3.2. Calibration and detection limits of IC and IC-ICPMS
We generated calibration curves with the IC in standard mode (without chloride removal), and with the coupled IC-ICPMS setup with chloride removal, using both the conductivity detector in the IC and the SEM detector in the ICP-MS. In all cases, the data show good linearity over the concentration ranges that were tested, with correlation coefficients (r2) better than 0.999 (Fig. 3, 4). As shown in Fig. 3, the performance of the conductivity detector in the IC is not negatively impacted by the presence of the Cl-removal setup. Good linearity was obtained with and without chloride removal, as illustrated here with phosphite and nitrate. This observation demonstrates that the Cl-removal may also be useful for analyses of ions such as nitrate that cannot be analysed by ICP-MS. Furthermore, it gives confidence that the clean-up column does not introduce random error performs as expected.
Regarding detection limits, the coupling of the IC to the ICP-MS resulted in significant improvement, as expected. In standard mode without chloride removal (Fig. 3a), we were able to confidently quantify phosphite peaks down to ca. 0.32 µmol/L (10 ppb P) using the conductivity detector alone. With the IC-ICPMS setup and the SEM detector in the ICP-MS, a resolvable peak was obtained for as little as 0.003 µmol/L (0.1 ppb P, Fig. 4a). Slightly lower concentrations may still be detectable. Our results are thus similar to the detection limit of 0.002 µmol/L reported by Ivey & Foster (2005)[4], despite a significantly smaller sample loop (37 µL instead of 800 µL). Reproducibility in our IC-ICPMS setup was 13% at 0.032 µmol/L.