Leg 118: Geochemical Processing Report
Note: A complete revision of all of the processed data from this leg was performed before putting the data online. This may have resulted in minor depth discrepancies between the published geochemical data and the online database version, particularly before Leg 128.
The following is simply an account
of the steps generally carried out in geochemical processing. Because no
geochemical data paper was published in the Scientific Results volume of Leg
118, there is no mention of specific procedures and/or problems encountered
during the processing of Hole 735B. Information about the geochemical units
encountered in this hole is found in Robinson, P. T., Von Herzen, R. et al
(1989), Proc. ODP, Prelim. Results, 118: College Station, TX (Ocean Drilling
Program), 89-222.
Geochemical Tool String
The Schlumberger geochemical tool
string consists of four logging tools: the natural gamma-ray tool (NGT) the
compensated neutron tool (CNT), the aluminum activation clay tool (AACT), and
the gamma-ray spectrometry tool (see figure below). The natural gamma-ray tool
is located at the top of the tool string, so that it can measure the naturally
occurring radio nuclides, Th, U, and K, before the formation is irradiated by
the nuclear sources contained in the other tools below. The compensated neutron
tool, located below the natural gamma-ray tool, carries a low-energy
californium source (252Cf) to activate the Al atoms in the
formation. The aluminum activation clay background radiation is subtracted out
by the aluminum activation clay tool below and a reading of formation Al is
obtained (Scott and Smith, 1973).
The gamma-ray spectrometry tool, at
the base of the string, carries a pulsed neutron generator to bombard the
borehole and formation and an NaI(Tl) scintillation detector, which measures
the spectrum of gamma-rays
generated by neutron-capture reactions. Because each of the elements measured
(silicon, iron, calcium, titanium, sulfur, gadolinium, and potassium) is
characterized by a unique spectral signature, it is possible to derive the
contribution (or yield) of each of them to the measured spectrum and, in turn,
to estimate their abundance in the formation. The GST also measures the
hydrogen and chlorine in the borehole and formation, but the signal for these
elements is almost entirely due to seawater in the borehole, and they are hence
of little value.
The only major rock-forming elements not measured by the geochemical tool string are magnesium and sodium; the neutron-capture cross-sections of these elements are too small relative to their typical abundance for the tool string to detect them. A rough estimate of Mg+Na can be made by using the photoelectric factor (PEF) measured by the lithodensity tool. This measured PEF is compared with a calculated of PEF (a summation of the PEF from all of the measured elements). The separation between the measured and calculated PEF is, in theory, attributable to any element left over in the formation (i.e., Mg, and Na). Further explanation of this technique is found in Hertzog et al. (1989).
Data Reduction
The well log data from the
Schlumberger tools have been transmitted digitally up a wireline and recorded
on the JOIDES Resolution in the Schlumberger Cyber Service Unit (CSU). The results
from the CSU have been processed to correct for the effects of drilling fluids,
logging speed, and pipe interference. Processing of the spectrometry data is
required to transform the relative elemental yields into oxide weight
fractions. The processing is performed with a set of log interpretation
programs written by Schlumberger that have been modified to account for the
lithologies and hole conditions encountered in ODP holes. The processing steps
are summarized below:
Step 1: Reconstruction of relative
elemental yields from recorded spectral data
The first processing step uses a
weighted least-squares method to compare the measured spectra from the
geochemical spectrometry tool with a series of standard spectra in order to
determine the relative contribution (or yield) of each element. Whereas six
elemental standards (Si, Fe, Ca, S, Cl, and H) are used to produce the
shipboard yields, three additional standards (Ti, Gd, and K) can be included in
the shore-based processing to improve the fit of the spectral standards to the
measured spectra (Grau and Schweitzer, 1989). Although these additional
elements often appear in the formation in very low concentrations, they can
make a large contribution to the measured spectra, because they have large neutron-capture
cross-sections. For example, the capture cross-section of Gd is 49,000 barns,
that of Si 0.16 barns (Hertzog et al., 1989). Gd is, therefore, included in the
calculation of a best fit between the measured and the standard spectra.
The recomputed yields are loaded in
the file 735B-yields.dat.
Step 2: Depth-shifting
Geochemical processing involves the integration of
data from the different tool strings; consequently, it is important that all
the data are depth-correlated to one reference logging run. A total gamma-ray
curve (from the gamma-ray tool, which is run on each tool string) is usually
chosen as a reference curve, based on cable tension (the logging run with the
least amount of cable sticking) and cable speed (tools run at faster speeds are
less likely to stick).
The reference logging run at Hole
735B was the LDT/CNTG/NGT logging string.
Step 3: Calculation of total
radioactivity and Th, U, and K concentrations
The third processing routine
calculates the total natural gamma radiation in the formation as well as
concentrations of Th, U, and K, using the counts in five spectral windows from
the natural gamma-ray tool (Lock and Hoyer, 1971). This resembles shipboard
processing, except that corrections for hole-size changes are made in the
shore-based processing of these curves. A Kalman filter (Ruckebusch, 1983) is
applied to minimize the statistical uncertainties in the logs, which would
otherwise create erroneous negative readings and anti-correlation (especially
between Th and U). At each depth
level calculations and corrections also were performed for K contained in the
mud. This K correction is
particularly useful where KCl is routinely added to the borehole fluid to
inhibit clay swelling.
The outputs of this program are: K
(wet wt %), U (ppm), and Th (ppm), along with a total gamma-ray curve and a
computed gamma-ray curve (total gamma-ray minus U contribution).
The processed gamma-ray data are
loaded in the file 735B-ngt.dat.
Step 4: Calculation of Al
concentration
The fourth processing routine
calculates an Al curve using four energy windows, while concurrently correct
for natural activity, borehole fluid neutron-capture cross-section, formation
neutron-capture cross- section, formation slowing-down length, and borehole
size. Porosity and density logs are needed in this routine to convert the wet
weight percent K and Al curves to dry weight percent.
A correction is also made for Si
interference with Al; the 252Cf source activates the Si, producing
the aluminum isotope, 28Al (Hertzog et al., 1989). The program uses
the Si yield from the gamma-ray spectrometry tool to determine the Si
background correction. The program outputs dry weight percentages of Al and K,
which are used in the calculation and normalization of the remaining elements.
Step 5: Normalization of elemental
yields from the GST to calculate the elemental weight fractions
This routine combines the dry weight
percentages of Al and K with the reconstructed yields to obtain dry weight
percentages of the GST elements using the relationship:
Wi = F Yi/Si
where
| Wi | = | dry weight percentage of the i-th element |
| F | = | normalization factor determined at each depth interval |
| Yi | = | relative elemental yield for the i-th element |
| Si | = | relative weight percentage (spectral) sensitivity of the i-th element |
The normalization factor, F, is a
calibration factor determined at each depth from a closure argument to account
for the number of neutrons captured by a specific concentration of rock
elements. Because the sum of oxides in a rock is 100%, F is given by
F (Sum(Xi Yi / Si)) + XK WK + XAl WAl = 100
where
| Xi | = | factor for the element to oxide (or carbonate) conversion |
| XK | = | factor for the conversion of K to K2O (1.205) |
| XAl | = | factor for the conversion of Al to Al2O3 (1.899) |
| WK | = | dry weight percentage of K determined from natural activity |
| WAl | = | dry weight percentage of Al determined from the activation measurement |
The sensitivity factor, Si, is a
tool constant measured in the laboratory, which depends on the capture
cross-section, gamma-ray production, and detection probabilities of each
element measured by the GST (Hertzog et al., 1989).
The factors Xi are simply element to
oxide (or carbonate, sulfate) conversion coefficients and effectively include
the O, C or S bound with each element.
In processing the GLT data the correct choice of Xi is important in the
closure algorithm described above and requires geological input. In most
lithologies the elements measured by the tool occur in silicates where the
compositions can be expressed completely as oxides.
Step 6: Calculation of oxide
percentages
This routine converts the elemental
weight percentages into oxide percentages by multiplying each by its associated
oxide factor (Table 1).
The oxide weight percentages are
loaded in the file 735B-oxides.dat.
Table 1. Oxide/carbonate factors
used in normalizing elements to 100% and converting elements to
oxides/carbonates.
| Element | Oxide/carbonate | Conversion factor |
| Si | SiO2 | 2.139 |
| Ca | CaO | 1.339 |
| Fe | FeO* | 1.358 |
| K | K2O | 1.205 |
| Ti | TiO2 | 1.668 |
| Al | Al2O3 | 1.889 |
| Mg | MgO | 1.658 |
References
Grau, J. and Schweitzer, J.S.
(1989). Elemental concentrations from thermal neutron capture gamma-ray spectra
in geological formations. Nuclear Geophysics 3(1): 1-9.
Hertzog, R., Colson, L., Seeman, B.,
O'Brien M., Scott, H., McKeon, D., Grau, J., Ellis, D., Schweitzer, J., and
Herron, M. (1989). Geochemical logging with spectrometry tools. SPE Formation
Evaluation, 4(2): 153-162.
Lock, G. A. and Hoyer, W. A. (1971).
Natural gamma-ray spectral logging. The Log Analyst, 12(5): 3-9.
Ruckebusch, G. (1983). A Kalman
filtering approach to natural gamma-ray spectroscopy in well logging. IEEE
Trans., Autom. Control, AC-28: 372-380.
Scott, H. D. and Smith, M. P.
(1973). The aluminum activation log.
The Log Analyst, 14(5): 3-12.
For further information or questions about the
processing, please contact:
Cristina
Broglia
Phone:
845-365-8343
Fax:
845-365-3182
E-mail:
Cristina Broglia
Trevor
Williams
Phone:
845-365-8626
Fax:
845-365-3182
E-mail:
Trevor Williams