# A Unified Method for Triple Oxygen Isotope Analysis of Sulfate, Water, and Organics

## TL;DR

A new automated method called TORCH enables accurate triple oxygen isotope analysis of sulfate, water, and organic materials using a unified approach.

## Contribution

The TORCH method unifies triple oxygen isotope analysis across different materials with high precision and direct reference scaling.

## Key findings

- The TORCH method achieves reproducibility of ±14 per meg for Δ′17O and ±0.3‰ for δ18O.
- Consistent triple oxygen isotope values are obtained across different sample types and techniques.
- Quantitative oxygen extraction and accurate scaling are essential for reliable isotope analysis.

## Abstract

Triple oxygen isotope
analyses of sulfate, water, and organic materials
provide vital insights into geochemical processes. Here, we present
the fully automated TORCH method (Triple Oxygen isotope analysis using
Reduction, Conversion, and High-precision laser spectroscopy), which
enables the direct normalization of various materials to the VSMOW-SLAP
reference scale using a unified method. The method integrates high-temperature
thermal conversion elemental analysis (TC/EA), high-voltage glow discharge
CO-to-CO2 conversion, and tunable infrared laser direct
absorption spectrometry for simultaneous δ18O and
Δ′17O determination from CO2. We
applied the TORCH method to international and in-house standards of
Barite (IAEA-SO-5, IAEA-SO-6), benzoic acid (IAEA-601), cellulose
(Sigma-Aldrich 0742213, labeled Cellulose-KOSI), and water (VSMOW2,
SLAP2, USGS46), demonstrating quantitative oxygen yields and a reproducibility
reaching ±14 per meg for Δ′17O and ±0.3‰
for δ18O (1 SD). We show that consistent triple oxygen
isotope values across different sample types and analytical techniques
are achieved when quantitative oxygen extraction and accurate scaling
are assured. The TORCH method delivers flexibility across varied matrices,
paving the way for high-accuracy triple oxygen isotope analysis of
geologically and environmentally relevant sample types.