New insights into the source of the boron memory effect and implications for accurate boron isotope measurements.

Abstract

The boron (B) memory effect is a concern for B isotope analysis in inductively-coupled plasma mass spectrometry and a potential cause of poor data comparability between laboratories. It is widely assumed that the memory resides in water droplets on the surface of the spray chamber. However, even without the use of the spray chamber, background subtractions are still required to generate accurate data, therefore additional causes for the memory effect exist, which are investigated here. Different parts of the mass spectrometer were examined to pinpoint the source of a particularly high B background. After identifying the torch as the source of the background, different parts of the torch were soaked in dilute nitric acid, which was analyzed for B over time. B was leached out of the tip of the outer quartz tube of the torch in a fashion similar to borosilicate glass, which suggests the incorporation of B into the silica structure of the torch at high temperatures. Running 3% nitric acid washes effectively reduces the background. B background compositions change based on the solutions run beforehand, therefore different blank subtraction methods generate systematic differences. A new background subtraction method that utilizes B isotope ratios improved the precision by up to 0.14‰. The addition of a water wash step before sample elution led to smaller eluent volumes and improved matrix matching without causing a B breakthrough. An important part of the B memory derives from the torch glass, which incorporates B from sample solutions at high temperatures. Multiple nitric acid washes, matrix matching, blank subtraction, and standard sample bracketing generated accurate B isotope analyses with background/signal ratios as high as 10%, without the need for hazardous chemicals as washes. B isotope values of two sediment standards that represent average post-Archean continental crust were reported.