High-Precision MC-ICP-MS Measurements of δ11B: Matrix Effects in Direct Injection and Spray Chamber Sample Introduction Systems

Abstract

Multi-collector inductively coupled plasma mass spectrometry (MC-ICP-MS) is increasingly being adopted for measurements of δ11B with several approaches being adopted to overcome inherent issues of high blank or memory during sample introduction. Here we examine the advantages of direct injection of samples versus use of a cyclonic spray chamber introduction system and the effects of different matrices. Consistent with previous reports, direct injection yielded faster washout times and reduced memory compared to the spray chamber; however, signal intensity drifted over time, requiring periodic retuning, and the system was not robust to bubbles in the sample introduction line. The more conventional spray chamber approach, due to its larger surface area, required long washout times to reduce memory to acceptable levels, but the system was stable over time and robust to air introduction. We assessed the sensitivity of both direct injection and spray chamber sample introduction methods to matrix induced biases, in particular the effects of anions and different acid types (nitric versus hydrochloric acid) as well as the sensitivity to differing acid concentrations in sample-standard bracketing. Chloride had the largest influence on the measured δ11B values, with direct injection exhibiting shifts in δ11B of up to ∼−4‰ for changes in hydrochloric acid concentration from 0.1 to 0.2 N. In contrast, similar variations in nitric acid concentrations between sample and standard had comparatively little effect (<0.5‰). In addition measurements made via direct injection were more sensitive to the presence of sulfate. Thus, while direct injection has much reduced memory it is generally a less robust measurement system compared to spray chamber sample introduction. However, we show that by operating below matrix-critical thresholds both direct injection and spray chamber methods yield comparable values, with direct injection having the advantage of faster washout times, allowing increased sample throughput and analysis of smaller sample volumes.