Natural chromite as a reference material for LA-ICP-MS analyses: development and calibration

Abstract

This study comprehensively characterizes the natural chromite reference material UG1-W and establishes a methodological framework for its quantitative LA-ICP-MS analysis. Integrated mineralogical and geochemical analysis confirms exceptional homogeneity and validates optimized calibration approaches for UG1-W. High-resolution BSE imaging combined with automated mineralogy (TIMA) reveals a chromite–plagioclase dominated microstructure (80.8% chromite and 17.3% plagioclase) with minor accessory phases. Multi-analytical methods (LA-ICP-MS, EPMA, ICP-OES, ICP-MS, and TXRF) are employed to verify the homogeneity of both major and trace elements. Systematic calibration assessments yield three key findings: (1) persistent matrix-induced analytical biases occur when using synthetic glass standards; (2) calibration strategies yield divergent results despite employing internal standards; (3) matrix-matched calibration achieves superior accuracy for chromite analysis, exhibiting relative deviations below 5% against certified values. Collectively, this work establishes UG1-W as a homogeneous chromite reference material and unequivocally demonstrates the necessity of matrix-matched standardization for accurate LA-ICP-MS analysis of chromite. These findings significantly improve the measurement accuracy for refractory mineral systems and provide a robust analytical framework for geochemical studies of chromite-bearing lithologies.

---

1. Introduction

Chromite, characterized by a spinel-type structure (general formula: AB₂O₄), serves as a critical petrogenetic indicator in ultramafic systems due to its compositional variability that records magmatic evolution processes.^1–5 The mineral's remarkable resistance to post-magmatic alteration (distinct from primary silicate phases) preserves pristine geochemical signatures even within metamorphosed terranes. This resilience makes chromite indispensable for reconstructing original melt characteristics where other petrogenetic indicators have been obscured.^6–8 Its characteristically low silica content reflects an ultramafic provenance and stability within mantle-derived melts.⁹ This diagnostic depletion, coupled with systematic variations in Mg–Fe–Al–Cr ratios, enables precise constraints on primary magma compositions and mantle source characteristics, particularly in Archean crust–mantle systems, where it acts as a robust archive of Precambrian geodynamic processes.^10–14

While inductively coupled plasma mass spectrometry (ICP-MS) provides an effective multi-element technique for bulk geochemical analysis, chromite presents dissolution challenges during chemical separation.¹⁵ X-ray fluorescence (XRF) offers rapid major element analysis with minimal sample preparation, achieving precision within ±1–2 wt% for bulk compositions, but its utility is constrained by poor trace element sensitivity.¹⁶ Pre-treatment for bulk analysis is susceptible to the introduction of extra contaminants. Furthermore, discriminating variations in elemental composition among different mineral phases within chromite, such as primary chromite and its secondary alteration products, poses significant analytical challenges. This issue is particularly pronounced when chromite contains inclusions, as conventional chemical analytical methods are prone to producing distorted data in such heterogeneous samples. Consequently, the analysis of chromite's major and trace elements has significantly advanced through microanalytical techniques. Major elements are predominantly quantified using electron probe microanalysis (EPMA), offering high precision and spatial resolution (∼1–5 μm).^17–20 For trace elements, laser ablation inductively coupled plasma mass spectrometry (LA-ICP-MS) has become essential. It possesses absolute advantages in its superior spatial resolution at the micrometer (μm) scale, low contamination, affordability, high efficiency, as well as minimal water-related spectral interferences.^21–24 Despite these advances, the complex physicochemical processes governing laser ablation of solid samples hinder a complete mechanistic understanding, particularly concerning elemental decoupling during LA-ICP-MS analysis.^25–28 Elemental fractionation may arise from non-stoichiometric ablation or incomplete vaporization of large particles in the plasma.²⁹ Matrix-matched reference materials are fundamental for ensuring analytical accuracy and precision in in situ microanalysis. They serve critical functions in instrument calibration, data quality assurance, and interlaboratory reproducibility.³⁰ Current methodologies for trace element quantification in chromite predominantly rely on synthetic glass standards (e.g., the National Institute of Standards and Technology (NIST) 61X series glass standard, USGS GSX-1G series reference materials and the BHVO-2G, BCR-2G, and BIR-1G reference standards).³¹ These standards are produced by high-temperature fusion of pre-mixed elemental oxides or natural rock powders and establish trace element gradients spanning four orders of magnitude (∼0.5–500 μg g⁻¹). However, inherent matrix discrepancies arise from the elevated SiO₂ and depleted Cr₂O₃ contents in these glasses relative to natural chromite. This matrix disparity underscores the pressing need for chromite-specific reference materials that preserve both the major element stoichiometry (Fe²⁺–Mg–Cr³⁺–Al³⁺ spinel structure) and trace element partitioning behavior inherent to magmatic chromitites.

The primary objective of this study is to evaluate calibrated strategy differences specific to chromite analysis using LA-ICP-MS. To achieve this, a compositionally homogeneous natural chromite material, derived from mantle-derived chromitites and rigorously characterized via multi-technique validation protocols (including EMPA, ICP-OES, ICP-MS, LA-ICP-MS, and TXRF), was utilized. Unlike conventional glass standards synthesized through high-temperature fusion, this chromite specimen preserves natural compositional patterns and trace element integrity, having undergone strictly non-thermal treatment to prevent post-crystallization modification. Widely used calibration and quality assurance reference materials for LA-ICP-MS, including NIST SRM 61X series, USGS basalt glasses, MPI-DING references were employed to characterize elemental fractionation during chromite analysis. Our findings significantly improve measurement accuracy for refractory mineral systems and providing a robust analytical framework for geochemical studies of chromite-bearing lithologies.

2. Experimental

2.1 Sample description and preparation

The developed reference material, UG1-W, was sampled from the UG1 chromitite layer (Upper Group 1) at the Impala Platinum Mine in the western limb of the Bushveld Complex, South Africa. Stratigraphically positioned within the Upper Critical Zone of the Rustenburg Layered Suite, the UG1 layer is bounded below by a plagioclase-poor pyroxenite unit and capped by a norite–anorthosite sequence.³² The UG1 chromitite layer is composed predominantly of massive and semi-massive chromitite. BSE imaging and TIMA phase mapping (Fig. 1) revealed that the UG1-W sample consists mainly of chromite (80.85%), plagioclase (17.31%), and minor accessory minerals such as olivine, orthopyroxene, and biotite. The chromite crystals are euhedral to subhedral and relatively small with 0.1–0.5 mm in size. Plagioclase occurs as an interstitial phase between chromite grains.

Fig. 1 BSE image (a) and TIMA mineral phase map (b) of a hand specimen fragment of chromitite.

The hand specimen UG1-W was sectioned into 16 parts along the A, B, C and D axial directions, as shown in Fig. 2. Two fragments were micro-drilled from each section, yielding a total of 32 particles for in situ micro-analysis. Furthermore, a total of 16 positions were randomly chosen across the entire specimen for bulk analysis, ensuring an even distribution. The sampling procedure employed a micro-drill with a diameter of 2.0 mm. Prior to drilling each position, the drill bit was thoroughly cleaned using dust-free paper and anhydrous ethanol three times. The epoxy mounts were meticulously polished multiple times until achieving a flat and smooth surface free from any discernible scratches.

Fig. 2 The whole hand specimen of chromitite UG1-W described in this study.

2.2 Analytical techniques

2.2.1 Automated mineral quantitative analysis system. Mineralogical characterization was conducted using a TESCAN Integrated Mineral Analyzer (TIMA) system at the State Key Laboratory of Critical Mineral Research and Exploration (SKLCMRE), Institute of Geochemistry, Chinese Academy of Sciences (IGCAS). The TIMA platform integrates a thermionic tungsten-filament SEM (VEGA II LSU) with four symmetrically arranged silicon drift EDX detectors, controlled using TIMA software (v1.5.50) for automated phase mapping. High-resolution mapping employed a 15 mm working distance with 3 μm pixel resolution, 25 kV accelerating voltage, and 4.76 nA beam current (corresponding to 18.2 μA emission current), ensuring sufficient X-ray generation while minimizing sample charging. Carbon-sputtered thin sections were analyzed, and the total scan required approximately 14 hours with 1000 counts collected per analysis point.

2.2.2 Electron probe X-ray micro-analyzer (EPMA). Major elemental analyses were conducted on a JEOL JXA8100 EPMA at the SKLCMRE in IGCAS. Calibrations were carried out using the standards provided by the Structure Probe, Inc. (SPI Supplies). Matrix effects were corrected using the ZAF software provided by JEOL. The working conditions were at an accelerating voltage of 15 kV and a beam current of 10 nA, with a 1 μm beam spot and a counting time on peak of 10–30 s.

2.2.3 Total reflection X-ray fluorescence (TXRF). Total-reflection X-ray fluorescence (TXRF) analyses were conducted using an S4 T-STAR spectrometer (Bruker AXS Microanalysis GmbH, Berlin, Germany) at Bolan Testing Technology (Shanghai) Co., Ltd. The system configuration included a molybdenum-target X-ray tube operating at maximum power (50 kV accelerating voltage, 1000 μA tube current, and 50 W total power), a multilayer monochromator for beam conditioning, and a silicon drift detector demonstrating energy resolution < 149 eV (FWHM) at the Mn Kα emission line (5.9 keV).

Sixteen chromite-hosted particles were randomly isolated from the UG1-W specimen surface. Monomineralic chromite grains were manually extracted from each particle under binocular microscopic observation, with individual aliquots averaging ∼50 μg in mass. The samples underwent mechanical comminution in an agate mortar until achieving particle sizes below 50 μm (corresponding to 300 mesh in the Tyler standard sieve classification). Analytical conditions for chromite quantification comprised: 50 kV excitation voltage, 1000 μA tube current, 200 μm aluminum primary beam filter, 1000 s acquisition time, and ambient atmospheric conditions. Arsenic (As) served as the internal standard element, with a spiked concentration of 1000 μg g⁻¹.

2.2.4 X-ray photoelectron spectroscopy (XPS). XPS analysis was performed using a Thermo Fisher Scientific K-Alpha Plus spectrometer at Guizhou University. The system is equipped with a monochromatic Al Kα X-ray source (1486.67 eV) operating at 72 W, along with an EX06 ion gun capable of delivering energies from 200 eV to 4 keV for sample cleaning and depth profiling. The instrument features a 180° double-focusing hemispherical analyzer with a 128-channel detector and a dual-beam charge neutralization system. The analyzer was oriented at 90° relative to the sample surface, with an acceptance angle of ±30°. The spectrometer work function was set to 4.2 eV, and the base pressure in the analysis chamber was maintained at 10⁻⁹ mbar. All measurements were conducted at room temperature using a spot size of 500 μm, with the analyzed area aligned accordingly.

2.2.5 Bulk analytical techniques. Nine fragments were randomly obtained along different directions of the hand specimen (UG1). Among them, six fragments were used for ICP-MS analysis, and three particles were used for ICP-OES analysis. For each fragment, 10 mg of monomineralic chromite was selected, and each sample underwent independent digestion procedures.
SKLCMRE. Approximately 10 mg of finely pulverized sample was weighed into a PTFE digestion vessel, followed by sequential addition of 1 mL of HF and 0.5 mL of HNO₃. The sealed vessel was placed within a stainless steel pressure jacket and heated in a temperature-controlled oven at 200 °C for 35 hours. After cooling to ambient temperature, the resultant digest was carefully evaporated to near dryness on a hotplate. Subsequently, 0.5 mL HNO₃ and 1.5 mL hydrochloric acid (HCl) were added to the residue for secondary digestion at 140 °C for 24 h. The sample was reconstituted using a rhodium internal standard solution (1 μg mL⁻¹), 2 mL deionized water, and 0.5 mL HNO₃. The final mixture was heated at 150 °C for 24 h to ensure complete dissolution. It was ensured that no precipitate was present in the digestion solution. The detailed procedure was described by Qi et al.³³

The major element composition of UG1-W was analyzed using ICP-OES (Agilent 720) at the SKLCMRE. BCR-1 and BHVO-1 were used as monitoring samples during ICP-OES analysis.

Trace elements in the chromite separates were determined using a PlasmaQuant Elite ICP-MS spectrometer (Analytik Jena, Germany) at the SKLCMRE, IGCAS. Sample introduction was accomplished via a MicroMist concentric pneumatic nebulizer (Glass Expansion, USA) coupled with a cooled double-pass Scott-type spray chamber. Analyses were conducted in standard mode for a suite of eight trace elements (⁴⁵Sc, ⁴⁹Ti, ⁵¹V, ⁵⁵Mn, ⁵⁹Co, ⁶⁰Ni, ⁶⁶Zn, and ⁶⁹Ga). Plasma conditions were optimized and monitored regularly, with the following performance criteria maintained: oxide production rates (CeO⁺/Ce⁺) < 1.5% and double-charged ion formation rates (Ce²⁺/Ce⁺ < 1.5%) < 1.5%. Quantification was performed using external multi-element standard solutions, and ¹⁰³Rh was employed as an internal standard to correct for instrumental drift and matrix effects. Analytical accuracy and precision were verified through the concurrent digestion of international basaltic whole-rock reference materials (BHVO-2 and BCR-2) with the unknown samples. This procedure yielded a reproducibility of better than 5% RSD for all reported elements.

IGCAGS. Chromite separates underwent pre-treatment procedures at the Laboratory of Isotope Geology, Institute of Geology, Chinese Academy of Geological Sciences. Approximately 10 mg of the fine ground sample was dissolved in sealed Teflon vessels jacketed in steel acid digestion bombs in a 2 : 1 mixture of HF and HCl with several drops of HClO₄ at 190 °C for 72 h. The dissolved chromite samples were transferred to Savillex screw-top beakers and evaporated to dryness at 120 °C. A 1 : 1 volume mixture of concentrated HNO₃ and high-purity water (18 MΩ) was added, and the mixture was refluxed at 150 °C for 1 day. It was ensured that no precipitate was present in the digestion solution. Finally, the solution was diluted accurately to 50 g with 2% HNO₃ (containing 10 ng per g In) for trace element measurement.³⁴

Solution ICP-MS analyses were applied to determine trace elements using an Agilent 7900 ICP-MS (Agilent Technologies, USA). Sample introduction was carried out using a MicroMist concentric nebulizer (Glass Expansion, USA) coupled with a cooled double-pass Scott-type spray chamber and a peristaltic pump to maintain consistent aspiration conditions. Argon served as the nebulizer, auxiliary gas, and plasma gas. The analysis was operated in He mode to mitigate polyatomic interferences. Instrument control, data acquisition, and processing were performed using the ICP-MS MassHunter software package (Agilent Technologies). ¹⁰³Rh was used as an internal standard to correct for instrumental drift and matrix effects. Detailed operating parameters are summarized in Table S1. The analytical accuracy and precision were monitored by digesting international whole-rock reference materials (W-2a and BCR-2) in parallel with the unknown samples. The measured values for all elements agreed with recommended values to within 10%, confirming satisfactory method performance.

2.2.6 LA-ICP-MS. Intercomparison among participating laboratories is essential. Detailed analytical parameters used in different laboratories are listed in Table S2. To minimize potential data bias introduced by the calibration strategy, all datasets in this study were uniformly processed using the ICPMSDataCal software.^35,36 Besides, acquired data underwent reduction using the following methods: (1) internal standardization using Al as the internal standard elements (Al-internal method), (2) internal standardization using Fe as the internal standard (Fe-internal method), (3) normalization to 100% total oxides using Al as the normalization element (Al-normalized method), and (4) normalization to 100% total oxides using Fe as the normalization element (Fe-normalized method).
KLCMRE. Major and trace element concentrations were determined using an ASI RESOLution-LR-S155 laser microprobe equipped with a Coherent Compex-Pro 193 nm ArF excimer laser at the State Key Laboratory of Critical Mineral Research and Exploration (SKLCMRE), Institute of Geochemistry, Chinese Academy of Sciences. An Agilent 7700x ICP-MS was used to acquire ion-signal intensities. The ablated aerosol was mixed with Ar (900 mL min⁻¹) and He (350 mL min⁻¹) within the ablation cell before exiting. The instrument was tuned with NIST 610 before measurement to achieve optimal performance, characterized by maximum sensitivity and minimized rates of double-charged (<0.3%) and oxide ions (<0.3%). The pulse/analogue (P/A) factors were optimized prior to analysis. Quantitative analysis was performed using 40 μm beam spots, 5 Hz pulse frequency, and a fluence of 5 J cm⁻², incorporating a background acquisition of ∼20 s (gas blank) and ∼45 s of data acquisition. Elemental contents were calibrated against multiple reference materials (GSE-1G, BCR-2G, BIR-1G and BHVO-2G).³⁷ The preferred values of element concentrations for the reference glasses were obtained from the GeoReM database (https://georem.mpch-mainz.gwdg.de/). The calibration strategy adopted multi-external calibration with content-weighted reference materials fitting to mitigate matrix effects. An in-house quality control material Cr_inhouse (developed by Dr Dany Savard from QUAC; unpublished) served as a quality control standard, and the certified values for this reference material are provided in Table S3.
USTC. A Coherent ArF excimer UV 193 nm laser ablation system (GeoLas pro) coupled to an Agilent 7700e quadrupole ICP-MS was used to acquire the ion-signal intensities at the Key Laboratory of Crust-Mantle Materials and Environments, University of Science and Technology of China (USTC). Analyses were performed using a laser beam diameter of 40 μm, a repetition rate of 5 Hz, and a fluence of 5.0 J cm⁻². BIR-1G, BHVO-2G, BCR-2G were used as external standards for trace element calibration. The calibration approach employed multi-external standardization with content-weighted fitting of reference materials to compensate for matrix effects. NIST 610 was analyzed as a quality control standard.
SYSU. LA-ICP-MS analysis was conducted at the Guangdong Provincial Key Laboratory of Marine Resources and Coastal Engineering (SYSU) with a 193 nm ArF excimer laser ablation system (NWR HE) coupled with an Agilent 7900 ICP-MS. Analyses were performed with a beam size of 40 μm, a repetition rate of 5 Hz, and a laser fluence of 5 J cm⁻². The ablation sequence comprised 18 s of background, 45 s of sample ablation and a 20 s washout period. The external element standard references GSE-1G and BHVO-2G were analyzed for every ten samples. The calibration strategy adopted multi-external calibration with content-weighted fitting of reference materials. NIST610 was employed as a quality control standard.
CUGB. An Agilent 7900 ICP-MS instrument (Agilent Technologies, USA) coupled to a 193 nm ArF excimer laser system (RESOlution S155 LR, ASI, USA) was employed to analyze chromite at the Mineral Laser Microprobe Analysis Laboratory (Milma Lab), China University of Geosciences, Beijing (CUGB). BCR-2G was used as the calibration material with a laser beam diameter of 40 μm and a repetition rate of 5 Hz.³⁸ NIST610 was used as a quality control reference material.

3. Results and discussion

3.1 Elemental composition

Thirty-two chromite grains were selected for quantitative elemental analysis, specifically focusing on element concentrations using LA-ICP-MS and EMPA. Two additional particles underwent LA-ICP-MS mapping to assess spatial heterogeneity (Fig. S1). The data demonstrated remarkable compositional consistency in both major (Fe, Cr, Mg, and Al) and trace elements (Sc, Ti, V, Mn, Co, Ni, Zn, and Ga).

Major element mass fractions were determined by inductively coupled plasma optical emission spectroscopy (ICP-OES) and EMPA (Table 1). Trace elements were quantified using solution nebulization inductively coupled plasma mass spectrometry (SN-ICP-MS) and TXRF (Table 2). Each SN-ICP-MS data represent the mean of three independent digestions. TXRF data represent the mean of sixteen replicate analyses for each element. During single-mineral separation, grains with well-developed crystal forms and minimal visible inclusions were preferentially selected. This selection was conducted with the aim of justifying the assumption that the analyzed elements are predominantly hosted within the chromite structure, with negligible contributions from minor impurity phases or inclusions during bulk analysis. The chromite sample (UG1-W) exhibits Cr concentrations ranging from 28.42 to 29.78% (m m⁻¹), Fe from 20.82 to 21.77% (m m⁻¹), Mg from 5.05 to 5.58% (m m⁻¹), and Al from 9.41 to 9.98% (m m⁻¹), with the corresponding mean oxide composition (according to EPMA data) of 42.52% Cr₂O₃, 18.63% Al₂O₃, 27.44% FeO and 9.05% MgO. Trace element concentrations (Sc, Ti, V, Mn, Co, Ni, Zn, and Ga) range from tens to thousands of μg g⁻¹.

Table 1 The mass fraction of major elements determined by ICP-OES and EMPA (% m m⁻¹)

	Recommended value (% m m^−1)	EPMA (n = 50)			ICP-OES (n = 3)
		Mass fraction (% m m^−1)	1 SD	% RSD	Mass fraction (% m m^−1)	1 SD
Cr2O3	42.52	42.52	0.49	1.16	—	—
FeO	27.44	27.61	0.22	0.78	27.26	0.57
MgO	9.05	8.94	0.17	1.85	9.15	0.98
Al2O3	18.63	18.31	0.26	1.44	18.94	0.61

---

Table 2 Measurement results for trace elements (μg g⁻¹) determined using bulk chemical analysis

	Recommended value (μg g^−1)	SN-ICP-MS (SKLCMRE)		SN-ICP-MS (IGCAGS)		TXRF
		Mass fraction (μg g^−1)	1 SD	Mass fraction (μg g^−1)	1 SD	Mass fraction (μg g^−1)	1 SD
Sc	7.43	7.31	0.15	7.54	1.45	—	—
Ti	4842.91	4866.04	1.41	4783.49	0.59	4879.2	5.6
V	2805.05	2927.00	0.30	2720.25	0.46	2767.9	5.5
Mn	1792.21	1768.00	0.61	1808.63	0.54	1800.0	1.5
Co	323.05	314.14	0.20	333.81	0.41	321.2	1.3
Ni	986.34	960.54	0.57	1023.49	0.40	975.0	1.8
Zn	978.79	1011.51	3.60	931.86	1.11	993.0	8.1
Ga	61.08	68.90	1.04	58.40	0.96	55.9	3.8

---

3.2 Elemental homogeneity verification

Chemical homogeneity is defined as variations in element mass fraction that do not exceed the measurement uncertainty of the analytical method.³⁹ Homogeneity of major elements (Cr, Fe, Mg, and Al) within chromite grains was evaluated using EPMA. Measurement precision for all major elements achieved better than 2% relative standard deviation, consistent with EPMA repeatability. EMPA results demonstrated excellent agreement with bulk ICP-OES data (Table 1), collectively confirming chemical homogeneity for major element compositions in the chromite sample. For trace element homogeneity assessment, LA-ICP-MS analyses were conducted across multiple grains in the KLCMRE. All measurements were calibrated using multi-external standardization using certified reference materials (GSE-1G, BIR-1G, BHVO-2G and BCR-2G). ⁵⁷Fe served as the internal standard element. The raw data of LA-ICP-MS can be obtained from SI Appendix 2. To minimize uncertainties from iron oxidation state assumptions, the Fe²⁺/∑Fe ratio in chromite was constrained to 0.76 based on charge balance calculations using EPMA data (see the calculation formula in SI Appendix 3), which was further validated by X-ray photoelectron spectroscopy (XPS, Fig. S2). A total of 480 analytical spots were analyzed in the KLCMRE, evenly distributed across 32 grains (15 spots per grain in a gridded array on resin-mounted samples). Elemental distribution profiles for the sample UG1-W revealed no detectable domains of significant elemental heterogeneity. Quantitative evaluation confirmed RSDs for all elements across 480 spots remained below 5%. In this study, a homogeneity index (H) was additionally applied to assess trace element homogeneity.^40,41,46 This index, defined as the ratio of measurement uncertainty to total combined uncertainty, provides a statistical metric, where H = 1 indicates homogeneity within analytical uncertainty and H >3 denotes significant chemical heterogeneity.⁴⁷ The H index is interpretable as a specialized case of the F-test when degrees of freedom approach infinity. Analysis of trace elements using this index confirmed that these elements exhibited homogeneity within analytical uncertainties (Fig. 3).

Fig. 3 Homogeneity of UG1-W for various elements. Relative standard deviation (% RSD) of repeat analyses versus average signal uncertainty. All datasets were compiled from 480 LA-ICP-MS data.

3.3 Calibration strategies

3.3.1 Glass standard-based calibration in LA-ICP-MS. LA-ICP-MS calibration, incorporating both internal and external calibration, can correct matrix-dependent elemental fractionation. According to Goldschmidt's geochemical classification, Al is categorized as a lithophile element, whereas Fe is classified as a siderophile element. Capitalizing on these contrasting geochemical affinities, Al and Fe were strategically employed as normalization elements to research matrix variations between glass standards and natural chromite. Al-normalized sensitivity ratios (Al-NSRs) and Fe-normalized sensitivity ratios (Fe-NSRs) were calculated for the chromite sample UG1-W relative to multiple glass standards (Fig. 4).³⁵ Under conditions free from matrix-induced elemental fractionation, these ratios were expected to approximate 1.³⁵ Although most Al/Fe-NSR values clustered around 1, greater variability was observed for Sc, Cr, Fe, Ni, Zn, and Ga. Notably, Fe-NSRs obtained using NIST SRM 612 showed significant deviations compared to those derived from USGS/MPI-DING glasses. The divergent ablation behavior may be attributed to three factors: (1) the measurements of ⁵⁷Fe in this work may have been complicated by spectral overlap from Ca-based polyatomic interferences (⁴⁰Ca¹⁶O¹H); (2) although Fe is present at comparable trace-level concentrations in NIST SRM 612 and NIST SRM 610, all other elements are present at lower concentrations in NIST SRM 612; and (3) the contrasting optically transparent NIST glasses and natural-composition USGS/MPI-DING glasses likely originate from contrasting physicochemical properties, including absorption characteristics that influence the ablation yield and plasma transport efficiency;^42,43 Additionally, substantial deviations observed during standardization using ATHO-G may originate from its exceptionally low trace element concentrations (e.g. Cr, Co, and Ni).

Fig. 4 Normalized sensitivity ratios (NSRs) of the chromite sample (UG1-W) relative to standard glasses (NIST SRM 61X, USGS and MPI-DING glasses). NSR = (normalized sensitivity of UG1-W)/(normalized sensitivity of standard glass). Fe-NSRs denotes values normalized to Fe, while Al-NSRs represents values normalized to Al.

Meanwhile, elemental sensitivity ratios (ESRs) were employed to compare internal elements for a suite of silicate reference materials, including NIST SRM 61X series glasses, MPI-DING glasses, and USGS basaltic glasses.⁴⁴ The ESR for element x relative to the internal reference element r_x is mathematically defined as:

where I denotes the peak intensity, f is the isotopic abundance, C is the concentration of the element, subscript i means the measured isotope of element x, and r_i is the measured isotope of the internal normalizing element r_x. Both ESR and NSR values reflect the degree of matrix match between standards and samples, with greater deviations in these ratios indicating more significant matrix discrepancies. As shown in Fig. 5, positive correlations between sensitivity and Fe concentrations were observed for most elements in the USGS and MPI-DING glasses, suggesting that Fe enhances the ablation efficiency by enhancing laser absorption and promoting the generation of finer particles, thereby increasing the ionization efficiency.^42,45 It is noteworthy that all standard reference points of Sc fall on one side of the chromite sample data, indicating that none of the above standards are suitable for Sc calibration in chromite. This is primarily due to the fact that Si and Ca are major elements in the glass standards, where measured Sc signals are affected by oxide-based polyatomic interferences (e.g.²⁹Si¹⁶O and ⁴⁴Ca¹H). Whereas chromite contains low Si content, resulting in a spectral mismatch between standards and samples. These limitations make current standards inappropriate for external calibration of Sc in chromite. Similarly, Co measurements are susceptible to interferences from oxides and hydroxides of Ca, Na, and K (e.g.⁴³Ca¹⁶O, ²³Na³⁶Ar, and ⁴¹K¹⁸O). As for Cr, a key constituent in chromite, noticeable deviations arise when using NIST SRM 610, BCR-2G, or BHVO-2G as calibrants. These findings underscore the necessity of matrix-matched calibration for analyzing geological samples with complex mineralogy.

Fig. 5 Difference in ESRs for a variety of glass standards (SRM 61X series, MPI-DING glasses, and USGS basalt glasses). Normalized sensitivity ratios (ESRs) are plotted with Al-normalized values on the abscissa and Fe-normalized values on the ordinate. Data markers encode sample provenance: USGS references (red squares), MPI-DING glasses (blue squares), NIST 61X glasses (gray diamonds), and natural chromites (green triangles).

To systematically evaluate the impact of calibration strategies on chromite trace element quantification by LA-ICP-MS, a series of analyses in three independent laboratories (USTC, SYSU, and CUGB) were deliberately designed with varying calibration strategies. At the same time, parallel experiments were additionally conducted at SKLCMRE. This experimental design achieved dual primary objectives: (1) to quantify the impact of standardization methodology on data accuracy and (2) to isolate instrument-specific variability through cross-validation. All laboratories adopted multi-external calibration with a content-weighted reference material fitting to mitigate matrix effects. Laboratory-specific protocols comprised: (i) CUGB: single external standard (BCR-2G), (ii) USTC: multi-external standards (BCR-2G, BHVO-2G, and BIR-1G), (iii) SYSU: extended multi-standards (GSE-1G and BHVO-2G), (iv) SKLCMRE: parallel analyses using the analytical parameters from the other three laboratories, yielding three independent datasets. All datasets were uniformly processed utilizing Al/Fe internal standardization and 100% oxide normalization (Fig. 6). Parallel comparison demonstrated consistent data reliability across all analytical protocols, although systematic differences in elemental concentrations were observed. While inter-method variability remained limited for most elements, notable discrepancies emerged: (1) single standard calibration (BCR-2G) systematically underestimated concentrations for the majority of elements except Mg, Al, Ti, V, Mn, and Ni; (2) multi-standard calibration with Al internal standardization yielded universal underestimation; (3) all methods failed to accurately quantify Zn, indicating inherent limitations of silicate glass standards for chromite-hosted Zn analysis. These findings provide critical insight into discrepancies between LA-ICP-MS and bulk chemistry data. Matrix effects demonstrated element-specific severity, with maximal fractionation in low-melting-point elements. Zinc in chromite displayed markedly distinct fractionation behavior relative to glass reference materials (significantly higher fractionation index, Fig. 7), likely attributable to its unique physicochemical properties: low melting point (419.5 °C), high volatility, and elevated first ionization energy (9.39 eV). Although matrix fractionation has been reported to exhibit mass–load dependence,⁴³ our findings indicate that for Zn, laser-induced fractionation effect plays a critical role in Zn calibration.

Fig. 6 Comparative cross-validation of analytical data from three calibration strategies in three independent laboratories. Error bars (1 SD) are determined by repeated measurements of LA-ICP-MS in the same laboratory. The black solid line indicates the recommended values obtained from chemical methods. Solid blue squares denote data from other laboratories using the “Al-normalized method”; solid black squares represent data from other laboratories using the “Fe-normalized method”; solid yellow circles indicate data from other laboratories using the “Al-internal standard method”; solid pink circles denote data from other laboratories using the “Fe-internal standard method”. Corresponding hollow symbols represent data from our laboratory (KLCMRE) using the same reduction methods. The black background area denotes CUGB test data; the blue background area indicates USTC test data; the pink background area represents SYSU test data.

Fig. 7 Fractionation indexes for Zn with respect to Al and Fe for the ablation of different matrix materials.

3.3.2 Calibration by matrix-matched materials. Under fixed instrumental conditions (40 μm laser spot, 5 Hz repetition rate, and 5 J cm⁻² energy density), the laboratory-developed chromite quality control material Cr_inhouse was analyzed as an unknown sample with independent calibration against USGS silicate glasses and matrix-matched UG1-W chromite. Raw data were reduced following four established protocols detailed in Section 2.2.6, namely Al-internal standardization, Fe-internal standardization, Al-normalization, and Fe-normalization methods. To evaluate potential systematic biases between microanalytical results and recommended reference values, the relative deviation (RD) was employed, which was calculated as: RD = ((micro data − recommended data)/recommended data) × 100.⁴⁷ Analytical data are summarized in Fig. 8. Calibration using USGS silicate glasses yielded RDs of ≤±5% for major elements and <±10% for trace elements. In contrast, matrix-matched calibration employing UG1-W chromite achieved near-zero RDs for most elements, with the exception of Sc, Mn, and Ni. Notably, four data reduction methods produced indistinguishable results under matrix-matched conditions. These findings underscore the critical importance of matrix matching in minimizing analytical biases for chromite analysis. While non-matrix-matched USGS glass standards provided acceptable accuracy for major elements, matrix-matched UG1-W chromite significantly improved precision across both major and trace elements. The consistent performance of all normalization/internal standardization methods under matrix-matched conditions further validates the robustness of these data reduction strategies when reference and sample matrices are compositionally compatible.

Fig. 8 Relative deviation of different elements in laboratory-developed chromite material (Cr_inhouse) between recommended data and microanalytical techniques using USGS glass standards and UG1-W chromite.

4. Conclusions

This study establishes UG1-W chromite as a rigorously validated reference material for LA-ICP-MS quantification analysis through comprehensive mineralogical, elemental, and methodological characterization, confirming exceptional homogeneity essential for microanalytical standardization. Systematic evaluation reveals that matrix mismatch between chromite and synthetic glass standards induces severe elemental fractionation. Although calibration using glass reference materials yielded relative deviations within ±10% for most elements, matrix-matched calibration with UG1-W achieved near-zero deviations for both major and trace elements, underscoring the necessity of chromite-matrix reference materials for accurate trace element quantification. Collectively, this work establishes UG1-W as a reliable chromite reference material and provides a methodological framework to minimize matrix-induced errors in LA-ICP-MS. For geologically relevant analyses, matrix-matched calibration is prioritized over synthetic glass standards to ensure analytical accuracy, particularly for trace elements susceptible to matrix effects.

Author contributions

Dan Chen: writing – original draft, visualization, methodology, investigation, formal analysis, validation, supervision, methodology, data curation, conceptualization. Liemeng Chen: writing – review & editing, methodology, investigation. Zhenhui Hou and Dengfeng Li: resources, methodology. Tingguang Lan: writing – review & editing, visualization. Zhihui Dai: writing – review & editing, writing – original draft, supervision, resources, funding acquisition, conceptualization.

Conflicts of interest

The authors declare no competing financial interest.

Data availability

Data will be made available on request. Supplementary information: the calculation of Fe oxidation values and the raw data of LA-ICP-MS. See DOI: https://doi.org/10.1039/d5ja00299k.

Acknowledgements

This work was financially supported by the National Key R & D Program of China (2023YFF0804403), the Natural Science Foundation of China (42273083, 42077313), Special Fund of the State Key Laboratory of Ore Deposit Geochemistry (202303), The Guizhou Provincial Science and Technology Program (Qiankehe Platform-YWZ[2023]006), the Outstanding member of Youth Innovation Promotion Association CAS (No. Y2023105), the National Funding Program for Guiding Local Science and Technology Development (Guizhou [2024] 043). The authors sincerely thank Prof. Liangliang Zhang from China University of Geosciences (Beijing) and Dr Lin Xing from the Institute of Geology and Geophysics, Chinese Academy of Sciences, for their invaluable assistance in analytical samples.

References

1. T. N. Irvine, Can. J. Earth Sci., 1967, 4, 71–103.
2. C. Stowe, Econ. Geol., 1995, 90, 2092–2094.
3. H. Rollinson, Contrib. Mineral. Petrol., 2008, 156, 273–288.
4. P. Pagé and S.-J. Barnes, Econ. Geol., 2009, 104, 997–1018.
5. S. J. Barnes and P. L. Roeder, J. Petrol., 2001, 42, 2279–2302.
6. J. Proenza, F. Ortega-Gutiérrez, A. Camprubı, J. Tritlla, M. Elıas-Herrera and M. Reyes-Salas, J. South Am. Earth Sci., 2004, 16, 649–666.
7. A. H. Ahmed, S. Arai, Y. M. Abdel-Aziz and A. Rahimi, Precambrian Res., 2005, 138, 225–234.
8. J. M. G. Jiménez, T. Kerestedjian, J. A. P. Fernández and F. G. Linares, Geol. Acta, 2009, 7, 413–429.
9. A. M. H. Zaheer, N. Zaland and L. Shirzad, Journal of Mechanical, Civil and Industrial Engineering, 2025, 6(1), 1–10.
10. R. Mukherjee, S. K. Mondal, M. T. Rosing and R. Frei, Contrib. Mineral. Petrol., 2010, 160, 865–885.
11. S. K. Mondal, E. M. Ripley, C. Li and R. Frei, Precambrian Res., 2006, 148, 45–66.
12. D. M. Evans, Ore Geol. Rev., 2017, 90, 307–321.
13. S. J. Barnes and V. Y. Kunilov, Econ. Geol., 2000, 95, 1701–1717.
14. S. J. Barnes and Z. L. Tang, Econ. Geol., 1999, 94, 343–356.
15. V. Balaram, K. V. Anjaiah and A. Kumar, Asian J. Chem., 1999, 11, 949–956.
16. M. F. Gazulla, A. Barba, M. P. Gómez and M. Orduña, Geostand. Geoanal. Res., 2006, 30, 237–243.
17. P. Pagé and S. J. Barnes, Econ. Geol., 2009, 104, 997–1018.
18. M. F. Gazulla, A. Barba, M. P. Gómez and M. Orduña, Geostand. Geoanal. Res., 2007, 30, 237–243.
19. E. Kotzé and C. Gauert, S. Afr. J. Geol., 2020, 123, 551–572.
20. S. Sanchez-Ramos, A. Domenech-Carbo, J. V. Gimeno-Adelantado and J. Peris-Vicente, Talanta, 2008, 74, 1592–1597.
21. S. Q. Liu, S. Y. Jiang, W. Chen, C. Y. Wang, H. M. Su, Y. H. Cao, H. X. Zhang and W. T. Li, RSC Adv., 2023, 13, 13303–13313.
22. J. L. Meng, T. Lu, B. Fang, Y. Ni and B. W. Fan, Opt. Laser Technol., 2022, 152, 1080–1092.
23. P. Nadoll and A. E. Koenig, J. Anal. At. Spectrom., 2011, 26, 1872–1877.
24. Y. Kil, J. Anal. Chem., 2011, 66, 496–503.
25. R. E. Russo, X. Mao, H. Liu, J. Gonzalez and S. S. Mao, Talanta, 2002, 57, 425–451.
26. B. Hattendorf, C. Latkoczy and D. Günther, Anal. Chem., 2003, 75, 341–347.
27. J. Košler, M. Wiedenbeck, R. Wirth, J. Hovorka, P. Sylvester and J. Míková, J. Anal. At. Spectrom., 2005, 20, 402–409.
28. Z. Hu, Y. Liu, S. Gao, S. Hu, R. Dietiker and D. Günther, J. Anal. At. Spectrom., 2008, 23, 1192–1203.
29. Z. Hu, Y. Liu, L. Chen, L. Zhou, M. Li, K. Zong, L. Zhu and S. Gao, J. Anal. At. Spectrom., 2011, 26, 425–430.
30. K. Jochum and J. Enzweiler, Treatise Geochem., 2nd edn, 2014, vol. 15, pp. 43–70.
31. Y. W. Wang, H. Xie, B. Guo, Y. Shi and G. Zhou, Geol. J., 2020, 55, 2185–2207.
32. Y. R. Murthy, S. K. Tripathy and C. R. Kumar, Miner. Eng., 2011, 24, 375–380.
33. L. Qi, H. Jing and D. C. Gregoire, Talanta, 2000, 51, 507–513.
34. J. Wang, B. X. Su, B. Y. Gao, W. J. Li, M. M. Cui, X. Liu, Y. Luo, Q. Q. Pan, L. H. Jia, L. Su and K. Y. Chen, Geostand. Geoanal. Res., 2023, 47, 683–695.
35. Y. Liu, Z. Hu, S. Gao, D. Günther, J. Xu, C. Gao and H. Chen, Chem. Geol., 2008, 257, 34–43.
36. Y. Liu, S. Gao, Z. Hu, C. Gao, K. Zong and D. Wang, J. Petrol., 2009, 51, 537–571.
37. S. A. S. Dare, S.-J. Barnes and G. Beaudoin, Geochim. Cosmochim. Acta, 2012, 88, 27–50.
38. L.-L. Zhang, D.-C. Zhu, Q. Wang, Z.-D. Zhao, D. Liu and J.-C. Xie, Lithos, 2019, 326–327, 144–157.
39. S. Gilbert, L. Danyushevsky, P. Robinson, C. Wohlgemuth-Ueberwasser and N. Pearson, Geostand. Geoanal. Res., 2012, 37, 51–64.
40. D. Harries, Geochemistry, 2014, 74, 375–384.
41. V. G. Batanova, J. M. Thompson, L. V. Danyushevsky, M. V. Portnyagin, D. Garbe-Schönberg, E. Hauri, J. I. Kimura, Q. Chang, R. Senda and K. Goemann, Geostand. Geoanal. Res., 2019, 43, 453–473.
42. M. Guillong, K. Hametner, E. Reusser, S. A. Wilson and D. Günther, Geostand. Geoanal. Res., 2005, 29, 315–331.
43. I. Kroslakova and D. Günther, J. Anal. At. Spectrom., 2007, 22, 51–62.
44. M. Gaboardi and M. Humayun, J. Anal. At. Spectrom., 2009, 24, 1188–1197.
45. H.-R. Kuhn and D. Günther, J. Anal. At. Spectrom., 2004, 19, 1158–1164.
46. S. Gilbert, L. Danyushevsky, P. Robinson, C. Wohlgemuth-Ueberwasser, N. Pearson, D. Savard, M. Norman and J. Hanley, Geostand. Geoanal. Res., 2013, 37, 51–64.
47. S. Wu, A. Audetat, K. P. Jochum, H. Wang, J. Chen, B. Stoll, C. Zhang, Z. Bao, S. Yang, C. Li, X. Wang, C. Xu, L. Xu, C. Huang, L. Xie, Y. Yang and J. Yang, Geostand. Geoanal. Res., 2022, 46, 673–700.

---