Mass Spectrometer Development and Data Reduction:
Mass spectrometers measure the isotopic composition of geologic materials and form the backbone of a geoscientist's analytical toolkit. The precision of any measurement is in part dependent on the number of atoms counted. However, modern mass spectrometers count only ~2% of sample ions loaded into the instrument, at best. I am hoping to change that using an ion source that I have developed over the course of the past year.
We are currently installing our new ion source (called a cavity ion source) into the modern Triton mass spectrometer at Carnegie and will soon be making high precision measurements at the sub-ppm precision level. At first, we will focus on analyses of the decay products of short-lived isotope systems (e.g., 142Nd, 53Cr) which are extremely useful tracers of the processes operative on planetary bodies in the first 500 Ma of solar system formation.
Above: Here we are re-installing the 1970's era home-built 15-inch mass spectrometer at DTM. We have designed and tested our new cavity ion source on this mass spectrometer before installation on the Triton TIMS instrument. From left to right: Steve Shirey, Tim Mock, Rick Carlson.
The principle of increasing ion generation in a thermal ionization mass spectrometer is simple and has been tested since at least the 1970's. It consists of a hot tube of metal with a high work function (e.g., Ta, Re, W) with sample loaded into the rear of the tube, or cavity. When the sample evaporates, every atom has many chances to hit a hot wall of the cavity, which due to the low probability of ionization during any one contact, greatly increases the chances of ionization. Below is a schematic diagram of a cavity ion source produced by our collaborators at Oak Ridge National Lab in the early 2000's.
Above: A model showing the current uncertainty limits for making high-precision Nd-isotope measurements. The x-axis is showing the 142Nd signal size (basically the total number of ions counted) for an 8-hour run using a multi-dynamic collection routine. The y-axis shows the resulting 2SE precision in the Nd-isotope composition. Yellow dots are actual measurements while the curve is the theoretical precision limit imposed by counting statistics (shot noise). Essentially, we are at the theoretical precision limit, so to increase precision we must count more ions. The blue vertical lines are the number of ions generated by our cavity ion source, which suggest a 142Nd precision of below 1 ppm. This five-fold precision increase will likely transfer to other elements and isotopic systems, which we are currently testing.
Above: Our cavity ion source in operation and generating a Nd ion beam. The glowing rod running down the center of the image is the heated cavity which emits ions that are extracted towards the top of the image.
Left: Schematic diagram of the basic operating principles of a cavity ion source; From Burger et al., 2007. The ionizing cavity is heated by electron bombardment, with electrons provided by a hot filament.
Data reduction strategies:
I am also interested in data processing and creating techniques used to model complex data sets such as detrital zircon U-Pb and whole-rock radioisotope data. So far, my work to this end includes data reduction strategies aimed towards interpreting complex U-Pb and isochron datasets. We hope these techniques will be useful to many aspects of the geochemical community, and I will continue applying them to complex, multi-phase datasets in the future.
Psuedo-numerical modeling of discordant U-Pb datasets, from Reimink et al., 2016, Journal of the Geological Society
Output from a model of shale Re-Os data used to infer the ages and sources of Os in complex sedimentary systems. From Davies, Sheldrake, Reimink et al., 2018, Geochemistry, Geophysics, Geosystems