SPWLA 34th Annual Logging Symposium, June 13-16.

1993

MODELING THE EFFECTS OF GLAUCONITE ON SOME OPENHOLE

LOGS FROM THE LOWER SENONIAN IN EGYPT

J. G. Patchett and R. Wiley (Amoco Production Company)

Mostafa El Bahr (Gulf of Suez Petroleum Company) *

Abstract of glauconite on their response. A regression analysis

based on this model data included the effects of glauco-
Glauconite is commonly found in the Lower Senonian nite on predicting porosity from density/neutron poros-
of the Gulf of Suez (Nezzazat Group Matulla Forma- ity and Pe logging measurements with acceptable error.
tion) and the Egyptian Western Desert (Upper Bahariya The possible effect of glauconite on formation resistiv-
Formation) in sufficient quantities to have an impact on ity was investigated by determining the CEC of samples
density, neutron, Pe, and resistivity log response. Glauc- and modeling its effect on electrical properties.
onite is a complex mineral composed mainly of iron,
potassium, and magnesium hydrated aluminum sili- Introduction
cates. Due to its complex mineralogy in particular varia-
tions in iron content, it has a wide range of grain densi- Glauconite in the Lower Senonian Matulla Formation,
ties. Modeling the effects of glauconite on the responses Gulf of Suez (October Field), and the Upper Bahariya
of density, Pe and neutron porosity logs has been used to Formation, Western Desert (Razzak Field), presents an
improve the determination of porosity and lithology unusual formation evaluation problem. These forma-
from well log data. tions have glauconite-rich zones, exceeding 25% in the
Glauconites are iron-rich minerals often found in sand- October Field, which are of potential petroleum reser-
stones, limestones, and siltstones. Sandstones rich in voir quality. The one sample from Razzak Field was
glauconite are commonly known as “greensands.” 80% glauconite but was not reservoir quality rock. The
Because of its complex composition, in particular ferric objective of this work was to evaluate the effect of glau-
and ferrous iron, the grain density of glauconite ranges conite on the density, neutron (CNL), Pe, and resistivity RR
between 2.4 g/cm3 to 2.95 gm/cm3. The presence of iron logs and then, if possible, to develop a log interpretation
also significantly affects neutron porosity response due procedure to address these problems. Modeling the
to its large thermal neutron absorption cross section. response of the CNL, density, and Pe to known mineral-
These density and elemental effects have been observed ogies containing glauconite was central to this work.
on logs from wells in the October and Razzak Fields.
The effect of glauconites on resistivity can be apprecia- Five core samples were selected, porosity ranged from
ble since they can have significant cation exchange 3.5% to 25%, and permeabilities from less than 1 md to
capacity (CEC). 90 md (see Table 1). The wide range in expected miner-
alogy of glauconite and the variation in associated min-
This work evaluated potential problems associated with erals presents an interpretation problem. Iron present in
interpreting logs from glauconite-rich formations and glauconite has an effect on all three logging measure-
proposes log interpretation procedures to determine ments (CNL, density and Pe) because of its large ther-
porosity and water saturation in these lithologies. This mal neutron absorption cross section, high density, and
procedure is based on petrophysical measurements from relatively high atomic number. The effect is large for the
a well in the Gulf of Suez, October Field, and a well density and Pe and less on the CNL (3 to 4 p.u.). Hydro-
from the Razzak Field in the Western Desert. The effect gen present in glauconite has an additional larger affect
of glauconite on the actual glauconite-rich rocks on den- on the CNL, over 10 p.u. for 100% glauconite. Clays
sity/neutron porosity response was investigated in also found in these rocks have a potentially large affect
detail. Core measurements from samples from these two on the CNL log. Glauconite can also have significant
wells included grain density determination, X-ray dif- cation exchange capacity, which effects formation resis-
fraction analysis, elemental analysis by X-ray fluores- tivity.
cence, and qualitative elemental analysis by scanning
electron microscopy. Mineralogy was based on these A systematic investigation was developed for the Lower
analyses. Senonian interpretation problem. The steps included:
Density, neutron and Pe models developed for various
Determining mineralogy by using X-ray diffrac-
concentrations of glauconite helped quantify the effect
tion and x-ray fluorescence,

* This paper, slightly expanded herein, was presented Measuring grain density of well-dried reservoir
by Mostafa El Bahr at the Egyptian General Petro- core samples,
leum Conference in Cairo Egypt, November 7-10,
1992 and published in the transactions of that confer- Estimating elemental composition and grain den-
ence. sity of the glauconite(s) fraction of these samples,
 SPWLA 34th Annual Logging Symposium, June 13-16, 1993

4. Computer modeling of binary mixtures of pure olive green pellets, Upper Bracklesham Beds, Shep-
glauconite and quartz using the nuclear modeling herds Gutter, Hampshire (Sample 21 of Buckley et al.,
method of Wiley and Patchett, 1990, 1978) edited by Newman, 1987. These data were com-
bined with elemental analysis from x-ray fluorescence
5. Similar nuclear modeling of neutron porosity, log
(Powers, 1992) to estimate mineralogy (Calvert et al.,
density and Pe responses of cored samples,
1985). The mineralogy of the five samples is given in
6. Using the data produced during the modeling Table 1, along with their grain densities. porosities, and
stage to develop a regression equation to predict permeabilities. Also included in Table 1 are the elemen-
porosity for binary mixtures of sandstone and tal analyses for these samples
glauconite; ternary mixtures of sandstone, glauco-
nite, and kaolinite; and actual rocks recovered by Table 1
coring, and Sample Mineralogy and Grain Densities
I I I
-
7. Measuring cation exchange capacity and estimat-
ing its effect on formation resistivity.
J-r
This project was designed to evaluate the problems
associated with interpreting logs from these rocks and to
investigate possible log interpretation procedures to
determine porosity and water saturation. More core data
and possibly production tests would be needed to ade-
quately evaluate this work. However, the general
approach is believed applicable to many reservoirs con-
taining complex mineralogy. This approach, under
favorable conditions, allows porosity determination in
1
reservoirs which are mineralogically complex, even
though more minerals are present than can be resolved ii- 20 12.697

using the three independent logging measurements dis-

Elemental Analysis
cussed here.

RR
Occurrence and Significance of Glauconite
Glauconite usually occurs as green, amorphous grains
seldom larger than fine sand size. Sands containing a
high percentage of glauconite are called greensands. It
may also exist as pellets or films. Glauconite formation,
accompanied by greensand sedimentation events on a
worldwide basis, took place during Cambrian. Upper
Cretaceous and early Tertiary (Pettijohn et al., 1973).
Glauconite-rich rocks are commonly found in the Upper
Cretaceous of Egypt. Geochemical evidence suggests
glauconite forms in a partially restricted, marine. reduc-
ing environment (Cloud, 1955). Glauconite is very sta-
These samples classify as glauconite-glauconitic sand-
ble in sea water, can survive transportation in a marine
stones which are medium-grained and well-sorted with
environment, and is an indicator of slow sedimentation.
some carbonate cementation (calcite/dolomite) contain-
Glauconites are essentially iron-rich illites (micas) and
ing from 13% to 26% glauconite in the October Field
mixed layer illiteismectite. Their chemical composition
and 81% in the Razzak Field. Sample No. 5 is kaolini-
is comprised of iron, potassium, hydrated aluminum sil-
te-rich, 50%, with 10% chlorite. Three other samples
icates.
also have significant amounts of kaolinite and/or chlo-
rite. Thin section photographs and SEM pictures repre-
Determining Mineralogy, Grain Density, sentative of glauconite-rich zones from October Field
and Glauconite Properties are shown in Figure 2.

The first three steps in this project defined the mineral- The grain densities of Samples 2, 3, 4, and 5 were deter-
ogy and grain density of the glauconite-rich rocks found mined by Boyle’s Law methods. Insufficient specimens
in the Gulf of Suez and Western Desert. The properties were available to measure the grain density of Sample 1.
of the glauconite could then be estimated. Core and log Its grain density was estimated using grain densities of
data for the Gulf of Suez well along with four core sam- 2.65 gicm’, 2.71 g/cm”, and 2.86 g/cm3 for quartz, cal-
ple depths are shown in Figure 1. cite, and glauconite respectively.

First, the mineralogy of the core materials was deter- Finally, the glauconite grain densities of Samples 2,3,4,
mined using x-ray diffraction and x-ray florescence. The and 5 given in Table 2 were also estimated. These grain
x-ray diffraction patterns of the glauconites in these five densities were based upon the measured sample grain
samples were very similar to the glauconite described as densities, the grain densities of the other minerals, and

-2-
 SPWLA 34th Annual Logging Symposium, June 13-16, 1993

assuming the chlorite was magnesium-rich rather than easily separated from illite by the modeled Pe as shown
iron-rich. by the log density Pe cross plot, Figure 3-B.
Table 2 Subsequent modeling included various combinations of
Glauconite Densities a hypothetical pure quartz sandstone, a glauconite based
on those summarized in Table 2 using an average grain
density of 2.95 cm3/cc, the five core samples, and
kaolinite. Typical Gulf of Suez conditions of formation
salinity (150,000 ppm), pressure (17,236 kPa, 2500 psi),
and temperature (65.6”C, 150°F) were used. The signifi-
cance of the modeled responses on porosity determina-
tion will be discussed in the next section.

The average glauconite grain density of the four sam- The first Egyptian modeling example, Figures 4 A-C
ples is 2.95 s/cm’, which is within the range of shows data from a binary mixture of sandstone and the
2.4 g/cm3 to 2.95 g/cm3 grain densities given by Deer, typical glauconite. Shown are plots of log bulk density
Howie, and Zussman, 1966. The density of glauconite versus CNL, log bulk density versus Pe, and CNL ver-
will vary with the Fe:AI ratio but may be masked by the sus Pe. Notice the combination of 80% sandstone and
presence of adsorbed water, resulting in a broad range of 20% glauconite produces about a four p.u. shift, higher,
density values (Deer, Howie, and Zussman, 1962). on the CNL as compares to sandstone at 20% porosity.
The modeled density and Pe reflect the changes in these
Modeling of Neutron, Density, and Pe parameters with increasing glauconite content.

The principal objective of the nuclear modeling was first Plots of the mixtures of sandstone, glauconite and
to study the log response to the actual rocks recovered kaolinite are shown in Figures 5 A-C. These plots reflect
by coring and to estimate the log response of the glauco- the differences in matrix properties of the three minerals
nitic fraction of these rocks. Next, modeling served as a modeled. Only the density/Pe cross plot, Figure 5-B, has
data source for use with regression analysis to predict an approximately common set of porosity lines through
porosity from actual well log data. all mixtures, thus only this combination of two logging
measurements would yield adequate porosity determina-
The modeling procedure is described in detail by Wiley tion.
and Patchett, 1990. Modeling can be conducted at either RR
the mineral or elemental level. Typically, log responses The results of modeling sandstone and the actual
are modeled for formation conditions of salinity, pres- glauconite-rich rocks are shown in Figures 6 A-C. The
sure, and temperature. The modeling discussed below, interpretation of these figures is similar to the theoretical
mixtures of sandstone/glauconite/laolinite shown in
except for the quartz sandstone and kaolinite, used ele-
mental composition and measured grain density. The Figures 5 A-C.
modeling report includes grain density, fluid density
Determining Porosity Through Regression
(both water and hydrocarbons and their mixture, if
applicable), hydrogen index of all fluids, thermal neu- Analysis
tron absorption cross section (sigma) of both the rock
Regression analysis was used in this work to develop
and all fluids, and the elemental composition of the
predictive equations for use with binary mixtures of
rock. For each rock or mineral modeled, the log
quartz sandstone and glauconite, and for mixtures of
responses were calculated for porosities from 0 p.u. to
quartz sandstones and the five samples. Modeling with
50 p-u. The computed parameters included slowing--
regression was also conducted for mixtures of quartz,
down length, diffusion length, count-rate ratio, CNL
glauconite and kaolinite.
reading at formation temperature, CNL additionally cor-
rected for temperature effects on the neutron detectors, The report from the modeling program includes a list of
true bulk density, log bulk density (Z/A effects, etc., or the log responses of the CNL, log bulk density, Pe, and
what the density log would actually read), Pe, the porosity modeled. A predictive equation can be devel-
response of a generic epithermal neutron log, and for- oped by combining this list for several minerals or mix-
mation sigma. tures of minerals and then running a regression to pre-
dict porosity from all or different combinations of the
The modeled log responses of log bulk density, Pe, and
modeled logging data.
CNL for glauconite and several clay minerals are illus-
trated in Figures 3-A and 3-B. Two glauconites, similar This regression technique will be demonstrated by fit-
except for matrix density, are compared with sandstone, ting data for binary mixtures of various concentrations
illite containing 500 ppm boron, iron chlorite, and of sandstone and glauconite. This data, shown in
kaolinite at typical Gulf of Suez reservoir conditions. Figures 4 A-C, should have a high correlation because
The CNL responses for the modeled glauconites are they are mixtures of two minerals, similar to a density/-
similar to illite, less than kaolinite and much less iron neutron cross plot (Schlumberger, 1989, Chart CP-lc).
chlorite, Figure 3-A. The density log response for these The regression statistics for this example are given in
two glauconite minerals is dominated by their matrix Table 3. The correlation is less than perfect because the
density, 2.85 gm/cc and 3.05 gm/cc. Glauconite can be porosity lines are not absolutely parallel and the

-3-
 SPWLA 34th Annual Logging Symposium, June 13-16, 1993

response may not be completely linear. Two-dimen- results are summarized in Table 4. These statistical
sional crossplots can be qualitatively evaluated by men- results are for various mixtures of sandstone, glauconite,
tally lining up the porosity ticks of equal value. If these and kaolinite. The actual concentration percentages
lines are approximately coincident with the standard modeled are shown in Figures 5 A-C. The application of
crossplot porosity lines of equal value, then the porosity various response equations for this data can be evalu-
obtained with standard crossplot porosity charts is appli- ated by looking at Table 4. Comparing Figures 4-A and
cable. Any differences between these two sets of poros- 5-A shows the density/neutron crossplot chart (Schlum-
ity lines indicate errors in the standard porosity interpre- berger, 1989, Chart CP-lc), does not apply to the data in
tation. If they form straight parallel lines, a usable Figure 5-A, probably due to clay and/or other minerals.
porosity relationship exists, independent of the number Note that kaolinite looks much like sandstone on both
of minerals modeled. The user can compare different the density and Pe logs but has a large effect on the
crossplots in this manner to determine the most appro- CNL, at least as compared to sandstone. As expected,
priate combination (e.g., density/neutron, density/Pe, or the regression equation using the three logging curves
neutron/Pe). Regression statistics for the four variable predicts porosity with acceptable error (standard error
cases for these data, shown in Figure 7, can make this 1.04 p.u.). The equation using the density Pe also gives
judgement more objective. All combinations of density, good results, which is fortunate, but not necessarily
Pe, and CNL were used to predict porosity, Table 3. The expected.
porosity equation for the four variable case is given in
Table 4
the regression statistics. Also included in Figure 7 is the
Determination of the Porosity from Regression
porosity regression equation for the data given in
Statistics Mixtures of Sandstone, Kaollnlte, and
Table 6. Glauconite
The statistical data in Table 3 shows the combination of Independent Standard Error
W
all three or any two logging measurements would yield Variables (Porosity)
satisfactory results as expected. The density/Pe combi- CNL, DN, Pe 0.9951 1.036
nation is quite satisfactory. The 40% sandstone, 60% CNL, DN 0.9533 3.193
glauconite line, Figure 4-A, lies on the dolomite line
CNL, Pe 0.8469 I 5.782
(CP-lc) with similar porosity values. Thus, the standard
density/neutron crossplot, as noted by the sandstone and IDN, Pe I 0.9924 I 1.286 1
dolomites lines, would yield approximately correct CNL 0.8174 6.306
I
RR porosity except at low porosity values. This observation DN 0.8680 I 5.363
is verified by comparing porosity tick marks from the IPe I 0.0198 I 14.52 I
standard sandstone and dolomites with the model values
for the mixtures. The final example concerns the actual rocks recovered
by coring, shown in Figures 6 A-C. As can be seen in
Table 3
Table 1, individual samples contain up to six of the
Determination of the Porosity from Regression
Statistics Binary Mixtures of Sandstone and seven total minerals found in all five. Data for these five
Glauconlte samples are shown with data from a pure quartz sand-
stone. Since the five samples are more mineralogically
Independent Standard Error complex than the simple mixtures shown in Figures 4
RZ
Variables (Porosity)
and 5, more error is expected. The increased error is
CNL, DN, Pe 0.9982 0.630
confirmed by the statistics in Table 5.
CNL, DN 0.9978 0.687
CNL, Pe 0.9932 1.215 Sample 5, from Table 1, is not a glauconitic sandstone
DN. Pe 0.9875 1.648 as it contains 50% kaolinite by x-ray analysis. The sta-
CNL 0.8215 6.22 tistics in Table 6 were calculated with Sample 5
DN 0.8698 5.313
removed. The correlation and standard error are
improved; however, both relationships predict porosity
Pe 0.0373 14.52
with acceptable error when using all three logging
The concept of a simple three variable plot (e.g., curves or the density/Pe. The plots of the actual rocks,
density/neutron/porosity) can easily be extended to four shown in Figures 6 A-C, are similar to the combination
variables by adding the Pe. This four-variable approach of clay, quartz, and glauconite shown in Figures 5A-C,
should have a high correlation with most mixtures of indicating that a simple quartz/glauconite model is not
three minerals. One advantage of combination regres- adequate for these rocks. A regression was run for the
sion and the modeling of more mineralogically complex October Field data, sample 1 excluded, which gives
actual rocks is to evaluate the impact of additional min- results similar to that found in Table 5.
erals on the prediction of porosity. The four-\,ariable
model cannot be easily evaluated by visual examination
Comparison of Model Results with Core
and is probably best evaluated using regression statis- Porosity
tics.
The interval shown on Figure 1 is mineralogically more
The second example shown here illustrates this method- complex than the data in Tables 1 and 2 indicate. This
ology using the data in Figures 5 A-C. The statistical complexity is illustrated by observing the thin section

-4-
 SPWLA 34th Annual Logging Symposium, June 13-16, 1993

Table 5 However, they are within the range of expected values,

Determination of the Porosity from Regression quoted by Weaver, 1989, of from 5 to 12 meq/lOOg for
Statistics for Sandstone and Five Rock Samples those giauconites having up to 5% expandable clay and
Independent Standard Error up to 40 meq/lOOg for giauconites with up to 50%
R2 expandable layers. Thus, a good correlation should not
Variables (Porosity)
CNL, DN, Pe 0.9852 1.797 be expected between CEC and the amount of glauconite.
The two highest CEC values reported here are higher
than any reported by Waxman and Smits, 1968.

The possible impact of the exchange capacity on the cal-

culation of water saturation is shown in Figure 8, which
was constructed by assuming three water saturations
bN ~~ 1 0.9682 I 2.624 1
(lOO%, 50%, and 25%) and calculating resistivity based
Pe 0.0503 14.34
I on the method of Waxman and Smits, 1968. Three
porosity values were modeled (5%, 15%, and 25%)
Table 6 which bracket the expected porosity range for these for-
Determination of the Porosity from Regression mations. Resistivity was then calculated using these
Statistics for Sandstone and Four Rock Samples
assumptions and this resistivity used to calculate the
Independent Standard Error Archie water saturation. Waxman and Smits had no
R2
Variables (Porosity) glauconites in their data, so the application of their
CNL. DN. Pe 0.9930 1.237 method should be questioned. C,-C,,, plots as defined by
I I

0.9929 1.243 Waxman and Smits and many other authors should be
CNL, DN I I
examined after laboratory resistivity measurements are
CNL, Pe 0.9615 2.893
conducted to confirm the applicability or refute the use
DN, Pe 0.9856 1.767
of the Waxman and Smits model.
CNL 0.8918 4.841
2.747
The purpose of Figure 8 is to show the potential error in
DN 0.9652
water saturation calculated using the Archie equation
Pe 0.0436 14.34
which does not account for the excess conductivity due
to the shaly sandstone nature of the glauconite sand-
made from the core of the Matuila Formation. Note, the stones. Figure 8 is a plot of CEC versus the water satu- RR
pyrobitumen streak through the center of the thin sec- ration calculated using the Archie, 1941, equation for
tion shown in Figure 2. Pyrobitumen has been modeled the three water saturations and three porosity values dis-
to determine its affect on the density neutron cross plot. cussed above. Figure 8 shows the CEC of the five sam-
Pyrobitumen added to clean sandstone gives the appear- pies listed in Table 1. Both “m*” and “n’” for the Wax-
ance of gas or light hydrocarbons. Five percent pyrobi- man and Smits calculations and “m” and “n” for the
tumen, by weight, in clean sandstones at zero porosity Archie calculations were assumed to be 2.0. An Rw at
has a bulk density of 2.47 gm/cm3 and about two per- reservoir conditions of 0.028 ohm-m and a B of 12
cent neutron porosity on a limestone scale. Additional (l/ohm-m)/(equiv/liter), Waxman and Thomas, 1974,
minerals, not shown in Table 1 have also been observed were used. An average grain density of 2.75 gm/cm’
in similar rocks include feldspar, pyrite, zircon, and was used based on the data in Table 1.
hematite.
Figure 8 indicates that, even with the saline formation
The only well which has a core over a Giauconite rich
waters (150,000 ppm) found in the Gulf of Suez, the
interval does not include the Pe measurement. The clay/glauconite has a large impact on resistivity. This
October Field, well discussed here, had about 200 feet surprisingly large effect should be confirmed with iabo-
core from which the four samples analyzed here were ratory measured resisttivities (C, versus C, plots) and
obtained, Figure 1. Poor correlations were obtained measured “n” values.
between the core porosity and either the density neutron
cross plot porosity, Schlumberger, 1989, or the regres- Conclusions
sion porosity equation for the density neutron summa-
rized in Table 5. These results were not unexpected in Glauconite and other minerals contained in Lower
light of the generally poor lithoiogic resolution with the Senonian rocks have a significant impact on formation
density-neutron only. The question yet to be answered is density, Pe, neutron porosity and probably resistivity
how much additional information would be added when logs in potential petroleum reservoir rocks. Giauconite
using the Pe curve. The modeling conducted here and in
other areas suggests a major improvement in porosity * C, versus C, plots constructed from recently completed
predictions. laboratory resistivity measurements of additional sam-
ples give excess conductivities of from .03 s/m to .07 s/m.
Cation Exchange Capacity and Its Effect on CEC values calculated from this data range from
Resistivity 3 meq/lOO qms to 6.5 meq/lOO qms. CEC values deter-
mined by wet chemistry are not yet available, however,
The cation exchange capacities (CEC) of the five core this appears to be a significant issue and further investi-
samples are high for potential reservoir rocks, Table 1. gation is warranted.
 SPWLA 34th Annual Logging Symposium, June 13-16, 1993

has a large effect on the density and Pe logs while Acknowledgements

kaolinite and other clays have a large effect on the CNL.
We thank Gulf of Suez Petroleum Company and Amoco
The combination of modeling the response of nuclear
Production Company for their encouragement and per-
logging tools and the analysis of the output of this mod-
mission to publish this paper. We also thank G. R. Pow-
eling using regression analysis is a powerful tool for
ers for interpreting the X-ray data, R. E. Larese for thin
developing interpretation techniques for these logs. This
section analyses, and T. W. Mace for the CEC measure-
method requires that representative mineralogies be
ments.
sampled. The same approach could be taken using core
porosity and actual logging data; however, this would
take a large core data set to cover all mineralogies over a
References
wide porosity range.
Archie, G. E., 1942, “The Electrical Resistivity Log as
In general, determining porosity from well log data an Aid in Determining Some Reservoir Character-
requires two independent logging measurements for for- istics,” Petroleum Technology, Vol. 5, 55-62.
mations containing a mixture of two minerals. For these
examples, adequate determination of porosity was pos- Buckley, H. A., Bevan, J. C., Brown, K. M., Johnson,
sible with the combination of density and Pe even when L. R., and Farmer, V. C., 1978, Glauconite and
three or more minerals were present. When three or Celadonite: Two Separate Mineral Species, Miner-
more minerals were present in the examples, the addi- alog. Mag., Vol. 42, 373-382.
tion of the CNL can provide marginally improved Calvert, C. S., Palkowsky, D. A., and Pevear, D. R.,
answers. 1985, Clay Mineral Society Workshop, Pevear,
The standard density/neutron crossplot would provide D. R. and Mumpton, F. A., editors.
acceptable porosity values for the one sample from the Cloud, P. E., 1955, “Physical Limits of Glauconite For-
Western Desert or for binary mixtures of sandstone and mation,” Bulletin of the American Assoc. Petro-
glauconite. However, this result is not adequate evi- leum Geologists, Vol. 39, No. 4,484-492.
dence for a general statement about porosity interpreta-
tion in the area. Sample 1 had the highest CEC and the Deer, W. A., Howie, R. A., and Zussman, J., 1962, Rock
highest glauconite content of the five samples, suggest- Forming Minerals, Vol. 3, 35-41, Longinans, Lon-
RR ing a large glauconite impact on resistivity and thus don, England.
water saturation determination. The standard density--
neutron crossplot porosity calculation does not apply to Deer, W. A., Howie, R. A., and Zussman, J., 1966, An
the four October Field samples discussed herein. The Introduction to Rock Forming Minerals, 207,
crossplot porosity would, based on Figure 6-A, be 3 to 8 Longmans, London, England.
p.u. too high.
Newman, A.C.D., editor, 1987, Chemistry of Clays and
While the method described herein is promising, further Clay Minerals, 72-75, John Wiley & Sons, New
evaluation with a larger core and well log data set is York.
warranted. This includes the comparison of porosity
Pettijohn, F. J., Potter, P. E., and Siever, R., 1973, Sands
from the method- described with core-determined
and Sandstones, Springer-Verhag, New York, 228.
porosity. Furthermore, the impact of gas and invasion on
this methodology complicates the interpretation as it Powers, G. R., 1992, Private Communication.
does with any technique used for gas reservoirs. Inva-
sion has not been addressed here. Schlumberger, 1989, Log Interpretation Charts, Schlum-
berger Educational Services, USA.
The resistivity modeling using the Waxman and Smits
method shows that electrical effect of the glauconite rich Waxman, M. H., and Smits, L.J.M., 1968, “Electrical
rocks sampled here can be large and that they should be Conductivities in Oil-Bearing Shaly Sands,” Sot.
treated as shaly sandstones. The contribution of glauco- Pet. Eng. J., 107-122, June.
nite to excess conductivity is interpreted to be larger
Waxman, M. H., and Thomas, E. C., 1974, “Electrical
than clays, due to the low expected exchange capacity of
Conductivities in Oil-Bearing Shaly Sands - 1. The
the kaolinite and chlorite found in these rocks.
Relation Between Hydrocarbon Saturation and
The work discussed here is based on a small data set and Resistivity Index; 11. The Temperature Coefficient
should be continued to better quantify all work reported of Electrical Conductivity,” SPE Journal, No. 14,
here. The effect of glauconite and other associated min- 213-225, February.
erals on density, Pe, CNL, and resistivity is significant
Weaver, C. E., 1989, Clays, Muds, and Shales, Elsevier,
and should be investigated in greater depth. The poros-
The Netherlands, 90.
ity from the regression models should be compared with
porosity to verify the modeling results. The effect of
other minerals, such as siderite and hematite, on logging Wiley, R., and Patchett, J. G., 1990, “CNL Neutron Po-
measurements should also be investigated after more rosity Modeling,” the Log Analyst, Vol. 31, No. 3,
detailed mineralogical analysis has been conducted. 133-149.

-6-
 SPWLA 34th Annual Logging Symposium, June 13-16, 1993

About The Authors

Jay Patchett is a senior

research associate at Amo-
co’s Tulsa Research Center
where he has been employed
for 18 years. Previously he
worked 16 years for Amoco
as Regional Log Analyst in

in Mount Carmel, Illinois, as a field logging engineer

after graduation from the University of Tulsa, where he
received a degree in Geophysical Engineering.

Ralph Wiley is a Special

Research Associate with
Amoco Production Research
in Tulsa, Oklahoma. He
received a B.S. and M.S.
degree in Physics from the
University of North Texas.
Mr. Wiley was employed by
General Dynamics from
1964 to 1972. He was then
employed by The Wireline
Division of The Western Company until 1975. He then RR
joined Amoco Production Company where he has been
envolved in petrophysical research. He is a member of
SPWLA and SPE.

Mostafa El Bahr is Depart-

ment Head in the Petrophys-
its Department in the Gulf of
Suez Petroleum Company
(GUPCO), Cairo, Egypt. He
received a B.S. degree in
geology from Alexandria
University in 1978. He
worked for GUPCO as a log
analyst in the the operations
group until 1986 whereupon
he moved to the projects
group. He attended the Amoco petrophysics training
program working on the evaluation of a low permeabil-
ity zone in one of the largest fields in the Gulf of Suez.
He has worked on a number of projects concerning the
fields in the Gulf of Suez and the Western Desert. In
addition, he has participated in the evaluation of some
open acreage.

-7-
 SPWLA 34th Annual Logging Symposium, June 13-16, 1993

Shallow Resistivity
Bulk Density
1.95 2.95 0.200 2000
GM CC ohm m
POROSITY NEUTRON Deep Resistivity
45 LIMESTONE -15.0 $200
___--------- ________________2000
PERCENT ohm m

Figure 1. Open Hole Log Playback Showing Depths of the Four

Samples Taken From the October field

ZRXWOl.062 -8-
 E

cn /Sandstone Reservoir Conditions

1.80 1
Csrnf-ictfwm CP-1 r ..;i Kaolinite
2.00 -
;
Ski“
_..
2.20 .”
“. Iron Chlorite
.$_yl
._ 2.40 -
.Z i
E
$ 2.60 -(
I
Y ,: ,/
,..”

r 2 2.80 - j ,,.p CP-1 c Sandstone

7 Dolomite:
P’
CP-1 c Dolomite
? ‘Glauconit~ Glauconite 2.85 g/cm3
CP-lc G,ac___:‘_ / ...P <
L3.05 g/cm
-I 3.00 - IGUI Ill
;:.,/
” -.C’ Glauconite 3.05 g/cm3
g/cm5
r\ fir lllite 500 ppm Boron
~33
Iron Chlorite
3.20 - Kaolinite
- SS Reservoir Condition
3.40 _I, f r I $ 3
-10.0 0.0 10.0 20.0 30.0 40.0 50.0 60.0 70.0 80.0
CNL Limestone Porosity

Figure 3-A. Density - CNL Cross Plot of Glauconite, Clays, Dolomite and Sandstone
 Sandstone CP-1 c
I f
Sandstone Reservoir Condition
--- CP-1 c Sandstone
-“-* CP-1 c Dolomite
-..-_- Glauconite 2.85 g/cm3
-._- Glauconite 3.05 g/cm3
-.-._ lllite 500 ppm Boron
.-- Iron Chlorite
.. Kaolinite
- SS Reservoir Condition

Glauconite 2.85 gmkm3 2

w...i
>....+
...._
I;
3.00 ...a.*
Kaolinite Glauconite 3.05 gmkm3
3.20
Iron Chlorite

3.40 - t l------l
1.0 2.0 4.0 6.0 8.0 10.0 12.0 14.0 15.0
Pe

Figure 3-B. Density - Pe Cross Plot of Glauconite, Clays, Dolomite and Sandstone
 1.80

2.00
>r
.E
2 2.20 Sandstone
d

- Standard Sandstone CP- Ic

- - - Dolomite CP-lc

2.80

20% SS 80% Glau.

3.00 ......./,.,..
... oy,& SS , ()(yy, (-lau_
CP-lc
I I I I I I I I I I I I I I
-5.0 0.0 10.0 20.0 30.0 40.0 . 56.0 - 6010
CNL Limestone Porosity

Figure 4-A. Density-CNL Cross Plot from Nuclear Modeling of Binary

Mixtures of Quartz Sandstone and Glauconite; Good Porosity
Resolution Possible with any Two Logging Tools
 50 Sandstone CP-lc
1.80
r 100% ss
:;; - - Standard Sandstone CP- IC
3 :..,:..,?
I ‘.

% - - - Dolomite CP-1 e
- Quartz Sandstone Egypt
“..,
..A.:)
--- 80% SS 20% Glau.
“Y.,
..I.\..,

.....:..,
‘...
....-..... SS 40% G 1au .
.... 6OC!,,&
,: ; ,.
i..) ; _.,_;<, ‘...
40% SS 60% Glau.
.
.j:.. -
‘?,
20% SS 80% Glau.

.:..
.,...
j...
:: ... .. /

0% ss 100% Glau.
:‘:: ...... .. ............ ......
“..,
.i/ :t,,,
” ,“:
..>:...2 ,,,,,, i.
4..
.. d.

“..

3
..::..
I

0 2.60
: .; I.
. .... ::,,,

I;; .:.: I, : ,.,.,. .“.

‘:::
I
4 .........
...‘I.
-::
.... .:,
..R
2.80 .; :’
..i
i,
,....
‘.I,
;::.

80% SS 2 ..........

..::.
‘..,

.::.
.,.,., :..

05 40% ss
3.00 Dolomite CP-1 c 100% G
20% ss <:\..is

I I I I I I I I I I I I I I

1 .oo 1.50 2.00 3.00 4.00 5.00 6.00 7.00 8.00

Pe

Figure 4-B Density-Pe From Nuclear Modeling of Binary Mixtures of Quartz

Sandstone Glauconite; Good Porosity Resolution Possible with
any Two Logging Curves

E
 SPWLA 34th Annual Logging Symposium, June 13-16, 1993
RR
ad
-14-
ls801 70% SS 15% G 15% K
Sandstone
”
;
‘::
::..,:
. Dolomite CP-1 c
L
..A. ^ ... .

“:t;.,
....
““$>. SS Reservoir Condition
2.00 - -<..
... ...+
‘<<
50% SS 50% Glau.
4L.
i
50 % SS 50% Kaol.
70% SS 15% Glau. 15% Kaol. z
50% SS 30% Glau. 20% Kaol. @.
50% SS 20% Glau. 30% Kaol. “,
y: Z
“’ “i
:I?.
8.
2.40 - ......_;s,
j :.‘., ,::;,, 5
c. < .: ,,.,,.
<,
..... ...
‘Z..
‘Y:.
2
(D
.:;
....i:;.
‘?:.
......1.
2.60 - : ::
.:;

:.
::’ \

: .;
‘..,

:::.

10“.
-+ .:..
.>:::
50% K - A
.::,
.A. . . . . :,

\ t

-\
wW.
U” CS 30% G 20% K [:,.::.:!
I” “I 50% ss F;no,~
2.80 - 50% ss -,‘\
20%G 02
30% K Dolomite CP-1 c
3.00 I I I I
1.oo 2.00 30.0 4.00 5.00
Pe
Figure 5-B. Density-Pe Cross Plot From Nuclear Modeling of Mixtures of
Quartz Sandstone, Kaolinite and Glauconite; Excellent Porosity
Resolution Possible with these Logging Curves
 70% SS 15% G 15% K
2.00 -
0
50
2.50 - - I Dolomite CPl -c
B 50% SS 20% G-
3.00 - L I- I__LJ--‘--
<:: i’.;
c
I r,:;.: - -
Sandstone
vI
50% SS 30% G 20% K *.I Dolomite CP-lc
3.50 - :I’)
~~,~~
,,,,
j<...&J - - -
‘4
‘>p -_- SS Reservoir Condition
~,~~~.,~,..~.,~~~~~~~”
” ..._ 50% SS 50C!,Io
...........,A Glau.
r^‘<
i’s 50 y. SS 50% Kaol.
4.00 - ;:::i,)
f/,bf+
50% v 5oc)/o$~~*fi,,,#&*& ,......
bJ*+ - 70% SS 15% Glau. 15% Kaol.
$..._,
,>>>
bJ+@ 50% SS 30% Glau. 20% Kaol.
4.50 -
- 50% SS 20% Glau. 30% Kaol.
I I I I I I I I I I I I I
-5.0 0.0 10.0 20.0 30.0 40.0 56.0 ’ 66.0
CNL Limestone Porosity
Figure 5-C. Pe-CNL Cross Plot From Nuclear Modeling of Mixtures of
Quartz, Kaolinite and Glauconite; Poor Porosity Resolution with
these Porosity Curves
 1.80 - 5 - Sandstone
- - - Dolomite CP-lc
:,:I
.‘: --.- SS Reservoir Conditions
i::...<
,,,,,,
r,_:fJ,+ ~‘_,,,,‘fi Sample 1
A$, - Sample 2
-%? ............... Sam p 1e 3
“Lj.(j~
Sample 4
A
.g 2.20_ A, --- Sample 5
,,5p
E Y.
,:<<:r;;>>,
0” L
$.za
x 2.40
3 +A>
/\.:r>!AA
m I. \,_ 5
a.
>..&
.-B>*
; 2.60 -5
* yjs,,..
4
7,
Sandstone Sample 3
2.80 44

i, 4
0: Sample 2 Sample 1
3.00 - Dolomite CP-1 c
I I I I
4 nn n nn ?b #x\r\ A hh
I .vu L.UU 3.UU 4.uu

Pe

Figure 6-B. Log Bulk Density-Pe Cross Plot From Nuclear Modeling of Five Core
Samples and Quartz Sandstone From Gulf of Suez; Better Alignment of
Porosity Ticks Illustrates Importance of Pe to Porosity Interpretation

B
 1.oo-

2.00 - Sandstone

Dolomite Cp-1 c
3.00 -

B
YI 4.00 - Sample 2 :I;*,,,,_,,,;
,,,,,/_,
,i,,,,,.,_.,.
2,..,..,.....;
......_.L.;~.~
.I........
g.......i..

Sample 3
58
-- Sandstone
5.00, - - - Dolomite CP-lc
--- SS Reservoir Conditions
--- Sample 1
6.00 _ -
..-.
........T
Sample 2
..... Sample 3
Sample 4
Sample 1 - Sample 5
7.00
I I I I I I I I I I I I I I 1
-5.0 0.0 10.0 20.0 30.0 40.0 50.0 60.0 65.5
CNL Porosity Limestone

Figure 6-C. Pe-CNL Cross Plot From Nuclear Modeling of Five Core
Samples From Gulf of Suez Indicating Poor Porosity Resolution
 Multiple Regression Report : General Regression
PROJECT = NMOD GUPCO GLAUCONITE STUDY GUAL. SS MIXTURES 1

POR 1 = 685749E+02
+ 0.656923E+OO * CNL
R-SQUARE = 0.9982 + -0.831738E+OO * DN 1’
+ -0.253580E+02 * PE 1

Standard Error of
the Residuals= 0.6340E+OO

Note: The predicted porosity should be equal to or less

than +0.27 p.u. from true porosity, 67% of time

Multiple Regression Report :

Analysis of Variance

Source DF SS MS F PROB > F

Regression 3.0 0.656E+05 0.219E+05 0544E+05 OOOOE+OO

Residual 301 .O 0.121 E+03 0.402E+OO
Total 304.0 0.657E+05

Multiple Regression Report:

Curve Partial F PROB>F
Regression Dependent POR 1
Coefficient CNL 1 O.l74E+04 O.OOOE+OO
Analysis 1 0.539E+02 O.OOOE+OO
PDEN 1 0.808E+03 O.OOOE+OO
(least important
curve has largest
PROB>F value ) RR
Note: The higher the Partial F, the more significant the variable

Mean values of regression data

CNL 1 0.304861 E+02
1 0.397086E+Ol
FEN 1 0.237478E+Ol
POR 1 0.250796E+02
Breakdown of observation elimination for model
----------____----------~____
306 total possible points in model
305 points actually used in final model
----
1 total points excluded from final model

1 points excluded by missing data in POR

Sum of points excluded can be greater than total excluded from model
because of points excluded for multiple causes.

Regression Equation Table 6 Data

Porosity = 110.89 + 0.3639 * CNL + 0.2198 * PE - 42.00 * DN

Figure 7. Statistical Report from Regression Analysis, Figure 4 Data

and Porosity Regression Equation for Table 6 Data.
-21-
 150- Trues,,,,=25 200 - TrueS,=50 A 300 True&,,,=100
A
A A
A A A
125,. I
A A 250 / A

A 150,. A
A
gi 00..
z? s
A r
2 A
r A
2 A
: 200
n A
A n
n as1 00 '. n n n
A n n n
n E ti A
n
n i?? n
n a ml i150
A n l .
n 0 . p" A
0 a a n 0 .

1
A 0
. 0 n 0 .
error in S, 0
g: . 100 fr
___ s,
______ ______ mm__
_saiple_ _ _ _ _ _ _ _ _ _ _ _ ir”e

Sample - -
I
2- SaTpIe Sample Say pie
SaYp;e
tf .t3 0, tSa;lple 1 50
tSa;ple
““ql”t tSa;lple
i
0.0 10.0 26.0 360 40.0 0.0 10.0 20.0 30.0 40.0 0.0 10.0 20.0 30.0 40.0
CATION EXCHANGE CAPACITY
CATION EXCHANGE CAPACITY CATION EXCHANGE CAPACITY

Figure 8. The Effect of Cation Exchange Capacity on Water Saturation

Calculations for Glauconitic Rocks. Showing Error in Archie
Computed SW Due to Clay Effect.