Curve fitting for restoration of accuracy for overlapping peaks in gas chromatography/combustion isotope ratio mass spectrometry.

The effect of graded degrees of overlap on high-precision and -accuracy carbon isotope ratios determined by gas chromatography/combustion isotope ratio mass spectrometry (GCC/IRMS) is reported. Overlapping peaks of closely matched isotope ratio (difference delta 13CPDB < 1%) were analyzed by the conventional vertical drop summation algorithm and by curve fitting using the Levenberg-Marquardt algorithm. The conventional algorithm resulted in systematic bias related to degree of overlap even though precision was not noticeably affected. The exponentially modified Gaussian (EMG) and HaarhoffVanderLinde (HVL) functions were found to model GCC/IRMS peaks satisfactorily. Useful models over a wide range of overlap were obtained by applying consecutive HVL/HVL or HVL/EMG functions to overlapping peaks. Accuracy was improved in most cases and was never degraded. This study demonstrates the presence of subtle bias in isotope ratio determinations of overlapping peaks and the ability of automated curve fitting to compensate for these biases.