Evaluation of isobaric interferences of Zr on silver isotopes 107 and 109
The intensities of the Ag109, Ag107, and Zr90 isotopes were measured in the ionic Zr solution (10 ug/L). The intensity (cps) ratio of Ag107/Zr90 and Ag109/Zr90, respectively 0.393% and 0.023% indicated the rate of Zr interference on silver measurements. Calculated ratios should not be considered constant, as they are strongly associated with the instrument optimization, specifically regarding the cerium oxide formation. In this study, ICP-MS parameters were optimized according to the manufacturer's specifications: CeO/Ce ratio below 2%. The detection of isotope Ag107 was thus much more interfered by Zr than the 109 isotope as indicated by Guo and al. [10] for the determination of total Ag concentrations by regular mode. Given that the two natural isotopes of silver are in similar proportions, 52% for the 107 and 48% for 109, we thus benefited from using the isotopes choice as shown in figure 1 for the analysis of NP Ag standards. The 107 isotope is rather recommended and frequently used for both routine analyses in regular and single particle ICP-MS modes [7, 8, 13, 14]. In SP mode for NP analysis, the selection of Ag isotope should however be investigated with respect to the Zr interferences.
Analysis of a solution of NP-ZrO2 by the technique of SP-ICPMS.
Figures 2 shows the analysis of a solution of NP-ZrO2 using the isotope 90 of the Zr and the two Ag isotopes, 107 and 109, by the technique SP-ICP-MS. A study by Deguelde et al. [4] showed the feasibility of the analysis of colloidal ZrO2 by the SP-ICP-MS technique using its isotope 90. The figure 2a shows the distribution of NP ZrO2 analyzed in suspension solution (4 ug/L) by SP-ICP-MS technique using the isotope 90. The figure 2b reveals the analysis of this NP-ZrO2 solution using rather the isotope Ag 107 while figure 2c was for the Ag109. Figure 2b shows a series of “fooling” peaks appearing similarly to signals for Ag particles. In contrast, the peaks with the 109 isotope of silver were much less intense (Fig. 2c). If the solution was contaminated with NP-Ag, the SP-ICP-MS analysis would have been able to measure similar intensities with the two Ag isotopes that typically occur in similar proportions. The figure 1 (a,b) shows the analysis of NP-Ag solutions by SP-ICP-MS with the two natural silver isotopes. The pulse intensities are similar with the two isotopes. NPs-ZrO2 interfered with the detection of Ag particles by inducing false positives, especially with the isotope 107 (Fig 2b).
Natural water Analysis
Zirconium is naturally present in surface waters and mainly associated with the particulate and colloidal phase rather than the dissolved phase [11, 12, 15]. As previously mentioned Zr interferes with the analysis of Ag and could overestimate the true Ag concentration (Fig. 3). In an aquatic system under natural erosion processes, significant concentrations of suspended particles are typical in surface waters and Zr concentrations were evaluated at <1 µg/L (Table 1). Considering that to have an interference, a false positive, on the detection of a nanoparticle it is necessary that the interference comes from the particulate phase. The analysis of particulate Zr in natural waters by the SP-ICP-MS technique would thus produce signals that can be confused with silver nanoparticles; so there is a likelihood of a false positive.
Total silver concentrations measured with the Ag isotopes 107 and 109 were similar, despite the presence of 700 ng Zr total/L in the sample (Table 1). The difference between the two Ag isotopes was below the detection limit (1 ng/L). For the quantification by ICP-MS in regular mode (i.e., for total Ag concentration), it was therefore concluded that the contribution of Zr interferences is marginal on the determination of total silver concentration in these natural waters.
On the other hand for the NP quantification, Figure 3a shows different Zr intensity distribution for the natural water analysis with the SP-ICP-MS technique. Figures 3b and 3c show the measurements of silver particles in natural water evaluated with isotopes 107 and 109 of silver. For these last two figures, the results are presented as spherical particle sizes calculated from the intensities measured. Figure 3a shows that the natural water contains Zr particles with the presence of intense peaks corresponding to particulate Zr. Figures 3b and 3c distinguished the presence of silver particles by comparing 107 and 109 isotopes. Series of natural water samples were analyzed by the SP-ICP-MS technique and figure 4 shows results on the determination of NPs. For each assay, the number of silver particles larger than 30 nm (size threshold) was calculated for silver isotopes 107 and 109. The figure 3 highlights that there is much more detection with the 107 isotope when compared to 109. In an assay, 25 particles were detected using the isotope 107 while only 9 were detected with the isotope 109 (Fig. 4). If it would have been only silver particles, the number of detection should have been similar between the two isotopes utilized. Measurements with isotope 107 thus have false positives due to detection of 91Zr16O and 90Zr16O1H originating from particulate Zr.
Thereby, because of higher detection specificity, the isotope 109Ag was selected for NP-Ag analysis in water samples. In regular ICP-MS detection mode analysis (ie., for total Ag concentrations), the Zr isobaric interferences have minor impacts on measured concentrations and could be simply subtracted from the total concentration as long as the Zr concentration and its correction factor are known. On the other hand, in the case of analysis by SP-ICP-MS (i.e., for NP analysis), this signal correction procedure is not possible since the signal generated by a particle is random in time. With this technique, the results are expressed in a number of NP/ml rather than a mass of NP/ml. The interference generated by the particulate Zr would thus be read as a false positive. To minimize the presence of false positives a threshold of intensity was calculated to avoid these problems. With the analysis of the 90Zr by SP-ICP-MS, the maximum pulse intensity was measured for the particulate Zr signal (6 x 10 6), which was then converted to 109Ag intensity equivalent multiplying the intensity of it by the interference rate roughly corresponding to NP-Ag with size in the order of 27 nm. As a conservative approach, NP-Ag smaller than 30 nm were therefore not considered in the assessment. With the isotope 107, the signal generated by Zr particle would even correspond to a particle of 60 nm, justifying the setup of a safe detection threshold of 30 nm with the silver isotope 109.