Investigation of high temperature carbon reduction for safer, easier, and faster analysis of stable oxygen isotopes in silicates, oxides, and biogenic silica

Thumbnail Image

Authors

Johnson, Julie A.

Issue Date

2013

Type

Thesis

Language

Keywords

delta 18O , diatomite , high temperature carbon reduction , quartz , silicate , stable oxygen isotope

Research Projects

Organizational Units

Journal Issue

Alternative Title

Abstract

In the past 30 years, use of biogenic silica has been gaining popularity as a tool for paleoclimate investigations in regions that lack significant carbonate deposits. Coupled with this is a continuing demand for δ18O analysis of silicates and oxides for petrologic applications. Fluorination techniques, though producing very accurate and precise δ18O data, are time-consuming and hazardous to generate a large number of analyses in a short amount of time, especially those required for paleoclimate research, and ion microprobe is not suited for analyzing bulk samples. We report here the use of a commercially available oxygen-nitrogen (O/N) analyzer capable of high temperature carbon reduction (HTCR) to produce consistently accurate and precise stable oxygen isotope data for quartz and, in the future, other silicates, oxides, and biogenic silica. International quartz standard NBS-28 and four quartz laboratory standards, ARQ, NCSU, Snowbird, and SGS, are used to test the accuracy and precision of oxygen isotope analyses in the HTCR apparatus. All four lab standards were previously analyzed for δ18O by in-lab laser fluorination (LF) for comparison with values produced in this study. One-mg samples of quartz are wrapped in tin capsules and dropped via autosampler into a graphite crucible within the graphite-electrode furnace of the O/N analyzer. Temperatures >2500°C (power ≥4.0 kW) are used to drive the reduction of SiO2 in the presence of excess carbon (graphite) to produce carbon monoxide (CO) in a continuous flow of He (~250ml/min) that is piped directly to a stable isotope ratio mass spectrometer (IRMS). Total trial time for oxygen extraction and δ18O analysis is 7 minutes. In addition to methodology, the HTCR configuration was developed, including experiments with installing a gas chromatograph in-line between the O/N analyzer and IRMS and attaching an autosampler. Good precision and accuracy is observed in results from analysis of the quartz standards. NBS-28 yields a HTCR precision of ± 0.12 / (n=4). ARQ yields HTCR δ18O of +18.70 ± 0.08 / (n=3) vs. accepted LF δ18O of +19.0 /; NCSU yields HTCR δ18O of +11.25 ± 0.26 / (n=3) vs. accepted LF δ18O of 11.7 /; Snowbird yields HTCR δ18O of +15.57 ± 0.24 / (n=9) vs. accepted LF δ18O of +16.2 /; and SGS (crushed silica glass) yields HTCR δ18O of +11.63 ± 0.38 / (n=4) vs. accepted LF δ18O of +11.8 /. The O/N analyzer produces a mean oxygen yield of 99.86% quantitative conversion of quartz-oxygen to CO. Six diatomite samples from the Quincy Diatomite deposit (Menicucci, 2010) were analyzed and produced consistent δ18O values relative to VSMOW and normalized to NBS-28 ranging from +22.54 ± 0.36 / (n=6) to +27.25 ± 0.31 / (n=6) and produced oxygen yields >95%. Adularia, labradorite, zircon, and magnetite were also analyzed. Consistent δ18O values and >99% oxygen yields were produced for labradorite and adularia, but not for zircon or magnetite. These results demonstrate that a commercially available HTCR-capable instrument, coupled to an IRMS, can consistently produce accurate and precise δ18O data for quartz in very short trial times and without hazardous reagents necessary for laser fluorination. Even so, duplicates must be run for every individual sample due to inexplicable and apparently random drops or spikes in δ18O values that are not related to memory effect or methodology. With full automation and attention to the N2 contamination issues, this HTCR apparatus may attain comparable sample throughput for biogenic silica to that of carbonates.

Description

Citation

Publisher

License

In Copyright(All Rights Reserved)

Journal

Volume

Issue

PubMed ID

DOI

ISSN

EISSN