A single-column separation procedure for Sr, Nd, and Sm in small-size samples and high-precision isotope measurements using a TIMS with 10¹³ and 10¹² Ω amplifiers

Abstract

High-precision measurements of Sr, Nd, and Sm isotopes in precious and single mineral samples are significant in geochemistry and cosmochemistry. It is also a challenge for analytical techniques. In this study, a single-column separation procedure was developed using TODGA-normal resin to purify Sr, Nd, and Sm from microgeological samples (∼3 mg). The total chemical procedural blank for Sr, Nd, and Sm is less than 80, 7, and 3 pg, respectively; the recoveries are all greater than 91%. Methods have also been developed for the static measurement of Sr, Nd, and Sm isotopes in small size samples using a TIMS with Faraday cups and 10¹² and 10¹³ Ω amplifiers. The accuracy and precision of Sr, Nd, and Sm isotope ratios were analyzed using TIMS with three single-element calibration solutions, SRM 987 Sr, La Jolla Nd, and Alfa Aesar Sm, respectively. The results show that the repeatability of ⁸⁷Sr/⁸⁶Sr, ¹⁴²Nd/¹⁴⁴Nd, ¹⁴³Nd/¹⁴⁴Nd, ¹⁴⁹Sm/¹⁵²Sm and ¹⁵⁰Sm/¹⁵²Sm ratios at 30 ng each of Sr, Nd, and Sm is 0.710248 ± 11, 141 846 ± 25, 0.511852 ± 13, 0.516845 ± 17, and 0.275998 ± 9 (2 SD), respectively. Five geological rock reference materials and one meteorite sample were used to validate the developed chemical procedure and TIMS measurement methods for Sr, Nd, and Sm. The results indicate that Sr, Nd, and Sm isotope ratios of all samples are in agreement with the published data with a precision of 11–25 ppm (2 RSD). These methods are suitable for geochemistry and cosmochemistry studies for scarce samples.

---

1 Introduction

The ⁸⁷Rb–⁸⁶Sr, ¹⁴⁷Sm-¹⁴³Nd, and ¹⁴⁶Sm-¹⁴²Nd isotope systems are important tools for dating and geochemical tracing, which have been widely used in the field of geochemistry and cosmochemistry.^1–7 As long-lived radioisotope systems, both ⁸⁷Rb–⁸⁶Sr and ¹⁴⁷Sm-¹⁴³Nd can be applied to the chronological study of the whole history of planetary formation and evolution. On one hand, Rb, Sr, Sm, and Nd are all incompatible elements, but Rb is more incompatible than Sr, and Nd is more incompatible than Sm. Therefore, ⁸⁷Rb–⁸⁶Sr and ¹⁴⁷Sm-¹⁴³Nd isotope systems can be complementarily applied to date and trace geological processes such as magmatic activity, mantle–crust interaction, and high-grade metamorphism. On the other hand, the Rb–Sr isotope system is more easily disturbed by weathering or metamorphism, but the Sm–Nd system is not easily affected by these processes, making it a good complement for ⁸⁷Rb–⁸⁶Sr dating on the samples that experienced postcrystallization heating.^3,8–11 In addition, ¹⁴⁶Sm is a short-lived extinct nuclide, and the half-life of ¹⁴⁶Sm decay to ¹⁴²Nd is 103 Ma¹² (or 68 Ma¹³), providing a good time constraint for geological events during the first 500 Ma after solar system formation.^3,6,7,10,14 Extraterrestrial samples that are exposed to galactic cosmic rays may show alterations in Sm and Nd isotopic composition due to neutron capture. In particular, ¹⁴⁹Sm with large neutron capture cross sections is commonly treated as a sensitive monitor for this process.^10,15–19 Therefore, simultaneous high-precision analyses of Sr, Nd, and Sm isotopic ratios in the same sample are crucial for ⁸⁷Rb–⁸⁶Sr and ^{146, 147}Sm-^{142, 143}Nd isotopic system research in extraterrestrial samples.

For conventional Sr and Nd isotope analyses, the sample weight is typically more than 50 mg.^20–23 To eliminate matrix elements and isobaric interference elements for high-precision isotopic analysis, the purification of Sr and Nd commonly employed two-step,^23–26 three-step,^21,27–29 or four-step³⁰ chromatography methods, which are tedious and time-consuming. Furthermore, the complex separation scheme may introduce higher total procedural blanks. Li et al.^22,31 also reported a one-step separation scheme for Sr and Nd isotope analyses using a two-layered mixed-resin column. However, during their chemical procedure, the recovery of Sr was only 82%, and Nd was incompletely separated from other rare earth elements (REEs). Obviously, these methods are not suitable for Sr, Nd, and Sm isotope analyses of micro or scarce samples, such as meteorites, lunar and Martian rocks, and other scarce earth and extraterrestrial single minerals.

To effectively separate Sr and Nd of microgeological samples (<5 mg), one-step or two-step separation methods have been developed in recent years.^32–34 Nevertheless, four problems remain: (1) Nd was not separated from other REEs, such as Ce and Sm, which may cause serious isobaric interference of ¹⁴²Ce on ¹⁴²Nd and ¹⁴⁴Sm on ¹⁴⁴Nd during the measurement of ¹⁴²Nd/¹⁴⁴Nd and ¹⁴³Nd/¹⁴⁴Nd;³² (2) the recovery of Nd was only ∼40%;³⁴ (3) the total procedural blanks of Sr and Nd were high, making it difficult to obtain reliable analytical results;³³ and (4) Sr, Nd, and Sm were not simultaneously separated from the same aliquot.^32–34 Recently, a novel TODGA resin has been widely used to separate Nd,^{24,33,35–38} or Sr-Nd^33,37 from geological samples in the field of geochemistry, and it also has great potential for simultaneous separation of Sr, Nd, and Sm.

Both thermal ionization mass spectrometry (TIMS) and multiple collector inductively coupled plasma mass spectrometry (MC-ICP-MS) are commonly used to analyze Sr, Nd, and Sm isotopes.^{22,23,26,27,31} Although MC-ICP MS has the advantages of high ionization efficiency and short analysis time, TIMS is more suitable for high-precision isotope measurements due to its better stability, less molecular interference, and absence of the memory effect of the transmission tube. The analytical precisions of ⁸⁷Sr/⁸⁶Sr, ¹⁴³Nd/¹⁴⁴Nd, ¹⁴²Nd/¹⁴⁴Nd, ¹⁴⁹Sm/¹⁵²Sm, and ¹⁵⁰Sm/¹⁵²Sm were improved to 3 ppm,³⁹ 6 ppm,⁴⁰ 6 ppm,⁴⁰ 8 ppm,¹⁷ and 18 ppm,¹⁷ (2 RSD), respectively, when Faraday cups (FCs) were connected with conventional 10¹¹ Ω amplifiers on a TIMS. However, the sample loading amounts required are more than 500 ng for Sr, 500 ng for Nd, and 100 ng for Sm. Compared to the 10¹¹ Ω current amplifiers, the signal/noise ratios of a TIMS equipped with the 10¹² Ω and 10¹³ Ω amplifiers have been significantly improved by a factor of three to ten.^41–44 Theoretically, the enhanced TIMS configuration can obtain high-precision isotopic compositions even for much smaller sample loadings.

In this study, we use the TODGA resin to develop a single-column separation method for separating Sr, Nd, and Sm from microgeological samples (<3 mg). During the chemical procedure, Sr, Nd, and Sm are sequentially and completely separated from the sample matrices and isobaric interferences (such as Rb, Ce, and Gd). The single-column separation can be finished within 5 hours, significantly simplifying the chemical procedure and shortening the separation time. We also establish a high-precision isotope determination method on very small samples (30 ng of Sr, Nd, and Sm) employing static FCs with 10¹² and 10¹³ Ω amplifiers. To confirm the validity and repeatability of the presented method, five geological reference materials and a Jilin meteorite are analyzed. The results show that the ratios of ⁸⁷Sr/⁸⁶Sr, ^142,143Nd/¹⁴⁴Nd, and ^149,150Sm/¹⁵²Sm are consistent with the published data in the error range. Even if the sample size is less than one-tenth that of the conventional method, accurate and high-precision data can still be obtained by the developed methods.

2 Experimental section

The chemical procedures and the element content analyses are performed at the Guangxi Key Laboratory of Exploration for Hidden Metallic Ore Deposits, Guilin University of Technology. The major elements are measured by inductively coupled plasma-optical emission spectrometry (ICP-OES) (Avio 200, PerkinElmer, USA), and the trace elements are measured by inductively coupled plasma mass spectrometry (ICP-MS) (Agilent 7500cx, Agilent Technologies, USA). The isotopic compositions of Sr, Nd, and Sm are measured by TIMS (Triton Plus, Thermo Scientific, Germany) at the Guangzhou Institute of Geochemistry, Chinese Academy of Sciences (GIG, CAS).

High-purity reagents during the chemical separation include (1) G4 grade (trace elements < 100 pg g⁻¹) concentrated hydrochloric acid (HCl), nitric acid (HNO₃), and hydrofluoric acid (HF) made in Shanghai Aoban Technology Company Limited, China; (2) ultra-pure water obtained from an 18 MΩ cm⁻¹ grade Millipore purification system; (3) TaCl₅ powder made in Alfa Aesar China Co., Ltd (Shanghai, China); and (4) a W ribbon (99.995%, 0.04 mm × 0.7 mm), Re ribbon (99.99%, 0.025 mm × 1 mm), and Ta ribbon (99.99%, 0.025 mm × 1 mm) made by H. Cross Company, USA.

2.1 Reference materials and samples

The multi-element synthetic solutions of alkaline earth elements (MISA-01-1, 100 μg mL⁻¹) and rare earth elements (MISA-04-1, 100 μg mL⁻¹) manufactured by AccuStandard, Inc. (USA) are used to calibrate the optimal acid type and concentrations for separating Sr, Sm, and Nd.

Five international geological reference materials (GRMs) are chosen for this experiment, namely BIR-1a, BHVO-2, BCR-2, AGV-2, and GSP-2 (USGS, USA). These GRMs represent most of the igneous rock types on Earth, and are widely used for method validation and quality control in the fields of geochemistry.^20–23 The mass fraction of Sr, Nd, and Sm is shown in Table 1. For comparison, we also analyzed the Jilin meteorite, an ordinary chondrite (OC) that is the most abundant type of meteorite in the solar system.

Table 1 Detailed information of geological reference materials and a meteorite in this study

Sample	Lithology	Column chemistry (mg)	Mass fraction			Isotopic analysis
			Sr (μg g^−1)	Nd (μg g^−1)	Sm (μg g^−1)	Sr (ng)	Nd (ng)	Sm (ng)
a Note: element mass fractions are taken from the GeoReM database, http://georem.mpch-mainz.gwdg.de/. b Raczek et al. (2001).^62 c Data determined in our laboratory.
BIR-1a^a	Basalt	3	108.60	2.40	1.11	326	7	3
BHVO-2^a	Basalt	3	394.10	24.27	6.02	1182	73	18
BCR-2^a	Basalt	3	337.40	28.26	6.55	1012	85	20
AGV-2^a	Andesite	3	659.50	30.49	5.51	1979	91	17
GSP-2^b	Granodiorite	3	238.00	207.00	26.20	714	621	79
Jilin^c	OC (H5)	30	9.03	0.55	0.18	271	17	5

---

The isotope standard solutions of NIST SRM 987 for Sr (purchased from the National Institute of Standards and Technology, USA), La Jolla for Nd (provided by Prof. Castillo, UCSD, USA),⁴⁵ and Alfa Aesar for Sm (purchased from Alfa Aesar China Co., Ltd, Shanghai, China) are respectively used to set up the measurement method for Sr, Nd, and Sm isotope analyses by TIMS.

To ensure the homogeneity of the samples, approximately 30 mg GRMs and 90 mg Jilin meteorite are weighed and dissolved in 1 mL of concentrated HNO₃ and 2 mL of concentrated HF in a 7 mL Teflon vessel. The sealed vessel is heated at 120 °C for 5 days. The solution is then evaporated and redissolved with 2 mL of 6 M HCl. This step is repeated twice to ensure complete elimination of fluoride. Finally, the solution is dried and re-dissolved with 2 mL of 3 M HNO₃ for centrifugation, and the supernatant is separated and stored for subsequent column separation. Approximately one-tenth of the GRMs (3 mg) and one-third of the Jilin meteorite (30 mg) solution are loaded onto the TODGA column.

2.2 Column elution

A single-column separation scheme is employed for separating Sr, Nd, and Sm from small-size (3 mg) geological samples with TODGA-Normal resin (DN-B100-S, 50–100 μm, supplied by TrisKem International). The TODGA resin is pre-treated as described by Chu et al. (2019).³⁶ The pre-cleaned PFA column, with a length of 3 cm, an inner diameter of 4 mm, and a reservoir of 15 mL, is packed with 0.25 mL TODGA resin. Before sample loading, the TODGA column is first cleaned with 3 mL of 0.05 M HCl and 3 mL of ultra-pure water, and then conditioned with 3 mL of 3 M HNO₃. Subsequently, the sample solutions are loaded on the column. The matrix elements (such as Na, Mg, Al, K, Ti, Fe, Mn, and Ba) are eluted with 0.9 mL of 3 M HNO₃, and Sr is collected with 1.2 mL of 4 M HNO₃. After that, Ca is removed with 0.9 mL of 4 M HNO₃ and 1.2 mL of 0.1 M HNO₃, and light rare earth elements (La, Ce, and Pr) are eluted with 4 mL 2.5 M HCl. Finally, Nd and Sm are collected with 1.2 mL 1.2 M HCl and 2.5 mL 1.2 M HCl, respectively (Table 2). The whole procedure including pre-cleaning and pre-conditioning of the column takes a total time of ∼5 hours.

Table 2 The single-column separation protocol for Sr-Nd-Sm using TODGA resin

Procedure	Eluting reagent	Elution volume (mL)
Resin cleaning	0.05 M HCl	3
	Ultra-pure H2O	3
Preconditioning	3 M HNO3	3
Loading sample	3 M HNO3	0.2
Rinsing matrix	3 M HNO3	0.9 (0.3 × 3)
Collect Sr	4 M HNO3	1.2 (0.3 × 4)
Rinsing Ca	4 M HNO3	0.9 (0.3 × 3)
	0.1 M HNO3	1.2 (0.3 × 4)
Rinsing La, Ce, and Pr	2.5 M HCl	4.0 (0.5 × 8)
Collect Nd	1.2 M HCl	1.2 (0.3 × 4)
Collect Sm	1.2 M HCl	2.5 (0.5 × 5)
Rinsing HREE	1.2 M HCl	0.5
	0.05 M HCl	2.0 (0.5 × 4)

---

2.3 Mass spectrometry

The purified Sr, Nd, and Sm fractions of the five GRMs and the Jilin meteorite are evaporated to dryness and redissolved in 1 M HNO₃ for isotopic analyses by TIMS (Triton Plus). The TIMS is equipped with ten FCs and three types of amplifiers, including one 10¹¹ Ω, four 10¹² Ω, and five 10¹³ Ω.⁴⁵ The Sr isotopes are measured as Sr⁺ using a single W filament. Approximately 30 ng of Sr is loaded on pre-degassed W filaments using TaF₅ as the activator with a “sandwich” loading technique. The Nd isotopes are measured as Nd⁺ using double Re filaments. Approximately 30 ng of Nd is loaded on pre-degassed Re filaments without any activators. The Sm isotopes are measured as Sm⁺ using a single Ta filament. Similar to Sr loading, approximately 30 ng of Sm is sandwiched by the TaF₅ activator on pre-degassed Ta filaments.

2.3.1 Cup configuration. The cup configuration has been set up for a static measurement mode involving FCs with 10¹² Ω and 10¹³ Ω amplifiers as shown in Table 3. According to the different natural abundances of four Sr isotopes, three isotopes of ⁸⁶Sr, ⁸⁷Sr, and ⁸⁸Sr with higher abundance are collected by L1, C, and H1 FCs connected with three 10¹² Ω amplifiers, respectively; the ⁸⁴Sr with lower abundance and isobaric interference of ⁸⁵Rb are collected by L2 and L3 FCs connected with two 10¹³ Ω amplifiers, respectively. The setting principle of cup configurations during Nd and Sm isotope analyses is the same as that of Sr, which is also listed in Table 3. The gain calibration for each amplifier is performed each day, and the baseline and peak center are measured before the beginning of each run.

Table 3 The cup configuration for static measurement of Sr, Nd, and Sm isotopes using FCs by TIMS

Isotopes	Cup	L3	L2	L1	C	H1	H2	H3	H4
Sr	Mass	^84Sr	^85Rb	^86Sr	^87Sr	^88Sr	—	—	—
	Abundance (%)	0.56	—	9.86	7	85.58	—	—	—
	Amplifier	10^13Ω	10^13Ω	10^12Ω	10^12Ω	10^12Ω	—	—	—
Nd	Mass	^140Ce	^142Nd	^143Nd	^144Nd	^145Nd	^146Nd	^147Sm	^148Nd
	Abundance (%)	—	27.20	12.20	23.80	8.30	17.20	—	5.70
	Amplifier	10^13Ω	10^12Ω	10^12Ω	10^12Ω	10^13Ω	10^12Ω	10^13Ω	10^13Ω
Sm	Mass	^146Nd	^147Sm	^148Sm	^149Sm	^150Sm	^152Sm	^154Sm	^155Gd
	Abundance (%)	—	15.0	11.2	13.8	7.4	26.70	22.7	—
	Amplifier	10^13Ω	10^12Ω	10^13Ω	10^12Ω	10^13Ω	10^12Ω	10^12Ω	10^13Ω

---

2.3.2 Measurement parameters.
2.3.2.1 Strontium. The Sr isotopes are measured at 4000–4500 mA (1400–1500 °C). The typical ion beam intensity (⁸⁸Sr⁺) ranges from 0.1 to 3 V for all samples. Data acquisition consists of 10 blocks of 20 circles, with 4.194 s integration time and 12 s idle time per circle. The isobaric interference of ⁸⁷Rb on ⁸⁷Sr is corrected using a natural abundance ratio of ⁸⁷Rb/⁸⁵Rb = 0.3860. The instrumental mass fractionation of ⁸⁷Sr/⁸⁶Sr ratios is corrected using ⁸⁸Sr/⁸⁶Sr = 8.375209 with an exponential law.
2.3.2.2 Neodymium. During the Nd isotope measurement, the ionization filament is heated to ∼3000 mA (1300–1400 °C), and the evaporation filament is heated to 1200–2000 mA. The ¹⁴²Nd⁺ ion beam intensity of the La Jolla Nd standard is 0.1–0.3 V. For rock samples, the ¹⁴²Nd⁺ ion beam intensities for BHVO-2, AGV-2, and GSP-2 are ∼0.2 V and are <0.1 V for BCR-2, BIR-1a, and the Jilin meteorite. Each measurement consists of 10 blocks of 20 circles, with 8.389 s integration time and 12 s idle time per circle. The isobaric interference of ¹⁴²Ce (on ¹⁴²Nd), ¹⁴⁴Sm (on ¹⁴⁴Nd), and ¹⁴⁸Sm (on ¹⁴⁸Nd) are corrected using natural abundance ratios of ¹⁴⁷Sm/¹⁴⁴Sm = 4.8827, ¹⁴⁷Sm/¹⁴⁸Sm = 1.3336, and ¹⁴⁰Ce/¹⁴²Ce = 7.9584, respectively. Finally, the mass fractionations of both ¹⁴²Nd/¹⁴⁴Nd and ¹⁴³Nd/¹⁴⁴Nd ratios are corrected using ¹⁴⁶Nd/¹⁴⁴Nd = 0.7219 with an exponential law.
2.3.2.3 Samarium. In the course of Sm isotope analysis, the evaporation filament current is increased up to 3250–3550 mA (1400–1550 °C). The typical ¹⁵²Sm⁺ ion beam intensity of the Alfa Aesar Sm standard is 0.1–0.4 V. For rock samples, the ¹⁵²Sm⁺ ion beam intensities for all GRMs can be achieved at 0.3 V, except for BIR-1a and the Jilin meteorite, for which the ¹⁵²Sm⁺ signals are less than 0.1 V. Each measurement consists of 10 blocks of 20 circles, with 8.389 s integration time and 12 s idle time per circle. The isobaric interferences of ¹⁴⁸Nd (on ¹⁴⁸Sm), ¹⁵⁰Nd (on ¹⁵⁰Sm), ¹⁵²Gd (on ¹⁵²Sm), and ¹⁵⁴Gd (on ¹⁵⁴Sm) are corrected using ¹⁴⁸Nd/¹⁴⁶Nd = 0.3314, ¹⁵⁰Nd/¹⁴⁶Nd = 0.3255, ¹⁵²Gd/¹⁵⁵Gd = 0.0135 and ¹⁵⁴Gd/¹⁵⁵Gd = 0.1473, respectively. The mass fractionations of ¹⁴⁹Sm/¹⁵²Sm and ¹⁵⁰Sm/¹⁵²Sm ratios are corrected using ¹⁴⁷Sm/¹⁵²Sm = 0.56081 with an exponential law.

3 Results and discussion

3.1 Elution profiles of Sr, Nd, and Sm

We have developed a rapid single-column separation protocol to extract Sr, Nd, and Sm from small-size samples with TODGA resin. In the following section, we first test the elution profiles with mixed synthetic solutions (MISA-04-1 and MISA-01-1, 100 ng mL⁻¹) to calibrate the type, concentration, and volume of the eluent to separate Sr, Nd, and Sm. Then, different GRM samples and the Jilin meteorite with various lithologies and trace element contents are processed to verify the reproducibility of the chemical procedure.

3.1.1 Sr, Nd, and Sm purification of synthetic solutions and geological samples. The elution profile tested with mixed synthetic solutions is plotted in Fig. 1. During our study, the matrix element (Ba) and interference element (Rb) are first removed with 0.8 mL of 3 M HNO₃. The target element Sr is then collected with 1 mL of 0.1 M HNO₃. Some LREEs, such as La, Ce, and Pr can be completely removed with 4 mL of 2.5 M HCl. Then, the target elements of Nd and Sm are recovered with 1.2 mL of 1.2 M HCl and 2.8 mL of 1.2 M HCl, respectively. The other HREEs (from Eu to Lu) are finally rinsed with 2 mL of 0.05 M HCl. This procedure successfully separated Sr, Nd, and Sm from the same aliquot, but failed to evaluate the elution behavior of major elements (such as Na, Mg, Al, K, Ti, Mn, and Fe) on TODGA resin, which may cause serious matrix effects during the isotope measurements. Because the contents of major elements of mixed synthetic solutions are very low compared to those in the instrumental blanks of the ICP-MS, we do not plot them for comparison in Fig. 1.

Fig. 1 Elution profiles of Sr, Nd, and Sm. Separation is performed using 0.2 mL 100 ng g⁻¹ MISA synthetic solution and 0.25 mL TODGA resin.

During the separation of geological samples, the effect of major elements (including Na, Mg, Al, K, Ti, Mn, and Fe) should be evaluated, especially for Ca which is overloaded on the TODGA column for high-Ca samples;³⁷ hence, we have to adjust the concentration, volume and leaching sequence of the eluent. For the sake of brevity, we only take the dissolved sample solution containing 3 mg BIR-1a (high-Ca basalt) as an example to test the elution profile (Fig. 2a). Almost all of the matrix elements (including Na, Mg, Al, K, Ti, Mn, Fe, and Ba, except Ca) together with the interference element (Rb) are easily rinsed with 0.8 mL of 3 M HNO₃ due to their weak retention on the TODGA resin in 3 M HNO₃. Then, Sr is successfully separated from Ca with 2.1 mL of 4 M HNO₃ and 1.2 mL of 0.1 M HNO₃. Most of Ca (92%) is leached out in these steps, and the residual Ca together with La, Ce, and Pr is leached in 4 mL of 2.5 M HCl. Finally, Nd and Sm are collected with 1.2 mL of 1.2 M HCl and 2.5 mL of 1.2 M HCl, respectively. Using the improved procedure (Table 2 and Fig. 2), Sr, Nd, and Sm are successfully separated in the other GRMs (BHVO-2, BCR-2, AGV-2, and GSP-2). The collected intervals of Sr (1.2 mL of 4 M HNO₃), Nd (1.2 mL of 1.2 M HCl), and Sm (3 mL of 1.2 M HCl) of all GRMs have similar distribution patterns (Fig. 2b), demonstrating good reproducibility of our procedure.

Fig. 2 The repeatability of elution profiles for the separation of Sr, Nd, and Sm using the improved method for five natural rock samples (3 mg). ME.: matrix elements (such as Na, Mg, Al, K, Ti, Fe, Mn, and Ba).

3.1.2 Recovery and procedural blanks. The recovery is calculated using a formula: Y_E(%) = E_elu/E_es × 100%. Here, Y_E represents the element recovery; E represents the counts of the element Sr, Nd, or Sm determined by ICP-MS; elu represents the eluent; es represents the external standard solution, i.e., another original sample solution which is the same as the loading sample solution analyzed at the same time as the eluent.

During the 1.2 mL of 4 M HNO₃ elution step, the recovery of Sr is greater than 92%. Almost all of the Rb and Ca has been removed in this fraction. The recovery of Nd during the 1.2 mL of 1.2 M HCl elution step is greater than 91%. The complete separation of Nd from Ce and Sm allows us to avoid isobaric interference of ¹⁴²Ce and ¹⁴²Nd, respectively, on ¹⁴⁴Sm and ¹⁴⁴Nd, which is particularly important for the accurate determination of ¹⁴²Nd/¹⁴⁴Nd and ¹⁴³Nd/¹⁴⁴Nd. The recovery of Sm during the 2.5 mL of 1.2 M HCl elution step is greater than 93%. Almost no Nd and Gd is detected in this fraction. The total procedural blanks including sample digestion and chemical separation are less than 0.1% (80 pg of Sr, 7 pg of Nd, and 3 pg of Sm), which is negligible relative to the amounts of the analytes contained in 3 mg GRMs.³⁴

3.1.3 Separation protocols for ultra-low concentrations. Using the 0.25 mL TODGA resin, our single-column separation method can successfully separate Sr, Nd, and Sm from small-size (≤3 mg) geological samples. Yet, this chemical procedure cannot meet the need to analyze those samples with ultra-low concentrations of Sr, Nd, or Sm less than 1.0 μg g⁻¹, such as the Jilin meteorite. To obtain sufficient Nd and Sm for isotopic analysis by TIMS, the sample amounts loaded on the TODGA column must be increased. When the sample loading amount is 30 mg (Jilin meteorite), the column is highly overloaded, and most of the Sr is eluted in the first step by 3 M HNO₃ along with matrix elements, while the elution behaviors of Nd and Sm are barely affected (Fig. S1†). Therefore, a two-column separation method using both TODGA and Sr Spec resins is employed. Following the single TODGA column, the recovered Sr-containing fraction (in 3 M and 4 M HNO₃) is evaporated and redissolved in 0.2 mL of 8 M HNO₃ and then loaded on the pre-cleaned and pre-conditioned 0.14 mL Sr Spec column. As shown in Fig. S2,† the matrix elements (i.e., Ca and Ba) and isobaric interference element of Rb are completely removed with 1.2 mL of 8 M HNO₃, and Sr is collected with 0.6 mL of ultra-pure H₂O. For ultra-low-content samples with sample sizes that have to be increased to more than 3 mg, Sr, Nd, and Sm can be purified using the two-column method.

3.1.4 Comparision with previous separation protocols. Previous studies have developed different separation methods, such as one-step, two-step, three-step, and four-step methods to separate Sr, Nd, and other elements of interest from conventional geological samples (50–200 mg) and microgeological samples (<5 mg), as shown in Table 4. However, two main problems remained in these experiments: (1) multi-step column procedures are tedious and time-consuming; (2) Sm was not separated after the separation of Sr and Nd from the same aliquot of a small-size sample. Compared to previous separation methods (Table 4), the advantages of our new protocol include (1) the low sample consumption (∼3 mg) with the potential for isotope analysis of scarce samples; (2) the simple and time-saving separation procedure; (3) low consumption of resin and the reagent and the resultant low procedural blanks; (4) simultaneous separation of Sr, Nd, and Sm from the same aliquot with high recoveries (>91%).

Table 4 Comparison of the separation methods and measurement methods of Sr, Nd, and Sm in this study with previous studies^{21–23,30,32–34}

Reference	Separation methods	Analytes	Sample-size	Resin	Elution volume	Recovery	Blanks	Separation duration	Instrument	Amplifiers (Ω)
Li et al.^22	One-step	Sr, Nd, Hf	50–60 mg	1.6 mL AG50W × 12	53 mL	Sr: 71%	Sr: 170–200 pg	9 h	TIMS	10^11
				0.4 mL Ln spec		Nd (REE): 96%	Nd: 70–80 pg
Lei et al.^23	Two-step	Sr, Nd, Pb, Hf	100–200 mg	0.8 mL AG50W-X8	102 mL	Sr: 86%	Sr: 48 pg	7.5 h	TIMS	10^11
				1.5 mL Ln spec		Nd: 93%	Nd: 36 pg
				0.5 mL Sr spec
Yang et al.^21	Three-step	Lu, Hf, Rb, Sr, Sm, Nd	100 mg	2 mL Ln spec	168 mL	Rb: 90%	Rb: 50 pg	3 days	TIMS	10^11
				2 mL AG50W × 12		Sr: 90%	Sr: 100 pg
						Nd: 90%	Nd: 50 pg
						Sm: 90%	Sm: 50 pg
Pin et al.^30	Four-step	Nd	100–200 mg	0.25 mL TRU. Spec	24.7 mL	Nd: 80–90%	Nd: 30 pg	2 days	TIMS	10^11
				0.95 mL Ln2 spec
				0.95 mL TODGA
Misawa et al.^32	Two-step	Sr, REE	1–5 mg	80 μl RE resin	4.7 mL	Nd (REE): 90%	Nd: 3 pg	—	TIMS	—
				50 μl Sr resin		Sr: 93–98%	Sr: 36 pg
Li et al.^34	One-step	Nd, Sm	1–3.5 mg	1.7 mL Ln spec	46.6 mL	Nd: 40%	Nd: 8 pg	2–3 h	TIMS	10^11
						Sm: 90%	Sm: 3 pg
Retzmann et al.^33	One-step	Sr, Nd, Pb	5 mg	1 mL TODGA	66 mL	Sr: 90%	Sr: 350 pg	2 h	MC-ICP MS	—
						Nd: 91%	Nd: 2280 pg
This study	One-step	Sr, Nd, Sm	3 mg	0.25 mL TODGA	14.6 mL	Sr: 92%	Sr: 80 pg	5 h	TIMS	10^12 + 10^13
						Nd: 91%	Nd: 7 pg
						Sm: 93%	Sm: 3 pg

---

3.2 Isotopic compositions of Sr, Nd, and Sm measured by TIMS

With the advent of the 10¹³ Ω amplifier, several high-precision TIMS methods for Sr and Nd isotopes analyses have been reported for small-size samples,^42,46 but Sm isotope measurement using 10¹³ Ω amplifiers has not yet been reported. In this work, static measurement methods for Sr, Nd, and Sm isotopes in small size samples were set up using FCs connected to 10¹² and 10¹³ Ω amplifiers of a TIMS. We analyzed different signal intensities and analytical amounts for Sr, Nd, and Sm isotopes. These methods significantly improve the precision of Sr, Nd, and Sm isotopic ratios for small-size/low-content samples compared to the conventional methods using 10¹¹ Ω amplifiers.

Single element calibration solutions of NIST SRM 987, La Jolla Nd, and Alfa Aesar Sm are measured to evaluate the accuracy and precision of Sr, Nd, and Sm isotope ratios using our new cup configurations on the TIMS. Five processed GRM samples and one Jilin meteorite sample are also measured to evaluate the validity and feasibility of the presented method. All analytical results are shown in Table 5.

Table 5 The accuracy and precision of Sr, Nd, and Sm isotope ratios of calibration solutions and natural samples

	^87Sr/^88Sr	n	^142Nd/^144Nd	^143Nd/^144Nd	n	^149Sm/^152Sm	^150Sm/^152Sm	n
a Data are analyzed in this study; “n” is the number of repeated measurements of the same aliquot; “#” indicates one sample separated by the single resin column; Reference: range of published data, references are listed in square brackets below. The uncertainty is the last digit of 2 SD.
Calibration solutions
	NIST SRM 987		La Jolla Nd			Alfa Aesar Sm
Mean^a	0.710248 ± 11	18	1.141846 ± 25	0.511852 ± 13	24	0.516844 ± 14	0.275996 ± 6	24
Reference	0.710241–0.710268 [ref. 20, 22, 23, 26, 27, 31 and 47–49]	—	1.141833–1.141838 [ref. 15, 18, 19, 52 and 53]	0.511839–0.511853 [ref. 15, 18, 19, 49, 50, 52 and 53]	—	No data	No data	—
BHVO-2
#1	0.703477 ± 14	3	1.141830 ± 28	0.512971 ± 19	3	0.516853 ± 19	0.275996 ± 8	3
#2	0.703474 ± 19	3	1.141835 ± 43	0.512970 ± 9	3	0.516846 ± 16	0.275998 ± 15	3
Mean^a	0.703475 ± 15	6	1.141833 ± 33	0.512970 ± 13	6	0.516849 ± 18	0.275997 ± 11	6
Reference	0.703475–0.703493 [ref. 20, 22, 23, 27, 28, 31, 47–50 and 63]	—	1.141833–1.141838 [ref. 30, 38 and 55]	0.512957–0.512987 [ref. 20, 22, 27, 30, 38, 49 and 55]	—	0.516860 [ref. 59]	0.275988 [ref. 59]	—
BCR-2
#1	0.704998 ± 19	3	1.141865 ± 45	0.512633 ± 29	3	0.516845 ± 20	0.275998 ± 9	3
#2	0.705003 ± 10	3	1.141839 ± 36	0.512637 ± 23	3	0.516833 ± 23	0.275994 ± 8	3
Mean^a	0.705000 ± 15	6	1.141852 ± 46	0.512635 ± 24	6	0.516839 ± 23	0.275996 ± 9	6
Reference	0.705000–0.705032 [ref. 20, 22, 23, 27, 28, 31, 47–50 and 63]	—	1.141835–1.141854 [ref. 18, 38, 40, 55 and 56]	0.512620–0.512638 [ref. 18, 20, 22, 28, 38, 40, 49 and 56]	—	0.516853–0.516857 [ref. 18 and 59]	0.275990–0.275993 [ref. 18 and 59]	—
AGV-2
#1	0.703979 ± 5	3	1.141856 ± 52	0.512777 ± 24	3	0.516838 ± 14	0.275989 ± 8	3
#2	0.703978 ± 12	3	1.141860 ± 42	0.512768 ± 13	3	0.516847 ± 29	0.275995 ± 18	3
Mean^a	0.703979 ± 8	6	1.141853 ± 40	0.512771 ± 19	6	0.516842 ± 22	0.275992 ± 14	6
Reference	0.703970–0.703994 [ref. 23, 31, 39 and 47–50]	—	1.141836 [ref. 55]	0.512755–0.512790 [ref. 31, 49, 50 and 55]	—	No data	No data	—
GSP-2
#1	0.765035 ± 7	3	1.141855 ± 42	0.511368 ± 23	3	0.516856 ± 18	0.275989 ± 10	3
#2	0.765041 ± 9	3	1.141871 ± 43	0.511370 ± 16	3	0.516853 ± 8	0.275994 ± 15	3
Mean^a	0.765038 ± 10	6	1.141863 ± 42	0.511369 ± 18	6	0.516854 ± 12	0.275992 ± 13	6
Reference	0.764954–0.765177 [ref. 20 and 48–50]	—	1.141835 [ref. 55]	0.511348–0.511369 [ref. 27, 31, 49, 50 and 55]	—	No data	No data	—
BIR-1a
#1	0.703107 ± 6	3	1.141840 ± 45	0.513098 ± 10	3	0.516858 ± 15	0.275994 ± 7	2
#2	0.703104 ± 13	3	1.141854 ± 31	0.513093 ± 44	3	0.516857 ± 21	0.275993 ± 9	2
Mean^a	0.703107 ± 10	6	1.141847 ± 38	0.513096 ± 29	6	0.516858 ± 19	0.275994 ± 9	4
Reference	0.703104–0.703116 [ref. 20, 22, 26 and 31]	—	1.141835–1.141866 [ref. 30 and 56]	0.513078–0.513101 [ref. 20, 22, 26, 30, 31 and 56]	—	No data	No data	—
Jilin
#1^a	0.746757 ± 11	3	1.141841 ± 39	0.512586 ± 15	3	0.516855 ± 10	0.275999 ± 6	3
Reference	0.7464 [ref. 51]	—	1.141801 ± 6 [ref. 19]	0.512581 ± 3 [ref. 19]	3	No data	No data	—

---

3.2.1 Accuracy and precision of the Sr isotope. As shown in Fig. 3, the ⁸⁷Sr/⁸⁶Sr ratios of the NIST SRM 987 standard are determined at different ⁸⁸Sr⁺ ion beam intensities (0.1 V, 0.2 V, and 0.3 V), yielding an average ⁸⁷Sr/⁸⁶Sr value of 0.710246 ± 103 (2 SD, n = 16), 0.710242 ± 72 (2 SD, n = 18), and 0.710247 ± 63 (2 SD, n = 22), respectively. The reproducibility of ⁸⁷Sr/⁸⁶Sr ranges from 89 to 145 ppm (2 RSD), worse than the typical values (8–25 ppm, 2 RSD) in the literature.^{20,22,23,26,27,31,47–49} Therefore, we improve the ⁸⁸Sr signal to ∼3 V for better precision. Replicate analyses of NIST SRM 987 yield an average ⁸⁷Sr/⁸⁶Sr value of 0.710248 ± 11 (2 SD, n = 18), in good agreement with the reported data (Fig. 3).^{20,22,23,26,27,31,47–49} The reproducibility of 15 ppm (2 RSD) is comparable with 15–20 ppm (2 RSD) reproducibility at a ⁸⁸Sr signal of 5–16 V published in the literature.^22,26,48,49

Fig. 3 The accuracy and precision of ⁸⁷Sr/⁸⁶Sr ratios measured with SRM 987 solution at different ion beam signals and Sr amounts. The solid red circles are the data from this study; the solid black diamonds are published data, the source of which is shown in Table 5. The error bars for each data point in this study are internal precision of 2 SE. The error bars of published data are external precision of 2 SD; the black line is the average value of the published data. The precision and symbols in the following figures are represented as in this figure.

The determined ⁸⁷Sr/⁸⁶Sr ratios of five GRMs (BIR-1a, BHVO-2, BCR-2, AGV-2, and GSP-2) and the Jilin meteorite are shown in Table 5 and plotted in Fig. 4. The external repeatability of ⁸⁷Sr/⁸⁶Sr for all samples is better than 21 ppm (2 RSD). The average ⁸⁷Sr/⁸⁶Sr ratio of BHVO-2 is 0.703475 ± 15 (2 SD, n = 6), compatible with the value of 0.703479 ± 13 (2 SD, n = 2) measured at a ⁸⁸Sr signal of 5 V reported by Li et al.²⁰ The average ⁸⁷Sr/⁸⁶Sr ratio of BCR-2 is 0.705002 ± 13 (2 SD, n = 6), consistent with the value of 0.705000 ± 11 (2 SD, n = 21) at 4.5 V reported by Jweda et al.²⁸ The ⁸⁷Sr/⁸⁶Sr ratio of AGV-2 is 0.703979 ± 8 (2 SD, n = 6), identical to the value of 0.703979 ± 8 (2 SD, n = 4) at 5–7 V reported by Li et al.³¹ The average ⁸⁷Sr/⁸⁶Sr ratio of BIR-1a is 0.703107 ± 10 (2 SD, n = 6), compatible with the value of 0.703104 ± 15 (2 SD, n = 6) at 5 V reported by Li et al.²⁰ The average ⁸⁷Sr/⁸⁶Sr ratio of GSP-2 is 0.765038 ± 10 (2 SD, n = 6). It is noteworthy that the reported ⁸⁷Sr/⁸⁶Sr ratios of GSP-2 show a large variation (Fig. 4d), ranging from 0.764962 to 0.765177,^{20,27,48–50} which have been attributed to the sample heterogeneity.^27,50 Our obtained ⁸⁷Sr/⁸⁶Sr ratios of GRMs are consistent with these reported data, demonstrating the good accuracy of Sr isotope measurement (Fig. 4). Nevertheless, the average ⁸⁷Sr/⁸⁶Sr ratio of the Jilin meteorite is 0.746757 ± 11 (2 SD, n = 3), roughly approximate to a few published data (⁸⁷Sr/⁸⁶Sr = 0.7464 ± 6) with large analytical uncertainties.⁵¹

Fig. 4 The accuracy and precision of ⁸⁷Sr/⁸⁶Sr ratios measured with GRMs and the Jilin meteorite.

3.2.2 Accuracy and precision of the Nd isotope. The ¹⁴³Nd/¹⁴⁴Nd and ¹⁴²Nd/¹⁴⁴Nd ratios of the La Jolla Nd standards are measured at an ¹⁴²Nd⁺ ion beam intensity of 0.1–0.3 V, yielding average ¹⁴³Nd/¹⁴⁴Nd and ¹⁴²Nd/¹⁴⁴Nd values of 0.511852 ± 13 (2 SD, n = 24), and 1.141846 ± 25 (2 SD, n = 24), respectively, in good agreement with the published data (Fig. 5).^{15,18,19,49,52,53} The repeatability of ¹⁴³Nd/¹⁴⁴Nd is 26 ppm (2 RSD), which is compatible with the 23–29 ppm (2 RSD) repeatibility obtained by Weis et al.,⁴⁹ but worse than 4–14 ppm (2 RSD) repeatability obtained at a ¹⁴²Nd⁺ signal of 1.6–8 V in the literature.^{15,18,19,52,53} This repeatability meets the demand for the ¹⁴⁷Sm–¹⁴³Nd chronological study.²⁴ The external repeatability of ¹⁴²Nd/¹⁴⁴Nd is 22 ppm (2 RSD), which is worse than the repeatability of 3–7 ppm (2 RSD) at 1.6–8 V in the literature.^{15,18,19,52,53} However, the ¹⁴²Nd/¹⁴⁴Nd ratios are also generally reported utilizing the ε-notation, namely parts per ten thousand deviations from the average of the terrestrial standards (ε¹⁴²Nd = [(¹⁴²Nd/¹⁴⁴Nd)_sample/(¹⁴²Nd/¹⁴⁴Nd)_{La Jolla} − 1] × 10⁴). At this point, our external reproducibility of ε¹⁴²Nd is ±0.22 (2 SD), which is better than ±0.33 (2 SD) measured by Wang et al.³⁵ at ¹⁴²Nd⁺ signals more than 4.5 V using an MC-ICP MS. Nevertheless, this repeatability must be improved in the field of ¹⁴⁶Sm-¹⁴²Nd chronology, because the most prominent ε¹⁴²Nd anomaly recorded in the chondrites and Isua supracrustal belt on the Earth ranges from −0.40 to +0.12,^19,54 which requires the analytical error of ε¹⁴²Nd to be at least less than ±0.10 (2 SD).^{15,18,19,52,53}

Fig. 5 The accuracy and precision of Nd isotope ratios measured with La Jolla Nd solution. The black lines represent the average values of the data from this study.

The determined ¹⁴³Nd/¹⁴⁴Nd ratios of five GRMs and the Jilin meteorite are listed in Table 5 and plotted in Fig. 6. The external repeatabilities of ¹⁴³Nd/¹⁴⁴Nd are 47–57 ppm (2 RSD) for BIR-1a and BCR-2, and 25–37 ppm (2 RSD) for BHVO-2, GSP-2, AGV-2, and the Jilin meteorite. The average ¹⁴³Nd/¹⁴⁴Nd ratio of BIR-1a is 0.513096 ± 29 (2 SD, n = 6), compatible with the values of 0.513092 ± 13 (2 SD, n = 4) reported by Li et al.³¹ The average ¹⁴³Nd/¹⁴⁴Nd ratio of BCR-2 is 0.512635 ± 24 (2 SD, n = 6), consistent with the values of 0.512636 ± 24 (2 SD, n = 2) reported by Li et al.²⁰ The average ¹⁴³Nd/¹⁴⁴Nd ratio of BHVO-2 is 0.512970 ± 13 (2 SD, n = 6), compatible with the values of 0.512976 ± 14 (2 SD, n = 2) reported by Li et al.²² and 0.512974 ± 12 (2 SD, n = 4) reported by Hui et al.²⁷ The average ¹⁴³Nd/¹⁴⁴Nd ratio of GSP-2 is 0.511369 ± 18 (2 SD, n = 6), identical to the values of 0.511369 ± 3 (2 SD, n = 5) reported by Weis et al.⁴⁹ The average ¹⁴³Nd/¹⁴⁴Nd ratio of AGV-2 is 0.512771 ± 19 (2 SD, n = 6), compatible with the values of 0.512776 ± 4 (2 SE, n = 1) reported by Li et al.³¹ The average ¹⁴³Nd/¹⁴⁴Nd ratio of the Jilin meteorite is 0.512586 ± 15 (2 SD, n = 6), consistent with the values of 0.512581 ± 3 (2 SD, n = 3) reported by Gannoun et al.¹⁹ As shown in Fig. 6, our determined ¹⁴³Nd/¹⁴⁴Nd ratios of GRMs and the Jilin meteorite are all in good agreement with published data within the analytical uncertainties,^{18,20,22,27,28,31,38,40,49,50,55,56} demonstrating the accuracy of Nd isotope measurement.

Fig. 6 The accuracy and precision of ¹⁴³Nd/¹⁴⁴Nd ratios measured with GRMs and the Jilin meteorite.

The ¹⁴²Nd/¹⁴⁴Nd ratios of GRMs and the Jilin meteorite are reported utilizing the ε-notation and plotted in Fig. 7. The external reproducibilities of ε¹⁴²Nd for all samples are better than ± 0.40 (2 SD). The average ε¹⁴²Nd values of BIR-1a, BHVO-2, BCR-2, AGV-2, GSP-2, and the Jilin meteorite are 0.01 ± 0.33, −0.12 ± 0.29, 0.05 ± 0.40, 0.06 ± 0.35, 0.15 ± 0.37 (2 SD, n = 6), and −0.04 ± 0.34 (2 SD, n = 3), respectively. As shown in Fig. 7, the average ε¹⁴²Nd values of all samples are plotted within the ranges of analytical uncertainties (±0.22) determined on the La Jolla standard, indicating that no significant ¹⁴²Nd anomaly is detected in our study, which is also consistent with other published data.^30,38,55,56 However, the relatively large analytical uncertainty of ε¹⁴²Nd (±0.22) fails to distinguish an ε¹⁴²Nd anomaly less than 40 ppm.

Fig. 7 The values of ε¹⁴²Nd measured with GRMs and the Jilin meteorites in this study. The solid lines represent the mean ε¹⁴²Nd value and the dashed lines represent an external precision of 2 SD of the measured mean ε¹⁴²Nd value. The gray area represents an external repeatability of ±0.22 ε (2 SD) of the La Jolla Nd standard.

3.2.3 Accuracy and precision of the Sm isotope. The ¹⁴⁹Sm/¹⁵²Sm and ¹⁵⁰Sm/¹⁵²Sm ratios of the Alfa Aesar Sm standard are measured at an ¹⁵²Sm⁺ ion beam intensity of 0.1–0.4 V, yielding average ¹⁴⁹Sm/¹⁵²Sm and ¹⁵⁰Sm/¹⁵²Sm ratios of 0.516844 ± 14 (2 SD, n = 24) and 0.275996 ± 6 (2 SD, n = 24), respectively. Although no Sm isotope data of the same isotopic standard (Alfa Aesar Sm) have been reported in previous studies; our obtained ¹⁴⁹Sm/¹⁵²Sm and ¹⁵⁰Sm/¹⁵²Sm values showed a good agreement with the data determined using Ames Sm isotopic standard solution (Fig. 8).^{10,15–19,57,58} This might reflect the fact that the Sm isotopic composition of unirradiated samples on the Earth is invariant.⁵⁹ The repeatability of ¹⁴⁹Sm/¹⁵²Sm is 27 ppm (2 RSD), which is slightly worse than the reproducibility of 6–19 ppm (2 RSD) obtained at 1–6 V reported in the literature.^{10,15–19,57,58} The repeatability of ¹⁵⁰Sm/¹⁵²Sm is 23 ppm (2 RSD), better than 62 ppm (2 RSD) reported by Carlson et al.⁵⁸ and 29 ppm (2 RSD) reported by Gannoun et al.¹⁹

Fig. 8 The accuracy and precision of Sm isotope ratios measured with Alfa Asear Sm solution. Black diamonds represent published data determined using Ames Sm isotope standard solutions. The black lines represent the average values of the data in this study and literature.

The determined ¹⁴⁹Sm/¹⁵²Sm and ¹⁵⁰Sm/¹⁵²Sm ratios of GRMs and the Jilin meteorite are also reported as the ε-notation (εⁱSm = [(ⁱSm/¹⁵²Sm)_sample/(ⁱSm/¹⁵²Sm)_{Alfa Aesar Sm} − 1] × 10⁴), as shown in Fig. 9. The external repeatabilities of ε¹⁴⁹Sm and ε¹⁵⁰Sm for all samples are better than ±0.45 (2 SD) and ±0.50 (2 SD), respectively. The average ε¹⁴⁹Sm and ε¹⁵⁰Sm values in this study are identical to those of the Alfa Aesar Sm standard within the ranges of the external reproducibilities of ±0.27 (2 SD) and ±0.23 (2 SD) for ε¹⁴⁹Sm and ε¹⁵⁰Sm, respectively, suggesting the absence of significant ε¹⁴⁹Sm and ε¹⁵⁰Sm anomalies for the studied samples. Only a few ε¹⁴⁹Sm and ε¹⁵⁰Sm values of BCR-2 and BHVO-2 have been found in the literature, e.g., BCR-2 spans ε¹⁴⁹Sm and ε¹⁵⁰Sm ranges from −0.15 to 0.06 and −0.22 to 0.04, respectively,^18,59 while BHVO-2 has ε¹⁴⁹Sm and ε¹⁵⁰Sm values of −0.10 and −0.29,⁵⁹ respectively, consistent with the data determined in this study. Therefore, our Sm isotope measurement method can meet the demand to resolve the Sm isotopic shift caused by neutron capture correction in the field of cosmochemistry, e.g., the largest ε¹⁴⁹Sm and ε¹⁵⁰Sm anomaly of the irradiated lunar sample achieved from −43 to +81, respectively.^60,61

Fig. 9 The values of ε¹⁴⁹Sm (a) and ε¹⁵⁰Sm (b) measured with GRMs and the Jilin meteorite in this study. The gray areas in (a) and (b) represent an external repeatability of ±0.27 ε (2 SD) and ±0.23 ε (2 SD) of Alfa Aesar Sm, respectively.

4 Conclusion

We develop a simple and rapid single-column separation protocol for separating Sr, Nd, and Sm from microgeological samples (3 mg) with TODGA resin and optimize the cup configuration of a TIMS for high-precision isotope determination. The advantages of our separation scheme include (1) low sample consumption (∼3 mg); (2) efficient and time-saving (∼5 hours); (3) negligible procedural blanks (80 pg of Sr, 7 pg of Nd, and 3 pg of Sm); and (4) high recoveries (>91%) and simultaneous separation of Sr, Nd, and Sm from the same aliquot. In addition, a second Sr Spec column is required to further purify Sr for ultra-low concentration samples with Sr, Nd, and Sm less than 1 μg g⁻¹, such as the Jilin meteorite. For very small samples (30 ng of Sr, Nd, and Sm), a Triton Plus TIMS equipped with 10¹³ and 10¹² Ω amplifiers can also obtain high-precision Sr, Nd, and Sm isotope ratios. Repeated analyses on five GRMs and the Jilin meteorite yield ⁸⁷Sr/⁸⁶Sr, ^142,143Nd/¹⁴⁴Nd, and ^149,150Sm/¹⁵²Sm ratios compatible with the published data within uncertainty, demonstrating excellent validation and repeatability. Our new chemical separation scheme and isotope determination method have great potential for ⁸⁷Rb–⁸⁶Sr and ¹⁴⁷Sm-¹⁴³Nd isotopic dating and geochemical tracing in the field of geochemistry and cosmochemistry.

Author contributions

Gui-Qin Wang: conceptualization, methodology, writing – review & editing, data curation, funding acquisition; Yu-Ming Xu: chemical experiments, TIMS measurement, writing – review & editing, data curation; Zhen Yang: chemical experiments, TIMS measurement; Yuling Zeng: TIMS measurement; Feng Guo: supervision, reviewing and editing the manuscript.

Conflicts of interest

There are no conflicts of interest to declare.

Acknowledgements

The authors had the great honor of applying for the Chang'e-5 lunar samples allocated by the China National Space Administration (CNSA), for which we developed this method for high-precision measurements of Sr, Nd, and Sm isotopes in this sample, as well as in other similar small-size samples. This study was financially supported by the Strategic Priority Research Program of the Chinese Academy of Sciences (XDB 41000000) and the National Natural Science Foundation of China (42073061). We would like to thank Dr Zheng-Lin Li and Dr Yin-Hui Zhang, technicians at the Guangxi Key Laboratory of Exploration for Hidden Metallic Ore Deposits, Guilin University of Technology, for their help throughout the process of establishing the chemical separation and purification procedure. Thanks to the two anonymous reviewers for their constructive comments, which improved the quality of the manuscript. This is contribution no. IS-3398 from GIGCAS.

Notes and references

1. G. A. Snyder, L. A. Taylor and A. N. Halliday, Geochim. Cosmochim. Acta, 1995, 59, 1185–1203.
2. J. Edmunson, L. E. Borg, L. E. Nyquist and Y. Asmerom, Geochim. Cosmochim. Acta, 2009, 73, 514–527.
3. N. E. Marks, L. E. Borg, C. K. Shearer and W. S. Cassata, J. Geophys. Res.: Planets, 2019, 124, 2465–2481.
4. H. Hui, C. R. Neal, C.-Y. Shih and L. E. Nyquist, Earth Planet. Sci. Lett., 2013, 373, 150–159.
5. L. E. Borg, G. A. Brennecka and T. S. Kruijer, Proc. Natl. Acad. Sci. U. S. A., 2022, 119, e2115726119.
6. R. W. Carlson, L. E. Borg, A. M. Gaffney and M. Boyet, Philos. Trans. R. Soc., A, 2014, 372, 20130246.
7. L. E. Borg, J. N. Connelly, W. S. Cassata, A. M. Gaffney and M. Bizzarro, Geochim. Cosmochim. Acta, 2017, 201, 377–391.
8. M. Boyet, R. W. Carlson, L. E. Borg and M. Horan, Geochim. Cosmochim. Acta, 2015, 148, 203–218.
9. C. K. Sio, L. E. Borg and W. S. Cassata, Earth Planet. Sci. Lett., 2020, 538, 116–219.
10. L. E. Borg, J. N. Connelly, M. Boyet and R. W. Carlson, Nature, 2011, 477, 70.
11. L. E. Borg, A. M. Gaffney and C. K. Shearer, Meteorit. Planet. Sci., 2015, 50, 715–732.
12. N. E. Marks, L. E. Borg, I. D. Hutcheon, B. Jacobsen and R. N. Clayton, Earth Planet. Sci. Lett., 2014, 405, 15–24.
13. N. Kinoshita, M. Paul, Y. Kashiv, P. Collon, C. M. Deibel, B. DiGiovine, J. P. Greene, D. J. Henderson, C. L. Jiang, S. T. Marley, T. Nakanishi, R. C. Pardo, K. E. Rehm, D. Robertson, R. Scott, C. Schmitt, X. D. Tang, R. Vondrasek and A. Yokoyama, Science, 2012, 335, 1614–1617.
14. A. M. Gaffney and L. E. Borg, Geochim. Cosmochim. Acta, 2014, 140, 227–240.
15. K. Rankenburg, A. D. Brandon and C. R. Neal, Science, 2006, 312, 1369–1372.
16. A. D. Brandon, T. J. Lapen, V. Debaille, B. L. Beard, K. Rankenburg and C. Neal, Geochim. Cosmochim. Acta, 2009, 73, 6421–6445.
17. C. L. McLeod, A. D. Brandon and R. M. G. Armytage, Earth Planet. Sci. Lett., 2014, 396, 179–189.
18. L. E. Borg, G. A. Brennecka and S. J. K. Symes, Geochim. Cosmochim. Acta, 2016, 175, 150–167.
19. A. Gannoun, M. Boyet, H. Rizo and A. El Goresy, Proc. Natl. Acad. Sci. U. S. A., 2011, 108, 7693–7697.
20. C. F. Li, X. H. Li, Q. L. Li, J. H. Guo, X. H. Li and Y. H. Yang, Anal. Chim. Acta, 2012, 727, 54–60.
21. Y. H. Yang, H. F. Zhang, Z. Y. Chu, L. W. Xie and F. Y. Wu, Int. J. Mass Spectrom., 2010, 290, 120–126.
22. C. F. Li, J. H. Guo, Y. H. Yang, Z. Y. Chu and X. C. Wang, J. Anal. At. Spectrom., 2014, 29, 1467–1476.
23. H. L. Lei, T. Yang, S. Y. Jiang and W. Pu, J. Sep. Sci., 2019, 42, 3261–3275.
24. C. F. Li, X. C. Wang, J. H. Guo, Z. Y. Chu and L. J. Feng, J. Anal. At. Spectrom., 2016, 31, 1150–1159.
25. C.-k. Na, T. Nakano, K. Tazawa, M. Sakagawa and T. Ito, Chem. Geol., 1995, 123, 225–237.
26. C. Pin, A. Gannoun and A. Dupont, J. Anal. At. Spectrom., 2014, 29, 1858–1870.
27. J. J. Hui, H. M. Lee, G. E. Kim, W. M. Choi and T. Kim, Separations, 2021, 8, 213.
28. J. Jweda, L. Bolge, C. Class and S. L. Goldstein, Geostand. Geoanal. Res., 2016, 40, 101–115.
29. C. Pin and A. Gannoun, Anal. Chem., 2017, 89, 2411–2417.
30. C. Pin and A. Gannoun, J. Anal. At. Spectrom., 2019, 34, 310–318.
31. C. F. Li, Z. Y. Chu, J. H. Guo, Y. L. Li, Y. H. Yang and X. H. Li, Anal. Methods, 2015, 7, 4793–4802.
32. K. Misawa, F. Yamazaki, N. Ihira and N. Nakamura, Geochem. J., 2000, 34, 11–21.
33. A. Retzmann, T. Zimmermann, D. Profrock, T. Prohaska and J. Irrgeher, Anal. Bioanal. Chem., 2017, 409, 5463–5480.
34. C. F. Li, X. H. Li, Q. L. Li, J. H. Guo, X. H. Li and T. Liu, Anal. Chim. Acta, 2011, 706, 297–304.
35. Y. Wang, X. Huang, Y. Sun, S. Zhao and Y. Yue, Anal. Methods, 2017, 9, 3531–3540.
36. Z. Y. Chu, M. J. Wang, C. F. Li and Y. H. Yang, J. Anal. At. Spectrom., 2019, 34, 2053–2060.
37. Q. Guan, Y. Sun, Y. Yue, X. Liu and S. Zhao, Anal. Sci., 2019, 35, 323–328.
38. D. Wang and R. W. Carlson, J. Anal. At. Spectrom., 2022, 37, 185–193.
39. E. Yobregat, C. Fitoussi and B. Bourdon, J. Anal. At. Spectrom., 2017, 32, 1388–1399.
40. C. Pin and A. Gannoun, J. Anal. At. Spectrom., 2019, 34, 2136–2146.
41. J. M. Koornneef, C. Bouman, J. B. Schwieters and G. R. Davies, J. Anal. At. Spectrom., 2013, 28, 749–754.
42. J. M. Koornneef, C. Bouman, J. B. Schwieters and G. R. Davies, Anal. Chim. Acta, 2014, 819, 49–55.
43. G. Wang, Y. Zeng, J. Xu and W. Liu, J. Mass Spectrom., 2018, 53, 455–464.
44. F. Scheiner, L. Ackerman, K. Holcová, J. Rejšek, H. Vollstaedt, J. Ďurišová and V. Santolík, Geochem., Geophys., Geosyst., 2022, 44, e2021GC010201.
45. G. Wang, H. Vollstaedt, J. Xu and W. Liu, Geostand. Geoanal. Res., 2019, 43, 419–433.
46. J. M. Koornneef, I. Nikogosian, M. J. van Bergen, R. Smeets, C. Bouman and G. R. Davies, Chem. Geol., 2015, 397, 14–23.
47. C. F. Li, Z. Y. Chu, X. C. Wang, J. H. Guo and S. A. Wilde, Anal. Chem., 2019, 91, 7288–7294.
48. W. G. Liu, S. Wei, J. Zhang, C. Ao, F. T. Liu, B. Cai, H. Y. Zhou, J. L. Yang and C. F. Li, J. Anal. At. Spectrom., 2020, 35, 953–960.
49. D. Weis, B. Kieffer, C. Maerschalk, J. Barling, J. de Jong, G. A. Williams, D. Hanano, W. Pretorius, N. Mattielli, J. S. Scoates, A. Goolaerts, R. M. Friedman and J. B. Mahoney, Geochem., Geophys., Geosyst., 2006, 7, Q08006.
50. I. Raczek, K. P. Jochum and A. W. Hofmann, Geostand. Newsl., 2003, 27, 173–179.
51. B. Q. Zhu, Geochemical, 1978, 02, 148–156.
52. M. Boyet and R. W. Carlson, Science, 2005, 309, 576–581.
53. D. Upadhyay, E. E. Scherer and K. Mezger, J. Anal. At. Spectrom., 2008, 23, 561–568.
54. J. O'Neil, H. Rizo, M. Boyet, R. W. Carlson and M. T. Rosing, Earth Planet. Sci. Lett., 2016, 442, 194–205.
55. C. F. Li, X. C. Wang, Y. L. Li, Z. Y. Chu, J. H. Guo and X. H. Li, J. Anal. At. Spectrom., 2015, 30, 895–902.
56. A. Ali and G. Srinivasan, Int. J. Mass Spectrom., 2011, 299, 27–34.
57. R. Andreasen and M. Sharma, Science, 2006, 314, 806–809.
58. R. W. Carlson, M. Boyet and M. Horan, Science, 2007, 316, 1175–1178.
59. Q. R. Shollenberger, L. E. Borg, E. C. Ramon, M. A. Sharp and G. A. Brennecka, Talanta, 2021, 221, 121431.
60. L. E. Nyquist, H. Wiesmann, B. Bansal, C.-Y. Shih, J. E. Keith and C. L. Harper, Geochim. Cosmochim. Acta, 1995, 59, 2817–2837.
61. M. Boyet and R. W. Carlson, Earth Planet. Sci. Lett., 2007, 262, 505–516.
62. I. Raczek, B. Stoll, A. W. Hofmann and K. P. Jochum, Geostand. Newsl., 2001, 25, 77–86.
63. C. Paton, M. Schiller and M. Bizzarro, Astrophys. J., 2013, 763, L40.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: https://doi.org/10.1039/d3ja00252g

---