Arbeitsbericht NAB 12-47

Laboratory Characterisation of Cores (Geothermal Well Schlattingen SLA-1, Switzerland):Petrophysics and rock mechanics / physics

CSIRO Earth Science and Resource Engineering was requested by NAGRA to perform a broad laboratory characterization of rock cores recovered from the well SLA-1 located in Schlattingen (Switzerland). Fifteen (15) preserved cores in total have been sent to CSIRO (see details in the following section; Appendix B). All cores were X-ray CT scanned. Five (5) cores, namely SCR01 to SCR05, were tested using the scratching method to determine the unconfined compressive strength (UCS) as a function of depth. Three (3) plugs with dimension D = 38mm and L = 76mm were sub-cored from the lower ends of cores SCR01 to SCR03 to perform unconfined uniaxial stress loading to failure and directly determine their UCS. One plug was sub-cored from the lower end of SCR04 and subjected to the Fast Tracked Petrophysics testing workflow. One plug was sub-cored from the lower end of SCR05, X-ray CT scanned and subjected to NMR characterization. The ten (10) remaining cores, namely BD3, BD5, BD9, BD10, BD32, BD37, PEM (835.62m), EMPA (886.60), EMPA (929.86) and PEM (946.08), were sub-cored and all plugs subjected to the Fast Tracked Petrophysics workflow. Table 1-1 summarizes the data associated with the rock cores received by CSIRO.

On the occasion of a status review in December 2011, a comparison was conducted of CSIRO's semi-quantitative XRD measurements and Nagra's own mineralogical analyses on other core samples from the same borehole section of the Schlattingen SLA-1 well (Mazurek 2011). The comparison displayed marked deviations in clay mineral content of several CSIRO samples when compared with the samples from Nagra. In regards of these apparent discrepancies and the recent access to new parameters at CSIRO such as Cationic Exchange Capacity (CEC) from Methylene Blue method, XRD clay analysis using a PANanalytical diffractometer and Specific Surface Area (SSA) from ethylene glycol monoethyl ether method, an extension of the previous contract was proposed. The extension re-addresses the previous results on mineralogy and dielectrics and adds new results on exact amounts of different clay types available in each of the SLA-1 core to properly define the rock types between carbonates, marls and shales. The new XRD results reduced the uncertainties associated with the semi-quantitative XRD measurements. The remaining minor differences between Nagra's own analyses and the CSIRO measurements (CSIRO sample ID "H") can be attributed to the small scale spatial variability of the clay mineral content.

We have also added a significant amount of dielectric data which correlate very strongly with much of the mineralogical and textural information gathered so far. We have verified the NAGRA reported mineralogy for all of the samples provided except for Sample-H which has been identified as a carbonate with a total carbonate content of 87%. SSA for the SLA-1 samples typically ranges from ~10 to ~120 m2/g and this is consistent with the mineralogy from XRD. In fact determination of the analytical SSA, using published values of the SSA for the constituent minerals gives a 93% correlation with a close to 1-to-1 relationship.

Excellent correlations appear between dielectrics, CEC and SSA using paste samples, which suggests that mineralogy alone is the principal determinant. Surprisingly, this also applies to Vp, which would normally be considered as a texturally-controlled parameter. It is found to correlate strongly (up to approximately 86%) with dielectric response determined from paste, which has a significant portion of its textural quality destroyed during preparation. Ions occurring within the shale samples are the dominant cause of both real and equivalent imaginary dielectric permittivity below 100 MHz through Maxwell-Wagner processes and conduction. Above 100 MHz however, the mineralogy and water content dominate the measured dielectric response. This is also consistent with a realistic petrophysical shale model established from high resolution imaging. Paste measurement gives better correlations with textural and mineralogicalproperties than intact preserved shale measurement, although a simple correction based on CEC and water content can convert the preserved measurement into an equivalent paste measurement with nearly 97% correlation. This important relationship is a critical first step in establishing drill cutting based formation evaluation and furthermore a dielectric log to SSA and CEC algorithm.

The new results lead the dielectrics method to a new level of interpretations: it proves to be a very accurate tool to assess the mineralogy control on P-wave velocities, CEC and SSA on rocks BUT also on cuttings without the needs of specific sample preparations or preservations.