The influence of different referencing methods on the accuracy of δ(13) C value measurement of ethanol fuel by gas chromatography/combustion/isotope ratio mass spectrometry.

Rationale: Brazil is the largest producer of sugar cane bioethanol in the world. Isotope ratio mass spectrometry (IRMS) is the technique of choice to certify the origin/raw materials for ethanol production, but the lack of certified reference materials (CRMs) for accurate measurements of δ(13) C values traceable to Vienna Pee Dee Belemnite (VPDB), the international zero point for (13) C/(12) C measurements, certified and compatible with gas chromatography (GC)/IRMS instruments may compromise the accuracy of δ(13) C determinations.

Methods: We evaluated the influence of methods for the calibration and normalization of raw δ(13) C values of ethanol samples. Samples were analyzed by GC/C/IRMS using two different GC columns. Different substances were used as isotopic standards for the working gas calibration. The δ(13) C values obtained with the three methods of normalization were statistically compared with those obtained with elemental analyzer (EA)/IRMS, since the δ(13) C results obtained using EA are traceable to VPDB via the NBS 22 reference material.

Results: It was observed that both the isotopic reference material for CO2 calibration and the GC column have a major effect on the δ(13) C measurements, leading to a bias of almost 2-3 ‰ in the δ(13) C values. All three methods of normalization were equivalent in performance, enabling an improvement in the GC/C/IRMS accuracy, compared with the EA/IRMS reference values for the samples.

Conclusions: All the methods of CO2 calibration, chromatography and normalization presented in this work demonstrated several sources of traceability and accuracy loss for the determination of δ(13) C values in ethanol fuel samples by GC/C/IRMS. This work has also shown the importance of using proper CRMs traceable to VPBD that should be compatible and certified using GC/C/IRMS, ideally in a wide range of δ(13) C values. This is important not only for bioethanol fuel samples, but also for many analytes commonly analyzed by IRMS.