Modeling of Carbonate-Evaporite Lithology and Mineralogy from Whole-Rock Elemental Analyses: A Tool for Improving Petrophysical Evaluation and Chemostratigraphic Correlation

Michael C. Dix¹, Timothy Pearce², and Steve Jones^3
1 Halliburton - Sperry Drilling Services, Houston, TX
² Chemostrat Ltd, Llanfyllin, United Kingdom
³ X-ray Mineral Services Ltd, Colwyn Bay, United Kingdom

Obtaining practical, rock-based information from wells that are not cored is increasingly important to drilling programs worldwide. Analyses of core have traditionally provided benchmark determinations for lithology, mineralogy, sedimentary facies, and reservoir quality; results of core analyses are then used to calibrate petrophysical logs, and as primary input to digital reservoir models. In uncored wells, log data becomes the primary information source, supplemented by simplistic and often subjective lithology determinations from cuttings. Furthermore, drilling advances such as fixed-cutter (PDC) bits alter or obliterate rock textures in cuttings, making visual lithological identifications even more challenging. Clearly, innovations are required to glean as much objective geological information as possible from cuttings. One promising technique is the modeling of lithology and mineralogy from whole-rock elemental analyses, which are insensitive to destruction of rock texture. As a first test of this approach, we collected elemental (ICP and XRF) and mineralogical (enhanced XRD) data on several sets of core samples from carbonate-evaporite reservoirs. Improved XRD sample preparation and peak-modeling capabilities allowed more precise quantification of Mg-calcites, which is correlated to MgO measurements from elemental analysis. This resulted in better modeling of calcite and dolomite contents from elemental data. Anhydrite can be readily modeled from SO3, silt and clay content can be modeled from SiO2, Al2O3, TiO2, and K2O, and minor phases such as Fe2O3, MnO, P2O5, Cl, Sr, and Y provide additional constraints on mineralogy. Better mineralogical determinations are obviously valuable for matrix-density and porosity calculations from petrophysical logs, and for evaluation of gamma-ray response. In addition, since field-wide chemostratigraphic units in carbonates vary in their origin (depositional, diagenetic, or polygenetic), mineralogical constraints can aid in interpreting the zonation.