First attempt to determine oxygen isotopes in oxygen by MC-ICP-MS

Abstract

Oxygen is the key component of crustal and mantle rocks and fluids. The oxygen isotopic composition is a key tool to understand Earth's geological history and processes, such as continental formation, magmatic-hydrothermal processes, and crust–mantle interactions. The oxygen isotopic analysis is commonly implemented by Isotope Ratio Mass Spectrometry (IRMS) and Secondary Ion Mass Spectrometry (SIMS); however, its wide application is limited by the high cost and serious matrix effect. LA-MC-ICP-MS has been the method of choice for in situ isotopic analysis due to its relatively low cost, high analysis speed, high spatial resolution, and the low matrix effect. The determination of oxygen isotope using Ar-ICP has two limitations. Firstly, the exposure of Ar-ICP to the atmosphere may result in atmospheric interference, leading to an increase in the blank of oxygen isotopes and a reduction in the signal-to-blank (S/B) ratio. Secondly, the presence of doubly-charged Ar ions introduces interference that affects the accuracy of oxygen isotopic analysis. In order to investigate whether MC-ICP-MS can be used in the determination of oxygen isotopes, we attempt to use three approaches (¹⁸O¹⁶O/¹⁶O¹⁶O, ¹⁸O/¹⁶O and ¹⁸O¹H₂/¹⁶O¹H₂) to determine oxygen isotopes in oxygen, and the applicability of three approaches is assessed based on interference, peak width, sensitivity, and stability. With our built methods, the obtained long-term productivity of δ¹⁸O measured by ¹⁶O¹⁸O/¹⁶O¹⁶O was greater than 0.16‰ (2 SD), and the measured results for oxygen were consistent with those obtained by IRMS and MC-MIP-MS within the uncertainty limit. This demonstrates the feasibility of our method and also lays the foundation for the realization of in situ oxygen isotope analysis using LA-MC-ICP-MS.

---

1. Introduction

Oxygen is a major component of crustal and mantle rocks and fluids and is the most abundant element in the silicate Earth. Oxygen isotopes are considered an effective means to study geological processes¹ and have been widely used in frontier scientific research fields, such as continental formation and evolution, early Earth history, magmatic-hydrothermal processes and metallogeny, crust–mantle composition and interactions, paleo-oceanic and paleoclimatic changes, and planetary geology and cosmogony.^2–14 Traditionally, two primary methods are commonly applied to analyze the oxygen isotopes: Isotope Ratio Mass Spectrometry (IRMS) and Secondary Ion Mass Spectrometry (SIMS).^15–22

The IRMS analysis method is a bulk analysis method for oxygen isotopes and includes the conventional BrF₅ method^23–25 and laser microprobe BrF₅ method.^26–29 The conventional BrF₅ method is a reliable technique for accurate oxygen isotope analysis of whole rocks and individual minerals, including silicates, phosphates, and sulfates. It typically requires a sample volume of 5–15 mg to react with excess BrF₅ in nickel reaction tubes at specified temperatures and durations. Although the analytical accuracy of this method can reach ±0.05–0.1‰, certain refractory minerals demand higher temperatures and a longer duration for sample melting. For instance, garnet and olivine necessitate a high temperature of 690 °C maintained for 12 hours.^23,25 In contrast, the laser microprobe BrF₅ method replaces external heating of nickel tubes with direct laser heating, which reaches a high temperature of approximately 2000 K. This innovation enables oxygen isotopic analysis of some refractory minerals, significantly reduces sample volume requirements to less than 100 μg and also maintains good analytical precision (±0.05–0.1‰).^29–31 However, the thermal effects occurring during the laser heating process can lead to significant isotopic fractionation, thereby influencing the precision and accuracy of the analysis.^32–36 Furthermore, both methods require the use of strong corrosive reagents, posing potential safety hazards. Compared with the bulk analysis method by IRMS, SIMS can be capable of performing in situ microanalysis of oxygen isotopes, which can offer the high spatial resolution of 5–20 μm beam spot and 1–2 μm depth with high analytical precision (better than 0.4‰).³⁷ However, the analytical results obtained by SIMS are susceptible to a severe matrix effect, i.e., the difference in chemical compositions and morphological characteristics of the standard and sample.^{21,22,37–39} Additionally, the high maintenance and operation cost has also constrained the widespread use of SIMS for in situ analysis of oxygen isotopes,^{15,18,20–22,40} and it has only been applied to the in situ oxygen isotopic analysis of the simple single mineral (e.g., pyroxene, zircon barite and olivine).^22,40–43 In addition, the high sample preparation requirements for SIMS, which necessitate that samples be compatible with high vacuum conditions, can significantly limit its applicability, especially when dealing with loosely structured and volatile samples.⁴⁴ In particular, with the development of modern geosciences, more and more studies have shown that in situ oxygen isotope analysis plays a key role in revealing some major geoscientific issues. For example, in situ microanalysis of oxygen isotopes from ancient zircons, targeting pristine domains within individual crystals to avoid later alteration, indicated the existence of a hydrosphere and water–rock interaction on Earth before >4.1 Ga;^45,46 oxygen isotopic composition in diamonds and their inclusions, specifically the pristine domains within the crystals, illustrated the presence of cryogenically altered oceanic crust in deep continental areas.^5,6,47 High-resolution paleoclimate records obtained through in situ microanalysis of oxygen isotopes can elucidate the relationships between paleoclimate changes, biological explosions, and mass extinctions.^3,7,48 Therefore, it is essential to develop new techniques for in situ oxygen isotope analysis with high spatial resolution, high precision, and high accuracy.

Multi-Collector Inductively Coupled Plasma Mass Spectrometry (MC-ICP-MS) has become the method of choice for analyzing the isotopic composition by virtue of its advantages such as high precision and accuracy, fast analysis speed, a weak matrix effect and low operating cost.^49–54 In particular, since the ICP operates under atmospheric pressure conditions, it can be flexibly switched among multiple sample introduction methods (traditional nebulizers and spray chambers, membrane desolvation,^55,56 laser ablation system^49,57–60 as well as the direct introduction of gaseous samples.^61,62). Among them, LA-MC-ICP-MS, which combines the advantages of in situ sample introduction by a laser ablation system and the high-precision isotopic determination by MC-ICP-MS, has already accurately analyzed more than 23 isotopes.⁵¹ In particular, the successful determination of high-ionization-energy isotopes (e.g., C (10.4 eV),⁴⁹ S (10.0 eV)⁶³ and Cl (12.97 eV)⁵⁰) using LA-MC-ICP-MS has encouraged us to explore the in situ microanalysis of oxygen isotopes. Compared with C, S and Cl, O has a higher ionization energy of 13.6 eV, although it is lower than that of Ar (15.6 eV). Oxygen can be ionized in argon inductively coupled plasma (Ar-ICP), but its ionization efficiency is only about 0.1% based on the Saha equation.⁶⁴ Moreover, two additional challenges exist. First of all, the Ar-ICP is exposed to the atmosphere, and the determination of oxygen isotopes can be interfered with by the oxygen components in the air. Second, as Ar is used as the plasma gas, the precision and accuracy of oxygen isotopic analysis can be severely impacted by the interference from ³⁶Ar²⁺. Consequently, in view of the inherent characteristics of Ar-ICP, no attempts have been made to employ MC-ICP-MS for oxygen isotopic analysis. Therefore, it is necessary to first explore whether the determination of the oxygen isotopic composition can be achieved by MC-ICP-MS, which is the prerequisite for discussing the applicability of LA-MC-ICP-MS in analyzing the oxygen isotope.

In this study, with the application of the simplest oxygen gas as the analyzing sample, we attempted to determine the oxygen isotope composition of oxygen using MC-ICP-MS by employing three methods, ¹⁸O/¹⁶O, ¹⁶O¹⁸O/¹⁶O¹⁶O, and ¹⁸O¹H₂/¹⁶O¹H₂. The ionization efficiency of oxygen in Ar-ICP was assessed by evaluating the sensitivity of all three analytes. Meanwhile, interference and the peak width were analyzed and quantified to assess the impacts of atmospheric oxygen components and argon. From the perspective of isotopic analysis, we evaluated the isotopic analytical reproducibility, the stability of oxygen isotope ratios, as well as the measurement accuracy. Through systematic investigation, ¹⁸O/¹⁶O and ¹⁶O¹⁸O/¹⁶O¹⁶O were selected for oxygen isotopic analysis. Furthermore, the accuracy of the established method was evaluated by comparing the obtained results with those acquired from MAT 253 and MC-MIP-MS. This demonstrated the feasibility of our method and laid the foundation for the realization of in situ oxygen isotope analysis using LA-MC-ICP-MS.

2. Experimental design

A double-focusing Neptune Plus MC-ICP-MS system (Thermo Fisher Scientific, Germany), equipped with seven ion counters and nine Faraday cups, was used at the State Key Laboratory of Geological Processes and Mineral Resources, China University of Geosciences in Wuhan (GPMR). In our experiment, ¹⁸O/¹⁶O, ¹⁸O¹⁶O/¹⁶O¹⁶O and ¹⁸O¹H₂/¹⁶O¹H₂ were used for oxygen isotopic analysis, in which the high-mass particles (¹⁸O, ¹⁸O¹⁶O, and ¹⁸O¹H₂) were measured on an H4 cup and low-mass particles (¹⁶O, ¹⁶O¹⁶O, and ¹⁶O¹H₂) were measured on an L4 cup, and “dummy” masses (m/z = 16.990, 33.030, 19.042) were measured on the center cup. The H4 Faraday cup was equipped with a 10¹¹ Ω resistor, whereas the L4 Faraday cup was equipped with a 3 × 10⁹ Ω resistor due to the high blank intensity of ¹⁶O⁺ and (¹⁶O¹⁶O)⁺. To eliminate the interference of ³⁶Ar²⁺ in the determination of ¹⁸O⁺ and because two adjacent Faraday cups cannot receive both signals simultaneously (the mass difference between ¹⁸O⁺ and ³⁶Ar²⁺ is 0.0154 amu), we built a sub-cup configuration for the determination of ³⁶Ar²⁺, wherein a Faraday cup H4 was used, and a “dummy” mass (m/z = 16.975) was measured in the center cup. In addition to ³⁶Ar²⁺, there are also interferences from ⁴⁰Ar²⁺,³²S⁺ and ³⁴S⁺ in oxygen isotope measurements, thus the high mass resolution (m/Δm = 5500) was applied to eliminate related polyatomic ion interference. Two tanks of oxygen (>99.999%) were used as the bracketing standard and sample, respectively. The O₂ (0.1–2 mL min⁻¹) and He sweeping gas flows were controlled using a mass flow meter (MFC). High-purity He (99.999%) was used as a sweeping gas to clean residual oxygen from pipelines. The schematic of the MC-ICP-MS inlet system in the oxygen isotopic analysis can be seen in Fig. 1.

Fig. 1 Schematic of the MC-ICP-MS inlet system in the oxygen isotopic analysis.

The oxygen isotope composition of the oxygen was determined using the sample-standard bracketing method (SSB), a technique designed to correct for mass bias and instrumental drift during isotopic analysis. To minimize the influence of blank on the isotopic measurements,⁶⁵ the on-peak blank was analyzed prior to each sample and standard measurement, utilizing a block of 15 cycles with an integration time of 4.194 seconds. Analysis of each sample and standard was conducted by one block of 30 cycles, with the integration time of 4.194 s. A repeated analytical sequence of “blank, O₂-std, blank, O₂-sample, blank, O₂-std, blank…” was conducted. And offline calculations of the blank and analyte signals, time-drift correction, and isotopic analysis calibration were performed using Iso-Compass software.⁶⁶ To verify the accuracy of the proposed method, an IRMS (Thermo Scientific™ MAT253 Plus™) at GPMR was used as a comparative method to analyze oxygen isotopes, and the instrumental operating parameters for MC-ICP-MS and IRMS are summarized in Table 1. The oxygen isotopes of the sample are expressed as δ¹⁸O, which can be calculated using eqn (1)–(4).

r = (¹⁶O¹⁸O)⁺/(¹⁶O¹⁶O)⁺ (1)

r′ = ¹⁸O/¹⁶O = r/(2 + r) (2)

r′ = ¹⁸O/¹⁶O = (¹⁸O¹H¹H)⁺/(¹⁶O¹H¹H)⁺ (3)

where r is the measured ¹⁶O¹⁸O/¹⁶O¹⁶O ratio. r′ is the measured ¹⁸O/¹⁶O ratio (when ¹⁸O/¹⁶O was measured) or the ¹⁸O/¹⁶O ratio when converted from ¹⁶O¹⁸O/¹⁶O¹⁶O (when ¹⁶O¹⁸O/¹⁶O¹⁶O was measured) or the ¹⁸O/¹⁶O ratio when converted from ¹⁸O¹H¹H/¹⁶O¹H¹H (when ¹⁸O¹H¹H/¹⁶O¹H¹H was measured), is the measured ¹⁸O/¹⁶O ratio of sample oxygen, and is the measured ¹⁸O/¹⁶O ratio of standard oxygen. δ¹⁸O is the oxygen composition of the analyzed sample.

Table 1 Summary of the operating parameters for MC-ICP-MS

Instruments	Analyzed method	Operating conditions
Neptune Plus MC-MS system
Cup configuration	^18O^16O/^16O^16O	L4			C			H4
		^16O^16O			33.03			^16O^18O
	^18O^1H^1H/^16O^1H^1H	^16O^1H^1H			19.042		^18O^1H^1H
	^18O/^16O	^16O			16.990			^18O
Sub-cup configuration	^36Ar^2+				16.975			^36Ar^2+
Resistor of the Faraday cup		3 × 10^9 Ω			10^11 Ω			10^11 Ω
Cool gas flow rate		15 L min^−1
Aux gas flow rate		0.95 L min^−1
Sample gas flow rate		1.12–1.45 L min^−1
RF power		1050 W
Guard electrode (GE)		On
Extraction		−2000 V
Focus		−676 V
X-Defl		0.27 V
Y-Defl		−3.48 V
Shape		248 V
Rot quad 1		0.01 V
Source offset		−1.00 V
Foc quad 1		−19.89 V
Rot quad 2		0.00 V
Focus offset		50.00 V
Matsuda plate		0.01 V
Focus quad		−8.00 V
Dispersion quad		−1.00 V
Interface cones		Jet sample cone + X skimmer cone
Mass resolution		High (m/Δm = 5500)
Block × cycle		1 × 30 (sample); 1 × 15 (blank)
Integration time		4.194 s
MAT253 Plus IRMS
Cup configuration	^18O^16O/^16O^16O		C1	C2		C3
			^16O^16O	^16O^17O		^16O^18O
Resistor of the Faraday cup			3 × 10^8 Ω	3 × 10^11 Ω		1 × 10^11 Ω
High voltage			9.450 KV
Emission			1.2 mA
Electron energy			108.926 V
Trap			22.6 V

---

In our experiment, we integrated the existing oxygen isotopic analysis methods of SIMS, which determines negative ions such as ¹⁸O⁻/¹⁶O⁻, as well as IRMS, which analyzes positive oxygen molecules like ¹⁶O¹⁸O⁺/¹⁶O¹⁶O⁺, and combined the determination methods of other isotopes by MC-ICP-MS, including positive ions like ¹³C⁺/¹²C⁺, ³²S⁺/³⁴S⁺, and ³⁷Cl⁺/³⁵Cl⁺. Based on these, a comprehensive evaluation of three possible methods (¹⁸O/¹⁶O, ¹⁸O¹⁶O/¹⁶O¹⁶O and ¹⁸O¹H₂/¹⁶O¹H₂) was conducted to explore the feasibility of using MC-ICP-MS for oxygen isotopic analysis.

3. Results and discussion

Given that oxygen isotopic analysis via MC-ICP-MS has never been explored before, we referred to the determination methods of other isotopes using MC-ICP-MS (determination of positive ions, e.g., ¹³C⁺/¹²C⁺,⁴⁹³²S⁺/³⁴S⁺,^63,67 and ³⁷Cl⁺/³⁵Cl⁺)⁵⁰ as well as SIMS (determination of negative ions, ¹⁸O⁻/¹⁶O⁻)²² and IRMS (determination of positive oxygen molecules, ¹⁶O¹⁸O⁺/¹⁶O¹⁶O⁺).⁶⁸ Taking into account the high ionization capacity of ICP as well as the lens parameter in our instrument (with the extraction lens set at −2000 V), we opted to select ¹⁸O⁺–¹⁶O⁺ and ¹⁶O¹⁸O⁺–¹⁶O¹⁶O⁺ for oxygen isotopic analysis. Additionally, during the process of scanning the blank, we also detected H₂O⁺. Hence, currently, there are three analytes available for oxygen isotopic analysis. To explore the feasibility of MC-ICP-MS to determine oxygen isotopes, a comprehensive evaluation of three possible methods was conducted, including the interference, sensitivity, dynamic linear range, stability of oxygen isotope ratios, analytical precision and accuracy of oxygen isotope ratios.

3.1 Interference on oxygen ions

The peak shapes of ¹⁸O–¹⁶O, ¹⁶O¹⁸O–¹⁶O¹⁶O and ¹⁸O¹H₂–¹⁶O¹H₂ as well as the related interferences are scanned at an O₂ flow rate of 0 mL min⁻¹ with a high mass resolution (m/Δm = 5000) (Fig. 2). For the measurement of ¹⁸O/¹⁶O, ¹⁶O¹⁸O/¹⁶O¹⁶O, and ¹⁸O¹H₂/¹⁶O¹H₂, the main interferences are listed in Table 2. It can be seen that the determination of oxygen isotopes is susceptible to the interference of the plasma gas (Ar²⁺) (Fig. 2b and e), S blank (Fig. 2c) and the atmosphere (O–H related interference) (Fig. 2a and f). Therefore, it is necessary to select the appropriate peak center to avoid the interference and obtain the true oxygen signal intensity for oxygen isotope determination.

Fig. 2 Peak shapes of ¹⁶O⁺–¹⁸O⁺ (a), (¹⁶O¹⁶O)⁺–(¹⁶O¹⁸O)⁺ (c) and (¹⁶O¹H¹H)⁺–(¹⁸O¹H¹H)⁺ (e) with high mass resolution in MC-ICP-MS. In the legend of “18Mass × 420”, “34Mass × 200’’ and “20Mass × 500”, 420-, 200- and 500-folds were used to obtain the signal intensities of 18Mass, 34Mass and 20Mass, consistent with those of 16Mass, 18Mass and 32Mass to check the simultaneous collection of the two peaks. Part figures (a, c and e) in the dashed box are enlarged to locate more suitable peak centers (mass = 16.990, 33.030 and 19.042) (b, d and f).

Table 2 Data for analyzed isotopes, related interfering ions, and mass resolutions

Analyzed isotopes	Mass	Interfering ions	Mass	Required mass resolution	Peak width (amu)	Blank (V)	Signal-to-blank ratio	Sensitivity (V ml^−1 min^−1)
^16O	15.9949	^14N^1H2^+	16.0187	672	0.008	65.80 (^16O)	22.49	2430
^18O	17.9992	^36Ar^2+	17.9838	1170
		^16O^1H2^+	18.0106	1578
		^14N^1H4^+	18.0344	511
^16O^16O	31.9898	^32S^+	31.9721	1807	0.015	0.80 (^16O^16O)	1585	719
		^14N^18O	32.0022	2579
		^14N^16O^1H2^+	32.0136	1344
^16O^18O	33.9941	^34S^+	33.9679	1297
		^16O^16O^1H2^+	34.0055	2981
^16O^1H2	18.0107	^36Ar^2+	17.9838	670	0.008	0.43 (^16O^1H2)	32.39	7
		^18O^+	17.9992	1566
		^14N^1H4^+	18.0344	760
^18O^1H2	20.0132	^40Ar^2+	19.9812	625

---

In terms of ¹⁸O–¹⁶O, ³⁶Ar²⁺ and ¹⁶O¹H₂⁺ are two main interferences. The doubly-charged interference of ³⁶Ar²⁺ is located on the low-mass side and the polyatomic ionic interference of ¹⁶O¹H₂⁺ is located on the high-mass side. The interference of ¹⁶O¹H₂⁺ on ¹⁸O⁺ could be isolated as the mass center of the peak was chosen at the mass side of 16.990 with the high mass resolution (Fig. 2a and b). And for ³⁶Ar²⁺ (the mass difference between ¹⁸O⁺ and ³⁶Ar²⁺ is 0.0154 amu, and two nearby Faraday cups cannot receive both signals at the same time), a sub-cup configuration was built, and the ³⁶Ar²⁺ signal can be directly acquired through peak jumping and deducted accurately. As for ¹⁶O¹⁸O–¹⁶O¹⁶O, ³²S⁺ and ³⁴S⁺ are two main interferences, and fortunately both interferences are located on the low-mass side; thus, a mass center of 33.03 can be selected to avoid the interferences of ³²S⁺ and ³⁴S⁺ (Fig. 2c and d). For the determination of ¹⁸O¹H₂–¹⁶O¹H₂, the main interferences are ³⁶Ar²⁺, ¹⁸O⁺ and ⁴⁰Ar²⁺. Three interferences are all located on the low mass side, and a mass center of 19.042 was selected to avoid the interferences. Notably, the applied peak width of ¹⁶O¹⁸O/¹⁶O¹⁶O was approximately 0.015 amu (Fig. 2f), whereas those of ¹⁸O/¹⁶O and ¹⁸O¹H₂/¹⁶O¹H₂ were 0.008 amu (Fig. 2b and d). Therefore, ¹⁶O¹⁸O/¹⁶O¹⁶O was estimated to be the preferred method in view of the peak width of the mass shoulder, which is crucial for accurate isotope measurement as it allows for better resolution and less interference from adjacent peaks.^69,70

3.2 Sensitivity and dynamic linear range of oxygen ions

The Saha equation predicts a low ionization rate (approximately 0.1%) for oxygen in an Ar-ICP,⁶⁴ making high sensitivity a critical factor for achieving precise and accurate oxygen isotopic analysis. Besides, as the Ar-ICP is exposed to the atmosphere, the oxygen components in air generate a significantly high oxygen blank during isotopic analysis, which affects the accuracy of oxygen isotope measurements. Consequently, a comprehensive assessment of the sensitivity, blank levels, and signal-to-blank (S/B) ratios of the two methods is essential to ascertain their influence on the reliability of isotopic analysis.

The blank and sensitivity (sensitivity was obtained by the ratio of net signal intensity and oxygen gas flow rate) of ¹⁶O and ¹⁶O¹⁶O were examined as the O₂ flow rate increased from 0 to 2 mL min⁻¹ (Fig. 3). For the measurement of ¹⁸O/¹⁶O, the ¹⁶O blank can be as high as 64 V, which may be due to the exposure of the ICP in the atmosphere.⁷¹ At the sample gas flow rate of 1.29 L min⁻¹ and RF power of 1050 W, the ¹⁶O intensity increased as the O₂ flow rate increased from 0 to 0.5 mL min⁻¹, as demonstrated by the linear function of intensity versus the O₂ flow rate. The slope of the linear regression relationship was 2430, which can be demonstrated as sensitivity (i.e., ∼2430 V mL⁻¹ min⁻¹). However, beyond this flow rate, the ¹⁶O signal intensity will exceed the Faraday cup's threshold of 1666 V when using a 3 × 10⁹ Ω resistor. Thus, the applied O₂ flow rate was 0.5 mL min⁻¹, and the S/B ratio of this method was 22.49. For the analyzed method of ¹⁶O¹⁸O/¹⁶O¹⁶O, at the sample gas flow rate of 1.254 L min⁻¹ and RF power of 1050 W, the ¹⁶O¹⁶O intensity increased linearly with the O₂ flow rate of up to 1.8 mL min⁻¹, beyond which a nonlinear relationship was observed. The slopes of the linear regression relationship were 718, which can be demonstrated as the sensitivity (i.e., ∼680 V mL⁻¹ min⁻¹). This nonlinearity may stem from incomplete ionization of the O₂ sample under the current conditions.^72,73 To reduce the isotopic fractionation and obtain the best precision, the O₂ flow rate of 1.6 mL min⁻¹ was selected for the measurement of ¹⁸O¹⁶O/¹⁶O¹⁶O with the S/B ratio of this method being 1585.25.

Fig. 3 Relationship between the O₂ flow rate and the signal intensity (a) as well as the sensitivity (b) of ¹⁶O and (¹⁶O¹⁶O).

In order to select an appropriate analysis method, we proceeded with two perspectives: sensitivity and S/B ratio. Based on our experiment, ¹⁸O/¹⁶O is the preferred option in terms of sensitivity (2430 V ml⁻¹ min⁻¹). This observation is somewhat surprising, the first ionization energy of an O atom (13.62 eV) is higher than that of ¹⁶O¹⁶O (12.07 eV), which suggests that the sensitivity of ¹⁶O¹⁶O should be greater than that of ¹⁶O theoretically. The high sensitivity of ¹⁶O may be attributed to the high gas temperature of the ICP ion source, which can lead to the dissociation of most O₂ into O atoms, leaving a small fraction of undissociated O₂ molecules. However, from the perspective of S/B ratio, the S/N ratio of ¹⁶O (22.49) was 750 times lower than that of ¹⁶O¹⁶O (1585.25). Therefore, when considering the two aspects of sensitivity and S/N ratio, the choices of optimal methods to be tested are contradictory. To explore the interaction between the S/B ratio and signal intensity in isotopic analysis, a simulation experiment was performed. Two simulated datasets were created: one with low signal intensity but high S/B ratios and the other with high signal intensity but low S/B ratios, while keeping the isotope composition of blank (−12‰) and sample (0.5‰) constant. For each dataset, the blank intensity was fixed (0.005 V vs. 0.2 V), and the signal intensity (1–8 V vs. 21–28 V) was varied to achieve different S/B ratios (200–1600 vs. 105–140). Isotopic compositions were calculated without blank correction, and the relative error (RE) between the calculated and true sample compositions was assessed. Fig. 4a and b shows that for the group with a high signal intensity but low S/B ratio, the blank intensity had a more significant impact on isotopic measurements. For example, even at a signal intensity of 28 V with a low S/B ratio (140), the effect of blank on isotopic composition was considerable, with an RE as high as −29.94%. Conversely, for the high S/B ratio group, where the S/B ratio reached 1000 at a low signal intensity of 5 V, the effect of blank on isotopic composition (RE) was less than 2.5%. Here, despite an 82.1% reduction in signal intensity (from 28 V to 5 V) compared to the high signal intensity group, the effect of the blank can be reduced by 91.6%. Fig. 4c and d further examines the effect of 1% fluctuation in blank intensity. The results indicated that such fluctuations affected the high signal intensity but a low S/B ratio group more than the low signal intensity but a high S/B ratio group. At an S/B ratio of 1000, a 1% blank fluctuation caused only a 0.03% change in the calculated isotopic composition at a signal intensity of 5 V. However, for the high signal intensity group, the same fluctuation led to a 1.01% change in isotopic composition at 28 V due to the low S/B ratio (140).

Fig. 4 (a) and (b) Relationship between the signal intensity, S/B ratio and relative error of the two simulated datasets; the relative error was calculated between the calculated and true sample compositions. (c) and (d) Relationship between the signal intensity, S/B ratio and relative deviation of the two simulated datasets; the relative deviation was assessed by the sample composition before and after 1% blank fluctuation.

These simulations highlight the crucial balance between the signal intensity and S/B ratio in isotope analysis. The higher S/B ratios enhance the accuracy of isotope ratio measurements by reducing blank, which is especially vital in laser-based in situ microanalysis of samples with low elemental concentrations and low signal intensities. Achieving high S/B ratios is essential for reliable measurements in such scenarios. In summary, accurate determination of isotopic compositions necessitates not only sufficient signal intensity but also optimized S/B ratios to ensure high precision and accuracy across a variety of analytical conditions. In this context, ¹⁸O¹⁶O/¹⁶O¹⁶O, which exhibits a higher S/B ratio, is more suitable as the analyte of analysis. Additionally, the peak width of ¹⁸O¹⁶O/¹⁶O¹⁶O (0.015 amu) is wider than that of ¹⁸O/¹⁶O (0.008 amu) and the blank intensity of ¹⁶O¹⁶O (0.8 V) is lower than that of ¹⁶O (65.8 V). Therefore, considering the S/B ratio, peak width and blank, the use of ¹⁸O¹⁶O/¹⁶O¹⁶O is anticipated to be a more favorable choice for isotopic analysis.

3.3 Effect of sample gas and sampling depth on oxygen isotopic analysis

Accurate calibration of isotopic mass discrimination is essential for achieving precise isotope ratios with MC-ICP-MS.⁵³ For the oxygen isotope, the lighter weight and lack of stable isotope pairs as internal standards determine that the standard-sample bracketing (SSB) method can only be used. This method is simple and convenient to operate; however, a higher level of stability for the instruments is required because the fractionation factor of standard is assumed to be that of sample with this method.⁷⁴ The region of stable O isotopic composition can be explored by studying the radial and axial distributions of O ions in the ICP, which can be confirmed by determining the variation of O signal intensity and ¹⁸O/¹⁶O ratio with respect to the position of the torch tube (radial and axial directions) and the sample gas flow rate. It has been established that the radial position of the torch that provides the maximum analyte intensity also results in the most stable isotopic ratios.^75,76 In contrast, the axial position of the torch and sample gas flow rate may not be optimal for both the maximum signal intensity and stable isotopic ratios.⁵² The axial distribution of ions in the ICP can be achieved by changing the position of the Z-axis of the torch tube (Fig. 5a and c) or the sample gas flow rate (Fig. 5b and d). The Z-axis position is the distance from the outer end of the torch tube to the hole of the sampling cone. A more negative Z-axis value indicates the plasma is closer to the sample cone. The sample gas flow rate was also manipulated to investigate the effect on ion distribution at the axial position, as a higher flow rate means that the plasma is closer to the sample cone.

Fig. 5 Relationship between the Z position (a and c) and sample gas flow rate (b and d) of (¹⁶O)⁺, (¹⁶O¹⁶O)⁺ signal intensity and oxygen isotope ratio (¹⁸O/¹⁶O).

For investigation of the Z-axis position, as the Z-axis position varied from −1.7 mm to 2.0 mm, with the sample gas flow rate remaining constant, the signal intensity of the two analytes initially increased and then decreased, following a similar trend. Comparing the stability of the obtained ¹⁸O/¹⁶O ratio within ±10% of the highest signal intensity, the stability of the ¹⁸O/¹⁶O ratio obtained by measuring ¹⁸O¹⁶O/¹⁶O¹⁶O (0.002291 ± 0.000001, 2SD) is about 20 times higher than that of ¹⁸O/¹⁶O (0.002222 ± 0.000023, 2SD). To further explore the effect of the sample gas flow rate, the Z-position was kept constant, and as the sample gas flow rate increased from 1.0 L min⁻¹ to 1.25 L min⁻¹, more ions were extracted into the mass spectrometer, leading to an increase in O intensity. However, the amount of ions extracted is limited by the sample cone diameter⁷⁷ and the temperature of the ICP;^78,79 thus, the O signal will not continue to rise indefinitely as the carrier gas flow rate increases. In terms of isotope ratio, the obtained ¹⁸O/¹⁶O by measuring ¹⁸O¹⁶O/¹⁶O¹⁶O is more stable (0.002289 ± 0.000002, 2SD) than that obtained by measuring ¹⁸O/¹⁶O (0.002251 ± 0.000019, 2SD). Additionally, both the optimum sampling depth and sample gas flow rate for the maximum signal intensity of ¹⁶O¹⁶O can be aligned with the stable area of the isotope ratio, which is superior for tuning and isotopic analysis. Therefore, in terms of isotope ratio stability, compared with the ¹⁸O/¹⁶O ratio, the measured ¹⁸O¹⁶O/¹⁶O¹⁶O ratio is the preferred choice. And, in practical oxygen isotopic analysis, it is advisable to adjust the sample gas flow rate and the Z-axis position to where the maximum signal intensity is achieved.

3.4 Precision and accuracy of oxygen isotopes in oxygen

Analytical precision is dependent on the signal intensity,^52,80 as demonstrated by the exponential function of the precision for the ¹⁸O/¹⁶O isotopic ratios (1 SE) vs. the ¹⁶O⁺ and (¹⁶O¹⁶O)⁺intensities (Fig. 6a). The internal precision of ¹⁸O/¹⁶O was low when the (¹⁶O¹⁶O)⁺ and ¹⁶O⁺ signal intensities were less than 100 V. As the signal intensity increases from 100 to 1000 V, there is a rapid improvement in the within-run precision. Therefore, using (¹⁶O¹⁶O)⁺ and ¹⁶O⁺ signal intensities of more than 1000 V, where the internal precision of ¹⁸O/¹⁶O was better than 1 × 10⁷, a 3 × 10⁹ Ω resistor was recommended. To obtain long-term reproducibility, the δ¹⁸O value was repeatedly measured by analyzing ¹⁶O¹⁸O/¹⁶O¹⁶O and ¹⁸O/¹⁶O with an optimum O₂ flow rate, respectively. The δ¹⁸O value was repeatedly measured by analyzing ¹⁶O¹⁸O/¹⁶O¹⁶O and ¹⁸O/¹⁶O with O₂ flow rates of 1.6 and 0.5 mL min⁻¹, sample gas flow rates of 1.254 and 1.285 L min⁻¹, and Z-axis positions of 0.1 and −0.6 mm, respectively, and an RF power of 1050 W. The obtained δ¹⁸O values were 0.05 ± 0.14‰ (2 SD, n = 50) and 0.07 ± 0.17‰ (2 SD, n = 50) for ¹⁶O¹⁸O/¹⁶O¹⁶O and ¹⁸O/¹⁶O, respectively (Fig. 6b), and ¹⁸O¹⁶O/¹⁶O¹⁶O was preferred.

Fig. 6 (a) Relationship between the signal intensity of ¹⁶O, (¹⁶O¹⁶O) and the internal precision (1 SE) of ¹⁸O/¹⁶O for oxygen isotopic analysis. (b) Long-term reproducibility of δ¹⁸O by measuring ¹⁶O¹⁸O/¹⁶O¹⁶O and ¹⁸O/¹⁶O.

For the assessment of measurement accuracy of δ¹⁸O using the proposed method, the oxygen isotopes of two different oxygen tanks were analyzed. One oxygen tank was used as the bracketing standard and the other one as the sample. The δ¹⁸O values measured by ¹⁸O and ¹⁶O were determined at a sample gas flow rate of 1.285 L min⁻¹, an oxygen flow rate of 0.5 ml min⁻¹, and a Z-axis position of −0.6 mm. And the δ¹⁸O values measured by ¹⁸O¹⁶O and ¹⁶O¹⁶O were determined at a sample gas flow rate of 1.254 L min⁻¹, an oxygen flow rate of 1.6 ml min⁻¹, and a Z-axis position of 0.1 mm. The δ¹⁸O values of sample-1 were −0.03 ± 0.22‰ (2 SD, n = 10, measured by ¹⁸O and ¹⁶O) and 0.16 ± 0.16‰ (2 SD, n = 10, measured by ¹⁸O¹⁶O and ¹⁶O¹⁶O). In terms of test accuracy, the accuracy of using ¹⁸O¹⁶O/¹⁸O¹⁶O to determine the δ¹⁸O value in oxygen was preferred. The obtained values were consistent with those measured according to MAT 253 (0.13 ± 0.1‰; 2 SD, n = 10) and MC-MIP-MS (0.14 ± 0.22‰; 2 SD, n = 6),⁸¹ as shown in Fig. 7, which substantiated the validity of our method.

Fig. 7 Measured δ¹⁸O values of oxygen samples using MC-ICP-MS, IR-MS and MC-MIP-MS.

4. Conclusions

Due to the interference from Ar plasma gas, high blank intensity of O from the atmosphere, and the high first ionization energy, no attempt has been made to determine oxygen isotopes by MC-ICP-MS. However, high mass resolution and sub-cup configuration can be used to eliminate the interference of ³⁶Ar²⁺ on ¹⁸O⁺; the application of a 3 × 10⁹ Ω amplifier can be used to obtain the true signal intensity of O blank. Additionally, the applied ¹⁸O¹⁶O/¹⁶O¹⁶O method can be used to overcome the interference of Ar²⁺ and the ultra-high blank intensity of ¹⁶O. Based on the above improvement measures, the obtained long-term reproducibility of δ¹⁸O measured by ¹⁸O¹⁶O/¹⁶O¹⁶O was better than 0.16‰ (2 SD). The results obtained by MC-ICP-MS were consistent with those measured by IRMS and MC-MIP-MS, which demonstrated that MC-ICP-MS can be used for the accurate determination of oxygen isotopes. It is worth noting that the obtained precision by MC-ICP-MS was worse than that obtained by IRMS, which may be attributed to the high blank. Fortunately, with the use of a shielding gas to maintain positive pressure, preventing atmospheric air from entering the ICP area has been proven to reduce interferences caused by gases such as H, C, N, and O.⁸² Using the Atmospheric Interface for Reduced Dust (AIRD) device, optimized through simulations, significantly enhances the analytical capabilities of ICP-MS and effectively manages atmospheric interference. Another reason may be the low ionization rate of O in the Ar-ICP, so the stable He-ICP with high ionization ability may be promising. Therefore, there is still a long way to go for in situ microanalysis of oxygen isotopes using LA-MC-ICP-MS. However, we have already taken the first step to evaluate the capability of MC-ICP-MS to analyze oxygen. Further work will be continued to achieve this goal.

Data availability

The authors declare that the data supporting the findings of this study are available within the paper. Should any raw data files be needed in another format they are available from the corresponding author upon reasonable request. Source data are provided with this paper.

Conflicts of interest

There are no conflicts to declare.

Acknowledgements

This work was funded by the National Natural Science Foundation of China (41927803 and 42473038), MOST (Ministry of Science and Technology) Special Fund from the State Key Laboratory of Geological Processes and Mineral Resources (MSFGPMR01), and the Fundamental Research Funds for the Central Universities, China University of Geosciences (Wuhan).

References

1. J. Hoefs, Stable Isotope Geochemistry, Springer International Publishing, Cham, 2015.
2. B. Dhuime, C. J. Hawkesworth, P. A. Cawood and C. D. Storey, Science, 2012, 335, 1334–1336.
3. J. A. Trotter, I. S. Williams, C. R. Barnes, C. Lécuyer and R. S. Nicoll, Science, 2008, 321, 550–554.
4. C. J. Hawkesworth and A. I. S. Kemp, Nature, 2006, 443, 811–817.
5. R. B. Ickert, T. Stachel, R. A. Stern and J. W. Harris, Geochem. Perspect. Lett., 2015, 65–74.
6. D. J. Schulze, B. Harte, J. W. Valley, J. M. Brenan and D. M. D. R. Channer, Nature, 2003, 423, 68–70.
7. J. A. Trotter, I. S. Williams, A. Nicora, M. Mazza and M. Rigo, Earth Planet. Sci. Lett., 2015, 415, 165–174.
8. U. Ryb and J. M. Eiler, Proc. Natl. Acad. Sci. U. S. A., 2018, 115, 6602–6607.
9. I. N. Bindeman, A. Bekker and D. O. Zakharov, Earth Planet. Sci. Lett., 2016, 437, 101–113.
10. U. Wiechert, A. N. Halliday, D.-C. Lee, G. A. Snyder, L. A. Taylor and D. Rumble, Science, 2001, 294, 345–348.
11. R. C. Greenwood, J.-A. Barrat, M. F. Miller, M. Anand, N. Dauphas, I. A. Franchi, P. Sillard and N. A. Starkey, Sci. Adv., 2018, 4, eaao5928.
12. I. N. Bindeman, D. O. Zakharov, J. Palandri, N. D. Greber, N. Dauphas, G. J. Retallack, A. Hofmann, J. S. Lackey and A. Bekker, Nature, 2018, 557, 545–548.
13. Y.-F. Zheng, B. Fu, B. Gong and L. Li, Earth-Sci. Rev., 2003, 62, 105–161.
14. G. A. Henkes, B. H. Passey, E. L. Grossman, B. J. Shenton, T. E. Yancey and A. Pérez-Huerta, Earth Planet. Sci. Lett., 2018, 490, 40–50.
15. Q.-L. Li, X.-H. Li, Y. Liu, G.-Q. Tang, J.-H. Yang and W.-G. Zhu, J. Anal. At. Spectrom., 2010, 25, 1107–1113.
16. D. Rubatto, B. Putlitz, L. Gauthiez-Putallaz, C. Crépisson, I. S. Buick and Y.-F. Zheng, Chem. Geol., 2014, 380, 84–96.
17. D. Liu, S. A. Wilde, Y. Wan, S. Wang, J. W. Valley, N. Kita, C. Dong, H. Xie, C. Yang, Y. Zhang and L. Gao, Chem. Geol., 2009, 261, 140–154.
18. Y. Sun, M. Wiedenbeck, M. M. Joachimski, C. Beier, F. Kemner and C. Weinzierl, Chem. Geol., 2016, 440, 164–178.
19. A. K. Schmitt, M.-C. Liu and I. E. Kohl, J. Anal. At. Spectrom., 2019, 34, 561–569.
20. N. M. Ellis and B. H. Passey, Chem. Geol., 2023, 635, 121616.
21. S. Schmidt, A. Hertwig, A. K. Schmitt, K. Cionoiu, K. D. McKeegan, I. Bindeman, T. D. Rocco and A. Pack, J. Anal. At. Spectrom., 2024, 39, 439–446.
22. B. Li, M. Wiedenbeck, F. Couffignal, A. M. Álvarez-Valero, H.-M. Bao, C.-F. Fan, J. Han, G.-S. Jin, Y.-B. Peng, M. D. Syczewski, K. T. Tait, F. D. H. Wilke and U. G. Wortmann, Geostand. Geoanal. Res., 2024, 48, 179–205.
23. R. N. Clayton and T. K. Mayeda, Geochim. Cosmochim. Acta, 1963, 27, 43–52.
24. G. D. Garlick and S. Epstein, Geochim. Cosmochim. Acta, 1967, 31, 181–214.
25. H. P. Taylor Jr and S. Epstein, Geol. Soc. Am. Bull., 1962, 73, 461–480.
26. D. Elsenheimer and J. W. Valley, Chem. Geol. Isot. Geosci. Sect., 1992, 101, 21–42.
27. U. Wiechert and J. Hoefs, Geochim. Cosmochim. Acta, 1995, 59, 4093–4101.
28. Z. D. Sharp, Geochim. Cosmochim. Acta, 1990, 54, 1353–1357.
29. D. Elsenheimer and J. W. Valley, Geochim. Cosmochim. Acta, 1993, 57, 3669–3676.
30. D. Mattey, D. Lowry and C. Macpherson, Earth Planet. Sci. Lett., 1994, 128, 231–241.
31. J. W. Valley, J. R. Chiarenzelli and J. M. McLelland, Earth Planet. Sci. Lett., 1994, 126, 187–206.
32. S. M. Eggins, R. L. Rudnick and W. F. McDonough, Earth Planet. Sci. Lett., 1998, 154, 53–71.
33. Z. Li, Z. Hu, Y. Liu, S. Gao, M. Li, K. Zong, H. Chen and S. Hu, Chem. Geol., 2015, 400, 11–23.
34. H.-R. Kuhn and D. Günther, J. Anal. At. Spectrom., 2004, 19, 1158–1164.
35. M. Guillong, H.-R. Kuhn and D. Günther, Spectrochim. Acta, Part B, 2003, 58, 211–220.
36. Z. Chen, J. Anal. At. Spectrom., 1999, 14, 1823–1828.
37. N. T. Kita, T. Ushikubo, B. Fu and J. W. Valley, Chem. Geol., 2009, 264, 43–57.
38. G.-Q. Tang, X.-H. Li, Q.-L. Li, Y. Liu, X.-X. Ling and Q.-Z. Yin, J. Anal. At. Spectrom., 2015, 30, 950–956.
39. M. J. Whitehouse and A. A. Nemchin, Chem. Geol., 2009, 261, 32–42.
40. Q. Yang, X. Xia, W. Zhang, Y. Zhang, B. Xiong, Y. Xu, Q. Wang and G. Wei, Solid Earth Sci., 2018, 3, 81–86.
41. J. Isa, I. E. Kohl, M.-C. Liu, J. T. Wasson, E. D. Young and K. D. McKeegan, Chem. Geol., 2017, 458, 14–21.
42. B. Peng, M. He, M. Yang and Y. Shi, Crystals, 2023, 13, 987.
43. F. M. Deegan, M. J. Whitehouse, V. R. Troll, D. A. Budd, C. Harris, H. Geiger and U. Hålenius, Chem. Geol., 2016, 447, 1–10.
44. B. De Samber, R. De Rycke, M. De Bruyne, M. Kienhuis, L. Sandblad, S. Bohic, P. Cloetens, C. Urban, L. Polerecky and L. Vincze, Anal. Chim. Acta, 2020, 1106, 22–32.
45. S. J. Mojzsis, T. M. Harrison and R. T. Pidgeon, Nature, 2001, 409, 178–181.
46. D. Trail, P. Boehnke, P. S. Savage, M.-C. Liu, M. L. Miller and I. Bindeman, Proc. Natl. Acad. Sci. U. S. A., 2018, 115, 10287–10292.
47. A. D. Burnham, A. R. Thomson, G. P. Bulanova, S. C. Kohn, C. B. Smith and M. J. Walter, Earth Planet. Sci. Lett., 2015, 432, 374–380.
48. C. Huang, M. M. Joachimski and Y. Gong, Earth Planet. Sci. Lett., 2018, 495, 174–184.
49. W. Chen, J. Lu, S.-Y. Jiang, K.-D. Zhao and D.-F. Duan, Anal. Chem., 2017, 89, 13415–13421.
50. C. Toyama, J.-I. Kimura, Q. Chang, B. S. Vaglarov and J. Kuroda, J. Anal. At. Spectrom., 2015, 30, 2194–2207.
51. J. Lin, A. Yang, R. Lin, J. Mao, Z. Hu and Y. Liu, J. Earth Sci., 2023, 34, 1663–1691.
52. J. Lin, Y. Liu, X. Tong, L. Zhu, W. Zhang and Z. Hu, J. Anal. At. Spectrom., 2017, 32, 834–842.
53. L. Yang, Mass Spectrom. Rev., 2009, 28, 990–1011.
54. V. Balaram, W. Rahaman and P. Roy, Geosystems Geoenvironment, 2022, 1, 100019.
55. M.-Y. He, L. Deng, J. Liu, Z. dong Jin and T. Ren, RSC Adv., 2023, 13, 32104–32109.
56. T.-L. Yu, B.-S. Wang, C.-C. Shen, P.-L. Wang, T. F. Yang, G. S. Burr and Y.-G. Chen, Anal. Chim. Acta, 2017, 988, 34–40.
57. Z. Bao, C. Zong, L. Chen, D. Lei, K. Chen and H. Yuan, J. Anal. At. Spectrom., 2018, 33, 2143–2152.
58. W. Zhang, Z. Hu, Y. Liu, L. Feng and H. Jiang, Chem. Geol., 2019, 522, 16–25.
59. J. Lin, Y. Liu, Z. Hu, W. Chen, C. Zhang, K. Zhao and X. Jin, J. Anal. At. Spectrom., 2019, 34, 1145–1153.
60. W. Zhang, Z. Wang, F. Moynier, E. Inglis, S. Tian, M. Li, Y. Liu and Z. Hu, J. Anal. At. Spectrom., 2019, 34, 1800–1809.
61. I. Gunther-Leopold, B. Wernli, Z. Kopajtic and D. Gunther, Anal. Bioanal. Chem., 2004, 378, 241–249.
62. E. M. Krupp and O. F. X. Donard, Int. J. Mass Spectrom., 2005, 242, 233–242.
63. Q. Li, W. Zhang, Z. Hu, Y. Feng, H. Hu, T. Luo, J. Huang and X. Zeng, Chem. Geol., 2023, 636, 121651.
64. H. Niu and R. S. Houk, Spectrochim. Acta, Part B, 1996, 51, 779–815.
65. C. T. Gerritzen, S. Goderis, H. F. James and C. Snoeck, Spectrochim. Acta, Part B, 2024, 217, 106955.
66. W. Zhang, Z. Hu and Y. Liu, J. Anal. At. Spectrom., 2020, 35, 1087–1096.
67. J. Fu, Z. Hu, W. Zhang, L. Yang, Y. Liu, M. Li, K. Zong, S. Gao and S. Hu, Anal. Chim. Acta, 2016, 911, 14–26.
68. N. M. Ellis and B. H. Passey, Chem. Geol., 2023, 635, 121616.
69. N. Dauphas, A. Pourmand and F.-Z. Teng, Chem. Geol., 2009, 267, 175–184.
70. Y. He, S. Ke, F. Teng, T. Wang, H. Wu, Y. Lu and S. Li, Geostand. Geoanal. Res., 2015, 39, 341–356.
71. X. Yan, T. Tanaka and H. Kawaguchi, Appl. Spectrosc., 1996, 50, 182–187.
72. M. Wu, Y. Duan, Q. Jin and G. M. Hieftje, Spectrochim. Acta, Part B, 1994, 49, 137–148.
73. E. V. Polyakova, Yu. N. Nomerotskaya and A. I. Saprykin, J. Anal. Chem., 2020, 75, 474–478.
74. N. J. Pearson, W. L. Griffin and S. Y. O'Reilly, in Laser Ablation-ICP-MS In The Earth Sciences, ed. P. Sylvester, Mineralogical Association of Canada, Québec, 2008, pp. 93–116.
75. H. Andrén, I. Rodushkin, A. Stenberg, D. Malinovsky and D. C. Baxter, J. Anal. At. Spectrom., 2004, 19, 1217–1224.
76. J. Barling and D. Weis, J. Anal. At. Spectrom., 2012, 27, 653.
77. M. M. Fraser and D. Beauchemin, Spectrochim. Acta, Part B, 2000, 55, 1705–1731.
78. C. Agatemor and D. Beauchemin, Anal. Chim. Acta, 2011, 706, 66–83.
79. A. A. Mills, J. H. Macedone and P. B. Farnsworth, Spectrochim. Acta, Part B, 2006, 61, 1039–1049.
80. Z. Hu, Y. Liu, S. Gao, W. Liu, W. Zhang, X. Tong, L. Lin, K. Zong, M. Li, H. Chen, L. Zhou and L. Yang, J. Anal. At. Spectrom., 2012, 27, 1391.
81. J. Lin, Z. Liu, Y. Liu, W. Liu, X. Jiang, G. Zhao, L. Chen and Z. Hu, Anal. Chem., 2023, 95, 16877–16884.
82. X. Jiang, Y. Liu, W. Liu, J. Lin, Z. Liu, L. Chen, X. Zhu, W. Zhang and Z. Hu, J. Anal. At. Spectrom., 2024, 39, 2452–2460.

---