Quantifying the abundance and stable isotope composition of pyrogenic carbon using hydrogen pyrolysis

Abstract

RATIONALE: Pyrogenic carbon (C(P)) is an important component of the global carbon budget. Accurate determination of the abundance and stable isotope composition of C(P) in soils and sediments is crucial for understanding the dynamics of the C(P) cycle and interpreting records of biomass burning, climate and vegetation change in the past. Here we test hydrogen pyrolysis (hypy) as a new technique potentially capable of eliminating labile organic carbon (C(L)) from total organic carbon (C(T)) in a range of matrices in order to enable reliable quantification of both the C(P) component of C(T) and the stable carbon isotope composition of C(P) (δ(13)C(P)). METHODS: We mixed C(P) at a range of concentrations with common C(P)-free matrices (C(L) = cellulose, chitin, keratin, decomposed wood, leaf litter, grass and algae) and determined the amount of residual carbon not removed by hydrogen pyrolysis (C(R)) as a ratio of C(T) (C(R)/C(T)). Mixing C(P) with a unique δ(13)C value provided a natural abundance isotope label from which to precisely determine the ratio of C(P) to residual C(L) remaining after hypy. RESULTS: All C(P)-free matrices contained trace carbon after hypy, indicating that hypy does not remove all the C(L). However, there was a strong correlation between C(R)/C(T) and C(P)/C(T), viz. C(R)/C(T)= 1.02(C(P)/C(T)) + 4.0 × 10(-3), r(2) = 0.99, p <0.001, suggesting that only a small and reasonably constant fraction of C(L) remains after hypy. Uncertainties associated with the correction for contamination of C(R) by residual C(L) are minimal allowing for reliable determinations of both C(P) and δ(13)C(P) in many cases. CONCLUSIONS: Hydrogen pyrolysis appears to be a robust technique for estimating C(P) abundance and δ(13)C(P) across a range of materials. Nevertheless, caution is required in interpreting δ(13)C(P) values when C(P)/C(T) is low, with C(P)/C(T)>4% being required for the determination of the δ(13)C(P) values within an interpretable error under our experimental conditions.