Approach for quantifying rare Earth elements at low keV

Abstract

The challenges of analyzing bastnaesite (REECO3F) and hydroxylbastnaesite (REECO3OH) include beam sensitivity, quantification of light elements in a heavy element matrix, the presence of elements that cannot be analyzed with EPMA (H), and the use of x-ray lines whose physical constants are not well known. To overcome some of these challenges, Ca, La, Ce, Pr, Nd, and Sm were analyzed at 15 keV accelerating voltage and the light elements (C, O, F) were analyzed at 7 keV accelerating voltage. This approach is ideal for samples that are homogeneous within the volume analyzed. However, for the bastnaesite of interest, this solution is unsatisfactory as backscattered electron imaging reveals chemical variations at scales of less than 1 m (fig. 1). Monte Carlo simulations and wavescans of the REE M family of x-rays were evaluated to determine the best analytical approach. Monte Carlo simulations using Casino v2.42 were ran on a substrate of CeCO3F with a density of 5.00 g/cm3 at 15 keV and 7 kCV accelerating voltage. Depth distribution curves, φ(ρz), reveal x-ray generation at 15 keV for the Ce L shell x-rays approaches 1.0 m whereas the depth of Ce M shell x-ray generation at 7 keV is less than 0.4 m. Similarly, the K shell x-ray generation depth for C, O, and F at 7 keV is also less than 0.4 m. These simulations demonstrate an accelerating voltage of 7 keV or less and use of the REE M x-rays is necessary to acquire chemical information from the same volume of material for the light and heavy elements. Bastnaesite contains all REE and therefore one must consider the energy of the highest x-ray line of interest, namely the Lu M at 1.83 kV, to achieve an overvoltage (Eo/Ec) of over 2. Bastnaesite is also an insulator and requires a conductive coating, therefore 7 keV was used to avoid contributions of the coating material to the analysis given the above constraints. The USNM REE phosphate standards, Edinburgh REE glasses, CeO2, LaB6, and the metals of Sm, Dy, Gd, Er, and Yb were selected to evaluate the peak overlaps of the M family x-ray lines. The materials were coated with ~ 5nm iridium with a Leica ACE600 prior to analysis. Full spectrometer wavescans were collected at 7 keV accelerating voltage, 50 nA beam current, and 20 um beam diameter on a JEOL 8530F Plus using TAP and TAPL crystals. The step size was 0.109 mm along the length of the spectrometer with a dwell of 3 seconds at each step. M and M lines are broad and not individually resolved for the light REE but become more separated and sharper with increasing atomic number. M and M lines for each REE are also present across the spectrometer range further complicating peak interference corrections. The peak positions and overall shape for the Ce M and M lines vary with the coordination of the element. Significant shifts in position and shape were not observed when comparing the Ce M peaks of CeO2 and CePO4. Similar results were seen when comparing the available REE metal to the phosphate (Sm, Dy, Gd, Er, Yb). These wavescans suggest the use of the M or M lines if matrix matched standards are not available.