Journal of Natural Gas Science and Engineering 37 (2017) 29e38

Contents lists available at ScienceDirect

Journal of Natural Gas Science and Engineering

journal homepage: www.elsevier.com/locate/jngse

Petrophysical characterization of shale reservoir based on nuclear

magnetic resonance (NMR) experiment: A case study of Lower
Cambrian Qiongzhusi Formation in eastern Yunnan Province, South
China
Ang Li a, b, c, Wenlong Ding a, b, c, *, Ruyue Wang a, b, c, Jianhua He a, b, c, Xinghua Wang a, b, c,
Yaxiong Sun a, b, c, Yang Gu a, b, c, Nailin Jiao a, b, c
a
School of Energy Resources, China University of Geosciences (Beijing), Beijing 100083, China
b
Key Laboratory for Marine Reservoir Evolution and Hydrocarbon Abundance Mechanism, Ministry of Education, China University of Geosciences (Beijing),
Beijing 100083, China
c
Key Laboratory for Shale Gas Exploration and Assessment, Ministry of Land and Resources, China University of Geosciences (Beijing), Beijing 100083, China

a r t i c l e i n f o a b s t r a c t

Article history: In order to characterize the petrophysical properties of shale using NMR technique, eight shale samples
Received 17 May 2016 from the Lower Cambrian Qiongzhusi Formation in the eastern Yunnan province were measured by
Received in revised form porosity and permeability tests, field emission scanning electron microscopy (FE-SEM) and NMR
13 November 2016
experiment. Pore types were obtained from the shape and distribution of transverse relaxation time (T2)
Accepted 15 November 2016
spectrum. Residual porosity and movable porosity could be well estimated based on T2 spectrum area
Available online 16 November 2016
fraction. On the basis of Coates model, we proposed a regional Coates model to calculate the NMR
permeability of shale. A method for determining T2cutoff of shale samples was also expounded. Moreover,
Keywords:
Shale reservoir
the specific surface area distributions and pore size distributions could be obtained based on the
Nuclear magnetic resonance (NMR) mathematical equation of T2. Results show that T2 spectrums of shale samples at water-saturated con-
Petrophysical characterization dition can be divided into unimodal and bimodal T2 spectrums. Continuous bimodal T2 spectrums reflect
T2 spectrums the samples with good connectivity between small pores and large pores, whereas discontinuous
Pore characteristics bimodal T2 spectrums reflect that the connectivity between small pores and large pores is poor. Shale
samples with higher bound water content have a greater T2cutoff. The NMR permeability is close to gas log
permeability, which proves the applicability of regional Coates model. In eight shale samples, transition
pores account for the largest proportion, followed by mesopores, indicating that transition pores and
mesopores are the major sites for the accumulation of shale gas.
© 2016 Elsevier B.V. All rights reserved.

1. Introduction Reservoir quality has a significant influence on the accumulation of

The large-scale commercial development of shale gas in North
America has changed significantly the pattern of world oil and gas
supply. A growing number of countries around the world have
strengthened the exploration and development of this new energy
(Bowker, 2007; Clarkson et al., 2012; Tang et al., 2014; Ding et al.,
2015; Li et al., 2016). Shale as a reservoir of natural gas is charac-
terized by low porosity, low permeability and strong heterogeneity.
* Corresponding author. School of Energy Resources, China University of Geo-
sciences (Beijing), Beijing 100083, China.
E-mail address: dingwenlong2006@126.com (W. Ding).
shale gas. As an important index for evaluating reservoir quality,
shale pore structure has been widely studied by lots of scholars.
Previous studies have indicated that pores in shale can be divided
into four types on the basis of pore size: macropores (pore
diameter > 1000 nm), mesopores (1000 nm < pore
diameter > 100 nm), transition pores (100 nm > pore
diameter > 10 nm), micropores (10 nm > pore diameter) (Zhong,
2012; Curtis, 2002; Caldwell, 2006). Shale gas existing in organic-
rich shale is primarily in two states of free gas and adsorbed gas.
Free gas is preserved in natural fractures and intergranular pores,
and adsorbed gas is stored in the organic matter and clay particle
surfaces (Curtis, 2002). There are a mass of nano-sized and micron-

http://dx.doi.org/10.1016/j.jngse.2016.11.034
1875-5100/© 2016 Elsevier B.V. All rights reserved.
30 A. Li et al. / Journal of Natural Gas Science and Engineering 37 (2017) 29e38
sized pores in shale reservoir which has complicated pore struc-
ture, and some qualitative and quantitative techniques have been
used to characterize shale pores. For example, nano-CT imaging and
field emission scanning electron microscopy (FE-SEM) have been
utilized qualitatively to characterize the structure and morphology
of shale pores (Heath et al., 2011; Curtis et al., 2012; Tiwari et al.,
2013). Low-pressure nitrogen adsorption, small-angle neutron
scattering, high-pressure mercury intrusion and nuclear magnetic
resonance have proven to be effective methods to obtain the
quantitative parameters such as surface area, pore volume and
pore-size distribution (Ross and Bustin, 2007, 2009; Chalmers and
Bustin, 2007; Clarkson et al., 2013; Li et al., 2016). However, some
of these methods have certain limitations. For example, N2
adsorption can not finely characterize parts of macropores and
microfractures within shale. High-pressure mercury intrusion
experiment is generally used to analyze mesopores and macro-
pores. Besides, mercury intrusion may result in the damage of shale
pore structure, thus affecting the reliability of measurement result.
FE-SEM can only be used to observe the local pore characteristics of
shale samples, but it cannot reflect the spatial distribution of pores
and microfractures. Moreover, the primary structure of shale
samples would be destroyed and a lot of pores and microfractures
are produced artificially in the polishing process of rock section,
which leads to great errors for the results. However, NMR has ad-
vantages in studying shale reservoirs with the characteristics of
rapidity, undamage and accuracy. At present, a lot of scholars have
extensively applied NMR technique to the evaluation of carbonate
and sandstone reservoirs, but NMR experiments have not been
used fully in the study of shale reservoirs (Yao et al., 2010a; Yu,
2013; Xu et al., 2015; Tan et al., 2015).
When samples containing fluid are in a uniform static magnetic
field, the hydrogen protons in fluid are polarized to generate a
magnetic vector. At this time, hydrogen protons are stimulated by
frequency pulse to produce nuclear magnetic resonance phenom-
enon. After removing frequency pulse, we can obtain a signal
whose amplitude attenuates with time. Two parameters can be
used to measure the attenuation rate of nuclear magnetic signal:
longitudinal relaxation time (T1) and transverse relaxation time
(T2). Generally, transverse relaxation time spectrum is utilized to
study sample characteristics because the measurement of trans-
verse relaxation time is fast.
The major goals of this paper are to investigate the reservoir
characteristics of the Lower Cambrian Qiongzhusi shale from the
eastern Yunnan Province using NMR experiment. NMR T2 response
characteristics, pore types, porosity, permeability and pore struc-
ture parameters were analyzed based on NMR transverse relaxa-
tion time distributions. Then we compared the difference between
NMR permeability and routine permeability. Meanwhile, we
calculated the NMR T2 cutoff value (T2cutoff) according to the T2
spectrums before and after centrifugation. The results could be
helpful for the application of NMR technique in the study of shale
reservoirs and provide a new way to characterize quantitatively the
complicated pore structure of unconventional oil and gas
reservoirs.
2. Materials and methods
2.1. Samples
Eight shale core samples from the Lower Cambrian Qiongzhusi
Formation were collected in eastern Yunnan Province, located in
the southwestern margin of the Yangtze plate (Fig. 1). The TOC
content of shale samples ranges from 1.42% to 2.51% with an
average of 2.09%. All shale samples are over-mature with the vit-
rinite reflectance Ro ranging from 2.16% to 3.32%. Mineralogically,
quartz and clay minerals are the dominant minerals in shale sam-
ples. Quartz content is between 31.3% and 42.7%, with an average of
34.41%. Clay minerals content is in the range of 25.3%e36.4%, with a
mean value of 31.44%. Lithologically, shale samples are dominated
by silty shale based on the ternary diagram of mineral compositions
(Fig. 2a). Moreover, the Lower Cambrian Qiongzhusi Formation was
deposited in shallow-marine shelf, and quartz originated from
terrigenous clast rather than siliceous organism. Therefore, TOC
content has no apparent relationship with quartz content. Clay
minerals mainly contain illite (average of 49.13%), illite/smectite
mixed layer (average of 23%), chlorite (average of 20.5%) and a small
amount of kaolinite (average of 7.38%). The parameters of eight
samples are shown in Table 1. Several horizontal cylindrical core
plugs with a diameter of 2.5 cm were prepared parallel to the
bedding planes for each sample.
2.2. Porosity and permeability measurements
The porosity and permeability of shale samples were analyzed
by SGS Unconventional Petroleum Technical Testing Co., Ltd.
following the Chinese Oil and Gas Industry Standard SY/T5336-
2006 “Method of core routine analysis”. A KXD-II porometer was
used to measure the porosity of samples with a helium expansion
method, and permeability measurements were carried out using
dry nitrogen as the medium with an instrument of permeameter
(QT-2). Porosity and permeability experiments utilized a pressure
of 96.8 kPa and a temperature of 25
C.
2.3. NMR core analyses
After porosity and permeability measurements, all eight sam-
ples were dried in the drying oven for 24 h and were vacuumed till
the weight of samples no longer changed. Subsequently, samples
were saturated in the 8% KCl solution and weighed at regular in-
tervals. Shale samples were fully saturated when their weight was
no longer increased. NMR analyses were performed by SGS Un-
conventional Petroleum Technical Testing Co., Ltd. using a RecCore-
2500 instrument with a resonance frequency of 2.38 MHz and a
magnetic field strength of 1200 G, and we obtained the T2 spec-
trums of eight samples at water-saturated condition. The samples
were centrifuged at a centrifuge pressure of 300Psi to reach a
perfect irreducible water state, and then NMR experiments were
conducted again to obtain the T2 spectrums of all samples at irre-
ducible water condition. The measurement parameters used in
NMR experiments were as follows: echo spacing, 0.2 ms; waiting
time, 6s; numbers of scans, 64; echo numbers, 1024; experiment
temperature, 25
C and humidity, 55%.
2.4. FE-SEM observation
The FE-SEM imaging of shale samples was performed using the
Quanta 200F field emission scanning electron microscopy at the
China University of Petroleum (Beijing). Before FE-SEM observation,
one surface of each sample was polished using an argon-ion cross-
section polisher “TechnoorgSC-100”. After polishing, the polishing
surface was coated with gold film at a thickness of 10 nm to
enhance the conductivity. Back-scattering electron model was
chosen to observe the micromorphology of shale sample at various
magnification scales.
3. Results
3.1. Porosity and permeability of shale
The porosity and permeability of eight shale samples are
 A. Li et al. / Journal of Natural Gas Science and Engineering 37 (2017) 29e38 31

Fig. 1. Sampling site (a) and stratigraphic column of Qiongzhusi Formation in eastern Yunnan Province (b).

Fig. 2. Ternary diagrams of the mineral compositions of Qiongzhusi shale. (a): Relationship between TOC and mineral compositions; (b): Relationship between porosity and mineral
compositions.

presented in Table 1. The shale is characterized by tight lithology
and complicated pore structure, so the porosity and permeability
are significantly low. Additionally, porosity is affected by various
geological factors (such as mineralogy, sedimentation, diagenesis
and tectonism), which may result in the inapparent relationship
between porosity and mineral compositions (Fig. 2b). Helium
porosity of samples ranges from 1.40% to 3.90% with a mean value
of 3.17%. The permeability varies from 0.0023 to 0.0061mD with an
average of 0.0042mD. Fig. 3 shows that permeability has no
apparent relationship with porosity (R2 ¼ 0.0302).
32 A. Li et al. / Journal of Natural Gas Science and Engineering 37 (2017) 29e38

Table 1
Characteristic parameters of the shale samples from Qiongzhusi Formation in eastern Yunnan Province.

Sample TOC Ro Quartz Clay Potash Feldspar Plagioclase Calcite Dolomite Pyrite I/S It C Kao Porosity Permeability
(wt.%) (%) (wt.%) (wt.%) (wt.%) (wt.%) (wt.%) (wt.%) (wt.%) (wt.%) (wt.%) (wt.%) (wt.%) (%) (mD)

YN-1 2.23 3.30 35.7 25.3 4.0 24.3 3.8 / 6.9 24 49 20 7 3.76 0.0050
YN-2 2.25 2.26 32.2 33.7 4.6 22.3 3.1 / 4.1 23 54 19 4 2.78 0.0023
YN-3 1.78 2.23 31.8 33.1 4.3 21.7 5.4 / 3.7 12 55 29 4 3.15 0.0031
YN-4 1.42 2.16 31.7 35.5 5.1 22.7 / / 5.0 32 53 9 6 2.96 0.0039
YN-5 2.22 2.45 35.0 27.6 5.5 24.4 4.5 / 3.0 20 52 18 10 1.40 0.0044
YN-6 1.86 2.31 31.3 32.6 5.0 23.8 3.4 / 3.9 28 40 24 8 3.61 0.0061
YN-7 2.42 3.32 34.9 27.3 5.7 21.1 6.5 / 4.5 28 38 26 8 3.90 0.0036
YN-8 2.51 2.44 42.7 36.4 / 4.2 / 13.4 3.3 17 52 19 12 3.80 0.0048

Fig. 3. Relationship between porosity and permeability of shale samples.
3.2. NMR T2 distributions of shale
T2 distributions can be obtained by the mathematical inversion
of T2 decay curves, namely, raw NMR signals (Fig. 4). According to
the basic principle of NMR, there is a positive correlation between
the transverse relaxation time of hydrogen protons in the pore
water and pore diameter (Sun et al., 2012; Huang et al., 2015).
Namely, the T2 distributions reflect the pore size distributions:
small pores have short relaxation time and large pores or micro-
fractures have long relaxation time (Yao et al., 2010a,b; Tan et al.,
2015). The amplitude of T2 spectrum reflects the proportion of
pores with different diameters in shale. Therefore, shale with
stronger T2 spectrum amplitude at the long relaxation time would
have more large pores and better reservoir quality.
lines in Fig. 5). T2 spectrums of shale cores are divided into two
types: unimodal and bimodal T2 spectrums. Unimodal T2 spectrum
(YN-2, YN-3, etc.) has an isolated peak in a short relaxation time
which normally ranges from 0.01 ms to 10 ms. Bimodal T2 spectrum
has two peaks, and the amplitude of right peak ranging from 10 ms
to 100 ms is obviously smaller than that of left peak ranging from
0.01 ms to 10 ms. It suggests that not only small pores but also some
relatively large pores or fractures are well-developed in shale. In
addition, we divide the bimodal T2 spectrum into continuous
bimodal T2 spectrum (YN-6) and discontinuous bimodal T2 spec-
trum (YN-1 and YN-8) based on the shape of T2 spectrum. The
former indicates that the samples have good connectivity between
small pores and large pores (or fractures). While the latter indicates
that pores in shale are relatively closed, and the connectivity be-
tween small pores and large pores is poor.
3.2.2. NMR T2 spectrum at irreducible water condition
NMR measurements were performed again after centrifugation
and the T2 spectrums (red dotted line) at irreducible water condi-
tion are illustrated in Fig. 5. For some samples (e.g., YN-1 and YN-2),
the peaks at 0.01e10 ms have no apparent changes, but those of
some samples reduce (e.g., YN-5 and YN-6). The right peak of
bimodal T2 spectrum reduces and even disappears after centrifu-
gation. The reason for this case is that fluid in some small closed
pores can not flow out freely by centrifugation, so the signal
strength of hydrogen protons and the shape of T2 spectrum do not
change significantly. Whereas fluid in some large pores with
appropriate connectivity can flow out freely by centrifugation,
which results in the signal strength decaying and the amplitude of
T2 spectrum reducing. The right peak of YN-6 disappears after
centrifugation, indicating that microfractures could be well-
developed in shale sample, benefiting fluid migration.

3.2.1. NMR T2 spectrum at water-saturated condition 4. Discussion

The NMR measurements for 8 shale samples were performed to
obtain the T2 spectrums at water-saturated condition (yellow solid In this section, the characteristics of the transverse relaxation

Fig. 4. T2 decay curves of shale samples by NMR measurements.

 A. Li et al. / Journal of Natural Gas Science and Engineering 37 (2017) 29e38 33

Fig. 5. NMR T2 spectrums of shale samples.

time distributions of NMR technique are used to evaluate the
reservoir properties of shale. T2 distribution characteristics include
the number, area, shape, and position of T2 spectrum peaks, which
can be used to analyze pore types, residual porosity, movable
porosity, reservoir permeability and pore structure parameters of
the shale. Moreover, we also calculate the NMR T2 cutoff value.
4.1. Pore types
T2 spectrum and pore size distribution have the similar trend in
34 A. Li et al. / Journal of Natural Gas Science and Engineering 37 (2017) 29e38
NMR measurements, where each transverse relaxation time rep-
resents a pore diameter: long transverse relaxation time represents
large pores and short transverse relaxation time represents small
pores in shale (Kleinberg et al., 1993; Dunn et al., 2002; Yao et al.,
2010a,b). The minimal pore radius allowed fluid to flow out un-
der the action of centrifugal force can be calculated based on the
capillary pressure equation:
2scos q
r¼
P tion. As illustrated in Fig. 5, BVI and FFI are presented with the light-
where r is the minimal pore radius, which allows water to flow out
at the pressure P, mm; P is the centrifuging pressure, MPa; s is the
interfacial tension of shale and water; and q is the contact angle
between pore surface and water molecule.
Although different shale samples have different contact angles
q, in order to simplify the analysis, we use the contact angle be-
tween water molecule and pore surface measured by Yakov (2001)
in this study, about 73
. At room temperature, the interfacial ten-
sion of water is 0.072 N/m. Therefore, the minimal pore radius for
water to discharge is 20.33 nm under the action of centrifugal
pressure of 2.07 MPa, that is to say, the irreducible water includes
not only the clay bound water but also the water remained in the
pores and throats of radius less than 20.33 nm.
The peaks of eight shale samples distributed from 0.01 ms to
10 ms have small changes after centrifugation except sample YN-5,
suggesting that this kind of peak is associated to the small pores
which have large capillary pressure to prevent the discharge of
fluid. Whereas the peaks in the range of 10e100 ms decrease or
disappear by centrifuging, indicating that the increase of pore
diameter weakens the bound effect of capillary pressure on fluid. As
shown in Fig. 5, whether unimodal or bimodal T2 spectrum, pores
in shale samples are mainly adsorption pores, and seepage pores
with bimodal T2 spectrum are more developed than those in
unimodal shale. Through SEM observation, microfractures are
relatively well-developed in bimodal shale (Fig. 6). Microfractures
in YN-6 and YN-8 increase the connectivity among pores and
improve the permeability of shale. Especially YN-6, microfractures
connect small pores and large pores, which contribute to the
seepage of fluid. In contrast, for YN-1 with discontinuous bimodal
T2 spectrum, although there are a certain number of macropores or
microfractures in shale samples, poor connectivity makes the fluid
difficult to flow out. However, for YN-5, shale has good connectivity
among small pores, so the fluid in seepage pores can be expelled by
centrifuging. Given the above, shale with high porosity may not
have high permeability, and shale with low porosity may have good
permeability. It indicates that porosity has no apparent relationship
with permeability in eight groups of shale samples, which is just in
accordance with Fig. 3.
4.2. Residual porosity and movable porosity
Previous studies have indicated that the estimation of porosity
can be obtained accurately regardless of the mineralogy based on
NMR measurements because of the relaxation signal from the
hydrogen-containing fluid rather than the matrix in rock (Straley
et al., 1997; Coates et al., 1999). The signal strength of T2 spec-
trums at water-saturated condition is converted into NMR porosity
4N of shale, and the results are presented in Table 2. Yao et al.
(2010b) have proved that the NMR porosity represents the pro-
portion of pore volume occupied by bound water and free water, so
NMR porosity includes residual porosity 4R, corresponding to the
proportion of bound water and movable porosity 4M, correspond-
ing to the proportion of free water. The 4R and 4M can be calculated
by the following two equations:
4R ¼ 4N  BVIðBVI þ FFIÞ (2)
4M ¼ 4N  FFIðBVI þ FFIÞ (3)
where BVI is the bound fluid index obtained by the T2 spectrum
area at irreducible water condition; FFI is the free fluid index;
BVI þ FFI represents the sum of bound fluid and free fluid that can
be determined by the T2 spectrum area at water-saturated condi-
(1)
gray area in T2 spectrum at irreducible water condition and the
dark-gray area in T2 spectrum at water-saturated condition,
respectively.
Residual porosity and movable porosity calculated by Eq. (2) and
Eq. (3) are shown in Table 2. Residual porosity is between 1.81% and
3.21% with a mean value of 2.58%. Movable porosity ranging from
0.40% to 2.07% with an average of 0.90% is much lower than residual
porosity, and the movable porosities of all samples except YN-6 and
YN-8 are less than 1%. The proportions of BVI and FFI are obvious in
Fig. 5, BVI corresponding to light-gray area in the irreducible water
spectrum accounts for a larger proportion, while FFI corresponding
to dark-gray area in the 100% water spectrum occupies a small area.
It suggests that adsorption pores are dominated in shale reservoir
of study area, and the development of adsorption pores is unfa-
vorable for seepage.
4.3. T2 cutoff values
NMR T2 cutoff value (T2cutoff) is a relaxation time boundary that
divides bound fluid and free fluid in the T2 spectrum at water-
saturated condition. The part on the left of T2cutoff in the T2 spec-
trum represents the bound fluid in the adsorption pores, while the
part on the right of T2cutoff in T2 spectrum corresponds to the free
fluid in the seepage pores. Some scholars have conducted a large
number of tests to obtain the T2cutoff of sandstones and carbonates.
Schlumberger recommended that the T2cutoff of mid-high perme-
ability sandstone reservoir and carbonate reservoir approximated
33 ms and 92 ms, respectively (Xiao, 1998; Sun et al., 2012; Ding
et al., 2012). The results of Wang et al. (2001) demonstrated that
the T2cutoff of low-permeability sandstone reservoir had a wide
range of distribution with a mean value of 12.85 ms, which is only
about one-third of that of mid-high permeability sandstone reser-
voir. Accurate calculation of T2cutoff helps to apply NMR technique in
the evaluation of shale reservoir.
The method for determining T2cutoff of shale samples is shown in
Fig. 7. First, two T2 spectrums at water-saturated condition and
irreducible water condition respectively are converted into two
accumulative T2 spectrums. Maximum cumulative amplitudes of T2
spectrums at water-saturated condition and irreducible water
condition are calibrated as total porosity and residual porosity
respectively, so the difference between total porosity and residual
porosity means movable porosity, which corresponds to the
segment “H” in Fig. 7. Second, a horizontal line through residual
porosity is drawn, and this line and accumulative T2 spectrum at
water-saturated condition intersect at one point “A”. Third, a ver-
tical line is made through the point “A”, and the T2 value at the
intersection projected on the time axis is the T2cutoff. The T2cutoff of
samples calculated by the above method is shown in Table 2, and
T2cutoff is in the range of 0.54e4.28 ms with an average of 1.45 ms.
The T2cutoff of YN-1 is significantly greater than that of other sam-
ples, and the T2cutoff of other seven samples is all less than 4 ms. As
shown in Fig. 7, the higher the content of bound water in shale is,
the greater the T2cutoff is, namely, shale with higher bound water
content has a greater T2cutoff.
 A. Li et al. / Journal of Natural Gas Science and Engineering 37 (2017) 29e38 35

Fig. 6. FE-SEM images of microfractures in shale samples. Netted microfractures and continuous microfractures increase the connectivity among pores, thus improving the
permeability of shale (aee). However, some small microfractures with poor extensibility cannot connect the pores, which are unfavorable for seepage (f). a and b: sample YN-6; c
and d: sample YN-8; e and f: sample YN-1.

Table 2
NMR porosity, residual porosity, movable porosity, NMR permeability and T2cutoff of shale samples.

Sample 4N (%) BVI(BVI þ FFI) FFI(BVI þ FFI) 4R (%) 4M (%) T2cutoff (ms) kNMR (mD)

YN-1 3.96 0.81 0.19 3.21 0.75 4.28 0.0042

YN-2 3.11 0.87 0.13 2.71 0.40 1.13 0.0014
YN-3 3.44 0.87 0.13 2.99 0.45 1.10 0.0021
YN-4 2.77 0.68 0.32 1.88 0.89 0.54 0.0035
YN-5 2.59 0.70 0.30 1.81 0.78 0.83 0.0024
YN-6 4.23 0.51 0.49 2.16 2.07 0.73 0.0070
YN-7 3.42 0.79 0.21 2.70 0.72 1.04 0.0039
YN-8 4.35 0.73 0.27 3.18 1.17 1.93 0.0045

4.4. Permeability

Shale as a tight reservoir of natural gas is characterized by low

permeability. Estimating shale permeability is another important
function of NMR core experiment (Shao et al., 2009; Ding et al.,
2014). Based on the bound fluid volume, free fluid volume and
NMR porosity, constructing a model among these parameters
provides a good method to calculate the permeability of shale. A
new regional model is built on the basis of Coates permeability
model in this study. The parameters in this new model vary with
different regions, so the results are more reasonable. On the basis of
Timur formula, Coates established the Coates permeability model
through a large number of experiments in 1991 (Arnold et al., 2006;
Tian, 2010). Coates model can be described as follows:
Fig. 7. T2 cutoff value calculation for sample YN-6 with the NMR measurements at
water-saturated condition (Sw) and irreducible water condition (Sir).
36 A. Li et al. / Journal of Natural Gas Science and Engineering 37 (2017) 29e38

pore surface and hydrogen-containing fluid (e.g. water and kero-

4 4  FFI 2
N sene), so total relaxation time T2 is almost determined by the sur-
kNMR ¼ (4)
C BVI face relaxation time T2S (Coates et al., 1999; Wang, 2009; Sun et al.,
2012). When the bulk relaxation and diffusion relaxation of pore
kNMR is the permeability estimated by NMR experiment, C is a fluid are not considered, Eq. (7) can be simplified as:
constant. Coates model is a basic model for calculating permeability
using NMR. In order to make this model suitable for different re- 1 S
¼ r2 (8)
gions, it is changed in the form of “y ¼ mx þ b”. Regional Coates T2 V
model is as follows:
where r2 is the transverse surface relaxivity of rock, mm/ms. The r2
 0:5 !
FFI k 0:25 of shale as a constant is about 0.05 mm/ms based on the work of
¼ m NMR þb (5) Sondergeld et al. (2012); S and V are the surface area and volume of
BVI 4N
rock pores, respectively, mm2 and mm3. S/V is the specific surface
area that relates to the pore size, smaller pores have greater S/V
m and b are the slope and intercept of straight-line respectively,
values and shorter T2. On the contrary, the more large pores in the
which are determined by fitting analysis. (FFI/BVI)0.5 and (k0.25
NMR/4N)
rock, the lower the value of S/V is, and the longer the T2 is. Hence,
are regarded as the “x” and “y” of linear equation respectively to
the interaction between water molecules and pore surface can be
carry on linear fitting. (FFI/BVI)0.5 has a strongly linear correlation
characterized by T2 relaxation of water-saturated samples.
with (k0.25
NMR/4N), with a correlation coefficient of 0.883. Therefore,
For simplification purpose, the pore geometry of shale is
the Coates model of study area is:
regarded as cylinder because of the connectivity among shale
 0:5 ! pores, so S/V equals 2/r. Eq. (8) can be further transformed into the
0:25
FFI k
¼ 11:62 NMR  0:336 (6) relationship between T2 and shale pore size.
BVI 4N
r ¼ 0:1T2 (9)
The NMR permeability of shale samples calculated by Eq. (6) is
in the range of 0.0014e0.0070mD with an average of 0.0036mD Therefore, the specific surface area distributions of samples can
(Table 2). Fig. 8 is the cross-plot of NMR permeability and gas log be obtained according to Eq. (8). The specific surface area distri-
permeability of samples. All sample points except YN-5 are close to bution of YN-1 is shown in Fig. 9. The specific surface area ranges
the bisector, indicating that the regional Coates model is effective from 0.34 mm1 to 433.33 mm1 with an average of 48.31 mm1 and
for permeability calculation and NMR permeability satisfies the is mainly distributed in the range of 0.5e1.79 mm1 and
demands of research and production. 8.3e93.72 mm1. Moreover, the T2 distributions of eight samples are
converted into the pore size distributions based on Eq. (9) (Fig. 10).
In eight shale samples, transition pores account for the largest
4.5. Pore structure proportion, followed by mesopores, indicating that transition pores
and mesopores are the major sites for the accumulation of shale
The transverse relaxation characteristics of fluid in porous me- gas. Furthermore, the well-developed macropores and micro-
dia can be described by the following mathematical equation: fractures in shale samples with bimodal T2 spectrums also provide
more reservoir space for shale gas.
1 1 1 1
¼ þ þ (7)
T2 T2B T2D T2S
5. Conclusions
where 1/T2 is the total relaxation rate of pore fluid, ms1. T2B, T2D,
and T2S are bulk relaxation time, diffusion relaxation time and
In this paper, the petrophysical properties of 8 shale samples
surface relaxation time, respectively. In low-field NMR experiment,
from the Lower Cambrian Qiongzhusi Formation in eastern Yunnan
the applied magnetic field from NMR instrument is uniform, so the
Province are investigated using NMR experiments. Additionally, the
diffusion relaxation rate 1/T2D approaches zero. Moreover, the bulk
method for determining T2cutoff of shale has also been discussed
relaxation time of pore fluid T2B is so long that it contributes little to
based on the physical significance of T2cutoff. The following con-
total relaxation rate. Porous media are generally characterized by
clusions can be made:
complicated surface structure with strong interaction between

Fig. 9. The specific surface area distribution of sample YN-1 based on NMR
Fig. 8. The cross-plot of NMR permeability and gas log permeability of samples. measurements.
 A. Li et al. / Journal of Natural Gas Science and Engineering 37 (2017) 29e38 37

Fig. 10. Pore size distributions of eight shale samples by NMR measurements.

(1) T2 spectrums of shale samples at water-saturated condition
can be divided into two types: unimodal and bimodal T2
spectrums. Whether unimodal or bimodal T2 spectrum,
pores in shale samples are mainly adsorption pores, and
seepage pores in bimodal shale are more developed than
those in unimodal shale. Continuous bimodal T2 spectrum
reflects the samples with good connectivity between small
pores and large pores.
(2) According to the T2 spectrum area fraction, residual porosity
and movable porosity can be well estimated. The results
show that movable porosity is much lower than residual
porosity. Adsorption pores are dominated in shale reservoir
of study area, and the development of adsorption pores is
unfavorable for seepage.
(3) The method for determining T2cutoff of shale has been dis-
cussed based on the physical significance of T2cutoff. The
T2cutoff of shale samples is less than 4 ms except for YN-1. The
higher the content of bound water in shale is, the greater the
T2cutoff is. On the basis of Coates model, a regional Coates
model is proposed to calculate the permeability of shale
samples. The NMR permeability is close to gas log perme-
ability, so the regional Coates model is effective for perme-
ability calculation.
(4) The specific surface area distribution and pore size distri-
bution can be obtained based on the mathematical equation
of T2. The specific surface area is mainly distributed in the
range of 0.5e1.79 mm1 and 8.3e93.72 mm1. Transition
pores and mesopores contribute the major sites for the
38 A. Li et al. / Journal of Natural Gas Science and Engineering 37 (2017) 29e38
accumulation of shale gas. Moreover, the well-developed
macropores and microfractures in shale with bimodal T2
spectrum also provide more reservoir space for shale gas.
Acknowledgements
This research was supported by the National Natural Science
Foundation of China (Project Nos. 41072098 and 41372139) and the
Important National Science and Technology Specific Projects of
China (Nos. 2016ZX05046-003, 2011ZX05018-001-002 and
2011ZX05009-002-205). The authors would like to thank the staff
of all of the laboratories that cooperated in performing the tests and
analyses. We are also grateful to the reviewers, whose comments
improved the quality of this manuscript.
References
Arnold, J., Clauser, C., Pechnig, R., Anferova, S., Anferov, V., Blümich, B., 2006.
Porosity and permeability from mobile NMR core-scanning. Petrophysics 47,
306e314.
Bowker, K.A., 2007. Barnett shale gas production, Fort Worth Basin: issues and
discussion. AAPG Bull. 91 (4), 523e533.
Caldwell, R., 2006. Unconventional resources: are they for real. Scotia Newsl. 3, 1e2.
Chalmers, G.R.L., Bustin, R.M., 2007. The organic matter distribution and methane
capacity of the Lower Cretaceous strata of Northeastern British Columbia. Can.
Int. J. Coal Geol. 70, 223e339.
Clarkson, C.R., Jensen, J.L., Chipperfield, S., 2012. Unconventional gas reservoir
evaluation: what do we have to consider? J. Nat. Gas. Sci. Eng. 8, 9e33.
Clarkson, C.R., Solano, N., Bustin, R.M., Bustin, A.M.M., Chalmers, G.R.L., He, L.,
Melnichenko, Y.B., Radlin  ski, A.P., Blach, T.P., 2013. Pore structure character-
ization of North American shale gas reservoirs using USANS/SANS, gas
adsorption, and mercury intrusion. Fuel 103, 606e616.
Coates, G.R., Xiao, L.Z., Prammer, M.G., 1999. NMR Logging Principles and Applica-
tions. Gulf Publishing Company, Houston, Texas.
Curtis, J.B., 2002. Fractured shale-gas system. AAPG Bull. 86 (11), 1921e1938.
Curtis, M.E., Ambrose, R.J., Sondergeld, C.H., Rai, C.S., 2012. Microstructural inves-
tigation of gas shales in two and three dimensions using nanometer-scale
resolution imaging. AAPG Bull. 96, 665e677.
Ding, W.L., Dai, P., Zhu, D.W., Zhang, Y.Q., He, J.H., Li, A., Wang, R.Y., 2015. Fractures
in continental shale reservoirs: a case study of the Upper Triassic strata in the
SE Ordos Basin. Cent. China. Geol. Mag. 153 (4), 663e680.
Ding, W.L., Fan, T.L., Yu, B.S., Huang, X.B., Liu, C., 2012. Ordovician carbonate
reservoir fracture characteristics and fracture distribution forecasting in the
Tazhong Area of Tarim Basin, Northwest China. J. Pet. Sci. Eng. 86e87, 62e70.
Ding, Y.J., Guo, B.H., Yan, X.R., Li, J., Lu, Q., 2014. On identification shale reservoirs
validity and physical parameters quantitative evaluation method. Well Logging
Technol. 38 (3), 297e303.
Dunn, K.J., Bergman, D.J., Latorraca, G.A., 2002. Petrophysical NMR measurements.
Handb. Geophys Explor Seismol. Explor 32, 71e127.
Heath, J.E., Dewers, T.A., McPherson, B.J., Petrusak, R., Chidsey, T.C., Rinehart, A.J.,
Mozley, P.S., 2011. Pore networks in continental and marine mudstones: char-
acteristics and controls on sealing behavior. Geosphere 7 (2), 429e454.
Huang, J.G., Xu, K.M., Guo, S.B., Guo, H.W., 2015. Comprehensive study on pore
structures of shale reservoirs based on SEM, NMR and X-CT. Geoscience 29 (1),
199e205.
Kleinberg, R.L., Straley, C., Kenyon, W.E., Akkurt, R., Farooqui, S.A., 1993. Nuclear
magnetic resonance of rocks: T1 vs. T2. SPE Pap. 26470, 555e563.
Li, A., Ding, W.L., He, J.H., Dai, P., Yin, S., Xie, F., 2016. Investigation of pore structure
and fractal characteristics of organic-rich shale reservoirs: a case study of Lower
Cambrian Qiongzhusi formation in Malong block of eastern Yunnan Province,
South China. Mar. Pet. Geol. 70, 46e57.
Ross, D.J.K., Bustin, R.M., 2007. Shale gas potential of the lower Jurassic gordondale
member, northeastern British Columbia, Canada. Bull. Can. Pet. Geol. 55 (1),
51e75.
Ross, D.J.K., Bustin, R.M., 2009. The importance of shale composition and pore
structure upon gas storage potential of shale gas reservoirs. Mar. Pet. Geol. 26,
916e927.
Shao, W.Z., Ding, Y.J., Liu, Y., Liu, S.Q., Li, Y.Q., Zhao, J.H., 2009. The application of
NMR log data in evaluation of reservoir pore structure. Well Logging Technol.
33 (1), 52e56.
Sondergeld, C.H., Ambrose, R.J., Rai, C.S., Moncrieff, J., 2012. Micro-structure studies
of gas shales: SPE Paper 131771 presented at the SPE Unconventional Gas
Conference and Exhibition, Pittsburgh, Pennsylvania, February 23-25, 1e17.
Straley, C., Rossini, D., Vinegar, H., Tutunjian, P., Morriss, C., 1997. Core analysis by
low field NMR. Log. Anal. 38, 84e93.
Sun, J.C., Chen, J.P., Yang, Z.M., Liu, X.W., Liu, Y.J., 2012. Experimental study of the
NMR characteristics of shale reservoir rock. Sci. Technol. Rev. 30 (14), 25e30.
Tan, M.J., Mao, K.Y., Song, X.D., Yang, X., Xu, J.J., 2015. NMR petrophysical inter-
pretation method of gas shale based on core NMR experiment. J. Pet. Sci. Eng.
136, 100e111.
Tang, X., Zhang, J.C., Wang, X.Z., Yu, B.S., Ding, W.L., Xiong, J.Y., Yang, Y.T., Wang, L.,
Yang, C., 2014. Shale characteristics in the Southeastern Ordos Basin, China:
implications for hydrocarbon accumulation conditions and the potential of
continental shales. Int. J. Coal Geol. 128e129, 32e46.
Tian, Y., 2010. NMR logging permeability models and application. Inn. Mong. Pet-
rochem. Ind. 10, 25e27.
Tiwari, P., Deo, M., Lin, C.L., Miller, J.D., 2013. Characterization of oil shale pore
structure before and after pyrolysis by using X-ray micro CT. Fuel 107, 547e554.
Wang, S., 2009. Analysis of rock pore structural characteristic by nuclear magnetic
resonance. Xinjiang Pet. Geol. 30 (6), 768e770.
Wang, W.M., Ye, Z.H., Guo, H.K., 2001. Experimental studies of NMR properties of
continental sedimentary rocks. Chin. J. Magn. Reson 18 (2), 113e121.
Xiao, L.Z., 1998. NMR Imaging Logging and Rock NMR Properties and its Applica-
tions. Scientific Press, Beijing.
Xu, H., Tang, D.Z., Zhao, J.L., Li, S., 2015. A precise measurement method for shale
porosity with low-field nuclear magnetic resonance: a case study of the Car-
boniferousePermian strata in the Linxing area, eastern Ordos Basin, China. Fuel
143, 47e54.
Yakov, V., 2001. A practical approach to obtain primary drainage capillary pressure
curves from NMR core and log data. Petrophysics 42, 334e343.
Yao, Y.B., Liu, D.M., Cai, Y.D., Li, J.Q., 2010a. Advanced characterization of pores and
fractures in coals by nuclear magnetic resonance and X-ray computed tomog-
raphy. Sci. China-Earth Sci. 53, 854e862.
Yao, Y.B., Liu, D.M., Yao, C., Tang, D.Z., Tang, S.H., Huang, W.H., 2010b. Petrophysical
characterization of coals by low-field nuclear magnetic resonance (NMR). Fuel
89, 1371e1380.
Yu, B.S., 2013. Classification and characterization of gas shale pore system. Earth Sci.
Front. 20 (4), 211e220.
Zhong, T.X., 2012. Characteristics of pore structure of marine shales in South China.
Nat. Gas. Ind. 32 (9), 1e4.