limits of detection (LODs) that can be obtained using conventional nebuliza- tion with inductively coupled plasma atomic emission spectrometry (ICP-AES) are inadequate for the determina- tion of the low concentrations of some of the REE. This disadvantage can be overcome by using enhanced methods of sample introduction such as ultrasonic nebulization (USN) which results in a 10-fold improvement in the LODs of the REE.' In the present investigation the superior LODs for the REE using USN-ICP-AES are applied to determine low REE concentrations in geological samples and decomposition residues.*Presented at the 1993 European Winter Conference on Plasma Spectrochemistry Granada Spain January 10- 15 1993. In previous investigation^^.^ it was reported that chemi- cally resistant residues were observed when geological materials were decomposed in a poly(tetrafluoroethy1ene) (PTFE) open dish using a mixture of HF-HC104. These residues consist mainly of Zr- and Ti-bearing minerals and were subsequently decomposed using alkali fusions. Shol- kovitz demonstrated that the HF dissolution excluded a significant portion of the heavy REE (HREE Ho-Lu) whereas this fraction was decomposed using LiBOz fusion. In contrast fusion and HF dissolution techniques yielded similar concentrations for the light (LREE La-Nd) and middle REE (Sm-Dy).He concluded that shelf and slope marine sediments are more similar in REE composition to shales than previously rep~rted.~ Sholkovitz3 stressed the importance of these fractions in the interpretation of fluviatile influxes oceanic abundances and the behaviour of REE in diagenesis. Therefore the determination of REE in small amounts of residues is an additional analytical challenge when using ICP-AES. In the present study several methods of decom- position for the determination of the REE were evaluated using the following geological reference materials Cana- dian Certified Reference Material Project (CCRMP) SY-2 and SY-3 Silicate Rock; Community Bureau of Reference (BCR) 32 Morrocan Phosphate Rock; National Institute of Standards and Technology (NIST) Standard Reference Material (SRM) 120B Phosphate Rock (Florida) and 694 Phosphate Rock (Western); South African Bureau of Stan- dards (SABS) SARM 1 Granite SARM 2 Lujavrite and SARM 3 Syenite; and in-house reference materials.The following decomposition routines were studied HC1; HN0,-HC10,-HF-HCl; and sodium peroxide sintering and cold leaching. The resultant solutions were analysed by simultaneous and sequential ICP-AES after the major elements and fusion cations were excluded and the REE concentrated using a cation-exchange procedure described by Brenner et al.' and Watkins and N01an.~ Solutions were also analysed directly for trace and major elements which included La Y Yb Eu and Ce using procedures described by Brenner et a1.lo834 JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY SEPTEMBER 1993 VOL. 8 Experimental Decomposition Procedures Several dissolution routines were applied to several in- house and international geological reference materials phosphorites ferruginous and siliceous phosphates bitumi- nous and calcareous shales and silicate samples.Hydrochloric acid A sample weighing 2 g was stirred continuously with 75 ml of 2 mol 1-l HCl for 1 h at room temperature then filtered through a Whatman 40 filter-paper into a 50 or 100 ml calibrated flask. Solutions were made up to volume with 2 mol 1-l HCl. The residues were ashed overnight at 600 "C; the ashes were then digested to dryness in a Pt dish at 300 "C with 15 ml of HF and 10 ml of HN03. The salts were then dissolved in 2 mol 1-l HCl and made up to 50 ml.Calcareous and phosphatic fractions of the samples were dissolved using this procedure. Sodium peroxide sinter A sample weighing 0.5 g was decomposed by sintering with 2 g of sodium peroxide in a graphite or Zr metal crucible in an oven at 500 "C for 30 min. After cooling the sintered mass was disintegrated by treatment with water and the mass transferred into a beaker containing 25 ml of HCl dissolved and made up to 100 ml. The procedure used has been described previously.1o With this technique a total analysis was obtained. Mixed acid (hydrochloric-nitric-perchloric-hydrofluoric) procedure A sample weighing 1 g was treated with 30 ml of HF (38-40% v/v) and 10 ml of concentrated HN03 in a platinum or Teflon dish. An additional 25 ml of concen- trated HN03 were added after the mixture was heated to dryness.Residues were then heated consecutively with 20 ml of HCl and 20 ml of HC104 until the solution was clear. Heating was continued to dryness to remove perchloric acid. After cooling the salt residue was dissolved in 10 ml of 2 mol 1-l HC1. With this technique a total analysis was obtained. However resistant REE-bearing minerals might be only partially Cold sodium peroxide extraction This procedure was employed for the decomposition of samples containing high contents of organic matter. A 10-20 g sample of bituminous shale or carbonaceous clay was mixed for 24 h with 150 ml of a 10% m/v solution of sodium peroxide in order to decompose selectively the organic-bound fraction. The soluble fraction separated by centrifugation was evaporated to about 20-30 ml then treated by boiling with 50 ml of HN03 and 25 ml of H202 until the solution was colourless.The heating was continued to dryness and the solid residue was dissolved in 50- 100 ml of 2 mol 1 - l HCl. Hydrofluoric acid Chert samples containing very high amounts of silica were treated with HF in order to remove the silica and to extract the REE bound to t h e inorganic fraction. A sample weighing 10 g was treated with HF-HC104 (2+ 1 v/v) and the solution was heated to dryness. The residue was dissolved in 25-50 ml of 2 mol 1-l HCl. REE Chromatographic Separation The procedure employed is similar to that used by Watkins and Nolan9 and Brenner et aL7. Sample solutions produced from the decomposition procedures were loaded onto a column of DOWEX 50 Wx8 ion-exchange resin pre- equilibrated with 2 mol 1-l HCl.Matrix elements (Ca Mg Na K Fe and Al) were eliminated using 100 ml of mixed acid 3 mol 1-l HN03-2 mol 1-l HC1 (3+ 1) followed by :25-50 ml of 2 moll-' HC1. The REE eluted with 100 ml of '7.2 mol 1-l HN03 were determined by simultaneous and sequential USN-ICP-AES. Several REE (Y La Eu Yb and Ce) were also determined directly using the polychromator and a multi-element routine designed to determine the major minor and trace elements.lo REE Calibration Multi-element REE calibration standards were prepared by stepwise dilution of stock solutions prepared from Specpure REE oxides (Johnson Matthey UK). A graded series of calibration standards were prepared in which the LREE concentrations exceeded the concentrations of the HREE in accordance with their abundances in the materials under investigation (1-500 pg 1-l).This calibration scheme was usually adequate for most of the geological materials investigated. In certain cases calibration ranges were ex- tended or samples diluted. The step involving removal of HN03 was avoided. However the USN is sensitive to changes in the acid content" (Fig. 1). Consequently a matrix match calibration procedure was adopted where the standards were accurately matched to the samples by adding 7.2 mol 1-' HN03. 1.05 0 1 2 3 4 5 6 7 8 [HNOJmol 1.' Fig. 1 Effect of HN03 on REE (1 mg I-') spectral line intensity using ultrasonic nebulization with desolvation A Ho; B Y; C Gd; D Yb; E La; and F Lu Instrumentation and Operating Conditions A high resolution 1 m JY 38 sequential system was employed for the determination of all the REE.A JY 48,48 channel polychromator was employed for the simultaneous determination of Y La C Yb and Eu. The lack of significant spectral line interferences is attributed to the high resolution of the sequential spectrometer (6 pm in the first order). A Cetac USN was employed. The instrumental configuration and operating conditions are listed in Tables 1 and 2.JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY SEPTEMBER 1993 VOL. 8 835 Table 1 Instrumentation Monochromator Grating Dispersion Polychromator Grating Dispersion R.f. generator Torch Jobin Yvon JY 38 3600 grooves mm-I range 180-490 nm 0.27 nm mm-l 6 pm in the first order (25 pm slit-widths) Jobin Yvon JY 48 1 m 2550 grooves mm-l spectral range 170-450 nm 0.35 nm mm-I PlasmaTherm 2.5 kW < 10 W reflected Jobin Yvon Ryton demountable Results and Discussion Limits of Detection The spectral lines employed were listed in a previous investigation.' The LODs were determined by calculating the concentration of the analyte that yielded a signal twice the standard deviation of the blank signal (n = 10).Limits of detection of USN-ICP-AES are compared with ICP-MS and other ICP-AES literature values in Table 3. The similarity of the USN-ICP-AES LODs to the ICP-MS values quoted by JarvisI3 is noteworthy; Ce Pr La and Tb are 10-fold better using ICP-MS whereas Lu is 10-fold better by USN- ICP-AES. Taking into consideration that implementation of these LODs using ICP-MS is dependent on the limita- tions of total salt and acid concentrations in ICP-MS the performance of USN-ICP-AES is commendable.Quantification limits in the solid samples were calculated based on a 1 +49 dilution (superior to those reported by Jarvis and JarvisI2 and permit the determination of all the REE in most types of geological materials. These levels are a prerequisite when REE are to be determined in siliceous and bituminous materials and silicate rocks such as SARM 2 syenite. They can be improved by using larger amounts of sample. Precision and Accuracy In general the relative standard deviations (RSDs) (stan- dard deviatiodmean x 100 n= 5-1 0) varied as a function of concentration (Fig. 2) and at 1-100 mg kg-I were about lo% except for concentrations that were near the limit of quantification (RSDs 10-25%).For concentrations greater Table 2 ICP and USN operating conditions Pneumatic nebulizer Nebulizer gas flow rate/l min-l Ultrasonic nebulizer Desolvation temperature/"C Cooling Heating Outer Intermediate Sheath (Trassy-Mermet) Pneumatic USN Gas flow rate/l min-' Aerosol carrier gas flow rate/ml min-l Washout periodls Pneumatic (Meinhard) USN Integration period/s Pol ychromator Monochromator Meinhard TR-C-20 1.2 Cetac U 500AT 45 psi - 5-0 140 14 0.2-0.4 0.2 0.1 0.95 30 50 5 0.5 ~~~~~ ~~~ ~~~ than 100 mg kg-l RSDs improved with increasing concen- tration to < 1%. In order to estimate the accuracy of the present proce- dures several international reference materials were de- composed in triplicate employing sodium peroxide sinter and HF-containing mixed acid decomposition procedures to furnish total REE contents. It should be noted that all REE were determined Tm and Tb included which are usually too low to be determined by ICP-AES using a conventional sample introduction system.Recommended value^'^-'^ listed in Table 4 are in good agreement with the published values. In general the data have a variation of < 10% when compared with recommended values. Values are also listed for SARM 2 which has very low REE contents. In this case the dilution factor was 1 +24. Unfortunately the precision of the determinations of SARM 2 varied from 10-35%. The efficiency of the ion-exchange columns for quantita- tive preconcentration of the REE was ensured by analysing the original decomposition solutions for Ce and Yb (a LREE and a HREE respectively) using the JY 48 simulta- neous spectrometer. The correlation between these values and those obtained by column preconcentration and sequential USN-ICP-AES are illustrated in Figs.3 and 4. Table 3 LODs (20 obtained by ICP-MS USN-ICP-AES and ICP-AES using conventional nebulization; all data in pg 1-1 Element Lu Tm Y Gd Ho Tb DY Sm Er Yb Eu La Nd Pr Ce W avelengt h/nm 26 1.52 31 3.126 37 1.03 342.246 345.6 350.9 17 353.17 1 359.26 369.265 369.41 9 38 1.97 398.852 406.66 41 7.939 4 18.66 ICP-MS (Ref. 13) 0.05 0.01 0.1 0.1 0.04 0.03 0. I 0.2 0.06 0.06 0.06 0.08 0.2 0.09 0.1 Meinhard (Ref. 7) 0.08 0.3 0.4 0.3 0.2 2.3 0.1 0.9 0.6 0.4 0.5 1 2.1 1.6 5.9 USN (Ref. 7) 0.008 0.06 0.30 0.09 0.05 0.3 0.04 0.3 0.15 0.03 0.06 0.35 0.5 0.8 1.2 ICP-AES (Ref.1 2) 0.2 1.5 2. l t 0.56 0.65 1.8 l . l t 0.25t 0.26 2.6 5.2t 3-87 8.7 - - Quantification limit*/pg kg-* (USN) 4 30 150 45 25 150 20 150 75 150 300 175 250 400 600 *Quantification limit is 10 x LOD. Data are reported as concentrations in the solid sample based on a 1 +49 dilution. ?Different spectral line.836 JOURNAL OF ANALYTI.CAL ATOMIC SPECTROMETRY SEPTEMBER 1993 VOL. 8 Table 4 Total REE concentrations in reference materials determined (Det) by ICP-AES and USN; data in mg kg-l RSDs (O/O) are based on five separate determinations. Recommended data (Rec) are from Jarvi,s1*.l3 and G~vindaraju'~ Element Parameter Y La Ce Pr Nd Sm Eu Gd Tb DY Ho Er Tm Yb Lu Det Oh RSD Rec Det % RSD Rec Det Oh RSD Rec Det Oh RSD Rec Det O/o RSD Rec Det Yo RSD Rec Det Yo RSD Rec Det O/o RSD Rec Det % RSD Rec Det Oh RSD Rec Det O/o RSD Rec Det % RSD Rec Det Oh RSD Rec Det Oh RSD Rec Det % RSD Rec NIST 120B 167 5 172 95 5 88 115 4 115 18 8 17.4 75 6.8 75 17 15 16 3.8 10 3.6 19.5 13 18.9 2.2 15 2 16.3 13 17.2 3.8 13 3.9 1 1 8 11.7 1.35 14 1.1 10.9 7 10.9 1.85 12 1.7 BCR 32 268 6 262 117 16 115 33 15 35 13.8 11 60 8 68 14 8 12 3.5 11 3.5 19 12 19.5 2.1 15 3.3 16.5 10 4 10 14 12 - - - - 1.85 25 2 14.4 2 14.2 2.5 4 2.6 NIST 694 189 20 125 128 5 103 26 20 25 14 2 70 5 41 12 2 4.9 3 25 1.5 16 15 - - 1.6 7 1.2 15 10 9 3.5 9 11 6 1.2 3 9 4 7.2 2.1 0.97 - - - 17 SY-2 90 4 117 75 10 66 168 5 161 18 4 18.2 76 8 72.9 20 8 14.8 2.6 11 2.62 15.7 4.5 15.4 2.4 13 18.7 6 18.4 4.3 5 4.29 13.1 1 12.5 2.1 6 16.6 3 16.5 2.7 5 2.7 - - SY-3 64 5 0.8 685 1345 1 1330 2200 3 2340 170 2 225 710 15 749 134 6 126 18 9 20.5 126 9 112 14.2 8 114 12 130 21 2 28.7 67 7 76.8 9.3 8 63 9 67.2 7.9 11 8.3 - - SARM I 145 2 130 95 3 105 220 3 200 20 5 21.7 70 4 73 21 7 16 0.44 23 0.4 14 12 13.7 2.4 20 2.3 18 9 15 4.2 16 3.7 17 8 13 2.3 20 2 15 11 14 20 2.1 1.6 SARM3 SARM2 27 9 22 285 2 50 285 3 270 18 15 19 48 7 48 5.7 16 5 1.4 22 1.2 3.4 17 4 3.4 - - - 2.9 3.1 0.8 0.6 1.6 2.2 3.4 17 3 2.5 13 3 0.44 31 0.4 15 25 21 1.2 2.5 4.5 10 5 13 10 11.9 1.5 15 1.18 6.5 10 6 1 18 1 0.3 0.3 0.9 0.83 0.16 0.1 0.5 0.4 0.06 0.05 0.13 0.12 0.06 20 - - 20 20 - 20 - - 0.1 0.06 0.1 0.0 1 30 30 The correlation coefficients for Ce and Yb are 0.99636 and 0.998 1 respectively (n = 2 1 and 14 respectively) indicating that these elements can be determined directly without preconcentration.This procedure was employed to cross check data obtained by USN-ICP-AES and to ensure column performance. Reference materials NIST SRM 120B Phosphate Rock (Florida) and BCR 32 Morrocan Phosphate Rock were treated using the decomposition procedures described previously. The effect of sample decomposition on the partition of the LREE is illustrated in Figs. 5 and 6. Evidently the Y La and Ce values for the sodium peroxide sintering procedure are maximum. Thus these data indi- cate that significant differences in Y La Ce and Nd concentrations can also occur when sediment samples are submitted to different modes of sample decomposition. Preliminary results on the partition of the REE in the various fractions of the samples investigated are listed in Table 5 and Fig.7. Several phosphate and shale samples were treated with HCl and sodium peroxide and the residues analysed. While the interpretation of these trends is beyond the scope of this article it is obvious that the acid insoluble and organic fractions have different REE distribu- tions from those of the whole samples. These patterns could be controlled by the mineralogy of the fractions derived from different parent rocks alteration processes leading to REE fractionation secondary phases that preferentially accommodate REE etc. It is clear that the present technique is capable of determining very low REE concentrations. Nevertheless at the present time the precision of determination of the HREE were inadequate to resolve the question of whether they were also discriminated.Conclusion The advantage of the USN lies in obtaining superior LODs permitting small amounts of sample to be analysed. TheseJOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY SEPTEMBER 1993 VOL. 8 837 loo c c3 v) K 1 000 0 0 0 00 El 000 0 0 0 0.1 1 I I I 1 10 100 1000 [REEI/ mg kg-' Fig. 2 Relation between RSD (%) of determination and REE concentration 1 g of solid sample dissolved in 100 ml 0 Y 0 0 F Brn - Q I I 1 I I 1 10 100 1000 E [Cel (sequential USN-ICP-AES)/ mg kg" Fig. 3 Correlation of sequential USN-ICP-AES (with column preconcentration) and simultaneous direct determination of Ce I1 4 18.66 nm using ICP-AES 0 0 10 100 [Ybl (sequential USN-ICP-AES)/ mg kg-' Fig.4 Correlation of sequential USN-ICP-AES (with column preconcentration) and simultaneous direct determination of Yb I1 369.419 nm using ICP-AES include residues that were produced by treating the samples with selective extractions and those that are chemically resistant. Furthermore all the REE including Tb and Tm can be determined and samples containing low REE contents can be analysed. 200 150 0 Y 0 E 100 E w w 50 0 HCI a Mixed acid / Y La Ce Nd Fig. 5 REE concentrations in NIST 120 B Phosphate Rock (Florida) treated with various decomposition procedures ' /I HCI 300 a Mixed acid 250 69 Recommended yo) 200 Y 150 h w w E 100 50 O V f f f / Y La Ce Nd Fig. 6 REE concentrations in BCR 32 Morrocan Phosphate Rock treated with various decomposition procedures Phosphate - Bituminous shale Fig.7 Y:La Y:Ce and La:Ce ratios in various fractions of phosphates (P) bituminous shales and cherts from southern Israel; Res = residue Tot = total Org = organic fraction and NASC = North American Shale Composite838 JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY SEPTEMBER 1993 VOL. 8 Table 5 Y La and Ce concentrations in various fractions of phosphates bituminous shales and cherts from southern Israel Sample Y NIST 120 Phosphate Rock Residue* 25 Total? 167 Residue 24 Total 2 60 BCR 32 Phosphate Rock IB-7 Bituminous Shale Residue Organic$ Total Residue Total IB-8 Phosphorite 1.25 0.005 45 0.6 45 IB-9 Bituminous Phosphorite Residue 2.2 Organic Total 62.7 IB-10 Chert Residue Total 0.02 2.2 North American Shale Composite 27 La 8.6 94. 3 124 0.8 0.007 1 7 0.3 1 '7 0.4 0.0002 21 - 10.9 32 Ce 3.8 116 4 35 0.45 0.008 6 0.6 8 0.2 0.0002 9 0.2 0.8 73 Y:La Y:Ce 2.9 6.6 1.8 1.4 8.0 6.0 2.1 7.4 1.6 2.8 0.7 0.6 2.6 7.5 2.0 1 .o 2.6 5.6 5.5 11.0 3.0 7.0 - 0.1 2.4 2.8 0.8 0.4 - - La:Ce 2.3 0.8 0.8 3.5 1.8 0.9 2.8 0.5 2.1 2.0 1 .o 2.3 - 1.1 0.4 *Residue HCl insoluble residue.tTotal sodium peroxide or mixed acid procedure with HF. $Organic fraction soluble in sodium peroxide solution. A comparison of several dissolution procedures for the determination of REE in phosphates and related sedimen- tary rocks indicated that satisfactory data were obtained when a reliable total decomposition procedure such as a sodium peroxide sinter or a mixed HN03-HC104-HF-HCl procedure was employed. Data obtained from selective dissolution procedures indicate that significant concentra- tions of the REE are also located in the insoluble silicate fraction.The data produced in this study indicate that the REE concentrations in bituminous and siliceous fractions are very low. This study was sponsored by the Earth Science Administra- tion Ministry of Energy and Infrastructure Israel. Their financial assistance is sincerely appreciated. The authors are indebted to Dr. A. Bein Director of the Geological Survey of Israel for his constructive discussions. 0. Joffe per- formed cross check REE determinations using simulta- neous multi-element analysis. The assistance of I. Segal and R. Binstock in instrument operation was most valuable. T. Minster supplied a sample of bituminous shale for which we are thankful. References 1 De Baar H. J. W. German C. R. Elderfield H. and Van Gaans P. Geochim. Cosmochim. Acta 1988 52 1203. 2 Elderfield H. and Sholkovitz E. R. Earth Planet. Sci. Lett. 1987 82 280. 3 Sholkovitz E. R. Chem. Geol. 1990 88 333. 4 Sholkovitz E. R. Am. J. Sci. 1988 288 236. 5 Murray R. W. Buchholtzen-Brink M. R. Gerlach D. C. Russ G. P. 111 and Jones D. L. Geochim. Cosmochim. Acta 1991,55 1875. 6 Smedley P. Geochim. Cosmochim. Acta 1991 55 2767. 7 Brenner I. B. Binstock R. Dorfman E. and Halicz L. ZCP Inf Newsl. 1992 18 473. 8 Brenner I. B. Watson A. E. Steele T. W. Jones E. A. and Goncalves M. Spectrochim. Acta Part B 1981 36 785. 9 Watkins P. J and Nolan J. Chem. Geol. 1992 95 131. 10 Brenner I. B. and Eldad H. ICP In Newsl. 1986 12 243. 11 Brenner I. B. Segal I. Long G. and Dorfman E. presented at the 1993 European Winter Conference on Plasma Spectro- chemistry Granada Spain January 10- 1 5 1993. Paper No. 12 Jarvis K. E. and Jarvis I. Geostand. Newsl. 1988 12 1. 13 Jarvis K. E. Chem. Geol. 1990 83 89. 14 Govindaraju K. Geostand. Newsl. 1989 13 1. P 1-24. Paper 3/0 1 9 70E Received April 6 1993 Accepted May 5 1993