Examining the validity of the computed Al curve in elemental spectroscopy logs with x-ray fluorescence techniques in the Marcellus Shale
Elemental capture spectroscopy logs are used in open hole logging to obtain elemental concentrations in the subsurface. It provides measured concentrations of Si, Ca, Fe, S, Ti and Gd, while Al concentrations are computed from Si, Ca and Fe. The objective of this study was to run X-ray fluorescence (XRF) on cuttings from a vertical well to validate the spectroscopy data-in particular the Al curve. Al is one of a few elements used to study detrital sediment input. This brings into question how representative the computed curve is of the actual amount of Al in the rock. Cuttings samples were collected every 30’ over a 1680’ interval and a total of 58 samples were analyzed using XRF. The cuttings are not high resolution like the spectroscopy data; however comparisons were effective. Initial results showed concentrations of Si and Al were too high relative to the logs in some intervals and too low relative to the logs in other intervals. Typically the bias between core and log data should be in only one direction, but the overall shape of the curves from the cuttings matched the shape of the logs fairly well. With this in mind, it was noted that Barium concentrations were also extremely high, up to 15wt%. This suggests possible BaSO4 contamination from drilling mud which could give false XRF readings. To investigate further, cuttings samples were washed more thoroughly a second time and the analysis was rerun. The second washing reduced the amount of Ba in the samples, up to 10wt%; however, there were still significant concentrations of Ba and discrepancies between the cuttings and logs. Plotting the Ba values of both runs and the Ba/S ratio clearly demonstrated that the zone of high Ba coincided with the Ba/S ratio in barite suggesting that the contamination came from BaSO4 in the drilling mud. The high increase was tied to where the drilling mud was switched to OBM. The next step was to correct for BaSO4 in all the elemental data. After doing this, the bias was more consistent across most of the elements and the comparison between the cuttings and log was greatly improved. Several studies have shown that XRF techniques used for cuttings can give slightly biased results for Al and Si depending on the calibration used. Considering this, it would seem that the spectroscopy data quality looks very good and that the derivation of the Al curve from the measured Si, Ca and Fe seems to be a fair representation of measured Al concentrations in the rock.
AAPG Datapages/Search and Discovery Article #90195 © 2014 Eastern Section Meeting, London, Ontario, Canada, September 27-30, 2014