Evaluation of Inductively Coupled Plasma Tandem Mass Spectrometry for Determination of As in Agricultural Inputs with High REE Contents

In the present work, an inductively coupled plasma-tandem mass spectrometer (ICP-MS/MS) operating in single quadrupole and mass shift modes were evaluated for As determination in mineral fertilizers and agricultural gypsum with high contents of rare earth elements. For MS/MS mode, oxygen was the gas introduced into the octopole reaction system (ORS). Samples were prepared by microwave-assisted digestion with diluted nitric acid solution. The accuracy was checked by analysis of a fertilizer certified reference material (NIST SRM 695) and by addition and recovery experiments. When operating in single quadrupole mode, recoveries ranged from 59 to 151%; while values obtained by MS/MS mode varied from 81 to 105% when 0.30 mL min O2 was introduced into the ORS. Limits of detection for As in single quadrupole and AsO in MS/MS mass shift mode were 6 and 9 ng L, respectively.


Introduction
The growth of food production over the years contributed to increasing the application of agricultural inputs to soils.In addition to nutrients, agricultural inputs may also be a source of toxic elements as As, Cd, Cr, Hg and Pb at trace levels, which can accumulate in the soil and enter into the food chain, therefore their concentrations need to be monitored. 1The toxic effects of inorganic arsenic exposure on human health are well reported in the literature and some of the negative secondary effects on lung, liver, skin and bladder cancers, diabetes, cardiovascular and neurological diseases are among them. 2he minerals fertilizers (i.e., phosphate fertilizers) and soil conditioners as agricultural gypsum, a coproduct of phosphate fertilizers processing plants, are inputs that contain macronutrients (Ca, Mg, N, P, and S), micronutrients (such as Fe and Si), rare earth elements (REEs, such as Ce, Dy, Er, Eu, Gd, Ho, La, Lu, Nd, Pr, Sc, Sm, Tb, Tm Y, and Yb) and, in some cases, toxic elements (As, Cd, Hg and Pb) in their composition.][5] Inductively coupled plasma mass spectrometry (ICP-MS) has been employed for trace elements determination, such as As, due to its high sensitivity, multielemental capability indicating the possibility of carrying out the determination of several elements in the same run, wide linear range and lower limits of detection. 6,7[10] In addition to polyatomic interferences, double charged species may also affect the accurate determination of As.Some of these species, such as 150 Nd 2+ and 150 Sm 2+ , are easily produced in plasma when the sample contains REEs which can overlap the As + isotope at m/z 75. 8,9Correction equations, 11 collision and reaction interface (CRI), 12 and dynamic reaction cell (DRC) 13 are useful strategies to overcome these interferences.][16][17][18][19] Amaral et al. 20 verified the effectiveness of ICP-MS/MS with O 2 into the ORS 3 for the elimination of interferences caused by chloride ions 40 Ar 35 Cl + , 35 Cl 16 OH + and 35 Cl 16 O + on As + , Cr + and V + determination, respectively.The analytes were determined by their oxides with good accuracy, precision and sensitivity.Flórez et al. 21exploited a mixture of CH 3 F/He (1/9) as reaction gas in single quadrupole ICP-MS to determine As from monitoring at ratio m/z 89 (by reaction product AsCH 2 + between As + and CH 3 F) in biological fluids that have high Cl concentration.Bolea-Fernandez et al. 22 determined As and Se interference free by ICP-MS/MS using CH 3 F (a mixture of 10% CH 3 F and 90% He) as reaction gas, via AsCH 2 + and SeCH 2 + reaction products in different plant, animal and environmental reference materials.
Jackson et al. 23 used an ICP-MS/MS configuration with collision (He) and reaction (H 2 ) gases pressurized into the ORS 3 in single quadrupole mode and O 2 in MS/MS mass shift mode to determine As and Se in two certified reference materials (CRM) of plant tissues, both with high REEs contents.The authors noticed that the formation of AsO + and SeO + and the rejection of non-target ions in MS/MS mode were required in order to gain accuracy.In another study, Jackson 24 employed an ion chromatography-ICP-MS/MS mass shift mode to determine arsenocholine (AsC), arsenobetaine (AsB), dimethylarsine (DMA), monomethylarsine (MMA), As III and As V in CRMs of urine samples and the method developed provided better detection limits, high sensitivity and sample throughput when compared to others methods.
Inorganic arsenic was determined by hydride generation (HG-ICP-MS) and high performance liquid chromatography (HPLC-ICP-MS) in rice product samples using single quadrupole and MS/MS mass shift mode, with O 2 pressurized into the ORS 3 .Pétursdóttir et al. 25 compared the results obtained by both methods and suggested that HPLC-ICP-MS can be replaced by simple HG system which provide As concentrations in the sample extract from low µg L -1 levels.
Many countries regulate the maximum limit for As concentration in agricultural inputs.In Canada, the Federal Fertilizers Act regulates the maximum concentration as 75 mg kg -1 .In the United States, this concentration changes in different states, ranging from 13 to 112 mg kg -1 depending on the composition of inputs.][28] In the present work, the performance of an ICP-MS/MS operating in both single quadrupole and MS/MS mass shift mode, using oxygen as a reaction gas, was evaluated to overcome REEs based interferences on As determination in mineral fertilizers and agricultural gypsum samples.

Experimental
Reagents, solutions and samples All materials employed were decontaminated by immersion in 10% (v/v) HNO 3 for at least 24 h and rinsed with distilled-deionized water (resistivity > 18.2 MΩ cm), obtained from a Milli-Q  Water System (Millipore, Billerica, MA, USA).Standard reference solutions and spiked samples were prepared from a 1000 mg L -1 of As (Qhemis, São Paulo, SP, Brazil) stock solution.For sample digestion, HNO 3 obtained from a sub-boiling distillation system (Milestone, Sorisole, Italy) and H 2 O 2 30% (m/m) (Qhemis) were used.
Samples of mineral fertilizers and agricultural gypsum from different regions were supplied by the National Laboratory for Agriculture (Goiânia, GO, Brazil) and Embrapa Southeast Livestock (São Carlos, SP, Brazil).A fertilizer certified reference material, SRM 695, from the National Institute of Standards and Technology (NIST, Gaithersburg, MD, USA) was used to check the accuracy.

Instrumentation
All measurements were carried out in an Agilent 8800 inductively coupled plasma tandem mass spectrometer ICP-MS/MS (Agilent Technologies, Tokyo, Japan) equipped with ORS 3 located between two quadrupole mass analyzers.Liquid argon was provided by White Martins (Sertãozinho, SP, Brazil).Pure oxygen (≥ 99.999%, Air Products, São Paulo, SP, Brazil) was used as cell gas.The single quadrupole and MS/MS mass shift modes were evaluated.The operational parameters are presented in Table 1.Samples were digested using an Ethos 1 microwave oven (Milestone) and were analyzed in triplicate.

Sample preparation
Mineral fertilizers and agricultural gypsum sample masses (ca.200 mg) were accurately weighed and aliquots of 6 mL of HNO 3 7 mol L -1 and 2 mL of H 2 O 2 30% (m/m) were added to the vessels.The microwave oven two-step heating program was applied as follows: (i) 20 min ramp to 200 ºC; (ii) 20 min plateau at 200 ºC.After cooling down, the digested samples were transferred to polypropylene flasks and volumes were made up to 50.0 mL with distilleddeionized water.All samples were digested in triplicate.

Double charged ion monitoring and determination of elements at m/z 150 in agricultural inputs
The As + isotope at m/z 75 is overlapped by double charged species formed by 150 Sm 2+ or 150 Nd 2+ present in mineral fertilizer, agricultural gypsum and SRM 695 samples, which can cause an overestimation of the 75 As + concentrations.To check the presence of interfering species, the net signal intensities at m/z 150 were monitored in the blank solution 1% (v/v) HNO 3 and in the samples using single quadrupole mode (Figure 1).As it can be seen, the signal intensities of agricultural gypsum samples are 7 to 400-fold higher than those observed for mineral fertilizers and SRM 695, respectively.
In order to check if the high intensities observed at m/z 150 were originated from REE ions or oxide ions as 132 Ba 18 O + and 134 Ba 16 O + , since Ba is present at high concentrations, the concentrations of 150 Nd + and 150 Sm + in the samples were determined in the single quadrupole and MS/MS mass shift modes as 150 Nd 16 O + and 150 Sm 16 O + by introducing 0.30 mL min -1 O 2 into the ORS 3 .Results are shown in Table 2.It can be observed that in all samples, the determined concentrations for elements at m/z 150 ( 150 Nd + or 150 Sm + ) were higher than 100 mg kg -1 and, for agricultural gypsum samples, concentrations 8 to 11-fold higher than in mineral fertilizers were obtained.No statistical differences were found at a 95% confidence level (t-test) when both operation modes were considered.Addition and recovery experiments were performed for evaluating the accuracy of the measurements at m/z 150 by adding 40 µg L -1 of Nd and Sm to fertilizer and gypsum digests.Recoveries were between 82 to 89% and 90 to 93% when operating in single quadrupole and MS/MS mass shift modes, respectively.
Effect of double charged ions 150 Sm 2+ and 150 Nd 2+ on 75 As + analytical signal intensities The effects of 150 Sm 2+ and 150 Nd 2+ on 75 As + net signal intensities were evaluated in solutions containing 1.0 µg L -1 As in the presence of 0, 5, 10, 20, 40, 70 and 100 µg L -1 Nd and Sm (Figure 2).Signal intensities were obtained in single quadrupole ( 75 As + ) and MS/MS mass shift modes ( 75 As 16 O + , 0.30 mL min -1 into the ORS 3 ) and normalized to 1 when Nd and Sm were absent.It is important to emphasize that in the case of single quadrupole measurements, 75 As + was the monitored specie while in MS/MS mass shift mode, oxides were easily formed into the ORS 3 when O 2 was introduced, thus 75 As 16 O + was selected for measurements at the second quadrupole.
As can be seen in Figure 2, when Nd and Sm concentrations increased, the signal at m/z 75 in single quadrupole also increased up to 60% due to spectral interferences caused by double charged species.When the   MS/MS mass shift mode was applied, the observed gain was lower than 7%, thus evidencing that the addition of oxygen into the ORS 3 cell, the mass shift from 75 As + to 75 As 16 O + and the selection of the oxide specie in the second quadrupole (Q 2 ) are efficient strategies to eliminate double charged ions interferences caused by REEs.It is also important to consider that possible interfering species, such as 91 Zr + , at the same mass/charge ratio of 75 As 16 O + are rejected in the first quadrupole (Q 1 ) by allowing only m/z 75 ions to pass through.

Analytical performance and accuracy
The analytical performance of ICP-MS/MS for As determination was evaluated for 75 As + (single quadrupole mode) and 75 As 16 O + (MS/MS mass shift mode).Limits of detection and sensitivities for As in single quadrupole and MS/MS mass shift modes are shown in Table 3.When oxygen gas was introduced into the ORS 3 , it was observed a sensitivity reduction and higher limit of detection (LOD) values when compared to single quadrupole mode in the absence of cell gas, probably due to the ion beam defocusing and the efficiency of the reaction processes.Despite the lower background signals obtained in MS/MS mode, the effect on sensitivity is even more pronounced when the oxygen flow rate was increased from 0.30 to 0.50 mL min -1 .
The ICP-MS operated in single quadrupole mode and MS/MS mass shift mode (with different oxygen flow rates into the ORS 3 ) was evaluated for correction of interferences at m/z 75.The accuracy for As determination in mineral fertilizer samples was evaluated using NIST SRM 695 (Table 4).Using the unpaired t-test, no significant differences between the determined and certified values at a 95% confidence level were observed and the recoveries found varied from 93 to 95% in all cases.It can be inferred that no significant improvement in accuracy was achieved using MS/MS mass shift mode with O 2 into the ORS 3 , which can be explained considering the low net signal intensity for m/z 150 (Figure 1) in this particular material.
In order to evaluate the accuracy in samples containing higher contents of elements that produce interfering double charged species, addition and recovery experiments for As at two levels (Table 4) were performed for agricultural gypsum and fertilizer samples.Recoveries obtained in single quadrupole mode ranged from 59 to 151%, while values obtained by MS/MS mass shift mode varied from 81 to 105% and 81 to 107% for O 2 flow rates of 0.30 and 0.50 mL min -1 , respectively.Recoveries observed in single quadrupole mode presented deviations higher than 20%, which are related to overlaps caused by double charged ions on 75 As + .
The As concentrations in minerals fertilizers and agricultural gypsum were determined in different instrumental conditions (Table 5).In single quadrupole mode determined concentrations are higher than those determined by MS/MS mass shift mode and this difference is more pronounced in agricultural gypsum samples, which present concentrations of elements at m/z 150 ( 150 Nd + and 150 Sm + ) up to 10-fold higher than those found in fertilizers.A paired t-test was used to check the agreement of results obtained by two oxygen gas flow rates (0.30 and 0.50 mL min -1 ) and no statistical differences were found at a 95% confidence level.It was possible to verify the effectiveness of MS/MS mass shift mode using O 2 into the ORS 3 to overcome double charged ions interferences on 75 As + determination.

Conclusions
The use of MS/MS mass shift mode was effective to remove double charged interferences from REEs on As determination in mineral fertilizers and agricultural gypsum  As 16 O + ions by addition of oxygen gas into the ORS 3 cell is one of the main advantages of this instrument configuration, and allows the accurate determination of As in samples with high contents of REEs, such as 150 Nd 2+ and 150 Sm 2+ .It can cause false positive errors on As determination causing erroneous decisions about commercialization and application of agricultural products, which is critical considering environmental effects and legislations established for controlling these products.

Figure 2 .
Figure 2. Relative signal intensities on m/z 75 for 1 µg L -1 As and increasing concentrations of Sm and Nd in single quadrupole and MS/MS mass shift modes.

Table 2 .
Determination of 150 Nd + and 150 Sm + (mean ± standard deviation, n = 3) in mineral fertilizers and agricultural gypsum in single quadrupole and MS/MS mass shift modes (0.30 mL min -1 O 2 into the ORS 3 ) MS/MS: tandem mass spectrometry.

Table 1 .
Operating conditions for the Agilent 8800 ICP-MS/MS

Table 3 .
Limits of detection (LOD) and sensitivities for75As + in single quadrupole and MS/MS modes with different oxygen gas flow rates a Defined as the slope of analytical calibration curves (counts L s -1 µg -1 ).

Table 5 .
Determination in mg kg -1 of As (mean ± standard deviation, n = 3) in mineral fertilizers and agricultural gypsum using single quadrupole and MS/MS mass shift modes with different oxygen gas flow rates

Table 4 .
Determination of As (mean ± standard deviation, n = 3) in fertilizer (SRM 695) and in spiked samples of agricultural gypsum and mineral fertilizer in single quadrupole mode and MS/MS mass shift mode with different oxygen gas flow rates Sample Single quad mode ( 75 As + ) MS/MS 0.30 mL min -1 O 2 ( 75 As 16 O + ) MS/MS 0.50 mL min -1 O 2 ( 75 As 16 O + )