Analysis of sulfur in soil and plant digests using methane as a reaction gas for ICP-MS

The goal of this study was to develop and validate a method for analyzing sulfur in agricultural soil and plant samples using ICP-MS with methane as a reaction gas. Since sulfur quantitation by ICP-MS is challenging due to spectral interferences, it is typically performed using ICP-OES, requiring the use of two instruments for multi-element analysis. This increases both analysis time and cost.

To address this, the researchers evaluated the performance of methane in both single and triple quadrupole ICP-MS systems, focusing on sulfur-containing cluster ions (CH₂SH⁺) at m/z 47 and 49. The results showed that triple quadrupole ICP-MS offered superior interference removal and detection limits suitable for agricultural analysis. Methane proved to be a viable alternative to oxygen as a reaction gas, providing comparable results to ICP-OES and enabling simplified, single-instrument workflows.

The original article

Analysis of sulfur in soil and plant digests using methane as a reaction gas for ICP-MS

Sukhjeet Singh, Michael R. Mucalo, Megan N.C. Grainger 

Talanta, Volume 281, 2025, 126797

https://doi.org/10.1016/j.talanta.2024.126797

licensed under CC-BY 4.0

Selected sections from the article follow. Formats and hyperlinks were adapted from the original.

Inductively Coupled Plasma Optical Emission Spectroscopy (ICP-OES) is a popular choice of instrument for the measurement of S in soil and plants [[6], [7], [8]]. However, although this technique is relatively inexpensive and robust (allowing for high total dissolved solids), the detection limits for most elements are typically much higher than those achieved by Quadrupole Inductively Coupled Plasma Mass Spectrometry (Q-ICP-MS) [9]. This makes multi-element measurements (particularly those that include trace elements) less achievable from a single analysis. However, S is not currently routinely analysed by Q-ICP-MS despite the technique achieving low detection limits for most elements as it is a challenging element to quantify. One of the difficulties is that S has a high first ionisation potential of 10.36 eV which is one of the highest for elements analysed by ICP-MS and comes close to that of argon (15.76 eV) [10]. A consequence of this is that only approximately 14 % of S atoms form S⁺ ions in an argon plasma at 7500 K, thus lowering its sensitivity [11]. Furthermore, all S isotopes are subject to severe spectral interferences that are common across most matrices measured [12]. The most abundant isotope of S (³²S, 95.02 % abundant) cannot be measured due to a severe spectral interference from ¹⁶O₂⁺, while ³⁶S⁺ cannot be measured due to the irresolvable interference of ³⁶Ar ⁺ using a single quadrupole detector. To measure S by ICP-MS without use of collision or reaction gases, high-resolution equipment such as a double focusing sector field instrument (SF-ICP-MS) must be used. However, price and complexity are a barrier for routine analysis using SF-ICP-MS.

Reaction gases can be used to improve the performance of S measurement by Q-ICP-MS. Previous studies report the use of xenon to reduce the ¹⁶O₂⁺ interference on ³²S⁺ in water samples [13] and in biofuel samples [14] by way of charge transfer reactions, but it is an uncommon and expensive choice for a reaction gas. Oxygen is a more common reaction gas for a wide variety of elements, including S [[15], [16], [17], [18]]. For the ³²S⁺ ion, ³²S¹⁶O⁺ is measured at m/z 48 to avoid the ¹⁶O₂⁺ polyatomic interference at m/z 32. However, a large disadvantage of this reaction is that there are multiple interferences at m/z 48 (e.g., from ⁴⁸Ca⁺, ⁴⁸Ti⁺ and ³⁶Ar¹²C⁺) which are not able to be easily overcome using ICP-MS but can be addressed with triple quadrupole ICP-MS (ICP-MS/MS) due to the advantage of selecting masses at Q1 and Q2. Plant and soil samples will have high carbon loading and hence the ³⁶Ar¹²C⁺ interference presents a challenge. The use of ICP-MS/MS, has been demonstrated to be effective at reducing interferences on ³²S¹⁶O⁺ by removing some of the interferences at the first quadrupole [19,20].

There is currently no published research carried out on investigating the potential of using methane as an ICP-MS reaction cell gas for S measurement. It is clear that there will be a strong scientific and economic benefit by filling this research gap. Additionally, the use of methane may provide some flexibility with instrument set up due to the limited number of gas channels available on the instrument. Hence the aim of this study was to assess the suitability of methane as a reaction gas for S analysis in agricultural soil and plant samples to give users of ICP-MS more choice by providing an alternative reaction gas, offering greater flexibility with instrument set up. The potential to analyse S by ICP-MS instead of ICP-OES also reduces the workflow required for elemental analysis by providing the means to analyse samples once for all elements, thus giving both a time and cost saving to laboratories.

2. Materials and methods

2.3. Instrumentation

2.3.1. Q-ICP-MS

A NexION 2000P Inductively Coupled Plasma – Mass Spectrometer equipped with a nickel sampler cone, nickel skimmer cone, aluminium hyperskimmer cone, quartz cyclonic spray chamber and quartz torch with 2.0 mm injector (PerkinElmer, USA) was used with a 0.4 mL min⁻¹ SeaSpray U-Series nebuliser (Glass Expansion, Australia) and controlled by Syngistix software version 2.4. Instrument conditions are detailed in Table 1. Liquid samples were introduced via an ASX-520HS autosampler (Teledyne CETAC Technologies, USA) and an ESI OneFAST system with MP2 peristaltic pump (Elemental Scientific, USA). The PVC carrier and internal standard tubing had an internal diameter of 0.76 mm and 0.38 mm respectively (Pulse Instrumentation, USA). The internal standard, which contained 1 % isopropanol, was introduced via a seven-port valve.

2.3.2. ICP-MS/MS

A 8900 Triple Quadrupole ICP-MS equipped with a quartz cyclonic spray chamber, torch and 2.0 mm injector along with a nickel sampler, skimmer cone and an extraction Omega lens (Agilent Technologies, USA) was used and controlled by ICP-MS MassHunter Workstation version 4.5. Instrument conditions are detailed in Table 1. Liquid samples were introduced to the spray chamber via an SPS4 autosampler (Agilent Technologies, USA), PVC tubing (Pulse Instrumentation, USA) and a 0.05–0.1 mL min⁻¹ SeaSpray U-Series nebuliser (Glass Expansion, Australia). The internal standard was introduced inline via a t-junction.

2.3.3. ICP-OES

An iCAP 6000 Series, radial viewing Inductively Coupled Plasma Optical Emission Spectroscopy Instrument (Thermo Electron Corporation, UK) was controlled with iTEVA software version 2.8.0.97 using conditions detailed in Table 1. Liquid samples were introduced to the spray chamber via an ASX-520HS autosampler (Teledyne CETAC Technologies, USA) and PVC tubing (Pulse Instrumentation, USA) with an internal diameter of 0.89 mm. A 2.0 mL min⁻¹ SeaSpray U-Series nebuliser was attached to a quartz cyclonic spray chamber followed a ceramic demountable torch with 2.4 mm injector (Glass Expansion, Australia).

3. Results and discussion

3.4. Plants

S was quantified in digested plant samples to assess the performance of methane as a reaction gas with the optimised parameters. The Method Detection Limit (MDL) was determined by digesting and analysing (n = 9) a low S-containing sample (1020 ± 30 mg kg⁻¹ S) determined by ICP-OES. Table 4 summarises the MDL for each mode of analysis (refer to Supplementary Material 4 for individual analyses for plant samples). As expected, the MDL was highest for standard mode, followed by the ion cluster results at m/z 47 then 49 when analysed by Q-ICP-MS. The ICP-MS/MS results for 34 → 49 were similar to that for m/z 49 for the single quadrupole. The MDLs were an order of magnitude higher than reported for ICP-OES (150 and 50 mg kg⁻¹ S respectively), however, an exceptionally low detection limit is not critical due to S in plant samples being at least an order of magnitude higher than the reported MDL. Therefore, the method is deemed fit for requirement. It is worthwhile to note that extremely low detection limits are only required for trace determination and assessment of contamination – which is not required for determination of S, a macronutrient in this matrix.

Precision was assessed using a low (1020 ± 30 mg kg⁻¹ S) and a high (3420 ± 80 mg kg⁻¹ S) S containing plant sample. When analysing the sample containing low concentrations of S by Q-ICP-MS standard mode at m/z 34, the precision was higher than acceptable (47 % RSD). However, this sample was acceptable when using methane as the reaction gas (18 and 10 % RSD for Q-ICP-MS clusters at m/z 47 and 49 respectively) and when determined by ICP-MS/MS (11 %). For the sample containing 3420 ± 80 mg kg⁻¹ S, the Q-ICP-MS m/z 49 ion cluster result had poor precision (25 % RSD), however, this can be explained due to the presence of approximately 0.5 mg L⁻¹ titanium in the vial (100 mg kg⁻¹ in the dried plant sample) due to improper sample collection which resulted in soil contaminating the sample. The interference due to the presence of Ti was adjusted for by subtracting the apparent sulfur concentration caused by titanium for a given sample (determined by a single element Ti solution measured on each analysis sequence). Using such a correction likely caused additional imprecision due to the higher standard deviation that arises from combing multiple measurements. In comparison, the equivalent ion cluster measured by ICP-MS/MS (34 → 49) demonstrated excellent precision (7 % RSD) due to the titanium interference being removed at Q1 a correction not being required. It is noted that digestion using HNO₃ and H₂O₂ will not completely recover Ti; however, this partial recovery of acid extractable Ti will interfere with the analysis if present.

3.5. Soils

Digested soils provide a more challenging matrix than plants because they contain relatively high titanium and chloride in the sample digest and the S concentration in the analysis vial (at instrument once digested) is relatively low compared to digested plant samples. Therefore only ICP-MS/MS mass shift of 34 → 49 was compared to the Q-ICP-MS standard mode as Q-ICP-MS reaction mode as unable to produce viable data due to these interferences. The MDL results are summarised in Table 4 (the full dataset is available in Supplementary Material 4). As expected, the Q-ICP-MS standard mode had a high MDL due to no interference removal. The ICP-MS/MS result using methane as the reaction gas had a lower method detection limit compared to standard mode and are similar to the MDL for ICP-OES (53 and 23 mg kg⁻¹ respectively). The MDL is acceptable as the S concentration in soil samples far exceeds this value, hence there will be no difficultly in measuring S in soil samples using ICP-MS/MS. Precision was assessed using two low (200–300 mg kg⁻¹) and one high (1600 mg kg⁻¹) S containing soil samples. The precision for ICP-MS/MS methane results was slightly higher than the ICP-OES results, but still within ±20 %, which is acceptable. For the samples containing 200–300 mg kg⁻¹ S the precision was 11–16.7 % for ICP-MS/MS compared to 7.2–7.4 % for ICP-OES; for the sample containing 1600 mg kg⁻¹ S, the precision was 9.8 % compared to 2.2 %, respectively. The precision for ICP-MS/MS was much better than when using standard mode due to the previously discussed interferences. The results were < MDL for the 200–300 mg kg⁻¹ S sample and 20.8 % at 1600 mg kg⁻¹ S. With further refinement (perhaps using mass shift of the internal standard [42,43]), this method performance may improve and is a good starting point for other researchers.

4. Conclusion

Overall, this work has demonstrated promising results when using methane as a reaction gas to provide accurate S quantification in plant and soil samples; Table 5 summarises the best mode to use for each matrix. For plant digests, interferences from Ti⁺ and CCl⁺ are relatively low allowing acceptable MDL and precision using Q-ICP-MS with analysis of CH₂SH ⁺ at m/z 47. This provides an alternative option for S analysis in plants for laboratories that do not have access to ICP-MS/MS. Discretion should be applied for plant samples by routinely monitoring both Ti and Al to check the level of soil contamination in individual samples due to poor sample collection. As expected, ICP-MS/MS is the superior choice of instrument for detection of S in plants since Ti⁺ and CCl⁺ interferences are removed at the first quadrupole allowing for more accurate analysis. The method performance for the measurement of S in plant digests when using ICP-MS/MS produced acceptable results.

Analysis of S in soil digests was a slightly more complicated matrix due to the contribution of Cl from the HCl digest and higher levels of Ti in soil. Therefore ICP-MS/MS was the only suitable choice for analysis. There was a slight low bias in soil samples with low S (e.g., ILCP results). However, these are still above the pastoral soil S deficiency limit of 600 mg kg⁻¹, hence the method is deemed to be fit for purpose. It is likely that further optimisation of the ICP-MS/MS method could improve the accuracy.

Analysis of S by ICP-MS/MS instead of ICP-OES is advantageous due to the ability to analyse multiple elements over multiple orders of magnitude in a single analysis, hence producing a large amount of data in a short time-period. This will simplify sample workflows, reduce staffing and consumable costs, and provide quicker turnaround times. This work shows that methane is a viable option as a reaction gas for determination of S when using ICP-MS/MS for both agricultural plant and soil samples, and when using Q-ICP-MS at m/z 47 for plant digests. Thus, giving laboratories an alternative option for gas choice and more flexibility.