The chemical gamma-ray …
… a breakthrough in being confident in cuttings and data quality.
A common question and critics on the use of drill cuttings is, how representative are the cuttings for the drilled formation, due to concerns about hole cleaning and depth control.
The chemical gamma-ray can enhance the confidence in both, sample and data quality.
Background
With field samples and drill core samples a good understanding of how representative a sample for the sampled lithology/layer is, can be achieved. Drill/ditch cuttings samples, on the other hand, bear some uncertainties, as they usually represent a mix of lithologies over a particular drilled interval. For instance, a common sample frequency is every 10 feet, but with high rates of penetration (ROP) collection of samples may vary and be on longer intervals. Depending on the way the cuttings are collected from the ‘shakers’ at the rig site, they may represent a narrow spot sample, and interval of certain footage drilled, commonly ca. 10 feet, or longer intervals for instance during high ROPs.
A solution – the chemical gamma-ray
As a means for the sample and analytical data quality control, a theoretical gamma-ray curve based on the determined potassium (K), thorium (Th), and uranium (U) concentrations can be calculated. The calculated gamma-ray curve is then directly compared to a borehole logging tool gamma-ray log (logging while drilling (LWD) and/or wireline logs).
This aims in identifying issues with:
- Data quality
- Sample depth control
- Cavings and contamination from previously drilled intervals
… or even to compensate for GR tool failure.
Some basics – what is a gamma-ray log?
The gamma-ray (GR) log is a record the natural radioactivity of the rock sequences. It is sourced from the naturally occurring, radioactive isotopes of potassium, uranium, and thorium.
Radioactivity measured by GR tools is reported in American Petroleum Institute (API) units. These are based on a test pit/well at the University of Houston, Texas, made of specific concretes with known compositions and radioactivity contrasts.
In contrast to the GR, which records the combined radioactivity of K, Th, and U in API units, the spectral gamma-ray log shows the individual concentrations of each of these elements in % for K and ppm for U and Th.
The formula …
GR = 16K + 8U + 4Th
This is the general formula after Ellis & Singer (1987); K is entered in wt%, but Th and U in ppm. The resulting GR has the unit of API. If you have K, Th, and U from the geochemical analyses, you can easily reconstruct a ‘chemical’ GR log.
Why is this formula different from the one you may know?
You may come along a slight diverging formula in some publications, i.e., GR = 16.32K + 8.09U + 3.93Th. These factors are going back to the book by Rider (2002, p. 71) and are specific for a certain wireline tool at that time.
Important
K2O to K
Please keep in mind that many analyzers report potassium as K2O and not as K. In that case, you need to convert K2O to K by multiplying it with 0.8301!
The practical use of the chemical gamma-ray
Despite the difference in resolution (LWD/wireline logs usually < 1 ft depending on logging speed), the tool-derived GR log can be easily compared with the calculated GR log. Small discrepancies are expected, due to the nature of cuttings samples, i.e., a mix of lithologies as explained above. but should not exceed more than a few (< 10) API units.
Figure 1: Good match between chemical GR and GR.
Figure 1:
Figure 1 shows a theoretical GR log (black) overlain by a GR log calculated from K, Th, and U determined from cuttings samples (GR(calc.) in red). Note that due to the 10 feet resolution of the GR(calc.) small scale variations in the GR are not expressed in the GR(calc.) log; an example is the higher GR spike in the upper part of the first/upper shaly interval.
Figure 2: Chemical GR depth shift required.
Figure 2:
Figure 2 shows the same two logs, but note that the lower third of the GR(calc.) (red) shows a mismatch with the GR (black) in terms of depth. This can happen when, for instance, the sample depth is miscalculated (for example, the pump rate changed, but the new parameters have not been updated into the depth calculation), or sample bags/depths have been mislabeled. In the displayed example, a correction of the sample depth by -10 feet (upwards) for the affected section would solve the issue.
Figure 3: Chemical GR absolute values mismatch.
Figure 3:
The absolute values of the GR(calc.) (red) do not fit with the GR (black) readings. This can have several reasons:
- Mismatch in scale between GR and GR(calc.).
- One or more of the elements is overestimated (analyzer, due to:
- calibration
- very clean lithology, e.g., sandstone or limestone, where lower element concentrations are close to detection limits (common with XRF)
- GR needs adjustment, e.g., borehole environment corrections.
Figure 4: Chemical GR indicates ‘cavings’ (shale).
Figure 4:
A generally good match between the two GR logs has been achieved, except for an interval beneath the upper shaly section. Seven samples in the GR(calc.) (red) show higher API values than the tool GR (black). Moreover, the GR(calc.) shows a distinct downwards decreasing log trend until the calculated values fit again the GR log. This suggests that the seven samples in question probably contain material from the shaly section (so-called cavings), which results in higher calculated values. The GR(calc.) log trend (orange arrow) can be interpreted as indicating a ‘cleaning’ trend (less and less shaly material).
To correct this, the samples need to be picked (or sieved) for the right lithology.
Figure 5: Chemical GR indicates ‘cavings’ (sandstone).
Figure 5:
Like in Figure 4, a generally good match between the two GR logs has been achieved, but here the second (lower) shaly interval shows a mismatch between GR(calc.) (red) and GR (black). Six samples in the GR(calc.) show lower API values than the tool GR. This time, the GR(calc.) shows a distinct downwards increasing log trend until the calculated values fit again the GR log. This suggests that the six samples in question probably contain material (cavings) from the sandy section in the middle, which results in lower calculated values. The GR(calc.) log trend (arrow) can be interpreted as indicating a ‘cleaning’ trend (less and less sandy material).
To correct this, the samples need to be picked (or sieved) for the right lithology.
Figure 6: Chemical GR use to compensate for GR tool failure.
Figure 6:
If the chemical GR log shows a good match with the tool GR over different lithologies and a relatively long drilled section, confidence can be placed on cuttings samples for being representative of the drilled lithologies. With additional confidence in borehole cleaning (correct depth calculation; Figure 2) and hole conditions, i.e., no/minimized caving (Figures 4 & 5), the GR(calc.) may be used to compensate for the tool GR, e.g., after a tool failure. Under some circumstances, it may be decided to continue drilling without GR for cost or technical reasons, as pulling out the tool string can be time-consuming.
However, although initial GR(calc.) and GR logs show good correspondence, borehole conditions can change, and effects, as shown in Figures 4 & 5, may take place.
Caution when …
… naming it
In this article, the theoretical gamma-ray response calculated from the elements K, U, and Th, is called chemical gamma-ray, chemical GR, or GR(calc.). Other names such as ‘theoretical GR’, ‘synthetic GR’, ‘elemental GR’, and so on are in use by different companies and chemostratigraphers. However, be aware that the term ‘ChemoGR’ is a trademark.
Disclaimer:
The displayed GR and GR(calc.) logs are made up for demonstration purposes. The data do not represent real data for a particular formation. They are fictional and are manually generated for the displayed figures.
References in this Article
Ellis, D.V. and Singer, J.M. (2007): Well Logging for Earth Scientists. Springer Netherlands. 708 p. (Originally published by Elsevier, 1987) [Link]
Rider, M. (2002): The Geological Interpretation of Well Logs. 2nd Edition, RiderFrench. [Link]