AFG – a new Cenozoic columbite–tantalite natural reference material for LA-ICP-MS U–Pb geochronology

Abstract

Columbite–tantalite series minerals are widely distributed in granites, pegmatites, carbonatite–alkaline rocks, and peraluminous granites. U–Pb dating of columbite–tantalite minerals can provide robust temporal constraints on petrogenetic and mineralization events. Laser ablation inductively coupled plasma mass spectrometry (LA-ICP-MS) has become an important analytical technique for columbite–tantalite U–Pb dating, providing high spatial resolution geochronological data through relatively straightforward sample preparation. However, matrix effects between reference materials and unknown samples remain the principal limitation for accurate columbite–tantalite U–Pb dating. The matrix-matched reference materials are crucial for primary calibration or as quality monitoring in LA-ICP-MS columbite–tantalite U–Pb analysis. Although some columbite–tantalite U–Pb dating reference materials have been reported in previous studies, samples containing low amounts of common lead are extremely scarce. Moreover, reference materials covering diverse age ranges are critical for external calibration and quality control, particularly for young materials, given the analytical difficulties in measuring low radiogenic Pb content in young samples by laser ablation. In this study, a Cenozoic columbite–tantalite AFG was investigated as new reference material for LA-ICP-MS columbite–tantalite U–Pb dating. The average U and Pb concentrations in AFG are 1213 ± 112 µg g⁻¹ (2 s) and 3.51 ± 0.45 µg g⁻¹ (2 s), respectively. The isotope dilution thermal ionization mass-spectrometry (ID-TIMS) analysis of AFG yielded a weighted mean ²⁰⁶Pb/²³⁸U ratio of 0.00289 ± 0.000017 (2 s, MSWD = 2.7, n = 5) and a weighted mean ²⁰⁶Pb/²³⁸U age of 18.59 ± 0.06 Ma (2 s, MSWD = 3, n = 5). The results from four independent LA laboratories exhibit excellent homogeneity and yield concordant U–Pb ages. All LA-ICP-MS U–Pb analyses yield a weighted mean ²⁰⁶Pb/²³⁸U age of 18.62 ± 0.05/0.38 Ma (2 s, MSWD = 0.75, n = 288), which is consistent with the ID-TIMS result within analytical uncertainty. We propose AFG columbite–tantalite as a promising new reference material in Cenozoic U–Pb geochronological studies.

---

Introduction

Columbite–tantalite series minerals represent an important source of niobium and tantalum, occurring widely in granites,¹ pegmatites,^2,3 carbonatite–alkaline rocks,⁴ and peraluminous granites.⁵ Columbite–tantalite minerals usually exhibit moderate uranium concentrations (tens to hundreds of µg g⁻¹) for reliable U–Pb isotopic analysis,^6–8 which provides robust age constraints on mineralization events. Both bulk analysis⁹ and in situ techniques¹⁰ are employed for columbite–tantalite U–Pb geochronological analysis. Among geochronological analytical techniques, isotope dilution thermal ionization mass spectrometry (ID-TIMS) yields high-precision U–Pb data.^11,12 However, the method is limited by complex sample preparation procedures and potential contamination from uranium- or lead-rich mineral inclusions.^13,14 In contrast, laser ablation inductively coupled plasma mass spectrometry (LA-ICP-MS) and secondary ion mass spectrometry (SIMS) provide high spatial resolution geochronological data from zoned minerals with relatively straightforward sample preparation.^12,15–17 Consequently, these in situ techniques are widely employed for columbite–tantalite U–Pb dating.^8,10,18 However, the matrix effects between reference materials and unknown samples are the principal limitation for accurate columbite–tantalite U–Pb dating. Previous studies have demonstrated that systematic bias of approximately 10% was observed in columbite–tantalite U–Pb ages when using the matrix-mismatched zircon reference material 91 500 as the calibrating standard in LA-ICP-MS analysis.^6,19 Therefore, matrix-matched reference materials are essential for correcting instrumental drift and Pb/U fractionation in columbite–tantalite U–Pb age determination, and various columbite–tantalite reference materials have been developed for U–Pb age analysis.^6,8,18,20

Columbite–tantalite series minerals form continuous solid solutions through Fe–Mn and Nb–Ta substitutions, with significant chemical compositional variations among different end-members.⁹ Previous studies have demonstrated that chemical compositional differences among columbite–tantalite series materials also cause significant matrix effects during in situ U–Pb age determination, especially variations in the Ta/(Nb + Ta) ratio.^18,21 For example, Yang et al. observed approximately 16% age bias in U–Pb age results with increasing Ta/(Nb + Ta) differences between columbite–tantalite reference material and sample during LA-MC-ICP-MS analysis using a spot size of 10 µm.²¹ The results from SIMS analyses also revealed significant matrix effects corresponding to variations in Ta/(Nb + Ta) ratios of columbite–tantalite, whereas no significant effect was observed for Mn–Fe compositional variations.¹⁸ Therefore, the development of additional columbite–tantalite reference materials with varying end-member chemical compositions is crucial for achieving accurate U–Pb dating results.^2,8,18,20,22 Moreover, variations in radiation damage accumulation resulting from different uranium concentrations and crystallization ages within crystals is another potential source of matrix effects.^23,24 For instance, systematic biases of approximately 3–4% were observed during inter-calibration between different zircon reference materials.^25,26 Consequently, the development of accessory mineral U–Pb geochronological reference materials spanning various age ranges is crucial for external calibration and quality monitoring.^27–30 The reference materials with young U–Pb ages are particularly crucial for quality control in laser ablation analyses,^27,30,31 since accurately measuring low radiogenic Pb content in young samples by laser ablation is challenging. A review of previously published columbite–tantalite U–Pb dating reference materials reveals that existing reference materials predominantly span the 130–2050 Ma age range (Table 1 and Fig. 1), with young (Cenozoic) columbite–tantalite U–Pb reference materials being notably deficient.

Table 1 Summary of U–Pb isotopic and elemental compositions of commonly used columbite–tantalite reference materials. The elemental compositions for all samples represent average values^a,b

Ref. material name	U–Pb age (Ma) (2 s)	U (µg g^−1)	Pb (µg g^−1)	f 206 (%)	MnO (wt%)	FeO (wt%)	Nb2O5 (wt%)	Ta2O5 (wt%)	Mn^#	Ta^#	Ref.
a “—” indicates data not reported in the literature. b Mn^# = Mn/(Mn + Fe) (atomic ratios). Ta^# = Ta/(Nb + Ta) (atomic ratios).
AFG	18.59 ± 0.06	1068	3.25	0–1.0%	11.79	7.49	57.14	22.74	0.61	0.19	This study
Coltan 139	506.6 ± 2.4; 507.8 ± 1.3	1697	171	0–8.29%	8.99	10.38	62.58	12.67	0.47	0.11	2 and 8
NP-2	380.3 ± 2.4	242	14.4	—	6.54	10.5	42.7	38.1	0.39	0.35	18
CT3	2053.2 ± 1.3	386	215	—	0.94	14.35	6.25	75.46	0.06	0.88	8 and 18
Rongi	931.5 ± 2.5	82.7	14.8	0–1.74%	6.45	14.14	65.68	10.96	0.32	0.09	8 and 18
Buranga	905.2 ± 3.2	285	59.3	—	9.83	10.63	64.40	11.59	0.48	0.10	8 and 18
SN3	404.0 ± 1.3	676	41.8	0–0.76%	2.60	17.92	64.58	9.72	0.13	0.08	20
HND	136.2 ± 0.9	702	14.4	0–1.86%	8.69	11.83	60.17	17.53	0.43	0.15	20
RL2	135.7 ± 0.3	954	26.0	0–1.55%	9.72	8.04	31.85	49.33	0.40	0.45	20
DDB	202.0 ± 1.0	1770	62.6	0–4.65%	11.40	7.86	58.51	20.74	0.60	0.18	8
ZKW	203.0 ± 1.6	957	35.9	0–1.19%	12.67	5.88	50.80	29.95	0.69	0.27	8
OXF	262.85 ± 0.64	428	20.0	—	10.16	10.79	69.71	8.12	0.49	0.07	22

---

Fig. 1 Summary of U–Pb ages of commonly used columbite–tantalite reference materials.^2,8,18,20,22

An additional factor that constrains the accuracy and precision of U–Pb geochronological results is the possible occurrence of common lead within reference materials.^32,33 If reference materials used for external calibration contain significant common lead content, pre-correction for common lead is required before Pb/U fractionation correction to ensure accurate U–Pb results.³³ However, this pre-correction procedure may cause bias from inadequate or excessive correction,³⁴ and error propagation during common lead correction will degrade the precision of analytical results.^32–34 When reference materials containing common lead are used for quality control, a sufficient number of analytical points is required to maximize the spread of U/Pb ratios in an isochron to obtain accurate and precise ages.^35–37 Therefore, reference materials with low common lead content are crucial for rapidly and accurately obtaining LA-ICP-MS U–Pb dating results.³² As shown in Table 1, most of the reported columbite–tantalite U–Pb dating reference materials contain varying amounts of common lead, with only a few samples (i.e., CT3, Buranga, SN3, and OXF) yielding concordant U–Pb ages.

To overcome the limitations of existing columbite–tantalite reference materials, a new Cenozoic columbite–tantalite natural reference material (AFG) with extremely low common Pb content is reported in this study. The homogeneity of U–Pb ages in AFG was validated through both ID-TIMS and LA analytical techniques in multiple independent laboratories. Moreover, the chemical compositions and textural features of AFG crystal were also investigated.

Sample description and preparation

The AFG columbite–tantalite megacryst was obtained from a gem trader. According to the gem trader, the sample was collected from the Mawi Pegmatite in Nuristan, Afghanistan. The columbite–tantalite AFG is a giant black megacryst with a total weight of 655 g (Fig. 2a). The megacryst was initially sectioned into smaller fragments using a diamond-trimmed metal wire with a diameter of 0.3 mm. The fragments were crushed into small shards (500–5000 µm) for chemical compositions and U–Pb homogeneity evaluation. Five grains without fractures and inclusions were selected for ID-TIMS analysis. For in situ analyses, the shards were randomly selected, mounted in 1-inch epoxy resin, and polished for major and trace element contents, and LA-ICP-MS U–Pb age determinations.

Fig. 2 Representative photographs of columbite–tantalite AFG. (a) Specimen photograph, (b)–(d) back-scattered electron (BSE) images.

Experimental

The chemical compositions (i.e., major and trace element contents), textural features, and U–Pb isotope composition of AFG were determined using multiple analytical techniques. Major and trace elemental compositions were determined using electron probe microanalysis (EPMA) and LA-ICP-MS, respectively. The textural features were examined using back-scattered electron (BSE) imaging. U–Pb ages of the columbite–tantalite AFG were determined by ID-TIMS and LA-ICP-MS methods. The analytical uncertainties are quoted at the 2 s confidence level unless otherwise specified. Concordia diagrams for the U–Pb ages were constructed using the online IsoplotR software package.³⁸

Chemical compositions and textural homogeneity

Prior to elemental content analysis, the textural features of AFG columbite–tantalite were investigated using BSE imaging techniques. Experiments were performed with a high-vacuum scanning electron microscope (JSM-IT) at Wuhan Sample Solution Analytical Technology Co., Ltd, Wuhan, China (WHSS). Imaging parameters were set to 0.5–30 kV accelerating voltage and 72 µA tungsten filament current.

Major element measurement

Major elemental mass fractions of columbite–tantalite AFG were conducted with EPMA on a JEOL JXA-8230 instrument (JEOL Ltd, Japan) at Wuhan Sample Solution Analytical Technology Co. Ltd, China (WHSS). Experiments were performed with a beam current of 2 nA, accelerating voltage of 15 kV, and beam diameter of 1 µm. Peak counting times were 10 s for all elements with 5 s background counting. Data correction was performed using the ZAF procedure.

Trace element measurement

Trace element compositions were measured by LA-ICP-MS at the State Key Laboratory of Geological Processes and Mineral Resources, China University of Geosciences (CUGW), Wuhan. Experiments were performed using an Agilent 7900 ICP-MS instrument (Agilent Technology, Tokyo, Japan) coupled with a Coherent 193 nm excimer laser ablation system (Geolas HD, Göttingen, Germany). Pulse/Analog (P/A) factor calibration was conducted using standard tuning solution prior to LA analysis. Helium was used as the carrier gas and mixed with argon (makeup gas) via a “Y” junction downstream before entering the ICP torch.³⁹ Instrumental parameters were optimized to maximize the ²³⁸U⁺ signal while maintaining the ThO⁺/Th⁺ ratio below 0.3% and the Th/U ratio close to 100%. The laser was operated at a beam size of 44 µm, repetition rate of 5 Hz, and fluence of 4 J cm⁻². Each analysis consisted of approximately 20 s background acquisition (gas blank) and 50 s data acquisition. Elemental concentrations were calibrated against multi-reference materials (BCR-2G, BHVO-2G, and BIR-1G)⁴⁰ using Mn as the internal standard. The Mn mass fraction was determined from EPMA results. Two analyses of NIST SRM 610 were performed after every 5 sample analyses to correct for time-dependent drift in sensitivity and mass discrimination. Data processing was conducted using ICPMSDataCal software.⁴¹

U–Pb geochronology

ID-TIMS measurement. High-precision U–Pb ages of AFG were determined by isotope dilution–thermal ionization mass spectrometry (ID-TIMS) at the Tianjin Center of Geological Survey, China Geological Survey. The analytical process includes mineral selection, cleaning, digestion, separation and purification, and mass spectrometric measurement.⁴² Five AFG grains without visible cracks and inclusions were selected for analysis under a binocular microscope. The grains were sequentially cleaned with high-purity anhydrous ethanol, HNO₃, and ultrapure water to remove organic matter and impurities adhering to the surface. Then the samples were placed in a Teflon beaker, diluted with a mixture of ²⁰⁵Pb–²³⁵U isotopic tracer, and then fully dissolved with HF. U and Pb were separated using AG1 × 8 anion exchange resin. U–Pb isotopic compositions were determined with a Triton thermal ionization mass spectrometer (TIMS) (Thermo Fisher Scientific, Bremen, Germany). The initial Pb isotopic compositions employed for common Pb correction were derived from the two-stage evolutionary model of Stacey and Kramers.⁴³ The total procedural blanks were 20 pg for Pb and 1 pg for U in this experiment.

LA-ICP-MS measurement. In situ U–Pb age determination of AFG columbite–tantalite was conducted at four independent LA-ICP-MS laboratories: China University of Geosciences (Wuhan) (CUGW), China University of Geosciences (Beijing) (CUGB), Institute of Geology and Geophysics, Chinese Academy of Sciences (IGGCAS), and Wuhan Sample Solution Analytical Technology Co., Ltd (WHSS). Uncertainties for pooled LA-ICP-MS ages are reported as ±α/β, where α denotes internal uncertainty and β denotes propagated systematic uncertainty, including uncertainties of the primary reference material and the decay constant. According to the recommendations from Horstwood et al.,⁴⁴ uncertainties for pooled LA-ICP-MS ages are reported as ±α/β, where α denotes internal uncertainty and β denotes propagated systematic uncertainty calculated by quadratic addition. The systematic uncertainty includes uncertainties from the primary reference material, the decay constant, and the long-term variance of validation material.
LA-ICP-MS U–Pb dating in CUGW. The analyses were conducted using Agilent 7900 ICP-MS instrument (Agilent Technology, Tokyo, Japan) in combination with a 193 nm ArF excimer laser system (GeoLas HD, Coherent Inc., Göttingen, Germany). The analytical methods were similar to those described by Qing et al.²² The columbite–tantalite Coltan 139 (ref. 2) was used as the external calibration reference material, with columbite–tantalite reference materials OXF²² and SN3 (ref. 20) serving as quality control samples for U–Pb isotopic determination. Laser ablation was performed using a 44 µm beam diameter, 4 Hz repetition rate, and 4 J cm⁻² fluence. Detailed analytical parameters are shown in Table 2. The raw data were corrected using the software ICPMSDataCal.⁴¹ The obtained concordia ages of OXF and SN3 were 259.1 ± 2.7/5.8 Ma (2 s, MSWD = 0.7, n = 13) and 407.0 ± 5.5/9.8 Ma (2 s, MSWD = 2.7, n = 7), which are consistent with the recommended ages of 262.85 ± 0.64 Ma (2 s)²² and 404.0 ± 1.3 Ma (2 s)²⁰ within analytical uncertainty. The detailed LA-ICP-MS U–Pb analytical results are shown in Table S2.

Table 2 Summary of the operating parameters for LA-ICP-MS measurements

Laboratory	CUGW		CUGB	WHSS	IGGCAS
	Trace element measurement	U–Pb dating	U–Pb dating	U–Pb dating	U–Pb dating
Laser ablation system
Laser type and wavelength	ArF excimer 193 nm	ArF excimer 193 nm	ArF excimer 193 nm	ArF excimer 193 nm	ArF excimer 193 nm
Repetition rate	5 Hz	4 Hz	8 Hz	2 Hz	5 Hz
Used spot size	44 µm	44 µm	33 µm	44 µm	32 µm
Energy density	4 J cm^−2	4 J cm^−2	3 J cm^−2	5 J cm^−2	5J cm^−2
Ablation gas flow	600 mL min^−1	600 mL min^−1	300 mL min^−1	600 mL min^−1	600 mL min^−1
Ablation time	50 s	50 s	30 s	50 s	50 s
Washout time	20 s	20 s	20 s	10–15 s	20 s
ICP-MS instrument
ICP–MS type	Agilent 7900	Agilent 7900	Agilent 7900	Agilent 7900	Agilent 7500a
RF forward power	1550 W	1550 W	1400 W	1390 W	1390 W
Carrier gas flow	0.85 L min^−1	0.85 L min^−1	1.10 L min^−1	1.10 L min^−1	1.10 L min^−1
Cool gas flow	15 L min^−1	15 L min^−1	15 L min^−1	14 L min^−1	14 L min^−1
Isotopes measured (m/z)^+ dwell time	^7Li, ^9Be, ^11B, ^29Si, ^45Sc, ^49Ti, ^51V, ^53Cr, ^55Mn, ^57Fe, ^59Co, ^60Ni, ^63Cu, ^66Zn, ^71Ga, ^72Ge, ^85Rb, ^88Sr, ^89Y, ^91Zr, ^109Ag, ^111Cd, ^133Cs, ^137Ba, ^139La, ^140Ce, ^141Pr, ^143Nd, ^147Sm, ^151Eu, ^157Gd, ^159Tb, ^163Dy, ^165Ho, ^166Er, ^169Tm, ^173Yb, ^175Lu, and ^178Hf all 6 ms	^202Hg (6 ms), ^204Pb (6 ms), ^206Pb (20 ms), ^207Pb (25 ms), ^208Pb (15 ms), ^232Th (10 ms), ^238U (15 ms)	^202Hg (10 ms), ^204Hg (20 ms), ^204Pb (20 ms), ^206Pb (20 ms), ^207Pb (30 ms), ^208Pb (15 ms), ^232Th (10 ms), ^238U (15 ms)	^202Hg (6 ms), ^204Pb (6 ms), ^206Pb (20 ms), ^207Pb (25 ms), ^208Pb (15 ms), ^232Th (10 ms), ^238U (15 ms)	^202Hg (6 ms), ^204Pb (15 ms), ^206Pb (15 ms), ^207Pb (30 ms), ^208Pb (15 ms), ^232Th (10 ms), ^238U (10 ms)

---

LA-ICP-MS U–Pb dating in CUGB. Sample AFG was analyzed using a 193 nm ArF excimer laser ablation system (RESOlution S155-LR, ASI, Australia) coupled with an Agilent 7900 quadrupole ICP-MS (Agilent Technology, Japan). The Coltan 139 (ref. 2) reference material was used for external calibration. To monitor the quality of the analytical method, a secondary columbite–tantalite reference material DKLS²¹ was alternately analyzed as an unknown. Each analysis consisted of 20 s background acquisition followed by 30 s signal ablation, with laser parameters of 3 J cm⁻² fluence, 33 µm spot size, and 8 Hz repetition rate. Detailed analytical parameters are presented in Table 2. Data reduction was performed using Iolite V4.8.4 software.⁴⁵ The obtained lower intercept age of DKLS was 254.1 ± 1.9/5.5 Ma (2 s, MSWD = 1.5, n = 20) (Table S2), which shows good agreement with the recommended age of 250.2 ± 0.3 Ma (2 s).²¹
LA-ICP-MS U–Pb dating in WHSS. Experiments were conducted using an Agilent 7900 quadrupole ICP-MS (Agilent Technology, Japan) coupled with a 193 nm ArF excimer laser system (GeoLas HD, Coherent Inc., Germany). The columbite–tantalite SN3 (ref. 20) was employed as the primary reference material. Columbite–tantalite NP-2 (ref. 18) was analyzed simultaneously as an unknown to evaluate analytical accuracy. Laser ablation was performed with a fluence of 5 J cm⁻², a spot size of 44 µm, and a repetition rate of 2 Hz. Detailed analytical parameters are shown in Table 2. The data reduction was processed using the software ICPMSDataCal.⁴¹ The obtained lower intercept age of NP-2 was 381.1 ± 3.9/8.6 Ma (2 s, MSWD = 1.6, n = 6) (Table S2), which shows good agreement with the reference age of 380.3 ± 2.4 Ma (2 s).¹⁸
LA-ICP-MS U–Pb dating in IGGCAS. U–Pb age determination was conducted using an Agilent 7500a quadrupole ICP–MS (Agilent Technologies, Japan) coupled with a 193 nm excimer laser ablation system (Coherent Inc., Germany). Columbite–tantalite Coltan 139 (ref. 2) was the primary reference material and Coltan17 (ref. 8) was analyzed as a secondary reference material to monitor accuracy. Each analysis consisted of 20 s background acquisition followed by 50 s laser ablation with a 32 µm spot size, 5 Hz repetition rate, and 5 J cm⁻² fluence. Detailed analytical parameters are listed in Table 2. Data reduction and calculations were performed using ICPMSDataCal software.⁴¹ The obtained concordia age of Coltan 17 were 498.3 ± 5.5/11.4 Ma (2 s, MSWD = 1.6, n = 8) (Table S2), which agrees well with the published age of 502.7 ± 3.1 Ma (2 s).⁸

Results and discussion

Major and trace element mass fractions

BSE images of randomly selected AFG grains showed uniform greyscale intensities, with no obvious zoning or mineral inclusions observed (Fig. 2b–d). The results indicate that AFG grains are texturally homogeneous. Ten major element analyses of AFG columbite–tantalite are presented in Table 3. The Nb₂O₅ contents range from 56.26 wt% to 57.91 wt%, with an average value of 57.14 wt%. The Ta₂O₅ mass fractions range from 22.19 to 23.39 wt%, averaging 22.74 wt%. The average values of MnO and FeO are 11.79 wt% and 7.49 wt%, respectively. The AFG columbite–tantalite exhibits relatively low Ta/(Nb + Ta) ratios of 0.19–0.20 and high Mn/(Mn + Fe) ratios of 0.61–0.62 (Table 3), indicating that AFG represents the manganocolumbite end-member (Fig. 3a). In contrast, other commonly used columbite–tantalite reference materials (i.e., Coltan 139,² SN3,²⁰ and OXF²²) display lower Ta/(Nb + Ta) and Mn/(Mn + Fe) ratios, corresponding to the ferrocolumbite end-member (Fig. 3a). The Ta/(Nb + Ta) ratios of Coltan 139, SN3, and OXF are 0.11, 0.08, and 0.07, and the Mn/(Mn + Fe) values are 0.47, 0.13, and 0.49, respectively.

Table 3 Major element results of AFG columbite–tantalite from EPMA analysis^a

No.	01	02	03	04	05	06	07	08	09	10	Average	2sd
a “—” indicates below the detection limit.
TiO2 (wt%)	0.49	0.49	0.66	0.46	0.66	0.59	0.59	0.46	0.51	0.63	0.55	0.15
MnO (wt%)	12.01	11.88	11.71	11.80	11.44	11.88	11.84	11.86	11.62	11.81	11.79	0.31
FeO (wt%)	7.54	7.43	7.46	7.63	7.53	7.47	7.51	7.47	7.48	7.37	7.49	0.13
ZrO2 (wt%)	0.09	0.10	0.10	0.02	0.05	0.03	0.10	0.10	0.00	0.12	0.07	0.08
Nb2O5 (wt%)	57.35	56.44	57.54	56.62	56.26	57.45	56.80	57.52	57.91	57.54	57.14	1.07
SnO2 (wt%)	0.10	0.09	0.13	0.13	0.22	0.06	0.14	0.11	0.12	0.17	0.13	0.09
Ta2O5 (wt%)	22.84	22.69	22.73	22.58	22.19	23.39	22.66	22.39	22.96	22.97	22.74	0.63
WO3 (wt%)	0.09	0.03	0.08	0.32	0.18	—	—	0.12	0.11	0.18	0.11	0.18
Total (wt%)	100.52	99.15	100.41	99.55	98.53	100.86	99.63	100.02	100.71	100.80	100.2	1.49
Ta/(Nb + Ta)	0.19	0.19	0.19	0.19	0.19	0.20	0.19	0.19	0.19	0.19	0.19	0.004
Mn/(Mn + Fe)	0.62	0.62	0.61	0.61	0.61	0.62	0.62	0.62	0.61	0.62	0.61	0.008

---

Fig. 3 Comparison of elemental compositions of columbite–tantalite AFG, Coltan 139, SN3, and OXF. (a) Columbite–tantalite quadrilateral diagram, data for Coltan 139, SN3, and OXF are from ref. 8, 20 and 22, (b) Chondrite-normalized REE patterns. Average values of REE are shown for clarity.

The trace element mass fractions of columbite–tantalite reference materials (i.e., AFG, Coltan 139,² SN3,²⁰ and OXF²²) were determined with LA-ICP-MS. The obtained U, Th, and Pb mass fractions of AFG are 1068 ± 401 µg g⁻¹ (2 s), 24.1 ± 9.63 µg g⁻¹ (2 s), and 3.25 ± 1.18 µg g⁻¹ (2 s) (Table 4). The Chondrite-normalized⁴⁶ REE patterns of AFG, Coltan 139, SN3, and OXF are presented in Fig. 3b. All columbite–tantalite reference materials exhibit depletion in LREE and enrichment in HREE. The AFG sample contains low REE contents at 0.862 µg g⁻¹, which is significantly lower than Coltan 139 (1060 µg g⁻¹) and SN3 (40.7 µg g⁻¹) (Tables 4 and S1).

Table 4 Trace element results of AFG columbite–tantalite from LA-ICP-MS analysis

Sample	AFG	2sd (n = 50)	Coltan 139	2sd (n = 10)	OXF	2sd (n = 10)	SN3	2sd (n = 10)
Y (µg g^−1)	0.592	3.46	2040	76.6	1.03	0.467	78.6	13.8
Hf (µg g^−1)	199	80	341	15.1	41.4	14.1	108	7.81
U (µg g^−1)	1068	401	1773	308	412	46.1	791	241
Th (µg g^−1)	24.1	9.63	67.1	11.5	6.46	1.25	0.842	0.279
Pb (µg g^−1)	3.25	1.18	138	22.5	17.0	2.40	49.8	17.8
La (µg g^−1)	0.013	0.132	0.168	0.132	0.004	0.009	0.005	0.014
Ce (µg g^−1)	0.004	0.008	3.94	0.476	0.017	0.008	0.053	0.035
Pr (µg g^−1)	0.002	0.009	1.89	0.222	0.004	0.006	0.017	0.012
Nd (µg g^−1)	0.007	0.025	24.5	2.142	0.027	0.050	0.197	0.268
Sm (µg g^−1)	0.023	0.159	98.9	3.769	0.020	0.030	0.732	0.350
Eu (µg g^−1)	0.003	0.014	0.125	0.070	0.006	0.007	0.013	0.021
Gd (µg g^−1)	0.016	0.082	262	7.20	0.017	0.018	3.57	0.824
Tb (µg g^−1)	0.015	0.164	74.4	2.78	0.005	0.007	2.17	0.383
Dy (µg g^−1)	0.034	0.115	394	16.4	0.064	0.051	17.9	3.14
Ho (µg g^−1)	0.010	0.050	39.8	1.52	0.008	0.013	2.55	0.490
Er (µg g^−1)	0.126	0.904	69.4	2.19	0.022	0.018	5.76	0.847
Tm (µg g^−1)	0.018	0.022	10.1	0.486	0.003	0.005	0.874	0.094
Yb (µg g^−1)	0.531	3.15	71.2	3.29	0.019	0.029	6.11	0.913
Lu (µg g^−1)	0.060	0.019	9.34	0.303	0.003	0.004	0.728	0.069
ΣREE	0.862		1060		0.218		40.7

---

U–Pb geochronology

ID-TIMS U–Pb results. Five fragments of AFG without visible inclusions and fractures were selected for ID-TIMS analysis. The results are shown in Fig. 4 and Table 5. The obtained U and Pb mass fractions of AFG from ID-TIMS analyses are 1213 ± 112 µg g⁻¹ (2 s) and 3.51 ± 0.45 µg g⁻¹ (2 s). The difference between ID-TIMS and LA analytical results suggests potential elemental heterogeneity within AFG. The AFG columbite–tantalite contains relatively low common lead with the ²⁰⁶Pb/²⁰⁴Pb ratios ranged from 415.68 to 927.43 (Table 5). The weighted mean ²⁰⁶Pb/²³⁸U ratio of AFG is 0.00289 ± 0.000017 (2 s, MSWD = 2.7, n = 5). The obtained apparent ²⁰⁶Pb/²³⁸U ages range from 18.48 Ma to 18.75 Ma, yielding a weighted average ²⁰⁶Pb/²³⁸U age of 18.59 ± 0.06 Ma (2 s, MSWD = 3, n = 5) (Fig. 4).

Fig. 4 Isotope dilution thermal ionization mass spectrometry (ID-TIMS) results of columbite–tantalite AFG.

Table 5 U–Pb isotopic data of AFG columbite–tantalite obtained with ID-TIMS analysis

No.	Weight/mg	Content/µg g^−1^a			Isotope ratio^b							Age (Ma)						Rho (7/5–6/8)
		Pb^c	Pbc^c	U	^206Pb/^204Pb	^206Pb/^238U^d	Err%	^207Pb/^235U^d	Err%	^207Pb/^206Pb^d	Err%	^206Pb/^238U^e	1σ	^207Pb/^235U^e	1σ	^207Pb/^206Pb^e	1σ
a Total blanks were less than 20 pg for Pb and less than 1 pg for U. b The ^206Pb/^204Pb ratios are the raw instrumental values without any correction, while the ^206Pb/^238U, ^207Pb/^235U, and ^207Pb/^206Pb ratios are the values corrected for blank, initial lead, and mass fractionation. c Pb and Pbc are total and common Pb, respectively. d Common Pb was corrected based on the Pb evolution model of ref. 43. e The relevant calculation constants λ238 and λ235 from Jaffey et al. (1971)^47 and ^238U/^235U ratio from Hiess et al. (2012).^48
1	0.00174	3.74	0.13	1280	558.48	0.002900	0.45	0.01854	0.73	0.04638	0.50	18.67	0.08	18.66	0.1	16.31	12.1	0.730
2	0.00229	3.39	0.07	1205	927.43	0.002912	0.32	0.01896	0.63	0.04720	0.52	18.75	0.06	19.07	0.1	58.39	12.3	0.582
3	0.00198	3.40	0.11	1189	673.82	0.002885	0.40	0.01842	0.63	0.04631	0.43	18.57	0.07	18.53	0.1	12.50	10.4	0.735
4	0.00158	3.82	0.16	1266	415.68	0.002875	0.44	0.01904	0.92	0.04802	0.73	18.51	0.08	19.15	0.2	99.09	17.2	0.619
5	0.00241	3.22	0.09	1125	649.11	0.002871	0.37	0.01960	0.59	0.04951	0.41	18.48	0.07	19.71	0.1	171.02	9.6	0.724

---

LA-ICP-MS U–Pb results. To assess the homogeneity of AFG at the laser ablation analysis scale (∼30 µm), U–Pb isotopic compositions were analyzed in multiple laser ablation laboratories (i.e., CUGW, CUGB, IGGCAS, and WHSS). Considering the extremely low common Pb content in AFG, all LA-ICP-MS U–Pb analytical spots were plotted on the concordia diagram (Fig. 5 and 6), and the weighted mean ²⁰⁶Pb/²³⁸U ages was calculated without applying common Pb correction. Laser ablation analyses were conducted on both randomly selected small shards and a polished slice at CUGW (Fig. 5 and Table S2). Two analytical sessions on small shards produced concordia ages of 18.69 ± 0.12/0.40 Ma (2 s, MSWD = 0.53, n = 44) and 18.64 ± 0.13/0.39 Ma (2 s, MSWD = 0.54, n = 46), respectively (Fig. 5a and b). The corresponding weighted mean ²⁰⁶Pb/²³⁸U ages were 18.69 ± 0.14/0.40 Ma (2 s, MSWD = 0.51, n = 44), 18.66 ± 0.14/0.40 Ma (2 s, MSWD = 0.53, n = 46), respectively. The line profile analyses produced a concordia age of 18.59 ± 0.06/0.38 Ma (2 s, MSWD = 1.0, n = 124) and a weighted mean ²⁰⁶Pb/²³⁸U age of 18.58 ± 0.06/0.38 Ma (2 s, MSWD = 1.0, n = 124) (Fig. 5d). The U–Pb results obtained from grain analyses and line profile measurements of AFG show excellent agreement with ID-TIMS results within analytical uncertainty.

Fig. 5 U–Pb results of AFG columbite–tantalite obtained with LA-ICP-MS analysis at the China University of Geosciences (Wuhan). (a) and (b) Concordia plot obtained from two sessions of shard analyses, (c) reflected light image of the line profile (three orange lines), (d) concordia plot obtained from line profiles analyses.

Fig. 6 U–Pb results of AFG columbite–tantalite obtained with LA-ICP-MS analysis. (a) Concordia diagram obtained from CUGB, (b) concordia diagram obtained from WHSS, (c) concordia diagram obtained from the IGGCAS, (d) histogram of the ²⁰⁶Pb^/238U ages from all LA-ICP-MS analyses.

U–Pb isotopic measurements were also conducted on different AFG grains at CUGB, WHSS, and IGGCAS. Twenty-three analyses at CUGB yielded a concordia age of 18.73 ± 0.17/0.41 Ma (2 s, MSWD = 0.84, n = 23) and a weighted mean ²⁰⁶Pb/²³⁸U age of 18.69 ± 0.18/0.41 Ma (2 s, MSWD = 0.68, n = 23) (Fig. 6a). A total of 30 analyses performed at the WHSS yielded a concordia age of 18.70 ± 0.20/0.42 Ma (2 s, MSWD = 0.54, n = 30) with a weighted mean ²⁰⁶Pb/²³⁸U age of 18.65 ± 0.21/0.43 Ma (2 s, MSWD = 0.57, n = 30) (Fig. 6b). Analyses at IGGCAS produced a concordia age of 18.70 ± 0.25/0.45 Ma (2 s, MSWD = 0.69, n = 21), with a weighted mean ²⁰⁶Pb/²³⁸U age of 18.71 ± 0.28/0.47 Ma (2 s, MSWD = 0.29, n = 21) (Fig. 6c). The consistent concordant ages obtained from different laser ablation laboratories further validate the extremely low common lead composition in AFG columbite–tantalite. We estimated the common Pb contribution using f₂₀₆=²⁰⁶Pb_common/²⁰⁶Pb_total.⁴⁹ The calculated f₂₀₆ values for samples AFG are less than 1.0%, although there are a few analyses with f₂₀₆ values up to 2.0%. A total of 288 LA-ICP-MS U–Pb analyses exhibit a Gaussian distribution and yield a weighted mean ²⁰⁶Pb/²³⁸U age of 18.62 ± 0.05/0.38 Ma (2 s, MSWD = 0.75, n = 288) (Fig. 6d), which is consistent with the ID-TIMS result of 18.59 ± 0.06 Ma (2 s, MSWD = 3, n = 5). Numerous LA analyses demonstrate that AFG columbite–tantalite exhibits homogeneous U–Pb distributions at the 30–40 µm analytical scale.

Evaluation of matrix effects between AFG and other reference materials

Previous studies have reported that matrix effects may occur among columbite–tantalite minerals during high resolution LA-ICP-MS U–Pb dating with spot size of 10 µm, mainly due to differences in Nb, Ta composition.^18,21 Nevertheless, other study indicated that matrix effects among columbite–tantalite samples with Ta/(Nb + Ta) ratios ranging from 0.09 to 0.88 are insignificant when using larger spot sizes (32 µm).⁸ In this study, U–Pb analyses of the AFG was performed in four independent LA laboratories (i.e., CUGW, CUGB, IGGCAS, and WHSS) with spot sizes from 32 to 44 µm (Table 2). The columbite–tantalite reference materials Coltan 139 and SN3 were used as the external calibration standards, while OXF, DKLS, and Coltan 17 were employed as quality control samples. The Ta/(Nb + Ta) and Mn/(Mn + Fe) ratios of these reference materials range from 0.07 to 0.35 and from 0.13 to 0.49, respectively. The obtained U–Pb ages of AFG and all monitoring materials are consistent with their recommended values within analytical uncertainty (Fig. 5 and 6). Therefore, under the large spot sizes employed in this study (Table 2), compositional variations among columbite–tantalite samples result in negligible matrix effects between AFG and other reference materials for LA-ICP-MS U–Pb age determination.

Conclusions

This study assesses the homogeneity and suitability of the columbite–tantalite crystal AFG as a reference material for in situ microbeam U–Pb geochronology. The major element mass fractions indicate that AFG represents the manganocolumbite end-member. The obtained U and Pb mass fractions of AFG are 1213 ± 112 µg g⁻¹ (2 s) and 3.51 ± 0.45 µg g⁻¹ (2 s), respectively. The U–Pb isotopic compositions of AFG were determined using both ID-TIMS and LA-ICP-MS analyses. ID-TIMS analysis yielded a weighted mean ²⁰⁶Pb/²³⁸U ratio of 0.00289 ± 0.000017 (2 s, MSWD = 2.7, n = 5) and a weighted mean ²⁰⁶Pb/²³⁸U age of 18.59 ± 0.06 Ma (2 s, MSWD = 3, n = 5). AFG columbite–tantalite contains extremely low common lead content and yields concordant U–Pb ages during laser ablation analysis. A total of 288 U–Pb analyses from four independent LA laboratories show homogeneous U–Pb age distributions in AFG, with a weighted mean ²⁰⁶Pb/²³⁸U age of 18.62 ± 0.05/0.38 Ma (2 s, MSWD = 0.75, n = 288). These results demonstrate that AFG can serve as a new Cenozoic columbite–tantalite reference material for LA-ICP-MS U–Pb geochronology.

The AFG crystal has a substantial sample with a total mass of 655 g, ensuring long-term availability in the U–Pb community. Aliquots of the columbite–tantalite AFG are available in batches to interested laboratories upon request to the corresponding author.

Author contributions

Liyuan Qing: data collection, investigation, visualization, original draft. Tao Luo: methodology development, data interpretation, original draft, review and editing. Jiarun Tu: data collection, data analysis, original draft, resources provision. Wen Zhang: statistical analysis, methodology, original draft. Hongtao Shen: investigation, resources provision. Xiaodong Deng: investigation, data interpretation, original draft. Zhaochu Hu: conceptualization, data interpretation, review and editing, supervision, resources provision.

Conflicts of interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Data availability

The data that support the findings of this study are available in the supplementary information (SI). Supplementary information: Table S1 presents the trace element results of columbite–tantalite AFG, Coltan 139, SN3, and OXF obtained by LA-ICP-MS analysis, and Table S2 presents the U–Pb results of AFG columbite–tantalite and the quality control samples obtained by LA-ICP-MS analysis. See DOI: https://doi.org/10.1039/d5ja00391a.

Acknowledgements

This research was supported by the National Natural Science Foundation of China (Grant No. 42330104, 42373032 and 42473029), the Natural Science Foundation of Hubei Province (Grant No. 2024AFD372 and 2025CSA006) and the MOST Special Fund from the State Key Laboratories of Geological Processes and Mineral Resources. We are grateful to Dr Yueheng Yang and Dr Yang Li for providing reference materials Coltan 139, and Dr Rucheng Wang for providing SN3. We thank Dr Liangliang Zhang, Dr Yueheng Yang, Dr Wei Gao, and Dr Guangyu Shi for their assistance with LA-ICP-MS and EPMA analysis.

References

1. F. Melcher, T. Graupner, H. E. Gäbler, M. Sitnikova, T. Oberthür, A. Gerdes, E. Badanina and T. Chudy, Ore Geol. Rev., 2017, 89, 946–987.
2. F. Melcher, T. Graupner, H. E. Gäbler, M. Sitnikova, F. Henjes-Kunst, T. Oberthür, A. Gerdes and S. Dewaele, Ore Geol. Rev., 2015, 64, 667–719.
3. Y. G. Feng, T. Liang, R. Linnen, Z. Zhang, Y. Zhou, Z. L. Zhang and J. G. Gao, Lithos, 2020, 362–363, 105461.
4. Y. C. Ying, W. Chen, J. Lu, S. Y. Jiang and Y. H. Yang, Lithos, 2017, 290–291, 159–171.
5. B. Wei, C. Y. Wang, Z. H. Zhao and H. T. Bao, Lithos, 2020, 370–371, 105648.
6. X. D. Che, F. Y. Wu, R. C. Wang, A. Gerdes, W. Q. Ji, Z. H. Zhao, J. H. Yang and Z. Y. Zhu, Ore Geol. Rev., 2015, 65, 979–989.
7. X. D. Deng, J. W. Li, X. F. Zhao, Z. C. Hu, H. Hu, D. Selby and Z. S. de Souza, Chem. Geol., 2013, 344, 1–11.
8. M. Yang, Y. H. Yang, R. L. Romer, X. D. Che, R. C. Wang, F. Y. Wu, G. C. Fei, Y. Deng and T. Wu, J. Anal. At. Spectrom., 2023, 38, 1816–1829.
9. R. L. Romer and J. E. Wright, Geochim. Cosmochim. Acta, 1992, 56, 2137–2142.
10. S. R. Smith, G. L. Foster, R. L. Romer, A. G. Tindle, S. P. Kelley, S. R. Noble, M. Horstwood and F. W. Breaks, Contrib. Mineral. Petrol., 2004, 147, 549–564.
11. R. L. Romer and S. A. Smeds, Precambrian Res., 1994, 67, 141–158.
12. U. Schaltegger, A. K. Schmitt and M. S. A. Horstwood, Chem. Geol., 2015, 402, 89–110.
13. R. L. Romer and B. Lehmann, Econ. Geol., 1995, 90, 2303–2309.
14. J. R. Tu, H. Y. Zhou, Y. R. Cui, G. Z. Li, J. Z. Geng and J. Zhang, Rock Miner. Anal., 2024, 43, 533–545.
15. C. X. Zhang, Y. X. Xiong, J. L. Fu, W. Zhao, L. Shu, Z. S. Li, J. Xu, C. X. Zhou and F. Hong, Rock Miner. Anal., 2025, 44, 369–383.
16. X. Yang, S. X. Gao, D. Wang, X. F. Qiu, X. R. Tong, D. Y. Ju and Z. F. Zhang, At. Spectrosc., 2024, 422–437.
17. Y. S. Huang, W. H. Zhao, Y. Liu, Y. H. Yang, G. Q. Tang, Y. Li, X. H. Li, H. Zhao and Q. L. Li, J. Anal. At. Spectrom., 2022, 37, 1109–1118.
18. H. Legros, J. Mercadier, J. Villeneuve, R. L. Romer, E. Deloule, M. V. Lichtervelde, S. Dewaele, P. Lach, X. D. Che, R. C. Wang, Z. Y. Zhu, E. Gloaguen and J. Melleton, Chem. Geol., 2019, 512, 69–84.
19. M. Á. Galliski, A. von Quadt and M. F. Márquez-Zavalía, Lithos, 2021, 386–387, 106001.
20. L. Xiang, R. C. Wang, R. L. Romer, X. D. Che, H. Hu and Z. M. Tang, Geostand. Geoanal. Res., 2023, 47, 609–628.
21. S. Yang, L. L. Zhang, R. Wang, D. C. Zhu, J. C. Xie, Q. Wang and W. T. Xu, J. Anal. At. Spectrom., 2024, 39, 2421–2432.
22. L. Y. Qing, T. Luo, Z. C. Hu, Y. Li, J. R. Tu, L. L. Zhang, W. Zhang and K. Q. Zong, J. Anal. At. Spectrom., 2024, 39, 2066–2077.
23. C. M. Allen and I. H. Campbell, Chem. Geol., 2012, 332–333, 157–165.
24. E. Marillo-Sialer, J. Woodhead, J. M. Hanchar, S. M. Reddy, A. Greig, J. Hergt and B. Kohn, Chem. Geol., 2016, 438, 11–24.
25. L. P. Black, S. L. Kamo, C. M. Allen, D. W. Davis, J. N. Aleinikoff, J. W. Valley, R. Mundil, I. H. Campbell, R. J. Korsch, I. S. Williams and C. Foudoulis, Chem. Geol., 2004, 205, 115–140.
26. E. Marillo-Sialer, J. Woodhead, J. Hergt, A. Greig, M. Guillong, A. Gleadow, N. Evans and C. Paton, J. Anal. At. Spectrom., 2014, 29, 981–989.
27. T. Luo, Q. L. Li, X. X. Ling, Y. Li, C. Yang, H. L. Wang, X. P. Xia, S. B. Zhang, L. Xu, X. M. Liu, X. D. Deng and Z. C. Hu, J. Anal. At. Spectrom., 2021, 36, 2216–2226.
28. Q. D. Wei, H. Wang, Y. H. Yang, J. H. Yang, C. Huang, S. T. Wu and L. W. Xie, Sci. China: Earth Sci., 2020, 63, 1780–1790.
29. X. H. Li, W. G. Long, Q. L. Li, Y. Liu, Y. F. Zheng, Y. H. Yang, K. R. Chamberlain, D. F. Wan, C. H. Guo, X. C. Wang and H. Tao, Geostand. Geoanal. Res., 2010, 34, 117–134.
30. L. J. Duan, L. L. Zhang, D. C. Zhu, J. C. Xie, Q. Wang, S. T. Wu, W. T. Xu, S. Kamo and G. H. Shi, Int. J. Mass Spectrom., 2023, 493, 117121.
31. M. P. Eddy, M. Ibañez-Mejia, S. D. Burgess, M. A. Coble, U. G. Cordani, J. DesOrmeau, G. E. Gehrels, X. H. Li, S. MacLennan, M. Pecha, K. Sato, B. Schoene, V. A. Valencia, J. D. Vervoort and T. T. Wang, Geostand. Geoanal. Res., 2019, 43, 113–132.
32. T. Luo, M. F. Li, X. D. Deng, J. R. Tu, H. T. Shen, T. Kapitany, B. N. Hu, W. Zhang and Z. C. Hu, Chem. Geol., 2025, 674, 122562.
33. D. M. Chew, J. A. Petrus and B. S. Kamber, Chem. Geol., 2014, 363, 185–199.
34. T. Luo, H. L. Wang, S. B. Zhu, L. Y. Qing and Z. C. Hu, Rock Miner. Anal., 2025, 44, 51–62.
35. A. R. C. Kylander-Clark, Geochronology, 2020, 2, 343–354.
36. L. L. Zhang, D. C. Zhu, Y. H. Yang, Q. Wang, J. C. Xie and Z. D. Zhao, At. Spectrosc., 2021, 42, 335–348.
37. G. C. Sun, M. Y. Chen, J. J. Huo, Y. G. Li, R. X. Chen, G. Q. Liu, L. Q. Dai, Z. F. Zhao, H. O. Gu, P. Gao and S. B. Zhang, At. Spectrosc., 2025, 46, 185–197.
38. P. Vermeesch, Geosci. Front., 2018, 9, 1479–1493.
39. T. Luo, Z. C. Hu, W. Zhang, D. Günther, Y. S. Liu, K. Q. Zong and S. H. Hu, J. Anal. At. Spectrom., 2018, 33, 1655–1663.
40. K. P. Jochum, M. Willbold, I. Raczek, B. Stoll and K. Herwig, Geostand. Geoanal. Res., 2005, 29, 285–302.
41. Y. S. Liu, Z. C. Hu, S. Gao, D. Günther, J. Xu, C. G. Gao and H. H. Chen, Chem. Geol., 2008, 257, 34–43.
42. J. R. Tu, Z. B. Xiao, H. Y. Zhou, S. Q. An, G. Z. Li, Y. R. Cui, W. G. Liu and H. M. Li, Ore Geol. Rev., 2019, 109, 407–412.
43. J. S. Stacey and J. D. Kramers, Earth Planet. Sci. Lett., 1975, 26, 207–221.
44. M. S. A. Horstwood, J. Košler, G. Gehrels, S. E. Jackson, N. M. McLean, C. Paton, N. J. Pearson, K. Sircombe, P. Sylvester, P. Vermeesch, J. F. Bowring, D. J. Condon and B. Schoene, Geostand. Geoanal. Res., 2016, 40, 311–332.
45. C. Paton, J. Hellstrom, B. Paul, J. Woodhead and J. Hergt, J. Anal. At. Spectrom., 2011, 26, 2508–2518.
46. W. F. McDonough and S. S. Sun, Chem. Geol., 1995, 120, 223–253.
47. A. H. Jaffey, K. F. Flynn, L. E. Glendenin, W. C. Bentley and A. M. Essling, Phys. Rev. C, 1971, 4, 1889–1906.
48. J. Hiess, D. J. Condon, N. McLean and S. R. Noble, Science, 2012, 335, 1610–1614.
49. D. M. Chew, P. J. Sylvester and M. N. Tubrett, Chem. Geol., 2011, 280, 200–216.

---