Upload
dinhkien
View
217
Download
1
Embed Size (px)
Citation preview
SPWLA 54th
Annual Logging Symposium, June 22-26, 2013
1
DETERMINATION OF FORMATION ORGANIC CARBON CONTENT
USING A NEW NEUTRON-INDUCED GAMMA RAY
SPECTROSCOPY SERVICE THAT DIRECTLY MEASURES CARBON
Jorge Gonzalez, Richard Lewis, James Hemingway, Jim Grau, Erik Rylander and Ryan Schmitt, Schlumberger
Copyright 2013, held jointly by the Society of Petrophysicists and Well Log
Analysts (SPWLA) and the submitting authors.
This paper was prepared for presentation at the SPWLA 54th Annual
Logging Symposium held in New Orleans, Louisiana June 22-26, 2013.
ABSTRACT
Over the past few decades, many techniques
have been developed for the log evaluation of
organic-rich rocks (ORR). More recently, ORR
have gained economic interest across the globe,
not only as the source rocks for conventional
reserves, but also as potential reservoirs
themselves. Quantifying total organic carbon
(TOC) and the complex mineralogy of these
rocks are two key components for a robust
characterization.
A number of well-log-derived TOC indicators
based on measurements such as gamma ray,
spectral gamma ray, resistivity, density, sonic,
neutron, and nuclear magnetic resonance (NMR)
have been developed over the years with
different levels of success depending on the
specific ORR and where it occurs within a basin.
An accurate TOC estimation faces two main
challenges: 1) the necessity for a correlation of
log data with core to utilize the most
representative model and 2) extensive log
interpretation. The process of an empirically
derived TOC has been hindered by the lack of a
consistent direct organic carbon measurement.
The introduction of a new neutron-induced
capture and inelastic gamma ray spectroscopy
tool using a very high-resolution scintillator and
a new type of pulsed neutron generator allows us
to measure new elements which, in turn,
increases our ability to properly interpret
lithologically complex formations. In addition to
measuring elements derived from capture
spectroscopy, this new service is capable of
measuring reliable inelastic yields, including
carbon, at reasonable logging speeds. This
carbon measurement allows the calculation of a
stand-alone quantitative TOC value that does not
require local calibrations or multitool
interpretation. This is possible because of the
successful combination of capture and inelastic
gamma ray spectra. Inorganic carbon in
carbonate is estimated by using other elements
from this logging tool (Ca, Mg, Mn, Fe) to
estimate total carbonate, and this value is
subtracted from the measured total carbon to
give TOC.
INTRODUCTION
For nearly the entire history of the logging industry,
fluid saturation calculations have been primarily
based on resistivity measurements which were
combined with porosity and lithology information. In
conventional reservoirs, the water saturation
calculations produced reasonable estimations of
actual water saturation and, consequently, oil and gas
recovery factors. However in cases where formation
water salinity is either too fresh or unknown the
resistivity calculations are unreliable. This created a
need for fluid saturation measurements that were
either independent of salinity or not nearly so reliant
upon knowing the exact formation water salinity.
Notable techniques are typically based on high-
frequency dielectric permittivity, nuclear magnetic
resonance, or carbon-oxygen ratio from neutron-
induced inelastic gamma spectroscopy.
This paper will focus on the application of inelastic
gamma spectroscopy in combination with capture
gamma spectroscopy, to measure the carbon content
of the formation to estimate its total organic carbon
(TOC) content.
Inelastic gamma spectroscopy has been used
with increasing success for well over 30 years
because of its ability to respond directly to
elemental carbon, which can then be related
directly to oil saturation using either a porosity
GG
SPWLA 54th
Annual Logging Symposium, June 22-26, 2013
2
log or elemental oxygen. Because of the nature
of this measurement, it works in either open or
cased wellbores. Historically, we have used the
term ―water saturation‖ when talking about fluid
saturations because we were measuring water
and assuming that what was not water was either
gas or oil. Carbon-based measurements such as
the (TOC) measurement described here are used
to calculate ―oil saturation‖ because we are
actually measuring the carbon rather than water.
The hydrocarbon volume is derived based on
knowledge of density of the hydrocarbon phase.
Neutron-induced gamma spectroscopy was
studied for potential use in the oil field in the
1960s (Tittman et al., 1960) and introduced as a
commercial service in the mid-1970s (Lock et
al., 1974; Ward 1974, Hertzog 1978). Gamma
rays resulting from thermal neutron capture are
used to identify and quantify elemental
concentrations of the major components making
up common sedimentary rocks (Figure 1).
Fig.1 When a neutron has slowed to a thermal
energy of, typically, 0.025 eV, by inelastic and
elastic interactions, it can be absorbed by the
nucleus. The excited nucleus will emit capture
gamma rays which will have an energy(s)
specific to the element that emitted the gamma
ray. Recording gamma ray count rate as well as
gamma ray energy allows one to estimate the
relative abundance of specific elements.
Weight percent of elements such as silicon,
calcium, iron, and sulfur can be used to quantify
quartz, carbonates, clay, evaporites, and heavy
minerals (Herron and Herron, 1996). These
measurements have been used in both open and
cased wellbores. By focusing on the gamma rays
generated by inelastic neutron interactions, we
can identify elements such as carbon and oxygen
(Figure 2).
Fig.2 When a high-energy neutron bounces off
a nucleus, it excites it into giving off inelastic
gamma rays. The energy(s) of these gamma rays
depends on the element emitting the gamma ray.
The unique gamma ray energy response for
carbon inspired the early attempts to calculate
oil saturation directly from the carbon
measurement. As seen in Figure 3, there are
several common elements that can be identified
based on their unique spectra. Since several
elements produce gamma rays in the same
manner as carbon, it is necessary to solve for
each element that contributes gamma rays to the
inelastic spectrum. Historically, it was the
carbon and oxygen yields which were used to
calculate oil saturation. The inelastic elemental
yields from the other elements were not used for
analytical purposes; they were simply calculated
to ensure that gamma rays generated from
elements other than carbon or oxygen did not
contribute to the carbon or oxygen yields. We
will demonstrate that with this new generation
tool we do use other inelastic elemental yields to
normalize the inelastic and capture gamma ray
spectra to quantify mineralogy and the carbon
content of the formation.
This latest generation tool, which uses a
lanthanum bromide scintillator, greatly increases
ones ability to resolve the details of a gamma ray
energy spectrum. The lanthanum bromide
scintillator offers many advantages over the
previous generation bismuth germinate (BGO)
detectors or sodium iodide, which were used
before the BGO detector. The spectral resolution
is more than twice as good as what could be
achieved with BGO, and the count rate handling
GG
SPWLA 54th
Annual Logging Symposium, June 22-26, 2013
3
capabilities are improved by a factor of nearly
10. Not only is it possible to measure elements
that were previously too difficult to detect, but
one can now accurately measure elemental
carbon at higher logging speeds over long
intervals.
It is this improved measurement of many
elements from both capture and inelastic
reactions that allows us to compute an absolute
formation carbon concentration, and in turn
calculate TOC in the formation within the region
of investigation of the measurement. At this
point it is important that we define what we are
actually measuring and how we derive TOC
from these basic measurements. It is also
important that we understand what TOC actually
means in this context. TOC in the context of this
measurement includes all forms of organic
carbon that are not part of the inorganic rock
matrix. This includes carbon associated with
kerogen, bitumen, oil, gas and coal. However, it
does not include carbon within the inorganic
minerals (e.g. carbonate).
Fig.3 By comparing the recorded gamma ray
spectrum to a set of elemental spectra standards
it is possible to determine the best combination
of standards that match the recorded spectra.
These calculated elemental yields are the best-fit
combination of the percentage yield of gamma
rays from each element. These elemental yields
can then be used to determine the relative
abundance of each element.
LITERATURE REVIEW
TOC is the weight percent of carbon that resides
within the organic portion of the rock.
Historically, TOC has been used to equate
directly to just the carbon in the kerogen fraction
of the rock, but, in practice in direct carbon
measurements, it also includes the portion of the
carbon that resides within bitumen and residual
oils (Rylander et al., 2013). Generally speaking,
the carbon density of gas is too low to be
detectable, although very high-pressure gas or
supercritical CO2 may be detectable in high-
porosity formations.
Quantification of TOC is critical to the
evaluation of organic shales. Organic shales are
typically defined by the quantity of TOC
present, which is usually from 1.5 to 2 wt%
(Herron, 1991). Thus, the quantification of TOC
is a valuable step in identifying potential organic
shale reservoirs, and the quantity of TOC can be
related to the reservoir quality of gas shales as
the producible pore fluid are within the organic
matter (Ambrose et al., 2010). Kerogen, the
organic matter where the TOC resides, has
petrophysical properties similar to pore fluids
(e.g., low bulk density, high neutron porosity,
low photoelectric factor), and it can be difficult
to differentiate from pore volume. An accurate
porosity estimation for an organic shale requires
an accurate TOC estimation.
Multiple models for the estimation of TOC from
logs are discussed below. The first four
measurements (gamma ray, uranium, density,
ΔLogR) are empirically derived and are based
on calibration to core measurements. They rely
on assumptions that matrix properties are
invariant through zones in the organic shale. The
last two measurements (NMR density,
geochemistry) are direct measurements.
Gamma ray. One petrophysical property for
which kerogen and pore fluids differ is their
gamma ray activity. Organic shales are
commonly associated with elevated gamma ray
activity. According to Schmoker (1981) factors
controlling this association include: 1) uranium
content of water at time of deposition; 2) type of
organic matter deposited; 3) water chemistry
near the water-sediment interface; and 4) rate of
GG
SPWLA 54th
Annual Logging Symposium, June 22-26, 2013
4
sediment deposition. The elevated gamma
activity is due to the concentration of uranium in
these sediments.
The gamma ray activity in some organic shales
can be high, with values in some Paleozoic
shales exceeding 1000 gAPI. Variations in gross
gamma activity primarily reflect differences in
uranium activity as the contribution from
potassium and thorium are relatively
insignificant in comparison. An empirical
correlation between gamma ray activity and
TOC content measured with core analysis must
be derived for each organic shale, and it may
need to be adjusted throughout a basin or field.
Some organic shales, especially those deposited
in the Mesozoic and Cenozoic eras and those
deposited in lacustrine environments, may have
moderate to low gamma ray activity. In these
cases, the use of gross gamma activity to
quantify TOC is not practical because of
relatively high contributions from the other two
radioactive elements, potassium and thorium.
A primary advantage of using the gamma ray
measurement for TOC quantification is that this
log is run in almost every well drilled. It can be
collected in practically any borehole
environment and is amenable to environmental
corrections. The primary disadvantage is that
gamma ray activity responds to the presence of
uranium, not kerogen, and the quantity of
uranium is also dependent upon differences in
kerogen type, water chemistry, and
sedimentation rate. In general, one can assume
that zones with higher gamma ray activity will
have higher TOC content, but this relationship is
commonly not linear, and it can vary abruptly
within one organic shale section. Finally,
differences in the quantity of clay or other
radioactive minerals can affect the potassium
and thorium content which, in turn, can lead to
changes in gross gamma activity that are not
related to changes in TOC content.
Uranium. Spectral gamma ray logs present the
advantage of direct uranium quantification (Fertl
and Rieke, 1980), negating the last issue
presented in the previous paragraph. An
empirical relationship between uranium content
and TOC still needs to be derived. And errors
can still arise due to the presence of other
uraniferous minerals like phosphates (Jacobi et
al., 2008) that could be misidentified as TOC.
The issues with sedimentation rate, etc. are still
valid.
Density. Gamma-gamma density logs acquired
over organic shales characteristically show a
lower value when compared to conventional
formations. A general rule would be that a gas
shale should have a bulk density less than 2.57
g/cm3 (equivalent to limestone porosity of 8 p.u.
with fluid density of 1 g/cm3) to be prospective.
Note that the actual porosity of the organic shale
will be less this estimate because a significant
portion of the anomaly is due to an incorrect
matrix density estimate for this calculation. This
is due to the low grain density of kerogen [1.1 to
1.4 gm/cm3 (Tissot and Welte, 1978)], in
contrast to a minimum of 2.65 g/cm3 for the
inorganic matrix. Algorithms have been
published that estimate TOC content based
solely on the bulk density of the formation. A
notable example is one published by Schmoker
and Hester (1983) that is based on core analysis
from the Bakken, Appalachian Devonian shales
and Woodford shales. This type of algorithm
implicitly assumes that any change in bulk
density is due to variations in kerogen content
and that the inorganic grain density is invariant.
The Schmoker equation assumes an inorganic
grain density of 2.69 g/cm3. This algorithm does
return a reasonable TOC estimate in clay-poor,
thermally mature organic shales—basically, the
shales used for this calibration—and a lithotype
that is common for organic shales. However, this
algorithm underestimates TOC content in clay-
or carbonate-rich shales (too low of a grain
density) and overestimates TOC in thermally
immature shales. Nevertheless, a density-based
estimate is simple to calculate; gamma-gamma
density logs are commonly acquired, and there is
limited need for environmental corrections. It
provides a reasonable first-pass estimate of TOC
content.
logR. The elevated apparent porosity values for
organic shale are not limited to gamma-gamma
density. Sonic and neutron logs commonly also
exhibit elevated apparent porosity in organic
shales. The compressional slowness of kerogen
is high, 109 µs/ft (Vernik et al., 1994), compared
GG
SPWLA 54th
Annual Logging Symposium, June 22-26, 2013
5
to an estimated 54 to 63 µs/ft for a shaly matrix,
parallel and perpendicular to bedding (Vernik et
al., 1994). The presence of kerogen will increase
the apparent sonic porosity if the matrix
slowness remains invariant. Neutron porosity
response in an organic shale is not as
straightforward as the responses for gamma-
gamma density and compressional sonic. The
neutron responds primarily to hydrogen in the
formation, and this element can reside in: a) pore
fluids (water, oil, or gas), b) clay-bound water,
c) hydroxyls within the clay mineral matrix, d)
bitumen and e) kerogen. The multiple sources of
hydrogen can lead to a complicated neutron
response in organic shales.
Compared to conventional shales, formation
resistivity is commonly elevated in organic
shales because of their low total water
saturations. This is a function of low effective
water saturation caused by the dominance of
hydrophobic kerogen-hosted pores (Rylander et
al., 2013), and a generally lower clay mineral
content, with associated lower clay-bound water,
in organic shales. Low-resistivity organic shales,
with resistivities less than 15 ohm-m and
neutron porosity greater than 30 p.u. (limestone
matrix), typically contain expandable clay
minerals.
Passey et al. (1990) developed an empirical
relationship to estimate TOC that is driven by
the elevated formation resistivity and elevated
apparent porosity typical in organic shales as
described above. A porosity curve, typically
sonic porosity, is overplotted with resistivity and
scaled so that they overlie, ―baseline,‖ in zones
with conventional shales above and/or below the
organic shale. The difference between these
curves, called LogR, is equated to TOC
content. An estimate of thermal maturity and
kerogen type is necessary to convert LogR to
TOC. This methodology is underlain by the
assumption that the baseline shale sections and
the organic shale have similar matrix properties
regarding resistivity and apparent porosity. This
resulting TOC estimate is very sensitive to the
thermal maturity estimate (Issler et al, 2002). If
there is a priori knowledge of the TOC content,
then the thermal maturity can usually be
adjusted to give a good TOC log representation.
Finally, the presence of expandable clays, with
low resistivities due to increased clay-bound
water, can lead to a significant underestimate of
TOC in zones where they are encountered.
LogR was later modified (Passey et al., 2010)
to estimate TOC for shales within the gas
window.
NMR density logs. The integration of NMR,
density, and geochemical logs can be used to
directly measure kerogen volume in an organic
shale. Simplistically, the low grain density of
kerogen leads to the density log identifying a
kerogen-rich zone as porous, yet the kerogen
will appear as matrix to an NMR log. The
difference in these two volumes can be equated
to kerogen volume (Figure 4). In practice, a
matrix density from a geochemical log is used to
calculate the density porosity because this
parameter does not include carbon in its
estimation (Herron and Herron, 2000). The
estimated density porosity will include the
kerogen volume.
Fig.4 Petrophysical representation of kerogen
volume estimate from integration of gamma-
gamma density, geochemical, and NMR logs.
The estimation is:
(1)
where
(2)
(3)
GG
SPWLA 54th
Annual Logging Symposium, June 22-26, 2013
6
and
= volume kerogen
= volume pore fluid
= apparent matrix density from geochemical
log (not corrected for organic carbon)
= kerogen density
= pore fluid density
= total NMR porosity
= hydrogen index of fluid
The conversion of kerogen volume to TOC, which
is in weight percent, is provided by Tissot and
Welte (1978):
(4)
= kerogen conversion factor (typically 1.2 to 1.4)
= bulk density
This methodology is generally robust. The
kerogen conversion factor is generally close to
1.2. It will decrease as the organic shale
thermally matures and the kerogen becomes
more carbon rich. Commensurately, the kerogen
grain density will increase as the shale thermally
matures. This methodology does require an
accurate total porosity from NMR, and some
organic shales do include fluids (e.g., clay-
bound water, bitumen) that relax faster than the
lowest T2 measured. An NMR logging tool
with a short echo spacing is advantageous (Hook
et al., 2011).
Geochemical logs.A geochemical log that uses
inelastic spectral measurements can directly
quantify the total carbon within the formation
plus mud (Herron, 1991; Herron et al., 2011;
Jacobi et al., 2008). The total carbon measured
with this technology (TC) needs to be
apportioned to estimate TOC:
Total Carbon = Inorganic Carbon + Organic
Carbon (5)
The inorganic carbon (TIC) can be quantified by
measuring calcium and magnesium with elastic
and inelastic spectral measurements and
calculating the amount of carbon that would be
bound into the appropriate inorganic minerals—
calcite and dolomite. The difference between
this value and the total carbon provides organic
carbon. In theory, this is an excellent way to
quantify organic carbon if the elemental
concentrations are accurate. There still are a few
issues to be aware of: 1) the calculated organic
carbon includes not only carbon in the kerogen,
but also carbon that resides in hydrocarbons in
the pore volume (e.g., gas, oil, bitumen) as
discussed previously and 2) some calcium and
magnesium may be in minerals other than
carbonates (e.g., clays) so the estimation of
inorganic carbon may not be straightforward.
The recently introduced neutron-induced capture
and inelastic gamma ray spectroscopy tool
allows the application of this last method to a
new level of accuracy and precision in the
industry. The combination of capture and
inelastic yields, including carbon, and required
borehole and formation corrections are described
below.
FROM SPECTRA TO TOC
While much of the basic tool analysis has been
described in Radtke, et al. (2012), we summarize the
more salient aspects here. The pulsed neutron source
used by the tool allows for a clean separation of
gamma rays produced by thermal neutron capture
from those produced by fast-neutron interactions
(mostly inelastic scattering). The spectra resulting
from the capture and inelastic reactions are analyzed
separately to produce two sets of elemental yields
(i.e., the fraction of the spectrum due to each
element). The capture set typically includes Si, Ca,
Mg, S, Al, Fe, Mn, Ti, K, Gd, Na, Ba, H, Cl, and tool
background. The inelastic set typically includes Si,
Ca, Mg, S, Fe, Ba, NaCl, C, O, and tool background.
Both sets of elemental yields are converted to
elemental weight fractions of the dry rock using a
simple transform:
(6)
where
= weight fraction of element i
= spectral yield of element i
= weight fraction sensitivity of element i
= yields-to-weights transform factor at depth d
GG
SPWLA 54th
Annual Logging Symposium, June 22-26, 2013
7
Although the transform equation has the same
form for both capture and inelastic elements, the
transform factor is governed by different
physics for the two sets of elements and must be
determined completely independently. For the
capture set we can robustly measure a sufficient
fraction of the commonly occurring matrix
elements to justify closure normalization. The
closure equation is expressed simply as:
(7)
where
= association factor of element i
The closure sum includes the elements Si, Ca,
Mg, S, Fe, Al, Ti, K, and Mn. The association
factors we use have been optimized from X-ray
fluorescence measurements on thousands of core
samples from around the world, resulting in an
accuracy of a few percent for the mineral
assemblages encountered in the oil field. The
beauty of closure normalization is that it
automatically accounts for the many
environmental parameters (e.g., borehole and
formation capture cross section, borehole size,
standoff, formation density, porosity) that alter
without needing to measure these parameters
or to quantify the perturbing effects of these
parameters using extensive calibration
measurements. However, acceptable accuracy
with closure normalization is achieved only with
accurate measurements of all of the elements
listed above. This is achieved for the capture
elements but not for the inelastic elements.
To determine the totally independent for the
inelastic elements one can simply insist on a
statistically weighted equivalence between the
elements measured robustly via both reactions.
This statistical weighting tends to be driven by
the common elements Si, Ca, and Mg. However,
once the inelastic is determined, it can be
applied to all of the elements measured via
inelastic, including, most importantly, carbon.
This process results in an extensive set of
elemental weight fractions, including carbon,
that all have a common reference, namely the
weight of the dry rock.
Corrections to carbon concentrations for OBM (oil in
the borehole). A complete description of the oil-
borehole correction is beyond the scope of this paper.
Suffice to say here that estimating the carbon
contribution from the borehole fluid is challenging
but the existing algorithm has been shown to be quite
accurate.
Corrections to carbon concentrations for inorganic
carbon. Assuming any carbon in the borehole has
been properly removed, the carbon concentration
measured represents total formation carbon, once
again relative to the total weight of the dry rock. To
get to TOC we must remove all of the inorganic
carbon found in the dry rock. The most common
minerals containing carbon in the oil field, together
with their diagnostic elements, include calcite (Ca),
dolomite (Ca, Mg), siderite (Fe), rhodochrosite (Mn),
ankerite (Ca, Mg, Mn, Fe), nahcolite (Na), and
dawsonite (Na, Al). With the extensive set of rock
matrix elements measured accurately by the new
spectroscopy tool (Si, Ca, Mg, Fe, Mn, Al, Na, S, Ti,
K), one can accurately estimate the concentrations of
the above carbon-bearing minerals, either from a
sequential process (Herron and Herron, 1996) or from
a more complete inversion. Since the carbon
concentration of these minerals is well known,
removing the inorganic carbon from the total carbon
is straightforward once the carbon-bearing mineral
fractions have been estimated.
CASE STUDY
In the case study, we present the TOC log of the new
spectroscopy tool of two wells from an
unconventional shale play in North America across
producible shale formations.
Figure 5 illustrates in a depth log the new carbon
outputs of the spectroscopy tool for an interval of
well 1. The improved measurement of carbon (TC)
combined with more robust estimation of mineralogy,
and therefore TIC, produces a new continuous TOC
log free from local calibrations and multitool
interpretations. TOC core from LECO analysis is also
displayed for comparison. The excellent match
between log and core TOC is generally quite good,
despite some local rapid TOC variations.
The entire section of wells 1 and 2 can be seen in
Figures 7 to 11 (tracks 1 and 2). Between the two
wells, a total of 68 sidewall cores have been collected
and TOC analyses performed. Crossplots (Figure 6)
GG
SPWLA 54th
Annual Logging Symposium, June 22-26, 2013
8
show a very good agreement with TOC core for both
wells.
Fig.5 Example of TOC determination technique
showing the new TC log and TIC log. In the left
track, TIC (in red) is subtracted from TC (in black) to
derive TOC. TOC from core using LECO analysis is
displayed with red dots.
Comparison with other models to estimate TOC.
Alternative models to quantify TOC described
earlier during the ―Literature Review‖ have been
calculated and their results compared to TOC
estimated with the new spectroscopy carbon
measurement presented in this paper. These
models are: density, uranium, LogR, and NMR-
density-geochemical. The comparisons for both wells
are illustrated in Figures 7 to 11 (depth log format)
and also in Figure 12 individually for each method
(crossplot format).
The common denominator of these alternative
methods is the need for calibration or extensive log
analyst interpretation. The exception to this is the
density model (Schmoker) and of course the
spectroscopy carbon measurement, the focus of this
paper. However, what we typically observe with the
density model is that depending on the environment
this is used in, it can lead to overestimation or
underestimation of TOC. In this case, we observed a
general overestimation of TOC for both wells (Figure
12-c).
Fig.6 Crossplot comparing core TOC and
geochemical spectroscopy, (a) well 1, (b) well 2. The
match core to log, is represented in the crossplot by
how close the trend line (in black) to the one to one
line (in red) is.
Although already mentioned, it is important to
reiterate that TOC quantification from uranium
(spectral gamma ray log) requires a local empirical
relationship to be derived. This is normally done with
core data. In this case, we have obtained relationship
with a priori knowledge (core) that while it seems to
provide reasonable results for well 1, it is clearly
GG
SPWLA 54th
Annual Logging Symposium, June 22-26, 2013
9
overestimating at the bottom of well 2 (Figure 11)
and missing the top peaks with high TOC content at
the top (Figure 12-b). The variability in this
comparison is typical for gamma ray/uranium.
LogR also requires a high input from the log
analyst. Setting the porosity-resistivity baseline is not
always straightforward and could, in some cases,
require an analyst with considerable experience using
the technique. Other important inputs to this
methodology are the thermal maturity (also known as
level of maturity, LOM) and the kerogen type. For
these two wells, the known LOM from core data
estimated TOC content (curve in pink of depth log
figures) that was much lower than core. In order to
match core a significantly lower LOM was input.
(curve in blue in depth log figures). The maturity
dependency of this method can also be seen in Figure
12-d.
The last model in these comparisons is the integration
of NMR, density, and geochemical logs. The
methodology is robust, as seen in both wells (Figure
12-d). The new spectroscopy tool contributes to this
method with an accurate and precise matrix density
as a result of improved mineralogy estimation. This
methodology also requires a priori knowledge of
kerogen maturity (conversion constant) and kerogen
density.
CONCLUSION
The application of a new neutron-induced gamma ray
spectroscopy measurement to estimate TOC in
organic shale rocks has been presented here.
From the examples shown in this paper, the proposed
method stands out as a reliable method to estimate
TOC without the need for local correlations and
extensive log interpretation. The alternative methods
to quantify TOC show a variable dependency of these
two factors. A noticeable approach is the integration
of NMR, density and geochemical logs which shows
reasonable good agreement for both wells.
The new pulsed neutron source in combination with
the also new lanthanum bromide scintillator allow for
accurately measure elemental capture and inelastic
yields including carbon. The improved carbon
measurement and mineralogy calculation provide a
better TOC quantification for, ultimately, a better
organic shale evaluation.
ACKNOWLEDGMENTS
The authors thank Energen Resources Corporation
and Bold Operating Ltd., for supporting field testing
of this tool in their wells and for allowing the use of
their well log data. We also want to thank the
Schlumberger development team in Sugar Land,
Texas, Princeton, New Jersey, and Cambridge,
Massachusetts, for bringing this tool from concept to
reality.
REFERENCES
Ambrose, R.J., Hartman, R.C., Diaz-Campos, M.,
Akkutla, I.Y., and Sondergeld, C.H., 2010, New
pore-scale considerations for shale gas in place
calculations: SPE 131772, SPE Unconventional Gas
Conference, 23–25 February, Pittsburgh,
Pennsylvania, USA,
Fertl, W.H., and Rieke, H.H, 1980, Gamma ray
spectral evaluation techniques identify fractured shale
reservoirs and source-rock characteristics: Journal of
Petroleum Technology, 31, 2053–2062.
Herron, S.L., 1991, In situ evaluation of potential
source rocks by wireline logs, Wireline source-rock
evaluation in the Paris Basin: AAPG Treatise of
Petroleum Geology: Handbook of petroleum geology,
source and migration processes and evaluation
techniques (AAPG, Tulsa), 127–134.
Herron, M.M., Grau, J.A., Herron, S.L., Kleinberg,
R.L., Machlus, M., Reeder, S.L., Vissapragada, B.,
Burnham, A.K., Day, R.L., and Allix, P., 2011, Total
organic carbon and formation evaluation with
wireline logs in the Green River oil shale: SPE
147184, SPE Annual Technical Conference and
Exhibition, 30 October–2 November, Denver,
Colorado, USA.
Herron, S. L., and Herron, M. M., 1996, Quantitative
lithology: An application for open and cased hole
spectroscopy: Transactions of the SPWLA 37th
Annual Logging Symposium, 16–19 June, New
Orleans, Louisiana, USA, Paper E.
Herron, S.L. and Herron, M.M, 2000, Application of
nuclear spectroscopy logs to the derivation of
formation matrix density, Paper JJ, SPWLA Annual
Logging Symposium, 4–7 June, Dallas, Texas, USA.
GG
SPWLA 54th
Annual Logging Symposium, June 22-26, 2013
10
Hertzog, R. C., 1978, Laboratory and field evaluation
of an inelastic neutron scattering and capture gamma
ray tool: SPE 7430, SPE Annual Technical
Conference and Exhibition, 1–4 October, Houston,
Texas, USA.
Hook, P., Fairhust, D., Ryalander, E., Bachman, N.,
Cary, S., Chatawanich, K., Taylor, T., 2011,
Improved Precision Magnetic Resonance
Acquisition: Application to Shale Evaluation, SPE
Annual Technical Conference and Exhibition, 30
October – 2 November, Denver, USA
Issler, D.R., Hu, K., Block, J.D., and Katsube, T.J.,
2002, Organic carbon content determined from well
logs: Examples from Cretaceous sediments of
Western Canada: Geological Survey of Canada, Open
File 4362, 19 p.
Jacobi, D., Gladkikh, M., LeCompte, B., Hursan, G.,
Mendez, F., Longo, J., Ong, S., Bratovich, M.,
Patton, G., and Shoemaker, P., 2008, Integrated
petrophysical evaluation of shale gas reservoirs: SPE
114925, SPE Gas Technology Symposium, 16–19
June, Calgary, Alberta, Canada.
Lock, G.A., and Hoyer, W.A., 1974, Carbon-oxygen
(C/O) log: Use and interpretation: Journal of
Petroleum Technology 26(9), 1044–1054.
Passey, Q.R., Creaney, S., Kulla, J.B., Moretti, F.J.,
and Stroud, J.D., 1990, A practical model for organic
richness from porosity and resistivity logs: AAPG
Bulletin, 74, 1777-1794.
Passey, Q.R., Bohacs, K.M., Esch, W.L.,
Klimentidis, R., and Sinha, S., 2010, From oil-prone
source rock to gas-producing shale reservoir—
Geologic and petrophysical characterization of
unconventional shale-gas reservoirs: SPE 13150, SPE
International Oil & Gas Conference and Exhibition,
8–10 June, Beijing, China.
Radtke, R. J., Lorente, M., Adolph, B., Berheide, M.,
Fricke, S., Grau, J., Herron, S., Horkowitz, J., Jorion, B.,
Madio, D., May, D., Miles, J., Perkins, L., Philip, O.,
Roscoe, B., Rose, D., and Stoller, C., 2012, A new
capture and inelastic spectroscopy tool takes
geochemical logging to the next level: Transactions of
the SPWLA 53rd Annual Logging Symposium, 16–20
June, Cartagena, Columbia, Paper AAA.
Rylander, E., Singer, P.M., Jiang, T., Lewis, R.,
McLin, R., and Sinclair, S., 2013, NMR T2
distributions in the Eagle Ford Shale: Reflections on
pore size: SPE 164554, SPE Unconventional
Resources Conference, 10–12 April, The Woodlands,
Texas, USA.
Schmoker, J.W., 1981, Determination of organic-
matter content of Appalachian Devonian shales from
gamma-ray logs:. American Association of Petroleum
Geologists Bulletin, 65(7), 1258–1298.
Schmoker, J.W., and Hester, T.C, 1983, Organic
carbon in Bakken Formation, United States portion of
the Williston Basin: AAPG Bulletin, 67(12), 2165–
2174.
Tittman, J., and Nelligan, W.B., 1960, Laboratory
Studies of a Pulsed Neutron-Source Technique in
Well Logging: Journal of Petroleum Technology, 63-
66.
Tissot, B.P., and Welte, D.H., 1978, Petroleum
Formation and Occurrence: Springer-Verlag, Berlin,
538 p.
Vernik, L, Liu, X, and Nur, A., 1994, Effect of
kerogen on velocity anisotropy in source rocks: SEG
Annual Meeting, 23–28 October, 323–326.
Ward, E., and Harry D., 1974, Laboratory and Field
Evaluation Of a Carbon/Oxygen (C/O) Well Logging
Sytem. Journal of Petroleum Technology. October,
1103-1110.
ABOUT THE AUTHORS
Jorge Gonzalez is a Petrophysicist currently
working in the in the Permian Basin in Midland as
an Associate Domain Champion for Schlumberger
Formation Evaluation, Wireline Services. Jorge has
been with Schlumberger for 7 years starting as a
Wireline Field Engineer for a short period and then
as a Log Analyst in the North Sea Data Services
Center. Jorge has received an MSc degree in
Mining Engineer from the Universidad Politecnica
in Madrid.
Richard Lewis is the Petrophysics Technical
Manager for the Schlumberger Unconventional
Resource Group, and he is located in Dallas. Rick
was a developer of the gas shale evaluation workflow
GG
SPWLA 54th
Annual Logging Symposium, June 22-26, 2013
11
that was initially fielded 10 years ago and which has
been applied to organic shales throughout North
America. In his current position, Rick manages a
group responsible for the continual improvement for
this workflow with application to lacustrine and
liquid-producing shales, and for its introduction and
application to the international market. He is also the
interface to the Schlumberger research and
engineering groups for the development of evaluation
technologies for organic shales. Prior to this
assignment, Rick was responsible for wireline
interpretation development for the central and eastern
United States. Rick has also worked for Shell Oil,
Battelle PNNL, and the U.S. Geological Survey. He
received a BS degree from UCLA and MS and PhD
degrees from Cal Tech, all in geology.
James Hemingway has been with Schlumberger
since January 1980, starting as a field engineer; since
then, he has held various log analyst and engineering
positions during his career. He is currently based in
Houston. Jim has authored many papers about pulsed
neutron logging and log interpretation. James
received a BS degree in chemistry from Emporia
State University in 1978 and a BS degree in chemical
engineering from Texas A&M University in 1979.
From 1982 to 1997 he held several assignments as a
log analyst, interpretation center manager, and
interpretation development engineer. Jim worked in
Ventura and Bakersfield from 1990 to 1997 as the
interpretation center manager. In 1997, he joined the
Formation Evaluation department at the
Schlumberger Sugar Land Product Center working on
the RSTPro* tool and "three-phase holdup"
interpretation techniques. He moved to Paris in 2001
as a new technology advisor responsible for teaching
the applications of new technology for formation
evaluation. In 2005, he was promoted to the position
of nuclear technology advisor. He has been based in
Houston since 2010 as a Petrophysics Advisor.
Jim Grau Jim Grau is a scientific advisor at
Schlumberger-Doll Research in Cambridge,
Massachusetts. He has been involved for over 30
years in all aspects of borehole elemental analysis
using nuclear spectroscopy techniques, including tool
design, data acquisition software, and spectral
analysis techniques. Jim received a BS in
engineering physics from the University of Toledo in
* Mark of Schlumberger
Toledo, Ohio, and an MS and PhD in experimental
nuclear physics from Purdue University in West
Lafayette, Indiana.
Erik Rylander is a Senior Log Analyst for the
Unconventional Resources Group of Oilfield
Services located in Dallas, TX. Erik is part of a
group responsible for the continual improvement of
shale analysis workflows and for their introduction
and application to the international market. Erik’s
current focus is the application of nuclear magnetic
resonance logs to liquid-bearing shale reservoirs.
Erik has been with Schlumberger for 18 years and
began as a wireline field engineer in West Africa and
the Gulf of Mexico. After an operations management
position, Erik entered into a product development
role at the Schlumberger product center in Clamart,
France, where he focused on the application of
dielectric logs. Eight years ago, Erik began focusing
on the petrophysical analysis of shale reservoirs in
Dallas, Texas. Erik has authored or coauthored seven
papers and holds two patents related to formation
testing. He received his BS degree from the
Colorado School of Mines in engineering.
Ryan Schmitt is a Segment Sales Engineer at
Schlumberger. He has spent the majority of his
career in Operations as a Field Engineer, gaining
operating proficiency in all of Schlumberger’s
reservoir evaluation logging tools. Ryan received a
BS in Electrical Engineering from Texas A&M
University in College Station, TX.
GG
SPWLA 54th
Annual Logging Symposium, June 22-26, 2013
12
Fig.7 Comparison of TOC measured by the new tool and comparison with core (track 2) in topmost interval of
well 1 of the unconventional shale play in North America. Also, comparison with alternative TOC estimation
models: NMR (track 3), uranium (track 4), density (track 5), and LogR (track 6). Two TOC ∆Log R
computations were done with different LOM; curve in pink represents TOC using LOM from core and curve in light
blue represent and adjusted LOM to match TOC content from core.
Fig.8 Comparison of TOC measured by the new tool and comparison with core (track 2) for second topmost
interval of well 1 of the unconventional shale play in North America. Also, comparison with alternative TOC
estimation models: NMR (track 3), uranium (track 4), density (track 5), and LogR (track 6). Two TOC ∆Log R
computations were done with different LOM; curve in pink represents TOC using LOM from core and curve in light
blue represent and adjusted LOM to match TOC content from core.
GG
SPWLA 54th
Annual Logging Symposium, June 22-26, 2013
13
Fig. 9 Comparison of TOC measured by the new tool and comparison with core (track 2) for second bottom-most
interval of well 1 of the unconventional shale play in North America. Also comparison with alternative TOC
estimation models: NMR (track 3), uranium (track 4), density (track 5), and LogR (track 6). Two TOC ∆Log R
computations were done with different LOM; curve in pink represents TOC using LOM from core and curve in light
blue represent and adjusted LOM to match TOC content from core.
Fig.10 Comparison of TOC measured by the new tool and comparison with core (track 2) for bottom-most
interval of well 1 of the unconventional shale play in North America. Also, comparison with alternative TOC
estimation models;: NMR (track 3), uranium (track 4), density (track 5), and ΔLogR (track 6). Two TOC ∆Log R
computations were done with different LOM; curve in pink represents TOC using LOM from core and curve in light
blue represent and adjusted LOM to match TOC content from core
GG
SPWLA 54th
Annual Logging Symposium, June 22-26, 2013
14
Fig.11 Comparison of TOC measured by the new tool and comparison with core (track 2) for well 2 of the
unconventional shale play in North America. Also, comparison with alternative TOC estimation models: NMR
(track 3), uranium (track 4), density (track 5), and LogR (track 6). Two TOC ∆Log R computations were done
with different LOM; curve in pink represents TOC using LOM from core and curve in light blue represent and
adjusted LOM to match TOC content from core.
GG
SPWLA 54th
Annual Logging Symposium, June 22-26, 2013
15
Fig.12 TOC for well 1 & 2 comparing core TOC versus; (a)NMR-spectroscopy, (b)uranium TOC, (c) density and (d)
∆Log R TOC models. Two TOC ∆Log R computations were done with different LOM; dots in pink represent TOC
using LOM from core and dots in light blue represent and adjusted LOM to match TOC content from core. For each
method, the match core to log, is represented in the crossplot by how close the trend line (in black) to the one to
one line (in red) is.
Well 1 –a) Well 1 –b)
Well 1 –c) Well 1 –d)
Well 2 –a) Well 2 –b)
Well 2 –c) Well 2 –d)
GG