Calculating standard captured γ spectra of formation elements
Pet.Sci.
Calculating standard captured spectra of formation elements
Wu Wensheng 0
Xiao Lizhi 0
Zhang Lijuan 0
Niu Wei 0
0 State Key Laboratory of Petroleum Resource and Prospecting, China University of Petroleum , Beijing 102249 , China
Multielement gammaray spectrum logging requires standard spectra of individual elements for its interpretation. Since the standard element spectra are usually derived using Monte Carlo simulation be observed using a NaI(Tl) detector) of elements H, Si, Ca and Fe from each element and its oxide. To compare the standard spectra from the elements and oxides, we operated three simulations of sandstone, limestone and mixed formation of sandstone and limestone each with ten different porosities, and used the two kinds of standard spectra to analyze the mixed spectra modeled from sandstone and limestone formations. The results show that the standard element spectra from oxides have more prominent energy peaks than the standard spectra from pure elements. The calculated formation element contents are close to the theoretical values when the standard element spectra from oxides are used to analyze the formation mixed spectra. Therefore, the formation element standard spectra should be calculated from oxide models in the analysis of neutron captured spectra by logging tools.
Standard spectrum; spectrum analysis; neutron captured spectrum; formation element

Multielement spectrum logging can effectively identify
lithology and mineralogy in formations, and plays an
important role in exploring complex reservoirs
(Hertzog et al,
1989; Chapman et al, 1987; Grau and Schweitzer, 1989; Grau
et al, 1990)
. Accurate and effective full spectrum analysis
logging, and the spectrum analysis method requires simulated
There are two ways to acquire standard element spectra,
one is instrument measurement in a standard calibration
model, and the other is numerical simulation
(Nguyen et al,
1996; AlGhorabie, 2006; Shi et al, 2002; Xiang and Guo,
2006)
. As numerical simulation can not completely take the
real tool and well conditions into account, it is better to obtain
the standard spectra with a spectrum tool in a model well.
Owing to a variety of reasons, almost all publications do not
describe elements or compounds used to synthesize standard
element spectra in the model well. Generally, either a pure
element or its oxide can be used to obtain standard element
spectra by Monte Carlo modeling.
With the Monte Carlo numerical simulation program
MCNP(5C)
(Pei and Zhang, 1980)
, we studied the differences
of spectrum analysis using standard spectra calculated either
from pure elements or their oxides.
2 Numerical simulation model
The calculation model is a cylinder with a height of 1 m
and a radius of 70 cm, as shown in Fig. 1. The diameter of
the logging tool with a pulsed neutron source is 45 mm. The
source emits 14 MeV neutrons into the formation with a pulse
detector 40 cm away from the source and the detector is 10
cm in length and 4 cm in diameter. The tool is pressed closely
up against one side of the borehole. The space between the
detector responses are considered in the calculation.
Detector
Source
Borehole
Formation
A twostep simulation is used in the calculation. First is
the simulation for neutrons. When photons are generated,
the position where photons are generated is treated as the
departure position. The energy and motion direction of the
photons are determined by random sampling of the known
energy distribution and motion direction distribution, thus
the second step simulation begins
(Hendriks et al, 2002)
.
The advantages of this twostep simulation are significant
whole energy, single escape and double escape peaks, short
computation time, and low calculation error generally less
than 0.1%.
3 The simulation for standard element
spectra in formation
elements
With the above calculation model and calculation
techniques, the next most critical step is filling the element
in the model for standard spectrum simulation. Since the
standard element spectrum is the detector response of the
logging tool to a transient nuclear reaction of atomic nuclei
in the borehole, pure substances are first considered to be
filled in the model to simulate element standard spectra. In
computation, the pure element in the model is respectively H,
in Fig. 2. The inelastic scattered gamma counts are deducted
from the captured spectra counts in Fig. 2. Because the
spectra are obtained between two neutron pulses, the spectra
gamma rays.
Ca
Si
Fe
10
y
its 1
n
e
t
n
i
ive0.1
t
a
l
e
R
0.01
1E3
H
0
1
2
3
4
5
6
7
8
EnergyMeV
Fig. 2 shows that there is a characteristic peak of H at
2.23 MeV; two whole energy peaks of Si appear at 3.54 MeV
and 4.93 MeV; Ca peaks are at 1.94 MeV and 6.42 MeV; Fe
peaks are at 7.46 MeV and 5.92 MeV
(Huang, 1985)
. In the
figure, the first peak on the left is the backscattered peak.
through the detector crystal without being detected. They then
interact with the atoms of the materials behind the detector
back at the detector, are detected and form the backscattered
lowenergy peak. As well, backscattered photons produced
in shielding materials also contribute to the backscattered
peak. The energy of backscattered photons is always around
200 keV, so it is easy to identify the backscattered peak in the
spectrum.
3.2 The simulation for the formation filled with
element oxide
Considering the element in the strata is not a pure element,
but an oxide form, for example, element H occurs as H2O,
and Si occurs as SiO2 2O,
SiO2, CaCO3, and FeO, and the calculated captured spectra
are shown in Fig. 3.
100
ity 10
s
n
e
t
in 1
e
v
it
a
l
e
R 0.1
0.01
0
4 Comparison of energy spectrum analysis
results
4.1 Sandstone with different porosities
In order to compare the spectrum analysis results based on
the above two types of element standard spectra, it is assumed
that there is a sandstone reservoir with different porosities
from 1% to 45% corresponding to ten points, and the pore is
SiO2
Si
6
The program MCNP (5C) is used to simulate the captured
analysis, the element standard spectra are isolated from
the oxide spectra by mathematical processing, and these
element standard spectra are called oxide standard spectra.
The standard spectra assuming pure elements are called
pure element standard spectra. To conveniently perform
mathematical processing, the standard spectra and formation
mixed spectra are expressed as 256dimensional vector P and
the components are normalized according to Eq. (1)
256
With oxide standard spectra and pure element standard
spectra, a full spectrum leastsquare method is used to analyze
the mixed spectra for the ten points in Table 1 and obtain the
relative yields of H and Si. In order to calculate the weight
percentage of each element with the oxide model, each point
in Table 1 is calibrated to get the normalization factor F
and sensitivity factor S, and then the weight percentages of
elements at each point are calculated by Eq. (2) (Pang et al,
0.01
0
(2)
2005; Pang and Li, 2006):
Wt j
F y j / S j
value is equal to the theoretical value.
Fig. 5 shows that the H content obtained from a pure
element standard spectrum is higher than the corresponding
theoretical value. The content of H based on oxide standard
spectrum is lower than the corresponding theoretical value for
low porosities, and higher than the corresponding theoretical
value for high porosities. On the whole, the analysis results
using the oxide standard spectrum are closer to the diagonal
dashed line than those of the pure element standard spectrum,
which demonstrates that spectrum analysis with the standard
spectrum calculated from the oxide can provide a more
accurate estimate of the hydrogen content.
Fig. 6 shows that the content of Si calculated from
the oxide standard spectrum is extremely close to the
corresponding theoretical value, and the content derived
from the pure element standard spectrum is lower than the
corresponding theoretical value. The lower the formation
porosity, the higher the deviation degree.
Therefore, according to the spectrum analysis results in
sandstone formation, the computation accuracy of element
content with oxide standard spectrum is better than that with
pure element standard spectrum.
4.2 Limestone with different porosities
To compare the spectrum analysis results for oxide
standard spectrum and pure element standard spectrum, it is
assumed that there is a limestone reservoir with the porosity
0.375
0.360
from 1% to 45% corresponding to ten points, and the pores
value, it is still close to it. The content of Ca from a pure
element standard spectrum is significantly higher than the
corresponding theoretical one.
Therefore, from the spectrum analysis results in the
limestone formation, the accuracy of the element content
calculated from the oxide standard spectrum is better than
that from the pure element standard spectrum.
The analysis results with an oxide standard spectrum and
pure element standard spectrum for the mixed spectra at ten
points in Table 2 are shown in Fig. 7 and Fig. 8.
Fig. 7 shows that the intersections of calculated content
of H using an oxide standard spectrum and the theoretical
content are almost all located on the diagonal line, which
means the calculated value is almost the same as the
theoretical one. The content of H obtained from the pure
element standard spectrum is much lower than the theoretical
one, and the higher the formation porosity, the larger the
difference between the calculated value and the theoretical
value. These show that the accuracy of H content calculated
from the oxide standard spectrum is relatively high.
In Fig. 8, although the content of Ca obtained from an
oxide standard spectrum is slightly higher than the theoretical
Pure element standard spectrum
Oxide standard spectrum 0.26 0.28
Theoretical value
0.30
0.32
4.3 Mixed formation of sandstone and limestone with
different porosities
To further compare the spectrum analysis results in
complex formations with the above two types of standard
spectra, we suppose that there is a reservoir with the lithology
of limestone and sandstone with porosity from 2% to 10%
corresponding to nine points, and the formation pores are
CH, CSi and
CCa in Table 3 represent the weight percentages of H, Si and
Ca respectively.
Fig. 9 and Fig. 10 show the analysis results with oxide
standard spectrum and pure element standard spectrum for
the mixed spectra at nine points in Table 3.
As shown in Fig. 9, when the content of Si is higher
than 0.2, the calculated value using pure element standard
spectrum is higher than the theoretical one, and the higher
the content of Si, the higher the deviation degree from the
theoretical value. When the content of Si is less than 0.2,
the content of Si obtained from a pure element standard
spectrum is lower than the theoretical value, and even
negative. However, the intersections of the calculated content
of H using oxide standard spectra and the theoretical content
are almost all distributed along the diagonal line, which
means the calculated value is extremely close to the model
composition. Fig. 10 shows information similar to that of Fig.
9.
1.0
0.8
0.6
e
lua0.4
v
d
ltae0.2
u
c
l
a
C0.0
0.2
0.4
Si
The same can be seen from the analysis results of mixed
formation spectra that the computation accuracy of element
content using the oxide standard spectrum is better than that
with a pure element standard spectrum.
The standard spectra of formation elements can be
calculated using either pure elements or their oxides.
Compared with the standard spectra using a pure element,
the standard spectra using the oxide have more prominent
characteristic energy peaks. It can be seen from the analysis
results of different formation models that the element content
calculated with oxide standard spectrum is close to the
theoretical value. Therefore, element standard spectra should
be calculated from the oxides.
Acknowledgements
This paper is financially supported by National Natural
Science Foundation of China (Grant No. 41074101) and
Science Foundation of China University of Petroleum
(Beijing) (No. KYJJ20120512).
Al Ghorabie F H H. Development of a computer code using the EGS4 Monte Carlo simulation system to evaluate the response of a NaI(Tl) detector to photons with energies below 300 keV . Applied Radiation and Isotopes . 2006 . 64 ( 1 ): 85  92
Chapman S , Colson J L , Everett B , et al. The emergence of geochemical well logging . The Technical Review . 1987 . 35 ( 2 ): 27  35
Grau J A and Schweitzer J S. Elemental concentrations from thermal neutron capture gammaray spectra in geological formations . Nuclear Geophysics . 1989 . 3 ( 1 ): 1  9
Grau J A , Schweitzer J S and Hertzog R C. Statistical uncertainties of elemental concentrations extracted from neutron induced gammaray measurements . IEEE Transactions on Nuclear Science . 1990 . 37 ( 6 ): 2175  2178
Hendriks P H G M , Maucec M and De Meijer R J. MCNP modeling Applied Radiation and Isotopes . 2002 . 57 ( 3 ): 449  457
Hertzog R , Colson L , Seeman O , et al. Geochemical logging with spectrometry tools . SPE Formation Evaluation . 1989 . 4 ( 2 ): 153  162 ( SPE16792 )
Huang L J. The Principle of Radioactive Logging . Beijing: Petroleum Industry Press. 1985 . appendix (in Chinese)
Nguyen H V , Campbell J M, Couchell G P , et al. Programs in C for and unfolding of continuous gamma spectra . Computer Physics Communications . 1996 . 93 ( 23 ): 303  321
Pang J F and Li M. Analytical methods with neutrongamma ray spectrometry from formation elements . Journal of Isotopes . 2006 . 19 ( 2 ): 70  74 (in Chinese)
Pang J F , Li M and Yan Z G. New analyzing method & software and applications of inelastic spectra from C/O spectrometry logs . Nuclear Physics Review . 2005 . 22 ( 1 ): 67  71 (in Chinese)
Pei L C and Zhang X Z. The Monte Carlo Method and Its Application in Particle Transport . Beijing: Science Press. 1980 . 5 50 (in Chinese)
Shi H X , Chen B X , Li T Z , et al. Precise Monte Carlo simulation of gammaray response functions for a NaI(Tl) detector . Applied Radiation and Isotopes . 2002 . 57 ( 4 ): 517  524
Xiang D and Guo L Y. Monte Carlo simulation of gamma spectra . Journal of Mathematical Medicine . 2006 . 19 ( 3 ): 229  231 (in Chinese)