Isobaric interference removal for selected radionuclides using nitrous oxide and ammonia with inductively coupled plasma tandem mass spectrometry

Isobaric interference removal for selected radionuclides using nitrous oxide and ammonia with inductively coupled plasma tandem mass spectrometry Shaun T. Lancaster, * a Ben Russell, b Thomas Prohaska ac and Johanna Irrgeher ac The determination of long-lived radionuclides by inductively coupled plasma tandem mass spectrometry (ICP-MS/MS) is a well-established approach. However, such determinations can still be hindered by isobaric interferences from stable isotopes of neighbouring elements. As such, investigations towards novel gas cell approaches for removing interfering ions are required in order to improve the reliability of the analysis. Nitrous oxide (N 2 O) is a reaction gas that has been well studied for stable isotope analysis. Studies towards its applicability to radionuclide analysis have so far been limited. Here, the use of N 2 O, as well as a mixture with ammonia (NH 3 ), have been evaluated for determinations of 10 radionuclides of interest for nuclear decommissioning: 41 Ca, 63 Ni, 79 Se, 90 Sr, 93 Zr, 93 Mo, 94 Nb, 107 Pd, 135 Cs, and 137 Cs. Single element solutions of stable isotope analogues of the radionuclides, as well as solutions of the interfering ions, were used to observe the reactions with the ICP-MS/MS reaction cell gases. Abundance-corrected sensitivities were used to assess the achievable separation factors and sensitivities for the determination of the radionuclides of interest. The N 2 O/NH 3 gas mixture was found to provide a significant enhancement in the removal of isobaric interferences, as well as instrument detection limits (given in brackets), compared to N 2 O alone for determinations of 41 Ca (0.50 pg g −1 (0.0016 Bq g −1 )), 79 Se (0.11 pg g −1 (5.4 ×10 −5 Bq g −1 )), 90 Sr (0.11 pg g −1 (0.56 Bq g −1 )), 93 Mo (0.12 pg g −1 (0.0044 Bq g −1 )), 135 Cs (0.1 pg g −1 (7.5 ×10 −6 Bq g −1 )), and 137 Cs (0.1 pg g −1 (0.33 Bq g −1 )). Shaun T:Lancaster Shaun Lancaster currently holds a permanent position at Montanuniversit¨ at Leoben, Austria. He earned his PhD in 2021 in conjunction with the University of Aberdeen, Scotland, and instrument manufacturer P S Analytical, England. His work focused on method development for mercury speciation and subsequent analysis of methylmercury in the livers of Scottish birds of prey, as well as the development of novel atomic uorescence based instrumentation for continuous monitoring of mercury in effluent wastewater streams. Following this, he began his postdoc at Montanuniversit¨ at Leoben working on the development of analytical methodology using ICP-MS/MS and XRF for the analysis of complex electronic waste matrices to assist the recycling industry as part of the drive towards a circular economy. Since taking a permanent position, his future ambition is to utilize ICP-MS/MS reaction cell gas chemistry to unlock new analytical approaches to solve challenging applications. In particular, to remove spectral interferences on difficult-to-measure isotopes for stable isotope ratio and radionuclide determinations, as well as for isotope ratio determinations of oxygen, which have previously never been performed by ICP-MS/MS. a Department of General, Analytical and Physical Chemistry, Chair of General and Analytical Chemistry, Montanuniversit¨ at Leoben, Leoben, Austria. E-mail: shaun. [email protected]c.at; [email protected]; thomas. [email protected]c.at b Nuclear Metrology Group, National Physical Laboratory, Hampton Road, Teddington, UK. E-mail: [email protected] c Department of Physics and Astronomy, University of Calgary, Calgary, Canada Cite this: J. Anal. At. Spectrom.,2025, 40,3210 Received 30th June 2025 Accepted 10th September 2025 DOI: 10.1039/d5ja00254k rsc.li/jaas 3210 |J. Anal. At. Spectrom.,2025,40,3210–3220 This journal is © The Royal Society of Chemistry 2025 JAAS PAPER Open Access Article. Published on 07 October 2025. Downloaded on 11/5/2025 9:49:59 AM. This article is licensed under a Creative Commons Attribution 3.0 Unported Licence. View Article Online View Journal | View Issue Introduction Decommissioning of nuclear sites represents a high cost and long-term analytical challenge to return the site to a safe state for future reactor development or alternative use. This necessitates the accurate measurement of a range of radionuclides in complex and varied sample matrices. Of particular interest are the medium and long-lived radionuclides that represent the most signicant contribution to the long-term waste inventory that must be safely stored or disposed of. Examples include waste products generated from nuclear ssion of 235 U, such as 79 Se, 90 Sr, 93 Zr, 107 Pd, 135 Cs, and 137 Cs, as well as activation products such as 41 Ca, 63 Ni, 93 Mo, and 94 Nb formed by neutron capture in the concrete and steel casings of the nuclear reactor. 1 Efficient characterisation of radioactive waste streams is key for safe and cost effective waste management, and requires the use of rapid and robust analytical methodology. Inductively coupled plasma mass spectrometry (ICP-MS) has been for decades a frequently applied method of choice for the determination of long-lived radionuclides, providing more rapid analysis of samples compared to decay counting instrumentation and thereby contributing to faster and more cost-effective decommissioning. 1,2 Accurate quantitative determinations of long-lived radionuclides by ICP-MS are hindered primarily by isobaric interferences from naturally occurring stable isotopes of other elements, as well as polyatomic ions and peak tailing interferences from neighbouring masses, oen requiring timeconsuming offline chromatographic separations to resolve. Modern quadrupole-based ICP-MS systems are equipped with a reaction cell that can be used to resolve isobaric interferences based on the differences in reactivity of different elements. Target analytes can be determined using either a mass-shiapproach, where the analyte of interest is selectively reacted to form aproductionatahighermass-to-chargeratio(m/z), or an on-mass approach, where the interfering ion is selectively reacted and the analyte of interest is measured without altering its m/z. 3 With the advent of tandem mass spectrometry (MS/MS), only interferences at the m/zof the target analyte are of concern, as a quadrupole mass lter removes all other ions prior to entering the gas cell. 4 The most commonly utilized cell gases for interference removal are hydrogen (H 2 ), helium (He), oxygen (O 2 ), and ammonia (NH 3 ). 4 Of these, O 2 and NH 3 are commonly used for the removal of isobaric interferences for radionuclide determinations. 5,6 Both H 2 and He have high ionisation energies and have low reactivity with most elements, and are typically employed as collisional gases, which can only reduce polyatomic interferences (by kinetic energy discrimination). Reactive cell gases can be combined with He to improve the interference separation. The addition of an unreactive gas can enhance sensitivity and reactions through the collisional focussing effect, focussing the ions to the minimum of the effective potential of the quadrupole by reducing the radial kinetic energy of the ion. 7,8 Moreover, the addition of H 2 to NH 3 gas was observed to improve formation of M(NH 3 ) n product ions, rather than splitting the sensitivity between product ions with a lower number of hydrogen atoms, 9 demonstrating that combining two gases can lead to a different product ion formation. Uncovering new and improved interference removal pathways through ICP-MS/MS reaction cell chemistry can improve detection limits, sample throughput, and provide simpler methodology by minimizing the need for complex and time consuming chemical separations. Thus, further investigation into the use of alternative cell gases and cell gas combinations is necessary to expand options for end users. Nitrous oxide (N 2 O) is a reaction gas that has been extensively studied for stable isotope determinations, 10–13 however for radionuclides its use has been limited primarily to the removal of radioactive caesium isotopes from stable barium isotopes enabling 135 Cs/ 137 Cs source attribution, which is a valuable tool for long-term environmental monitoring and used following the Fukushima disaster. 14–16 Additionally, limited studies focus on mixtures of N 2 O with other gases, such as He, H 2 and NH 3 . 17,18 As such, this study aims to evaluate the use of N 2 O, as well as gas mixtures of N 2 O with NH 3 , as ICP-MS/MS cell gases for the removal of isobaric interferences on 10 radionuclides that are of importance in waste characterisation for nuclear decommissioning: 41 Ca, 63 Ni, 79 Se, 90 Sr, 93 Zr, 93 Mo, 94 Nb, 107 Pd, 135 Cs, and 137 Cs. To achieve this, elemental standards containing stable isotopes of the elements of interest were utilized as analogues to assess the reactivity of the elements compared to the reactivity of the isobaric interferences. Sensitivities and interference separation factors for the target radionuclides were calculated using the instrument response of the stable analogues and factoring the isotopic abundances. Experimental Reagents Nitric acid (HNO 3 ,w=65%, p.a. grade; Carl Roth GmbH, Karlsruhe, Germany) was puried using a sub-boiling distillation system (Savillex DST-4000, AHF Analysentechnik, Tübingen, Germany). Reagent grade I water (18.2 MUcm; MilliQ IQ 7000, Merck-Millipore, Darmstadt, Germany) was used for all acid dilutions. Vials and pipette tips were pre-cleaned by soaking overnight in diluted sub-boiled nitric acid (w=3%) before use. Single-element standards of caesium (Cs), copper (Cu), potassium (K), magnesium (Mg), molybdenum (Mo) (b=1000 mgmL −1 ; Certipur, Merck); aluminium (Al), barium (Ba), calcium (Ca), iron (Fe), niobium (Nb), palladium (Pd), strontium (Sr), zirconium (Zr) (b=1000 mgmL −1 ; Inorganic Ventures, Christiansburg, VA, USA); ytterbium (Y, b=1000 mg mL −1 ; High Purity Standards, North Charleston, SC, USA); selenium (Se, b=1000 mgmL −1 ; Peak Performance, CPI international, Santa Rosa, CA, USA); nickel (Ni, b=1000 mgmL −1 ; Alfa Aesar, Karlsruhe, Germany); and silver (Ag, b=10 mgmL −1 ; Inorganic Ventures) were used throughout this work. Potassium bromide (NORMAPUR grade; VWR, Vienna, Austria) salt was used to prepare standards for the analysis of bromine (Br). Instrumentation All work was carried out using a NexION 5000 ICP-MS/MS system (PerkinElmer, Waltham, MA, USA) equipped with This journal is © The Royal Society of Chemistry 2025 J. Anal. At. Spectrom.,2025,40,3210–3220 | 3211 Paper JAAS Open Access Article. Published on 07 October 2025. Downloaded on 11/5/2025 9:49:59 AM. This article is licensed under a Creative Commons Attribution 3.0 Unported Licence. View Article Online a quadrupole-based dynamic reaction cell (DRC). The applied bandpass of the reaction cell quadrupole can be modied using rejection parameter “a”(RPa) and rejection parameter “q” (RPq). Instrument parameters for the different measurement modes applied are given in Table 1. Argon (purity 5.0 ($99.999%); Linde Gas GmbH, Stadl-Paura, Austria) was used as the plasma gas. Nitrous oxide (medicinal grade; Linde Gas GmbH) and ammonia (purity 5.0 ($99.999%); Linde Gas GmbH) were used as reaction gases. Analytical measurement Standards were diluted in nitric acid (w=2%). The isotopes of stable analogues and interferences monitored in this study are listed in Table 2. Initial mass scans were carried out using 0.8 mL min −1 N 2 O, as well as a mixture of 0.8 mL min −1 N 2 O and 0.4 mL min −1 NH 3 , in order to identify product ions that formed for each element monitored. Further investigation was carried out by varying gas ow rates and monitoring selected product ions. Once the optimum gas ow rates had been determined for the removal of isobaric interferences, further interferences from elevated levels of Mg (as MgO), Al (as ArAl) and Fe (as ArFe) were assessed by introducing each interfering element as a 5 mgg −1 single element standard and monitoring the instrument response. Data processing. Sensitivities for radionuclides were determined from their stable analogues by scaling the isotopic abundances to 100% via eqn (1): kradionuclide ¼ðIstd IblkÞ wstdx(1) where kis the sensitivity in cps (ng −1 g −1 ), I std and I blk are the measured signal intensities (in cps) of the standard and the blank respectively, w std is the mass fraction of the standard in ng g −1 , and xis the isotopic abundance (as isotope amount fraction) as stated by the IUPAC Commission on Isotopic Abundances and Atomic Weights. 20 The separation factor between a target radionuclide and its interference is given as the ratio of the sensitivity of the target analyte to the sensitivity of the interference (k radionuclide / k interference ). The greater the separation factor, the better resolved the radionuclide measurement is. For stable isotopes, the sensitivity is based on the elemental concentration. Product ion formation has been calculated by the ratio of the sensitivity of the product ion of an element to the sensitivity of the same element achieved on-mass in the absence of a cell gas (k product ion /k on−mass, no gas ) and expressed as a percentage. Instrument detection limits were determined as three times the standard deviation of 10 replicate determinations of the blank signal (using w=2% HNO 3 ), divided by the sensitivity of the radionuclide. The sensitivity was determined by performing a calibration using 7 standards of the stable isotopes of each given element at the optimum cell gas conditions. The sensitivity was normalized to 100% abundance of the isotope to determine the sensitivity of radionuclide determinations, as in eqn (1). The concentrations of the standards were: 1–1000 pg Table 1 Operational parameters of the ICP-MS/MS system Parameter Standard mode N 2 O DRC mode N 2 O/NH 3 DRC mode Scan mode MS/MS MS/MS MS/MS Cell gas None N 2 ON 2 O and NH 3 RPa 0 0 0 RPq 0.25 0.45 0.45 Sample introduction Self-aspiration Self-aspiration Self-aspiration Nebulizer PFA MicroFlow PFA MicroFlow PFA MicroFlow Spray chamber Peltier cooled SilQ cyclonic spray chamber Peltier cooled SilQ cyclonic spray chamber Peltier cooled SilQ cyclonic spray chamber Spray chamber temperature 5 °C 5 °C 5 °C Interface cones Nickel Nickel Nickel RF power 1600 W 1600 W 1600 W Ar nebulizer gas ow 0.92–0.98 L min −1 0.92–0.98 L min −1 0.92–0.98 L min −1 Ar auxiliary gas ow 1.2 L min −1 1.2 L min −1 1.2 L min −1 Ar plasma gas ow 16 L min −1 16 L min −1 16 L min −1 Hyperskimmer park voltage 5 V 5 V 5 V OmniRing park voltage −185 V −185 V −185 V Inner target lens voltage 2 V 2 V 2 V Outer target lens voltage −7V −7V −7V Deector exit voltage −8V −8V −8V Differential aperture voltage −3.5 V −3.5 V −3.5 V Q1 AC rod offset −6V −10 V −10 V Q1 rod offset −2V 0V 0V Cell rod offset −33 V −5V −5V Axial eld voltage 0 V 250 V 250 V Cell entrance voltage −5V −8.5 V −8.5 V Cell exit voltage −2V −5.5 V −5.5 V Q3 AC rod offset −2.5 V −7V −7V Q3 rod offset −2V −10 V −10 V Dwell time 50 ms 50 ms 50 ms 3212 |J. Anal. At. Spectrom.,2025,40,3210–3220 This journal is © The Royal Society of Chemistry 2025 JAAS Paper Open Access Article. Published on 07 October 2025. Downloaded on 11/5/2025 9:49:59 AM. This article is licensed under a Creative Commons Attribution 3.0 Unported Licence. View Article Online g −1 for Ni, Sr, Zr, Nb, Mo, Pd, and Cs, 10–10 000 pg g −1 for Se, and 100–100 000 pg g −1 for Ca (due to the low abundance of the 44 Ca isotope). In order to convert the sensitivities and instrument detection limits from mass fractions to activities, the specic activity was calculated for each radionuclide via eqn (2): a¼lnð2ÞNA t1=2ma (2) where ais the specic activity in Bq g −1 of pure substance, N A is the Avogadro constant, t 1/2 is the half-life of the radionuclide, and m a is the atomic mass of the radionuclide. The mass fraction units (in ng g −1 ) were converted to units of activity, A(in Bq g −1 of sample), via eqn (3): A=w×a×10 −9 (3) where the factor of 10 −9 is included to convert between g (in Bq g −1 ) and ng (in ng g −1 ). Results and discussion Calcium-41 41 Ca (t 1/2 =1.002 ×10 5 years) is produced by neutron activation of 40 Ca (96.94% abundance) present in the concrete shield surrounding nuclear reactors, which makes up a considerable amount of waste during decommissioning. Determinations of 41 Ca has primarily been be carried out using liquid scintillation counting (LSC) and accelerator mass spectrometry (AMS) following chemical separations. 21–23 Recently, the rst determinations of 41 Ca by ICP-MS/MS were conducted by Russell et al., 24 where the authors suggests the use of an NH 3 /H 2 /He gas mixture for interference removal. By this approach, 41 Ca + does not react with NH 3 or H 2 , while interferences of 40 Ar 1 H + and 40 Ar + (from peak tailing) are removed by a charge transfer reaction with NH 3 . The authors reported a detection limit of 99 pg g −1 (0.32 Bq g −1 ) and a sensitivity of 3700 cps (ng −1 g −1 ) (1150 cps (Bq g −1 )) for 41 Ca. 24 However, the major drawback of the NH 3 /H 2 approach is that the stable 41 K isotope also does not react with either gas, therefore 41 Ca determinations are still subject to interferences in matrices containing high K levels. In literature, N 2 O has previously been demonstrated to remove argon-based interferences, as well as interferences of K, by mass-shiing to the CaO + product ion. 25,26 However, group 2 elements are known to also form hydroxide product ions with N 2 O due to impurities in the cell gas, 26 therefore peak tailing of stable 40 Ca may still be problematic. In this study, a mixture of N 2 O and NH 3 were tested for determinations of the 41 Ca. The addition of NH 3 to the N 2 O generally caused group 2 elements to primarily form the hydroxide product ion (+17 amu) instead of the oxide product ion, while no product ion formation was observed for the group 1 elements. The argon interference is also removed effectively, as both N 2 O and NH 3 react with Ar and ArH by charge transfer. For the determination of 41 Ca using the 41 Ca 16 O 1 H + product ion, the optimum gas mixture was found to be 0.4 mL min −1 N 2 O and 0.2 mL min −1 NH 3 , achieving a sensitivity of 50 000 cps (ng −1 g −1 ) (15 500 cps (Bq g −1 )) for 41 Ca (Table S1). This gas mixture allowed interferences from peak tailing of 40 Ca to be successfully removed, as the 40 Ca 16 O 1 H 2+ product ion was not observed to form. This was in contrast to using N 2 O alone and monitoring the 41 Ca 16 O + product ion (m/z=57) as impurities in the N 2 O gas caused a noticeable shiof 40 Ca + to 40 Ca 16 O 1 H + (m/ z=57). Importantly, the interference from 41 K + was eliminated due to the selective reaction of 41 Ca + , thus providing a distinct advantage over the current best available methodology. The instrument detection limit achieved for 41 Ca using the N 2 O/NH 3 gas mixture approach was 0.35 pg g −1 (0.0011 Bq g −1 ). Further investigation was also carried out to investigate the Table 2 List of radionuclides of interest and the isotope of the stable analogue and interferences measured in this study. Half-lives of the radionuclides of interest were sourced from the atomic and nuclear database of the Decay Data Evaluation Project. 19 Abundances for the interferences are given as the isotopic abundance of the interfering element. 20 Abundances for polyatomic interferences are given as the product of the isotopic abundances of the elements it contains Radionuclide Half-life (years) Stable analogue monitored Spectral interference Interference abundance (%) 41 Ca 100 200 1700 44 Ca 41 K 6.73 40 Ar 1 H 99.59 40 Ca tailing 96.94 63 Ni 98.7 2.4 60 Ni 63 Cu 69.17 79 Se 356 000 40 000 82 Se 79 Br 50.69 90 Sr 28.80 0.70 88 Sr 90 Zr 51.45 89 Y 1 H 99.97 93 Mo 4000 800 94 Mo 93 Nb 100 93 Zr n/a 93 Zr 1 610 000 60 000 94 Zr 93 Nb 100 93 Mo n/a 94 Nb 20 040 40 93 Nb 94 Zr 17.38 94 Mo 9.19 107 Pd 6 500 000 300 000 105 Pd 107 Ag 51.84 135 Cs 1 330 000 190 000 133 Cs 135 Ba 6.59 137 Cs 30.018 0.022 133 Cs 137 Ba 11.23 This journal is © The Royal Society of Chemistry 2025 J. Anal. At. Spectrom.,2025,40,3210–3220 | 3213 Paper JAAS Open Access Article. Published on 07 October 2025. Downloaded on 11/5/2025 9:49:59 AM. This article is licensed under a Creative Commons Attribution 3.0 Unported Licence. View Article Online effectiveness of the optimised N 2 O/NH 3 gas mixture approach for sample matrices that may contain high levels of Mg, as 25 Mg 16 O + is an additional interference at m/z41. Analysis of a 5 mgg −1 Mg standard indicated a minor interference of 17 cps (mg −1 g −1 ) Mg. Thus samples with excessive Mg levels may require dilution. Nickel-59/63. Neutron activation of stable Ni isotopes in steel reactor casings produce two long-lived Ni radionuclides: 59 Ni (t 1/2 =7.6 ×10 4 years) and 63 Ni (t 1/2 =98.7 years). Following chemical separations, 63 Ni can be easily measured using LSC. 27,28 The 59 Ni radionuclide is more challenging to measure by LSC due to the more abundant beta emissions of 63 Ni, but has been measured using X-ray emission techniques. 29 For ICPMS/MS analysis, the 59 Ni and 63 Ni radionuclide have isobaric interferences from stable 59 Co and 63 Cu isotope, respectively. Russell et al. 30 demonstrated previously that an NH 3 /H 2 /He gas mixture can be used to separate 63 Ni from the 63 Cu isobar. By this approach, 63 Ni reacts to form the 63 Ni(NH 3 ) 3+ product ion at a greater rate than the formation of 63 Cu(NH 3 ) 3+ , achieving a 63 Ni/ 63 Cu separation factor of 100 and detection limits of 0.25 pg g −1 (0.52 Bq g −1 ). The addition of H 2 to the NH 3 /He mixture allowed for greater formation of the M(NH 3 ) n product ions. The authors additionally noted that the use of H 2 signicantly improved the formation rate of the 63 Ni(NH 3 ) 3+ product ion, and reported an achieved sensitivity of 6100 cps (ng −1 g −1 ) (2.86 cps (Bq g −1 )). To date, there has been no successful ICP-MS/MS reaction cell based separation of 59 Ni from 59 Co reported in literature. While determinations of 59 Ni using N 2 O and a N 2 O/NH 3 gas mixture were attempted in this study, it was found that Co and Ni behaved very similarly under both conditions. As such, no separation of 59 Ni and 59 Co could be achieved and 59 Ni was not investigated further here. Although Ni and Cu are both relatively unreactive with N 2 O, minor formation of oxide product ions have been previously observed. 13 In this study, Ni and Cu formed oxide product ions (+16 amu) at 4.6% and 0.1% respectively, allowing for a potential interference separation route. The maximum separation factor of 13 600 was achieved using a N 2 Oow rate of 1.4 mL min −1 , achieving a 63 Ni sensitivity of 3640 cps (ng −1 g −1 ) (1.71 cps (Bq g −1 )). The obtained instrument detection limit using N 2 O was calculated to be 0.29 pg g −1 (0.62 Bq g −1 ). Slightly higher sensitivities could be achieved by reducing the N 2 Oow rate, but result in a less efficient separation of 63 Cu from 63 Ni (Fig. 1). For example, applying 1 mL min −1 N 2 O achieved a 63 Ni sensitivity of 5310 cps (ng −1 g −1 ) (2.49 cps (Bq g −1 )) and an instrument detection limit of 0.20 pg g −1 (0.42 Bq g −1 ), but a lower separation factor of 4960. Despite the slightly lower sensitivity and higher instrument detection limits obtained compared to the existing NH 3 /H 2 /He approach, the factor of 10 to 100 times greater separation factors achieved using N 2 O provides a signicant advantage. Moreover, samples containing elevated levels of Al may pose an additional challenge, as 36 Ar 27 Al + interferes on m/z63 and, in this study, was observed to form at a rate of 11 100 cps (mg −1 g −1 ) Al in the absence of a cell gas. When applying N 2 O at the optimum conditions, the rate of interference became negligible at <0.1 cps (mg −1 g −1 ) Al. When applying an N 2 O/NH 3 gas mixture, both Cu and Ni were observed to form product ions at +17 amu, +34 amu, and +51 amu (Fig. 2). While these 17 amu increments could correspond to either 16 O 1 Hor 14 N 1 H 3 ligands, it is likely that the 14 N 1 H 3 is forming, as the product ion formation rates aligned with previous literature using only NH 3 as a cell gas. 31 Cu formed the Cu(NH 3 ) 2+ at the highest rate (maximum of 7.4% at 0.4 mL min −1 N 2 O and 1.0 mL min −1 NH 3 ), while Ni was observed to form Ni(NH 3 ) 3+ at the highest rate (maximum of 4.4% at 0.2 mL min −1 N 2 O and 1.0 mL min −1 NH 3 ). By utilizing the Ni(NH 3 ) 3+ product ion with a gas mixture of 0.6 mL min −1 N 2 O and 0.4 mL min −1 NH 3 , a similar sensitivity of 3440 cps (ng −1 g −1 ) (equivalent to 1.62 cps (Bq g −1 )) could be achieved, however with a separation factor of 254 (Table S2) –approximately 50 times less efficient than when using N 2 O alone. Moreover, at these ow rates, slightly higher instrument detection limits of 0.58 ng g −1 (1.2 Bq g −1 ) were obtained, and the removal of the 36 Ar 27 Al + interference was less efficient (1.9 cps (mg −1 g −1 ) Al) than when using N 2 O alone. Hence, it would be recommended that the N 2 O approach would be optimal for determinations of 63 Ni. Selenium-79 79 Se (t 1/2 =7.6 ×10 4 years) is a product of 235 Ussion that is important to characterise during decommissioning due to its high environmental mobility. Given its long half-life, ICP-MS is well suited to the determination of the 79 Se radioisotope. In Fig. 1 (A) Calculated sensitivity of the 63 Ni radionuclide determined as the 63 Ni 16 O + product ion (blue, circle) and the obtained sensitivity of the 63 Cu interference determined as the 63 Cu 16 O + product ion (orange, square), and (B) the separation factor of 63 Ni/ 63 Cu. 3214 |J. Anal. At. Spectrom.,2025,40,3210–3220 This journal is © The Royal Society of Chemistry 2025 JAAS Paper Open Access Article. Published on 07 October 2025. Downloaded on 11/5/2025 9:49:59 AM. This article is licensed under a Creative Commons Attribution 3.0 Unported Licence. View Article Online addition to isobaric interferences from the stable 79 Br isotope, argon-based polyatomic interferences ( 40 Ar 38 Ar 1 H + , 38 Ar 21 H + , and peak tailing of 40 Ar 2+ ) present an additional challenge for the analysis of 79 Se by ICP-MS/MS. A recent publication utilized NH 3 gas with ICP-MS/MS to reduce 79 Br and argon-based interferences by charge transfer, following a chemical separation step. The authors achieved an instrument limit of detection of 1.2 pg mL −1 (5.7 ×10 −4 Bq mL −1 ). 32 The authors additionally reported a separation factor of 7 orders of magnitude using their combined chemical separation and ICP-MS/MS methodology. Here, bromine was observed to react very efficiently with N 2 O to form the BrO + product ion, with a maximum product ion formation of 77% at 0.6 mL min −1 . Selenium was observed to react with N 2 O to form the SeO + product ion, however the maximum product ion formation was 25% at a cell gas ow rate of 1.8 mL min −1 . At lower ow rates, the Se signal remained primarily on-mass, with a maximum of 96% signal intensity (relative to no cell gas) at an N 2 Oow rate of 0.6 mL min −1 .By utilizing the on-mass determination of Se, separation of 79 Se from the interfering 79 Br can be achieved. A separation factor of 188 000 was achieved in this study using 2.0 mL min −1 N 2 O, with a calculated sensitivity for 79 Se of 3140 cps (ng −1 g −1 ) (6670 cps (Bq g −1 )) (Fig. 3), with instrument detection limits of 0.51 pg g −1 (2.4 ×10 −4 Bq g −1 ). Higher sensitivities could be achieved by reducing the N 2 Oow rate, although at the cost of less efficient 79 Br interference removal. The argon-based polyatomic interferences were no longer observed at N 2 Oow rates of 1.2 mL min −1 and above. At 1.2 mL min −1 , the 79 Se sensitivity obtained was 7520 cps (ng −1 g −1 ) (16 000 cps (Bq g −1 )), with an instrument detection limit of 0.38 pg g −1 (1.8 ×10 −4 Bq g −1 ). Thus, the use of N 2 O provides an alternative cell-gas approach to the recently developed NH 3 approach. Applying the N 2 O/NH 3 gas mixture improved interference removal at lower ow rates, allowing for determinations of 79 Se at greater sensitivities. By applying 0.4 mL min −1 N 2 O and 0.2 mL min −1 NH 3 , a calculated sensitivity of 13 100 cps (ng −1 g −1 ) (equivalent to 27 800 cps (Bq g −1 )) was achieved with an improved interference separation factor of 4 600 000 (Table S3) and a calculated instrument detection limit of 0.17 pg g −1 (8.0 × 10 −5 Bq g −1 ). This is likely due to an efficient charge transfer reaction between 79 Br + and NH 3 . The use of NH 3 has been shown impede the sensitivity of Se at higher ow rates (by a charge transfer reaction). 33 In this case however, the low ow of NH 3 enhanced the removal of the 79 Br interference, but was not a high enough ow to impact the sensitivity of 79 Se. Moreover, it is possible that the N 2 O acts to enhance the sensitivity of 79 Se on-mass through the collisional focussing effect. 13 Therefore, the use of the N 2 O/NH 3 mixture offers an alternative approach for improved sensitivity and lower instrument detection limits, as well as a high interference separation factor of 6– 7 orders of magnitude that can compete with the chemical separation approaches. Strontium-90 Another ssion product of high concern is 90 Sr (t 1/2 =28.80 years), which is also known to be mobile in nature. Although this radionuclide can be determined using LSC following chemical separation, 34 ICP-MS determinations are growing more common due to advances in instrument sensitivity. The standard approach for interference removal for 90 Sr determinations by ICP-MS or ICP-MS/MS is the use of oxygen as a reaction gas. 35,36 In this approach, oxygen reacts efficiently with the interfering 90 Zr + (forming 90 Zr 16 O + and 90 Zr 16 O 2+ ), as well as 89 Y 1 H + , but does not react efficiently with 90 Sr, thus Fig. 2 Product ion formations of Ni (blue) and Cu (orange) with 0.6 mL min −1 N 2 O and 0.4 mL min −1 NH 3 applied as a reaction gas mixture. Fig. 3 (A) Calculated sensitivity of the 79 Se radionuclide determined on-mass (blue, circle) and the obtained sensitivity of the 79 Br interference determined on-mass (orange, square), and (B) the separation factor of 79 Se/ 79 Br. This journal is © The Royal Society of Chemistry 2025 J. Anal. At. Spectrom.,2025,40,3210–3220 | 3215 Paper JAAS Open Access Article. Published on 07 October 2025. Downloaded on 11/5/2025 9:49:59 AM. This article is licensed under a Creative Commons Attribution 3.0 Unported Licence. View Article Online allowing for an on-mass determination of 90 Sr. This established method is reported to provide instrument detection limits of 4 fg g −1 (0.02 Bq g −1 ). 36 The use of N 2 O provided separation of 90 Sr from 89 Y 1 H + interference by forming the 90 Sr 16 O + product ion (m/z=106), with the 89 Y 16 O 1 H + interference forming at <0.3 cps (ng −1 g −1 ) (of 89 Y). However, the 90 Zr interference also reacted with N 2 Oto form 90 Zr 16 O + . Higher ow rates of N 2 O lead to a higher separation of 90 Sr from 90 Zr, however at the cost of lower sensitivity (Fig. 4). Using only N 2 O, the maximum achieved separation factor for the removal of 90 Zr on 90 Sr was 10 600 at 3 mL min −1 , with a sensitivity of 2870 cps (ng −1 g −1 ) (equivalent to 0.562 cps (Bq g −1 )) for 90 Sr. The instrument detection limit under these conditions was calculated to be 0.27 pg g −1 (equivalent to 1.4 Bq g −1 ), which does not provide an advantage over the current best available methodology. Applying an N 2 O/NH 3 gas mixture and using the 90 Sr 16 O 1 H + product ion (m/z=107) provided a much greater separation from 90 Zr, with <0.01% product ion formation of Zr at +17 amu. The maximum separation factor of 334 000 was achieved at gas ow rates of 0.8 mL min −1 N 2 O and 0.5 mL min −1 NH 3 (Table S4). Product ion scans indicated that Zr preferentially formed higher-order product ions (primarily at +82 amu to +84 amu) using the N 2 O/NH 3 gas mixture, allowing for more effective removal on m/z107. Moreover, interferences from polyatomic 89 Y 1 H + and peak tailing of 89 Y + were observed to be more effectively reduced using the N 2 O/NH 3 gas mixture. By reducing the NH 3 ow rate to 0.2 mL min −1 and maintaining a N 2 Oow rate of 0.8 mL min −1 , a sensitivity of 108 000 cps (ng −1 g −1 ) (equivalent to 21.1 cps (Bq g −1 )) for 90 Sr was obtained (SrOH product ion formation of 38%), while the 90 Sr/ 90 Zr separation factor remained >100 000 (Table S4), giving a much greater performance than using N 2 O alone. The instrument detection limit for 90 Sr at these gas ow rates was calculated to be 0.015 pg g −1 (0.076 Bq g −1 ), which is similar to the detection limits achieved by the oxygen cell gas approach 35,36 and thus makes the N 2 O/NH 3 gas mixture approach a viable alternative for interference removal. Zirconium-93, molybdenum-93, and niobium-94 93 Zr (t 1/2 =1.61 ×10 6 years), 93 Mo (t 1/2 =4000 years), and 94 Nb (t 1/2 =2.004 ×10 4 years) are produced through neutron activation of stable 92 Zr (17.16% abundance), 92 Mo (14.65% abundance), and 93 Nb (100% abundance), respectively, that are contained within steel casings of nuclear reactors. Additionally, 93 Zr is also a high-yield ssion product in spent nuclear fuel. Following chemical separations, 93 Zr and 93 Mo can be determined by LSC, 37,38 whereas 94 Nb can be determined by gammaor X-ray spectrometry. 39 Given the long halflives, ICP-MS determinations are well suited to the analysis of these radionuclides. However, 93 Zr and 93 Mo share isobars with each other as well as 93 Nb, which provides an additional challenge as 93 Nb is the only stable isotope of Nb. ICP-MS/MS has previously been demonstrated to be an effective approach to analysing 93 Zr and 93 Mo. Petrov et al. 9 reported that an NH 3 / H 2 /He gas mixture could be utilized to effectively remove interferences of 93 Mo and 93 Nb on 93 Zr by measuring the 93 Zr(NH 3 ) 6 product ion, with separation factors of 10 5 achievable from both interferences. The authors report a sensitivity of approximately 20 000 cps (ng −1 g −1 ) (230 000 cps (Bq g −1 )) and a detection limit of 0.14 pg g −1 (1.3 ×10 −5 Bq g −1 )for 93 Zr. Russell et al. 30 subsequently demonstrated that the same NH 3 / H 2 /He gas mixture could be utilized to analyse 93 Mo using the 93 Mo(NH 3 ) 2 product ion. The authors reported a sensitivity of 1100 cps (ng −1 g −1 ) (31 cps (Bq g −1 )) and a detection limit of 45.6 pg g −1 (1.6 Bq g −1 ). The 94 Nb radionuclide has been much less studied using ICP-MS/MS compared to 93 Zr and 93 Mo and, to the authors knowledge, there is no literature on the best available reaction gas conditions. Nevertheless, as the 94 Nb radionuclide shares isobars with stable 94 Zr and 94 Mo isotopes, it has also been assessed using N 2 OandaN 2 O/NH 3 gas mixture in this study. Zr and Nb react with N 2 O to form primarily oxide (+16 amu) and dioxide (+32 amu) product ions, whereas Mo reacts very little with N 2 O. 13 Therefore, the removal of stable 93 Nb interference, as well as interferences from the 93 Zr radionuclide, using an on-mass determination of 93 Mo would be possible. Using a ow rate of 1 mL min −1 N 2 O, 93 Mo could be separated from 93 Nb and 93 Zr by factors of 750 and 1160 respectively, while providing a calculated sensitivity of 126 000 cps (ng −1 g −1 ) (equivalent to 3540 cps (Bq g −1 )) (Fig. 5). The instrument detection limit was calculated to be 0.017 pg g −1 (6.1 ×10 −4 Bq g −1 ). Additionally, the 93 Zr and 94 Nb radionuclides can be Fig. 4 (A) Calculated sensitivity of the 90 Sr radionuclide determined as the 90 Sr 16 O + product ion (blue, circle) and the obtained sensitivity of the 90 Zr interference determined as the 90 Sr 16 O + product ion (orange, square), and (B) the separation factor of 90 Sr/ 90 Zr. 3216 |J. Anal. At. Spectrom.,2025,40,3210–3220 This journal is © The Royal Society of Chemistry 2025 JAAS Paper Open Access Article. Published on 07 October 2025. Downloaded on 11/5/2025 9:49:59 AM. This article is licensed under a Creative Commons Attribution 3.0 Unported Licence. View Article Online separated from the 93 Mo radioisotope and the stable 94 Mo isotope respectively by mass-shiing and analysing the dioxide product ion. However, as Zr and Nb react very similarly with N 2 O, 13 these radionuclides cannot be separated from their stable isotope counterparts. Applying an N 2 O/NH 3 gas mixture, Mo was, again, found to react relatively little with the cell gas and most of the ions were transmitted on-mass. The Mo(OH) 2 product ion was formed at the highest rate (maximum of 3.2% formation). Nb and Zr preferentially reacted to form higher order product ions, primarily around +82 amu to +84 amu (Fig. 6). Both 14 N 1 H 3 and 16 O 1 H shithe mass by +17 amu, and can form ligands with and without the hydrogen atoms present (i.e. as MO + or MNH 2+ ), therefore it is difficult to predict exactly which higherorder complexes are being formed at these higher masses. Nevertheless, the distinct differences in reactivity between Mo and the interfering Nb and Zr isobars mean that separation is possible. However, the addition of NH 3 resulted in a slightly lower sensitivity. Using 0.6 mL min −1 N 2 O and 0.4 mL min −1 NH 3 achieved equivalent separation factors compared to using 1 mL min −1 N 2 O, with a calculated sensitivity for 93 Mo of 105 000 cps (ng −1 g −1 ) (equivalent to 2950 cps (Bq g −1 )) (Table S5), 17% lower than using N 2 O alone. However, the instrument detection limit was calculated to be 0.015 pg g −1 (5.5 ×10 −4 Bq g −1 ), which was similar to that obtained by using N 2 O alone. Further improvements to the removal of Zr and Nb could be achieved by varying the N 2 O and NH 3 ow rates, but at the cost of lower sensitivities for 93 Mo (Table S5). Separation of Zr and Nb was found to be possible using the N 2 O/NH 3 mixture, however with low separation factors. The product ions of Nb and Zr at +82 amu and +84 amu were found to form at different rates depending on the composition of the N 2 O/NH 3 mixture. For 93 Zr, the greatest separation factor achieved for the removal of 93 Nb was 80 by observing the product ion at +84 amu and using a cell gas mixture of 0.2 mL min −1 N 2 Oand1mLmin −1 NH 3 (Table S6). The high ammonia gas ow rate resulted in a relatively low sensitivity of 7670 cps (ng −1 g −1 ) (86 700 cps (Bq g −1 )), equivalent to 2.6% product ion formation. The instrument detection limit for 93 Zr was calculated to be 0.27 pg g −1 (2.4 ×10 −5 Bq g −1 ). For 94 Nb, the greatest separation factor achieved for the removal of 94 Zr was 25 by observing the product ion at +83 amu and using 1 mL min −1 N 2 Oand0.1mLmin −1 NH 3 (Table S7). In this case, the ow rates also corresponded to the maximum observed product ion formation of 9.6%, equivalent to a sensitivity of 19 900 cps (ng −1 g −1 ) (2830 cps (Bq g −1 )) for 94 Nb. The instrument detection limit for 94 Nb was calculated to be 0.23 pg g −1 (0.0016 Bq g −1 ). An additional concern for determinations of 94 Nb is the 40 Ar 54 Fe + interference in samples with elevated Fe content. In the absence of a cell gas, the interference formed at a rate of 82 cps (mg −1 g −1 )Fe.Attheow rates of 1 mL min −1 N 2 Oand0.1mLmin −1 NH 3 , the interference was reduced to 0.7 cps (mg −1 g −1 ) Fe. However, given the low separation factors and sensitivity achieved for 93 Zr and 94 Nb, the use of the N 2 O/ NH 3 gas mixture may have limited application for these radionuclides. Palladium-107 107 Pd (t 1/2 =6.5 ×10 6 years) is a ssion product found in spent nuclear fuel. Determinations by LSC are possible for 107 Pd, however given its very long half-life, ICP-MS determinations provide greater sensitivity compared to radiometric techniques. 40 For ICP-MS determinations, the radionuclide 107 Pd shares isobaric interferences with stable 107 Ag. Literature utilizing reaction gases for interference separation of 107 Pd is Fig. 5 (A) Calculated sensitivity of the 93 Mo radionuclide (blue, circle), the calculated sensitivity of the interfering 93 Zr radionuclide (grey, triangle) and the obtained sensitivity of the stable 93 Nb interference (orange, square), and (B) the separation factor of 93 Mo/ 93 Nb (blue, circle) and 93 Mo/ 93 Zr (orange, square). Fig. 6 Product ion formations of Zr (blue) and Nb (orange) under two different reaction gas mixtures: 0.2 mL min −1 N 2 O and 1.0 mL min −1 NH 3 (left); and 1.0 mL min −1 N 2 O and 0.2 mL min −1 NH 3 (right). This journal is © The Royal Society of Chemistry 2025 J. Anal. At. Spectrom.,2025,40,3210–3220 | 3217 Paper JAAS Open Access Article. Published on 07 October 2025. Downloaded on 11/5/2025 9:49:59 AM. This article is licensed under a Creative Commons Attribution 3.0 Unported Licence. View Article Online limited, with studies focussing on improving chemical separations prior to ICP-MS detection. 41 In combination with chemical separation, Weller et al. 40 utilized a propane/He gas mixture to separate 107 Pd from its isobaric interference by monitoring the Pd(C 2 H 2 ) product ion, resulting in detection limits of <2 pg g −1 (<4 ×10 −5 Bq g −1 ). The product ion formation was reported to be approximately 4%, which limits the available sensitivity, however the authors also report that similar detection limits were achieved for 107 Pd in the absence of a reaction gas. Both Pd and Ag react very little with N 2 O, with less than 1% product ion formation. 13 However, Ag does not form AgO + product ions (instead, favouring AgN 2 O + formation), whereas Pd does form PdO + product ions. 13 Therefore, determinations of the 107 Pd radionuclide may be possible. However, in this study, the maximum formation rate of the PdO + product ion was 0.15% at 0.7 mL min −1 N 2 O, equating to a sensitivity of 164 cps (ng −1 g −1 ) (8620 cps (Bq g −1 )) for 107 Pd. The calculated instrument detection limit was 14 pg g −1 (2.7 ×10 −4 Bq g −1 ). N 2 O shows potentially limited use for determinations of 107 Pd due to the low achievable sensitivities. Applying the N 2 O/NH 3 gas mixture, Pd and Ag reacted to form product ions at +17 amu, +34 amu, and +51 amu. While these 17 amu increments could correspond to either 16 O 1 Hor 14 N 1 H 3 ligands, it is likely that the 14 N 1 H 3 is forming, as the product ion formation rates aligned with previous literature using only NH 3 as a cell gas. 31 The Ag(NH 3 ) 3+ product ion formed at a lower rate than that of Pd(NH 3 ) 3+ . Using cell gas ow rates of 0.6 mL min −1 N 2 O and 0.8 mL min −1 NH 3 , a separation factor of 84 could be achieved, with a sensitivity of 3300 cps (ng −1 g −1 ) (173 000 cps (Bq g −1 )) (Table S8 and Fig. 7). The calculated instrument detection limit under these gas conditions was 1.1 pg g −1 (2.0 ×10 −5 Bq g −1 ). While the achieved sensitivity is 20 times greater than using N 2 O alone, the use of N 2 O alone achieved detection of 107 Pd (as PdO + ) free from 107 Ag interference, since the AgO + product ion was not formed. Therefore, depending on the mass fraction of Ag present in the matrix, it may be preferable to use the N 2 O approach over the N 2 O/NH 3 gas mixture approach. Caesium-135/137 135 Cs (t 1/2 =1.33 ×10 6 years) and 137 Cs (t 1/2 =30.018 years) are important ssion products used for environmental monitoring. Specically, 135 Cs/ 137 Cs ratios have been used as a forensic tool to identify sources of radioactive contamination, for example following the Fukushima nuclear disaster. 16 ICPMS/MS methodology is already established for this application, where N 2 O has already been used as a cell gas to separate 135 Cs + and 137 Cs + from the stable isobaric interferences of 135 Ba + and 137 Ba + respectively. 14–16 By this approach, Ba is massshied and the Cs can be determined on-mass. Additionally, Magre et al. 18 recently described the use of an N 2 O/NH 3 gas mixture for the determination of 135 Cs/ 137 Cs ratios with improved interference removal. The authors noted a sensitivity of 110 000 cps (ng −1 mL −1 ) (equivalent to 1 500 000 cps (Bq g −1 ) for 135 Cs and 34 cps (Bq g −1 )for 137 Cs) under wet-plasma conditions, with a factor of 3 times improvement using an Apex Udesolvating system. The authors reported instrument blank levels of <0.6 cps. Under optimum conditions, including the use of the desolvating system, the authors achieved detection limits of 1.66 fg g −1 (1.2 ×10 −7 Bq g −1 )for 135 Cs and 0.67 fg g −1 (0.0022 Bq g −1 )for 137 Cs. The N 2 O/NH 3 gas mixture was applied to the separation of Cs radionuclides from isobaric Ba interferences. At cell gas ow rates of 1 mL min −1 N 2 O and 0.1 mL min −1 NH 3 , separation factors of 223 000 for 135 Cs/ 135 Ba and 131 000 for 137 Cs/ 137 Ba were obtained, with sensitivities of 101 000 cps (ng −1 g −1 ) for both radioisotopes (Table S9) (equivalent to 1 370 000 cps (Bq g −1 ) for 135 Cs and 31.4 cps (Bq g −1 ) for 137 Cs). The instrument blank was found to be 0.7 cps for both 135 Cs and 137 Cs. The instrument detection limit under these conditions was calculated to be 0.017 pg g −1 (equivalent to 1.3 ×10 −6 Bq g −1 for 135 Cs and 0.056 Bq g −1 for 137 Cs). Compared to the use of N 2 O only (optimum ow rate =1 mL min −1 ), the N 2 O/NH 3 provided a factor of 2 times greater interference separation, while maintaining an equivalent sensitivity. Given the achieved sensitivity and blank levels, the results in this study are in good agreement with those produced by Magre et al., 18 though with higher observed detection limits due to the absence of a desolvating system. Conclusions This study demonstrates the promising applicability of N 2 O for the removal of isobaric interferences for radionuclides of interest in nuclear decommissioning, especially for 63 Ni, 79 Se, and 93 Mo. Moreover, the great power of combining two reactive cell gases to open new interference removal pathways using new product ions is highlighted. In particular, enhanced interference removal for the group 2 radionuclides ( 41 Ca and 90 Sr), as well as for 79 Se, 93 Mo, and 107 Pd were achieved using a novel N 2 O/NH 3 gas mixture approach. Hence, it can be recommended that both N 2 OandNH 3 are employed during ICP-MS/MS analysis. Such ndings may also be applicable to stable isotope analysis, therefore further investigation is required into the use of the N 2 O/NH 3 gas Fig. 7 Product ion formations of Pd (blue) and Ag (orange) with 0.6 mL min −1 N 2 O and 0.8 mL min −1 NH 3 applied as a reaction gas mixture. 3218 |J. Anal. At. Spectrom.,2025,40,3210–3220 This journal is © The Royal Society of Chemistry 2025 JAAS Paper Open Access Article. Published on 07 October 2025. Downloaded on 11/5/2025 9:49:59 AM. This article is licensed under a Creative Commons Attribution 3.0 Unported Licence. View Article Online