Matrix-matched calibration in LA-ICP-MS of silicate, phosphate and carbonate minerals: application of G-Probe samples 4 D O I: 1 0. 15 82 6/ ch im te ch .2 02 0. 7. 1. 01 Kiseleva D. V. Chimica Techno Acta. 2020. Vol. 7, no. 1. P. 4–12. ISSN 2409–5613 D. V. Kiseleva A. N. Zavaritsky Institute of Geology and Geochemistry of Urals Branch of Russian Academy of Sciences, 15 Akademika Vonsovskogo, 620016, Ekaterinburg, Russia Kiseleva@igg.uran.ru Matrix-matched calibration in LA-ICP-MS of silicate, phosphate and carbonate minerals: application of G-Probe samples Laser ablation (LA) sampling provides fast microelement ICP-MS analysis of a wide range of solid materials without their dissolution, thus decreasing contamination from water and reagents as well as reducing polyatomic isobaric interferences from acid solutions. However, the issue of matrix-matched calibra- tion becomes crucial for LA-ICP-MS due to differences in behaviour during laser interaction and evaporation of solid samples. There are several approaches to LA calibration: simultaneous supply of standard solutions into a spray chamber; calibration using a set of NIST 61х synthetic glasses and glasses prepared from natural rocks and minerals (basalt, nephelinite, etc.) or pressed synthetic sam- ples (calcium carbonates, phosphates and sulphides produced by USGS). A set of natural glasses for microanalysis is available from the International Association of Geoanalysts (IAG) in co-operation with the USGS. The G-Probe proficiency testing programme has been operating since 2008 and deals with solid samples for microanalysis (LA-ICP-MS, EPMA, EDS-SEM). A number of samples of dif- ferent compositions were distributed: BBM-1G and BSWIR-1G natural basaltic glasses, GSM-1 gabbro; NIST SRM-based basaltic and diabase glasses; GP-MACS synthetic pressed calcium carbonate, GP-MAPS phosphate and some others. The aim of the present work was to estimate the LA-ICP-MS analysis quality us- ing matrix-matched calibration with G-probe samples of various composition. All G-Probe samples were analysed using an ELAN 9000 Q-ICP-MS combined with a LSX-500 (Nd:YAG, 266 nm) laser ablation system. For silicate rocks, TB-1 basaltic glass was used for calibration; the remaining samples were analysed as unknowns. MAPS-4 calibration material were used for phosphate rock analysis. A combination of external matrix-matched calibration and internal normalisation was used for calculating element concentrations. LA-ICP-MS analysis quality was estimated using z-scores. Most of the results obtained were in a good agreement with assigned values. Keywords: laser ablation; inductively coupled plasma mass-spectrometry; geological glasses; matrix-matched calibration; internal standard Received: 23.12.2019. Accepted: 20.02.2020. Published: 31.03.2020. © Kiseleva D. V., 2020 5 Introduction Laser ablation (LA) sampling in induc- tively coupled plasma mass spectrometric (ICP-MS) analysis allows the rapid analysis of the trace element composition of solids to be carried out without their dissolution. At the same time, the contamination from the reagents used is minimised, and polya- tomic isobaric interferences arising due to the presence of acid solutions are signifi- cantly reduced [1]. LA-ICP-MS is widely applied in the studies of microobjects with high spatial resolution, individual mineral grains, and spatial distributions of  trace elements. However, the issue of matrix-matched calibration becomes crucial for LA-ICP-MS of solids due to differences in behaviour during laser interaction and evaporation of  solid samples, especially when using 193 nm excimer and 213 and 266 nm Nd:YAG lasers [2]. To date, there are several widely prac- ticed approaches to constructing calibra- tion curves in the analysis of solid samples of various compositions using laser abla- tion sampling: some include the simultane- ous supply of aqueous calibration solutions to the spray chamber of a mass spectrom- eter [3, 4]; others use synthetic glasses of the NIST SRM 61x series (manufactured by the National Institute of Standards and Technology, USA), as well as glasses made by  fusion of  natural rocks and minerals (basalt, nephelinite, etc.) or pressed syn- thetic non-silicate samples (calcium car- bonates and phosphates, sulphides, US Geological Survey, USA). The  first approach leads to  a  com- plication of  design features (for exam- ple, the  manufacturing, often in-house, of  chambers with additional inputs for solutions, the connection of an additional gas for spraying), as well as the formation of polyatomic isobaric interferences from the  solvents (water and acids). The  use of synthetic glasses for calibration provides a certain unification of the results obtained, but can be justified only in some cases (for example, for the analysis of silicate sam- ples), while their composition does not re- flect the natural sample composition with their wide range of trace element content [5]. For example, about 33 microelements with concentrations from 15 to  80 μg/g are certified in the most widespread NIST SRM 612 glass, which is insufficient when analysing the entire variety of rocks and minerals. Taking into account the  above-men- tioned, the most acceptable way to calibrate a mass spectrometer when analysing solid samples is to use matrix-matched calibra- tion samples, and NIST SRM 612 glass- es to  optimise the  analytic parameters of  the  device (adjusting the  interface, monitoring the level of oxides, etc.). The range of solid reference materials (RM) of natural composition has expanded significantly at the moment: The US Geo- logical Survey produces 4 natural basalt glasses and 1 nephelinitic glass, which are the melts of powdered BCR-2, BHVO-2, BIR-1, TB-1 and NKT-1 certified reference materials (SRM) [5], respectively, as well as  synthetic pressed MACS-3 calcium carbonate and MAPS-4 phosphate and MASS-1 polymetallic sulphide. The  glass making procedure [5] in- volves the  fusion of  the  powdered rock material in an oven in a platinum crucible at 1325–1645 °C for 2–6 hours with several stirrings with a platinum rod and subse- quent rapid cooling in  deionised water. The  pieces of  glass are then dried, fixed in epoxy resin and distributed to analytical laboratories. For pressed powdered sam- 6 ples [6], a specially developed procedure of trace element co-precipitation with pure calcium carbonate or phosphate in  a  re- action vessel is  used. The  resulting sus- pension is powdered to less than 40 μm, dried at  110 °С and pressed into pellets with a diameter of 19 mm. The homoge- neity and composition of the samples ob- tained is confirmed by a number of studies in USGS laboratories (XRF, LA-ICP-MS, EPMA, etc.). Thus, the samples described above are microanalytical reference ma- terials (MRM), that is, they have passed the  certification procedure and issued certificates with certified concentrations of major and trace elements. An expanded set of natural glasses for microanalysis is presented as part of the G- Probe interlaboratory testing programme, which combines the efforts of the US Geo- logical Survey and the International As- sociation of Geoanalysts (IAG). The  main idea of  the  programme is to analyse the samples of unknown com- position sent to  a  number of  interested laboratories around the world, after which the organisers provide the data on the com- position of  these samples and protocols summarising the results obtained by all par- ticipants, and evaluate the quality of anal- ysis in  a  particular laboratory. G-Probe programme has been operating since 2008 and specialises in solid samples for microa- nalysis (LA-ICP-MS, EPMA, SEM-EDS); the number of participating laboratories has been ranged from 12 to 30 in various rounds with those providing the LA-ICP-MS results from 6 to 18, which reflects the complexity and specificity of mass spectrometric analy- sis of natural geological materials with high spatial resolution [7]. Since 2008, 13 samples of  various composition and origin were distributed: natural basalt glasses based on samples taken by the USGS expeditions (BBM-1G, BSWIR-1G), gabbro (GSM-1); basalt glass based on the  geological NIST SRM 688; glass based on the  USGS W-2 diabase standard; SL factory soda-lime glass; syn- thetic pressed GP-MACS calcium carbon- ate and GP-MAPS calcium phosphate et al. Almost all G-Probe samples are prepared and analysed in USGS laboratories in ac- cordance with the procedures adopted for microanalytical standard samples, and very often they become certified reference ma- terials after processing the results provided by the G-Probe participants. Since 2008, the  laboratory of  Physi- cal and Chemical Methods of  Investiga- tion of Mineral Substance (IGG UB RAS) has been taking part in the G-Probe pro- gramme. The laboratory has all of the list- ed solid geological glasses and synthetic pressed pellets, as  well as  NIST SRM 612 glass. The aim of the present work is to esti- mate the LA-ICP-MS analysis quality using matrix-matched calibration with G-probe samples of various composition. Experimental The  following geological glasses and microanalytical reference materials were studied (Table 1). LA-ICP-MS trace element analysis was carried out using an ELAN 9000 quadru- pole mass spectrometer (PerkinElmer) with an  LSX-500 laser ablation system (Cetac, Nd:YAG laser with a wavelength of 266 nm). The LA-ICP-MS analysis algo- rithm included daily tuning and selection of  the  operating parameters of  the  mass spectrometer using multi-element calibra- tion solutions followed by laser ablation, while the  ELAN 9000 operating condi- 7 tions and gas flows were adjusted in  ac- cordance with the highest possible inten- sity signal at  a  minimum level of  oxides (ThO/Th≤0.8%) using NIST 612. The isotopes analysed were: 7Li, 9Be, 11B, 45Sc, 51V, 53Cr, 55Mn, 59Co, 60Ni, 65Cu, 66Zn, 71Ga, 74Ge, 75As,85Rb, 86Sr, 89Y, 90Zr, 93Nb, 95Mo, 109Ag, 111Cd,115In, 118Sn, 123Sb, 128Te, 133Cs, 135Ba, 139La, 140Ce, 141Pr, 146Nd, 147Sm, 153Eu, 157Gd, 159Tb, 163Dy, 165Ho, 167Er, 169Tm, 172Yb, 175Lu, 178Hf, 181Ta, 184W, 205Tl, 208Pb, 209Bi, 232Th, 238U. The following LA operational param- eters (Table 2) were used when analysing the samples of various compositions based on the previously obtained data taking into account the specific features of rock/min- eral laser evaporation [8, 9]. LA-ICP-MS data were acquired in runs of up to 15 analyses. Each run started and ended with two data acquisitions on a cal- ibration material. The  15  analysis limit ensured that calibration was performed on no more than about an  hourly basis, in order to monitor and correct for the drift of the inter-element sensitivities (i.e., ana- lyte-internal standard) with time. Most often, for the  quantitative LA- ICP-MS analysis, a combination of calibra- tion by RM and internal standardisation is used [10], when the ratio of the analyte and the internal standard intensities is taken as the analytical signal. Since the addition of internal standards is not easily and rou- tinely carried out with solid samples, natu- rally occurring elements are used as internal standards. These are elements that are found Table 1 G-Probe samples and microanalytical reference materials used for the study G-Probe round number Name Material Origin GP-5 BBM-1G Basalt USGS GP-6 BNV-1G Basalt NIST SRM 688 GP-8 SL-1G Soda-lime glass Corning Glass Works, 1976 GP-10 GP-MAPS Calcium phosphate USGS GP-11 DVA-1G Diabase W-2 USGS SRM GP-12 GSM-1G Gabbro San Marcos Mountains, Southern California Batholith, USA GP-14 BBRZ-1G Basalt BRP-1 SRM (University of Campanias, Brazil) GP-15 SyMP-1G Syenite SyMP-1 USGS SRM GP-16 BSWIR-1G Mid-Ocean Ridge Basalt (MORB) Southwest Indian Range, the border of the African and Antarctic Plates RM MAPS-4* Calcium Phosphate with trace elements USGS MRM RM TB-1G** Basalt TB-1 USGS SRM (Golden, Colorado, USA) *Used for calibration for the analysis of phosphate rock samples. **Used for calibration for the analysis of silicate rock samples. 8 in both the samples and calibration material, and for which the concentrations are known in both materials [10]. The concentration of the internal standard can be obtained from an analysis using an alternative meth- od (for example, EPMA, EDS-SEM) or from the known elemental stoichiometry when crystalline materials are analysed. Often, one of the major elements serves as the in- ternal standard containing in the sample in sufficiently large quantities (e.g. silicon for silicate rocks, calcium and phosphorus for carbonate and phosphate minerals, zir- conium, hafnium for zircons, etc.). In  this case, finding the  analyte con- tent in the sample is carried out according to the formula: ( ) ( )’ C C ’ a a as s cala cal I K I = ⋅ ⋅ (1) where ( )’ a a iss s sI I I= is  the  analyte rela- tive intensity normalised to  the  internal standard intensity in  the  sample, both background subtracted, ( )’a a iscal cal calI I I= is  the  analyte relative intensity normal- ised to  the  internal standard intensity the in the calibration material, both back- ground subtracted, Cacal is  the  analyte concentration in the calibration material, C Cis iss calK = is the coefficient taking into account the  ratio of  the  concentrations of the internal standard in the sample (C )iss and the calibration material .(C )iscal Thus, the silicate samples were analysed using a TB-1G basalt glass for calibration with internal silicon standardisation. Phos- phate sample was analysed using MAPS-4 synthetic phosphate RM for calibration and internal calcium standardisation. Table 2 LA and ICP-MS operational parameters during the analysis of silicate and phosphate samples LA parameters Phosphate rocks [8] Silicate rocks [9] Energy, mJ 0.9 0.9 frequency Hz 20 20 Spot diameter, microns 50 50 Pulse Duration, ns <10 <10 ICP-MS parameters RF power 1300 W Plasma Ar flow rate 16 l/min Auxiliary Ar flow rate 1.0 l/min Sample Ar flow rate 1 l/min Data acquisition parameters Dwell time 8 ms Quadrupole settling time 1.5 ms Sweeps for total analysis 480 (240 for background and 240 for ablation) Sweeps per reading 3 Points per peak 1 Total analysis time ~240 s 9 Results and discussion Scoring and statistical analysis in GeoPT is undertaken according to the ISO 13528 Standard relating to  statistical methods used in proficiency testing [11] based on the earlier recommendations of the IUPAC International Harmonised Protocol [12]. According to the GeoPT Protocol [13], the results of the analysis are evaluated us- ing z-scores in the form: z = (xi – xpt)/σpt (2) where xi is the result of the analysis of a par- ticular laboratory, xpt is the assigned value of the element content in the test sample, σpt is the corresponding standard deviation for proficiency testing (SDPT), or target precision, based on a  GeoPT fitness for purpose criterion. In the G-Probe programme, the values of element concentrations obtained dur- ing analysis in USGS laboratories, as well as  the  data from NIST SRM certificates (in case glass was fused from a standard sample), and results of  analysis of  bulk powder samples from previous rounds are taken as xpt assigned elemental content. In  the  Protocol of  the  G-Probe pro- gramme, a model of the standard deviation dependency on concentration is adopted as  an  σH estimate of  the  target precision in the form of the Horwitz function [14]: σpt = σH = 0.02 · xpt 0.8495 (3) where the  values of  concentrations and precision should be expressed in  mass fractions (for example, 1  ppm = 10–6, 1% = 0.01); the coefficient k = 0.02 corre- sponds to the results of the second category of results — applied geochemistry. Accordingly, a z-score outside the range ±3 implies that an  unacceptable source of bias may be present in the participant’s analytical system and that the  need for remedial action should be considered. Z-scores more extreme than ±2 carry the same message to a lesser degree [13]. However, an assumption has been made by [15] that the dependency of the stand- ard deviation on concentration in the form of the Horwitz function used in the Ge- oPT programme does not take into ac- count the main source of analytical errors due to the distribution of the component being determined in  the  sample, while determining trace elements additional er- rors in the measurement results contribute to  the  heterogeneity of  the  distribution. Thus, the dependence σH = f(x), based on approximation by the Horwitz function, is too strict for the determination of trace elements. The authors [15] have proposed the  following approximations of  the  de- pendency of the permissible standard de- viation for high and low concentrations: S = 0.005 · xpt 0.5 if xpt > 0.1% (4) S = 0.035 · xpt 0.8495 if xpt ≤ 0.1% (5) The results of LA-ICP-MS analysis and z-scores are given in Supplementary Ta- ble S1. The measurement result is the mean value of two measurements performed on two different sample fragments each con- sisting of two parallel measurements. For a number of samples studied (Ta- ble  1), the  values of  the  z-scores were calculated in  accordance with the  Pro- tocol of the G-Probe programme, as well as the z’ — scores using the dependency of the standard deviation for low concen- trations S = 0.035 · C0.8495 (C ≤ 0.1%) since almost all the concentrations of determined elements in studied samples were less than 0.1% [15]. Figure 1 shows the generalised distribu- tion of a number trace elements depend- ing on the z-scores (a) and the z’ — scores (b) for the samples studied. It can be seen that the results of the LA-ICP-MS analy- 10 sis of most trace elements are satisfactory, a certain amount falls outside the range ±3. When the quality assessment is performed using the  z’  — scores, a  greater number of  results are recognised as  satisfactory, and only some elements still fall outside the range ±3. Most often, during LA-ICP-MS analy- sis of geological samples, a number of ele- ments are of the greatest interest — these are the rare-earth elements (REE), yttrium, uranium, thorium and lead. To illustrate the  quality of  their analysis, z (z’) plots were constructed (Fig. 2). Figure  2 shows that when assessing the quality of analysis by the z’ — scores, the vast majority of the results of the de- termination of  REE, yttrium, uranium, thorium and lead fall within the  range not exceeding ±2 with the  insignificant number of outliers for SL-1G, SyMP-1G и BSWIR-1G, thus indicating the satisfac- tory results. The outliers for SL-1G and SyMP-1G samples can be explained by  its ma- jor element composition different from the TB-1G natural basaltic glass. SL-1G so- da-lime glass contains ~71% of SiO2, which is rather close to NIST SRM 612 glass (72% SiO2). SyMP-1G syenite (55.3% SiO2) has intermediate composition according to the igneous rock classification by silicon Fig. 1. Summary plots of element distribution versus z-scores (a) and z’ — scores (b) Fig. 2. z (a) and z’ (b) plots for REE, Y, U, Th, and Pb 11 dioxide content (52–63% of SiO2), while all natural basaltic, gabbro and diabase glasses have mafic composition, and their silica content is generally 45–52%. Thus, the difference in the composition of major elements especially if they are used as in- ternal standards, can lead to  the  biased results. When analysing igneous rocks composed of  silicate minerals, a  careful selection of  calibration materials should be performed taking into account their composition classified by silicon dioxide content — mafic, intermediate or felsic. Conclusions This study describes the  matrix- matched calibration approach to the anal- ysis of geological samples by LA-ICP-MS using the  samples provided by  G-Probe programme (International Association of Geoanalysts). A wide range of G-Probe samples with the  composition of  major elements (matrix) similar to natural geo- logical objects (silicates, phosphates and carbonates) can be used as  calibration samples especially in geoanalytical labo- ratories in order to correct for differences in behaviour during laser interaction and evaporation of  solid samples. The  con- ducted evaluation of the analysis quality for a number of geological samples using z-scores has proved that the combination of matrix-matched external calibration and internal standardisation in LA-ICP-MS mi- croanalysis of rocks and minerals allows satisfactory results to be obtained for most of the determined elements. Acknowledgements The study was carried out at the Geoanalytic Collective Use Centre and supported by RSF grant No. 16-17-10283. References 1. Jackson SE. The application of Nd:YAG lasers in LA ICP-MS. Ed. P. Sylvester. Laser ablation ICP-MS in the Earth Sciences: Principles and applications. Short Course Series. Mineralogical Association of Canada. 2001;29:29–45. 2. Jochum KP, Stoll B, Weis U, Jacob DE, Mertz-Kraus R, Andreae MO. Non-matrix- matched calibration for the multi-element analysis of geological and environmental samples using 200 nm femtosecond LA-ICP-MS: A Comparison with nanosecond lasers. Geostandards and Geoanalytical Research. 2014;38:265–92. DOI:10.1111/j.1751-908X.2014.12028.x 3. O’Connor C, Sharp B, Evans P. On-line additions of aqueous standards for calibra- tion of laser ablation inductively coupled plasma mass spectrometry: theory and comparison of wet and dry plasma conditions. J Anal At Spectrom. 2006;21:556–65. DOI: 10.1039/B600916F 4. Votyakov SL, Adamovich NN. O protsessakh lazernogo ispareniya i ispol’zovanii vodnykh standartov pri LA-ISP-MS-analize ryada mineralov [About the processes of  laser evaporation and using of  aqueous calibration solutions for LA-ICP-MS analysis of a number of minerals]. Litosfera. 2011;4:56–69. 12 5. Jochum KP, Stoll B. Reference materials for elemental and isotopic analyses by LA — (MC) — ICP — MS: successes and outstanding needs. Mineralogical Association of Canada. Short Course 40. 2008;40:147–68. 6. Jochum KP, Garbe‐Schönberg D, Veter M, Stoll B, Weis U, Weber M, Lugli F, Jentzen A, Schiebel R, Wassenburg JA, Jacob DE, Haug GH. Nano‐Powdered Calcium Car- bonate Reference Materials: Significant Progress for Microanalysis? Geostandards and Geoanalytical Research. 2018;43(4):595–609. DOI: 10.1111/ggr.12292 7. Adamovich NN, Kiseleva DV. Mezhlaboratornyy kontrol’ kachestva LA-ICP-MS mikroanaliza v ramkakh programmy testirovaniya geoanaliticheskikh laboratoriy G-Rrobe [Interlaboratory quality control of LA-ICP-MS microanalysis according to G-Probe proficiency testing programme for geoanalytical laboratories]. Ezhe- godnik-2011. Tr. IGG UrO RAN. 2012;159:222–3. 8. Adamovich NN. Vozmozhnosti metoda mass-spektrometrii s induktivno svyazan- noy plazmoy i lazernoy ablyatsiey dlya lokal’nogo opredeleniya mikroelementnogo sostava apatitov [The possibilities of LA-ICP-MS analysis of trace element composi- tion of apatites with high spatial resolution]. Ezhegodnik-2012. Tr. IGG UrO RAN. 2013;160:309–13. 9. Adamovich NN, Glavatskikh SP, Votyakov SL. Issledovaniya osobennostey ispareniya silikatov, fosfatov i sul’fidov kak osnova dlya sozdaniya metodik lokal’nogo opredele- niya mikroelementnogo sostava v metode mass-spektrometrii s lazernoy ablyatsiey [The studies of evaporation of silicates, phosphates, and sulphides as a basis for LA methodologies of trace element analysis with high spatial resolution]. Ezhegod- nik-2012, Tr. IGG UrO RAN. 2013;160:314–8. 10. Longerich HP, Jackson SE, Günther D. Laser ablation inductively coupled plasma mass spectrometric transient signal data acquisition and analyte concentration calculation. J Anal At Spectrom. 1996;11:899–904. DOI: 10.1039/JA9961100899 11. ISO 13528, 2015. Statistical methods for use in proficiency testing by interlaboratory comparison, Geneva: ISO; 2015. 12. Thompson M, Ellison SLR, Wood R. The international harmonised protocol for the proficiency testing of analytical chemistry laboratories. IUPAC Technical Report. Pure Appl. Chem. 2006;78:145–96. DOI:10.1351/pac200678010145 13. Thompson M. GeoPT™. Protocol for the operation of proficiency testing scheme. International Association of Geoanalysts. 2018. 14. Horwitz W, Kamps LR, Boyer KW. Quality assurance in the analysis of foods for trace constituents. J. AOAC Int. 1980;63:1344–54. 15. Kuznetsova AI, Zarubina OV. Mezhlaboratornyy kontrol’ kachestva pryamogo atomno-emissionnogo analiza s ispol’zovaniem serii gornykh porod programmy testirovaniya geoanaliticheskikh laboratoriy GeoPT [Interlaboratory quality con- trol of direct atomic emission analysis using the series of rock samples of GeoPT proficiency testing programme for geoanalytical laboratories]. Analitika i kontrol’. 2005;3(9):230–9.