Testing reproducibility of vitrinite and solid bitumen reflectance measurements in North American unconventional source-rock reservoir petroleum systems

An interlaboratory study (ILS) was conducted to test reproducibility of vitrinite and solid bitumen reflectance measurements in six mudrock samples from United States unconventional source-rock reservoir petroleum systems. Samples selected from the Marcellus, Haynesville, Eagle Ford, Barnett, Bakken and Woodford are representative of resource plays currently under exploitation in North America. All samples are from marine depositional environments, are thermally mature (T_max >445 °C) and have moderate to high organic matter content (2.9–11.6 wt% TOC). Their organic matter is dominated by solid bitumen, which contains intraparticle nano-porosity. Visual evaluation of organic nano-porosity (pore sizes < 100 nm) via SEM suggests that intraparticle organic nano-pores are most abundant in dry gas maturity samples and less abundant at lower wet gas/condensate and peak oil maturities. Samples were distributed to ILS participants in forty laboratories in the Americas, Europe, Africa and Australia; thirty-seven independent sets of results were received. Mean vitrinite reflectance (VR_o) values from all ILS participants range from 0.90 to 1.83% whereas mean solid bitumen reflectance (BR_o) values range from 0.85 to 2.04% (no outlying values excluded), confirming the thermally mature nature of all six samples. Using multiple statistical approaches to eliminate outlying values, we evaluated reproducibility limit R, the maximum difference between valid mean reflectance results obtained on the same sample by different operators in different laboratories using different instruments. Removal of outlying values where the individual signed multiple of standard deviation was >1.0 produced lowest R values, generally ≤0.5% (absolute reflectance), similar to a prior ILS for similar samples. Other traditional approaches to outlier removal (outside mean ± 1.5*interquartile range and outside F10 to F90 percentile range) also produced similar R values. Standard deviation values < 0.15*(VR_o or BR_o) reduce R and should be a requirement of dispersed organic matter reflectance analysis. After outlier removal, R values were 0.1%–0.2% for peak oil thermal maturity, about 0.3% for wet gas/condensate maturity and 0.4%–0.5% for dry gas maturity. That is, these R values represent the uncertainty (in absolute reflectance) that users of vitrinite and solid bitumen reflectance data should assign to any one individual reported mean reflectance value from a similar thermal maturity mudrock sample. R values of this magnitude indicate a need for further standardization of reflectance measurement of dispersed organic matter. Furthermore, these R values quantify realistic interlaboratory measurement dispersion for a difficult but critically important analytical technique necessary for thermal maturity determination in the source-rock reservoirs of unconventional petroleum systems.