High-precision gas chromatography-combustion isotope ratio mass spectrometry at low signal levels.

Precision and accuracy of gas chromatography-combustion isotope ratio mass spectrometry are investigated for sample levels down to about 5 pmol C in fatty acid methyl ester mixtures spanning 1000-fold in concentration. Precision and accuracy of isotope ratios diverge rapidly for conventional summation methods, and become unusable below 30 pmol material on column. At lower levels, mean isotope ratios were statistically different from reference values indicating bias as well as poor precision. In contrast, curve fitting, using the exponentially modified Gaussian line shape, gives improved precision for most peaks and useful results down to 3 pmol. The curve-fitting algorithm was also less sensitive to signal integration time than the summation method. These data indicate that curve fitting may be the method of choice for integration of noisy data when high-precision isotope ratios are desired.