Tribology International 144 (2020) 106119

Contents lists available at ScienceDirect

Tribology International
journal homepage: http://www.elsevier.com/locate/triboint

Effect of normal load and shearing velocity on the interface friction of

organic shale – Proppant simulant
H. He, L. Luo, K. Senetakis *
City University of Hong Kong, Yeung Kin Man Academic Building, Blue Zone 6/F, Kowloon, Hong Kong, China

A R T I C L E I N F O A B S T R A C T

Keywords: In the present study, experiments were conducted using an advanced micromechanical apparatus investigating
Shale rock the tribological behavior of interfaces between an organic shale against two different proppant simulants; one
Proppant composed of Leighton Buzzard sand (LBS) and the other composed of glass beads. A negative correlation was
Interface friction
observed between the coefficient of friction and the magnitude of normal load from organic shale-proppant
Hydraulic fracturing
interface shearing tests. At relatively low shearing velocities (0.2–0.4 mm/h) stick-slip shearing behavior was
found to be more prominent, and the change of shearing velocity had insignificant influence on the coefficient of
friction. Additional discussion is presented comparing these results from the present work with a previously
studied inorganic shale.

1. Introduction comprise key input parameters to be used in simulations using the

discrete element method (DEM), which is a popular
Benefited from the development and advancement in hydraulic micromechanical-based numerical tool used to model hydraulic frac­
fracturing techniques, the extraction and production of the unconven­ turing [e.g., 10–13]. Previous experimental studies focused mainly on
tional hydrocarbon that is trapped in shale gas/oil reservoirs was the shale-proppant interaction at the macroscopic level with only a few
boosted in recent decades [1–3]. One of the major difficulties of the recent studies published in the literature which examined the tribolog­
unconventional hydrocarbon extraction is caused by the low porosity ical behavior of the shale-proppant interface [14–17], and due to the
and permeability of the shale rocks; for example, Soeder [4] reported limited number of experimental works, the micromechanical
that the permeability of a typical shale was less than 8.34μd. High shale-proppant parameters adopted in DEM models often vary a lot
pressure fluid is pumped into the wellbores to create fractures to the among different numerical studies. For example, the coefficient of fric­
shale rock, and thus making pathways for the gas or oil in the reservoirs tion between the shale and proppant was taken as 0, 0.1 and 0.5 by
to flow to the vertical and horizontal wellbores during hydraulic frac­ Shimizu et al. [10], Zeng et al. [12] and Zhang et al. [13], respectively,
turing. To withhold the fracture created and maintain the flow of the in their DEM-computational fluid dynamics (CFD) coupled models.
gas/oil after the hydraulic pressure is released, proppants are mixed Other than the conventional proppants, including natural quartz
with the hydraulic fracture fluids and transported into the fractures by sand, resin coated sand and ceramic balls, researchers developed mul­
the pressurized fluids during the fracturing process [5–8]. During the tiple types of proppants with various shapes for the sake of transporting
process, the mechanical interaction between proppant and shale directly distance, flow-back resistance and gas/oil conductivity, like bauxite,
affects the proppant transport (i.e., whether the proppants can be walnut shells, hollow glass spheres and elongated and rod shaped
transported far enough and evenly distribute in the fracture) and proppant [18–20]. Shearing two types of quartz sand with different
proppant settlement (i.e., whether the proppants do not flow-back and particle shapes against a siliceous (inorganic) shale, He and Senetakis
withhold the normal stress induced by the fracture closure after the [17] reported that the micromechanical properties of the shale-proppant
pumping process) [8,9]. Though there is a number of influencing factors interface are notably affected by the shape of the proppant. Both Xiao
on the hydraulic fracturing problem, the interface micromechanical et al. [15] and He and Senetakis [17] reported on the influence of
properties of shale against proppant are key parameters to understand normal load on the coefficient of friction of the shale-proppant com­
fundamentally the mechanics of proppant-shale response and they also posite interface. The friction of the interface or gouge of shale and other

* Corresponding author.
E-mail addresses: huanhe6@cityu.edu.hk (H. He), linaluo2@cityu.edu.hk (L. Luo), ksenetak@cityu.edu.hk (K. Senetakis).

https://doi.org/10.1016/j.triboint.2019.106119
Received 26 August 2019; Received in revised form 21 November 2019; Accepted 12 December 2019
Available online 13 December 2019
0301-679X/© 2019 Elsevier Ltd. All rights reserved.
H. He et al. Tribology International 144 (2020) 106119

rock materials was found to be related to the shearing velocity [21–23,

among others], and the coefficient of friction of composite interfaces
was reported to be dependent to the shearing velocity [24,25]. The fluid
injection speed, which affects the velocity of the proppants during the
hydraulic fracturing process, was proved to affect markedly the travel
distance and distribution of the proppant in the fractures [9]. Kohli and
Zoback [23] and Sone and Zoback [26,27] highlighted that the miner­
alogy, especially the amount of clay and organic carbon of the shale
dominated many aspects of the mechanical properties of the shale.
Based on recent micromechanical experimental studies on geological
materials, researchers emphasized that the interparticle contact
response is affected mainly by the surface properties, including the
microhardness, surface roughness and material type [28–31] and that
plowing mechanisms can play an important role on the tribological
behavior of geological materials [32,33]. Based on the aforementioned
literature studies, the tribological behavior of the shale-proppant con­
tact can be closely related to the shape of the proppant, the magnitude of
the normal load, the shearing velocity, the mineralogy of the shale, as
well as the material surface characteristics including microhardness and
roughness. However, limited studies systematically investigated the ef­
fect of all these factors on the shearing behavior of shale-proppant
interfaces.
In this study, the micromechanical shearing behavior of a black shale
with high organic content against two types of proppant simulants, i.e.,
Leighton Buzzard sand (a natural quartz sand) and glass beads, was
investigated, with a particular focus on the effect of normal load and
shearing velocity. The elemental composition as well as the surface
characteristics including microhardness and roughness of the materials
were quantified in order to be linked to the observed trends from the
tribological tests. The results from the present study were also compared
with a recent work on the interface behavior of an inorganic shale
against proppant simulant by He and Senetakis [17].

2. Experimental study

2.1. Materials studied

A dark-colored shale rock (denoted as black shale, BS) with organic

matters was sheared against Leighton Buzzard sand (LBS) with sizes
between 1.18 and 2.36 mm (denoted as the “BL” interface) and glass Fig. 1. Photo and scanning electron microscope (SEM) images of (a) black
shale (BS); (b) white shale (WS); (c) the element percentage of the two shale
beads (GB) of 2 mm in diameter (denoted as the “BG” interface), to
rocks from energy dispersive spectroscopy (EDS) tests.
investigate the shale-proppant tribological properties. Through visual
comparison with the chart proposed by Krumbein and Sloss [34], the
sphericity and roundness of the LBS particles were quantified to be 0.8 black shale contains notably more organic matters than the white shale.
and 0.7, respectively, while those of glass bead spheres were equal to 1. A set of micro-indentation tests and optical surface profile tests were
He and Senetakis [17] studied the micromechanical behavior of a conducted on the smoothened surfaces of both types of shale rocks to
light-colored siliceous shale rock (denoted as white shale, WS) against characterize the micro-hardness and the surface roughness (the details
Leighton Buzzard sand (denoted as “WL” interface) of sizes between 2 of the apparatus and procedures of the surface characterization tests are
and 4 mm, and the shearing response of the WL interface was used as introduced by He et al. [25]). After the manual slickensiding, the surface
benchmark in this study. The surfaces that are parallel to the bedding roughness (Sq) of the two types of shale rock were close in magnitude,
planes of the shale rock pieces were manually slickensided on dry pieces and the roughness can be considered isotropic within the scales of the
of samples with 600Cw abrasive paper before being characterized and measurements. The average Sq of the BS and WS were 591.4 nm and
tested. Representative photos together with scanning electron micro­ 529.2 nm, respectively, based on 10 measurements on each type of
scope (SEM) images of the BS and WS are given in Fig. 1(a) and Fig. 1(b). shale. From at least 15 measurements on three to four specimens, the
It can be observed from the SEM images that the BS specimen is domi­ average Martens hardness of the BS rock was 0.40 GPa, and that of the
nated by clay flakes with limited particulate matter, while clear silt sized WS rock was 0.72 GPa. The difference in hardness and organic contents
particles that are bonded and connected by clay minerals could be of the two types of shale rocks is mainly due to their formation origin.
observed on the WS specimen. A set of energy dispersive spectroscopy For example, Wang et al. [35] reported that the black shale they
(EDS) tests were conducted to quantify the elements of the shale spec­ investigated, which contained 3.9–15.4% of total organic carbon (TOC),
imens and the results are presented in Fig. 1(c) in the form of bar charts. was extracted from the bottom of the formation, while the light-colored
The predominant elements of both BS and WS were oxygen and silicon siliceous shale was extracted from the top. The average surface rough­
(36.49 wt%), with a small amount of aluminum, potassium, as well as ness of the LBS and GB were 223 nm and 145 nm [29], which are both
some trivial elements like sodium, magnesium and calcium. However, much smaller in magnitude in comparison with the shale surfaces. The
carbon took up to 8.62 wt% in the BS specimen, while only trivial micro-hardness of the LBS and GB were measured to be around 4.59 GPa
amount of carbon was observed in the WS specimen. The color of the and 3.25 GPa, which are approximately an order of magnitude higher
shale pieces [35], as well as the SEM/EDS analysis indicate that the than that of the black shale surface.

2
H. He et al. Tribology International 144 (2020) 106119

2.2. Micromechanical apparatus used shale against LBS interface was investigated through two brands of tests.
Velocity-stepping type of shearing tests, which were adopted by previ­
A newly developed two-axis dynamic micromechanical testing ous research works on rock interfaces, for example Dieterich and Conrad
apparatus constructed and introduced by He et al. [25] at City Univer­ [36], Dieterich and Kilgore [37] and Kohli and Zoback [23], were per­
sity of Hong Kong was used to conduct the micromechanical experi­ formed on specimens BL5, BL6 and BL7 to investigate the effect of
ments. The schematic illustration of the apparatus is given in Fig. 2, shearing velocity on the frictional behavior of the specimens. The FN for
where its key components are illustrated. The vertically positioned tests BL5, BL6 and BL7 was equal to 1 N, 3 N and 5 N, respectively. The
loading system was designed to apply the designated normal load, while range of velocities covered for BL5 and BL6 were 0.4–40 mm/h and
the horizontal system was utilized to perform shearing on the interface. 0.2–40 mm/h, while the velocity range for BL7 was extended to 0.2–170
An image of a representative pair of BL specimen is given in the sub­ mm/h. For specimens BL8 and BL9, nine cycles of monotonic shearing
figure of Fig. 2, where it can be observed that the proppant simulant is were performed on individual shearing paths of the shale surface for
fixed onto the upper specimen mount and the shale is fixed onto the each pair of specimens under a constant normal load, and the velocity of
lower mount. The interaction between the proppant and the shale dur­ the shearing cycles varied between 0.1 mm/h to 4096 mm/h. The
ing the tests, i.e. the load and displacement in both normal and normal load magnitudes for BL8 and BL9 were 3 N and 1 N, respectively.
tangential directions, was monitored by high precision load cells (with a
repeatability of 0.01 N) and non-contact eddy current type of displace­ 3. Results and discussion
ment sensors (with a repeatability of 0.01 μm), and the sensor signal was
collected and transferred to the computer by a data logger with high 3.1. Effect of normal load on the shearing behavior
sampling rate capability (up to 20Hz). More technical details and cali­
brations of the apparatus have been introduced by He et al. [25]. The shearing response of a representative specimen, BL3, under four
different normal loads (0.5 N, 1 N, 3 N and 5 N) is illustrated in Fig. 3(a)
in terms of mobilized coefficient of friction against shearing displace­
2.3. Micromechanical testing program ment. The four shearing cycles were performed by shearing the same
pair of LBS and BS along four variant shearing paths next to one another.
Monotonic shearing tests were performed on three pairs of black The mobilized coefficient of friction reached a steady state value after
shale against glass beads specimens (denoted as BG1 to BG3) and four the initial regime of shearing displacement, where the shearing force
pairs of black shale against Leighton Buzzard sand specimens (denoted increased non-linearly with displacement. The shearing displacement
as BL1 to BL4) to investigate the effect of the normal load magnitude on required to reach steady-state sliding is denoted as the microslip
the shearing behavior of the specimens. The details of this set of tests are displacement [30]. As the normal load increased from 0.5 N to 5 N, the
listed in Table 1. For each pair of specimens in this set of tests, multiple microslip displacement of specimen BL3 increased from around 0.011
cycles of shearing were performed along different shearing paths of the mm–0.067 mm and the coefficient of friction decreased by around 12%,
shale surface at a constant shearing velocity of 0.4 mm/h, and the from 0.76 to 0.67. The tangential (shearing) stiffness against shearing
normal load (FN) of the shearing cycles varied from one another (ranging displacement plots of the four cycles of shearing of BL3 specimen are
from 0.5 N to 10 N). given in Fig. 3(b), where it can be observed that the stiffness gradually
The effect of shearing velocity on the frictional response of the black

Fig. 2. Schematic illustration of the dynamic apparatus and an image of a representative pair of black shale-Leighton Buzzard sand (BL) sample in contact.

3
H. He et al. Tribology International 144 (2020) 106119

Table 1
Details of monotonic shearing test results under a shearing velocity of 0.4 mm/h.
Interface Test Normal Coefficient Microslip Initial
type code load FN of friction μ displacement Tangential
(N) (mm) stiffness KTo
(N/mm)

BG1- 0.5 0.42 0.008 25

0.5
N
BG1- 1 0.38 0.016 31
1N

BG2- 0.5 0.40 0.009 14

Black 0.5
shale- N
Glass BG2- 1 0.36 0.016 38
beads 1N

BG3- 1 0.41 0.018 33

1N
BG3- 3 0.36 0.025 50
3N
BG3- 10 0.30 0.058 65
10 N

BL1- 0.5 0.70 0.008 30

0.5
N
BL1- 1 0.67 0.014 52
1N
BL1- 3 0.66 0.040 55
3N
BL1- 5 0.67 0.063 65
5N

BL2- 0.5 0.68 0.013 26

0.5
N
BL2- 1 0.74 0.024 40
1N
BL2- 3 0.72 0.051 49
3N
BL2- 5 0.71 0.070 50
Black 5N
shale- BL2- 10 0.69 0.155 56
Leighton 10 N
Buzzard
Sand BL3- 0.5 0.76 0.011 39
0.5
N
Fig. 3. Representative plots of (a) mobilized coefficient of friction against
BL3- 1 0.73 0.018 40
1N shearing displacement (b) tangential stiffness against shearing displacement of
BL3- 3 0.72 0.049 56 specimen BL3 at different normal loads.
3N
BL3- 5 0.67 0.067 54 exceptional cases for BL specimens), which confirms the discussion
5N
above for specimen BL3. The coefficient of friction of each test is plotted
BL4- 0.5 0.71 0.012 40 against the corresponding normal load in Fig. 4(a), where the average μ
0.5
value at each normal load is also given. The coefficient of friction of the
N
BL4- 1 0.69 0.018 52 BL interface is noticeably higher than that of the BG interface by a
1N magnitude of around 0.30–0.35 at a given normal load. Sandeep and
BL4- 5 0.66 0.065 55 Senetakis [30] reported that the average coefficient of friction of the
5N grain-grain type of contact between two glass beads was around 0.12,
BL4- 10 0.59 0.100 73
10 N
while the interparticle friction of the LBS particles was around 60%
higher than that of the glass beads equaling to 0.19. Besides the lower
surface roughness, the higher roundness of the glass beads is also
dropped to zero as the shearing reached the steady state condition. The believed to be part of the reason for the lower coefficient of friction
tangential stiffness at the initial stage of shearing (initial tangential observed for the BG interfaces in comparison with the BL interfaces. He
stiffness, KTo) increased from around 39 N/mm to 54 N/mm (increase of and Senetakis [17] stressed that more angular quartz sand particles,
almost 40%) as the normal load increased from 0.5 N to 5 N. with smaller local contact radius, would result in higher friction values
The details of all the shearing tests between the black shale-proppant when sheared against the siliceous shale surfaces, which agrees with the
interfaces, including the normal load, coefficient of friction, microslip observation of the current study on the black shale-proppant interfaces.
displacement and initial tangential stiffness, are summarized in Table 1. The average coefficients of friction of the white shale against
When comparing the shearing behavior of a given specimen (i.e. given Leighton Buzzard sand (WL) are plotted in Fig. 4(a) for comparison. The
shale sample against given proppant sample) under different normal friction of the BL interface is higher than that of the WL interface in the
loads, it is noticed that the μ values decreased, while the microslip range of 3–10 N of normal load. Since the surface roughness Sq values of
displacement and KTo increased as the normal load increased (with a few the two types of shale rocks are close, the difference in friction could be

4
H. He et al. Tribology International 144 (2020) 106119

N/mm as the normal load increased from 0.5 N to 10 N. In the range of

0.5 N–3 N of normal load, the average KTo of the BL interface was higher
than that of the BG interface, while at higher normal loads the two
curves converged. The relatively higher initial tangential stiffness
observed for the BL interfaces compared with the BG interfaces is
speculated to be attributed to the higher roughness and lower local
radius of the LBS particle surface compared with the glass beads surface,
which could probably resulted in higher level of contact maturing/sur­
face penetration causing adhesion force under relatively low normal
loads before starting of the shearing [38,39]. This extra adhesion force
caused by the difference in roughness and local contact radius became
insignificant when plowing dominated during shearing under higher
normal load. The initial tangential stiffness of the BL interface was found
to be around 50%–70% lower than that of shearing LBS particles against
the siliceous white shale.
Based on the SEM and EDS analysis, the black shale consists of clay
flakes with organic matter, while the white shale is composed of sili­
ceous silt particles with limited organic content. This compositional and
microstructural difference between the two types of shale rocks is
believed to be the reason for the different shearing behaviors observed.
It was proved by previous researchers that the mineralogy of shale rocks,
especially the content of clay and organic matter, notably affects their
mechanical properties, such as stiffness, anisotropy, strength, as well as
frictional behavior [23,26,27,40]. Compared with the white siliceous
shale, the lower hardness of the black shale, which was determined by its
mineralogy and microstructure, could result in higher level of plowing
during shearing, and the debris (clay dominant for the black shale or
siliceous particles dominant for the white shale) and the tribofilm
created due to shearing and plowing will, in turn, affect the shearing
behavior [15,17,41–43].

3.2. Effect of shearing velocity on the coefficient of friction

The results of the velocity-stepping experiments on specimens BL5

(FN ¼ 1N), BL6 (FN ¼ 3N) and BL7 (FN ¼ 5N) are illustrated in Fig. 5,
where the mobilized coefficient of friction, as well as the nominal
shearing velocity, are plotted against the shearing displacement. Due to
finite apparatus stiffness, a sudden rise of the mobilized coefficient of
friction could be observed when the shearing velocity increased, and a
dip was observed when the shearing velocity dropped, which is similar
Fig. 4. Effect of normal load on the (a) coefficient of friction (μ) and (b) initial to what was reported by previous studies, like Dieterich and Conrad [36]
tangential stiffness (KTo) of the shale-proppant interfaces. and Dieterich and Kilgore [37]. Due to the morphology of the surface
and stick-slip behavior, the curves fluctuated, especially during the
partly attributed to the higher hardness of the white shale compared shearing at relatively low speeds, as the shearing proceeded, which can
with the black shale, which could result in less asperity breakage and be attributed to the possible increased contact area under lower shearing
plowing on the white shale surface. Comparing multiple types of gran­ velocities [36]. The frictional stability can be quantified by the following
ular materials, Sandeep and Senetakis [29] observed lower friction from parameter (a-b) to illustrate the shearing velocity effect on the coeffi­
materials with lower hardness. He and Senetakis [17] captured clear cient of friction:
plowing tracks with a surface profiler after shearing an angular quartz Δμ
sand on the white shale surface, and they speculated that due to plowing ða bÞ ¼ (1)
lnðViþ1 =Vi Þ
and shale surface asperity breakage, the coefficient of friction of the WL
interface increased as the normal load increased. However, slight where Vi and Viþ1 are the shearing velocities under step i and step (iþ1),
negative correlations between coefficient of friction and normal load can and Δμ is the change in the steady state coefficient of friction of the two
be observed from most of the BL and BG specimens tested, based on the consecutive velocity steps (i.e., from Vi to Viþ1). Positive (a-b) values
results shown in Figs. 3(a), 4(a) and Table 1. The magnitude of the indicate velocity-strengthening response, i.e. the frictional force in­
decrease in average coefficient of friction of the BL and BG interfaces creases as the shearing velocity increases, while negative (a-b) values
were found to be 0.07 and 0.11 as the normal load increased from 0.5 N indicate velocity-weakening response. The average friction of the later
to 10 N. stage of each velocity step shearing was considered as the steady state
The initial tangential stiffness of all the tests are plotted against the coefficient of friction and it was taken into the calculation of Δμ. Mul­
corresponding normal load in Fig. 4(b), where the average values of tiple velocity steps were conducted on each of the three tests, and one (a-
initial tangential stiffness of the WL interface is also illustrated. The b) value can be calculated each time the velocity change takes place.
average KTo of the BG and BL interfaces are given in the dash lines in Among the total 20 (a-b) values derived, 14 of them were not larger than
Fig. 4(b), from which it can be observed that the initial tangential zero, and the maximum and minimum values were equal to 0.0026 and
stiffness of the BG interface increased from around 20 N/mm to 65 N/ 0.0069. Despite the fluctuation of the curves, for specimens BL7, BL8
mm and that of the BL interface increased from around 36 N/mm to 64 and BL9, the average (a-b) values were found to be 0.0035, 0.0007

5
H. He et al. Tribology International 144 (2020) 106119

Fig. 5. Coefficient of friction-displacement curves at varying shearing velocities: (a) specimen BL5 tested under 1 N of normal load; (b) specimen BL6 tested under 3
N of normal load; (c) specimen BL7 tested under 5 N of normal load.

and 0.0033, respectively, which indicate a slight velocity weakening specimen BL8 are plotted in Fig. 6(a). Fluctuation and discrepancies can
response. be observed among the curves; however, the elevated shearing velocity
Nine cycles of individual shearing cycles were performed on speci­ did not systematically alter the shearing response, including the initial
mens BL8 and BL9 under different shearing velocities (ranging from 0.1 tangential stiffness, the microslip displacement and the coefficient of
mm/h to 4096 mm/h) to further investigate the effect of shearing ve­ friction. The coefficients of friction of BL8 (FN ¼ 3N) and BL9 (FN ¼ 1N)
locity on the frictional behavior of the interfaces. Representative are plotted against shearing velocity in Fig. 6(b). In general, the co­
mobilized coefficient of friction against shearing displacement curves of efficients of friction of BL8 were lower than those of BL9, however, no

6
H. He et al. Tribology International 144 (2020) 106119

differences in mineralogy, surface microhardness and content of organic

matter of the black (organic) and white (inorganic) shale rocks. From the
velocity stepping tests, the BL interface tended to exhibit a slight ve­
locity weakening response, however, from a series of shearing tests
within a wider range of velocities, the μ values were not systematically
affected by the change of shearing velocity.

Acknowledgments

The work described in this paper was fully supported by the grants
from the Research Grants Council of the Hong Kong Special Adminis­
trative Region, China, project no. “CityU 11206617” and project no.
“CityU 11214218”.

References

[1] Boyer C, Clark B, Jochen V, Lewis R, Miller CK. Shale gas: a global resource.
Oilfield Rev 2011;23:28–39.
[2] DOE, NETL. Shale gas: applying technology to solve America’s energy challenges.
Washington, DC: US Departement of Energy (DOE) and National Energy
Technology Laboratory (NETL); 2011.
[3] Stevens P. The shale gas revolution: developments and changes. Chatham House
London; 2012.
[4] Soeder DJ. Porosity and permeability of eastern Devonian gas shale. SPE Form Eval
1988;3:116–24.
[5] Britt LK, Smith MB, Haddad ZA, Lawrence JP, Chipperfield ST, Hellman TJ.
Waterfracs: we do need proppant after all. In: SPE annual technical conference and
exhibition. Society of Petroleum Engineers; 2006.
[6] Montgomery CT, Smith MB. Hydraulic fracturing: history of an enduring
technology. J Pet Technol 2010;62:26–40.
[7] Barati R, Liang JT. A review of fracturing fluid systems used for hydraulic
fracturing of oil and gas wells. J Appl Polym Sci 2014;131. 40735.
[8] Wang H, Sharma MM. Modeling of hydraulic fracture closure on proppants with
proppant settling. J Pet Sci Eng 2018;171:636–45.
[9] Huang H, Babadagli T, Li HA, Develi K, Wei G. Effect of injection parameters on
proppant transport in rough vertical fractures: an experimental analysis on visual
models. J Pet Sci Eng 2019;180:380–95.
[10] Shimizu H, Murata S, Ishida T. The distinct element analysis for hydraulic
Fig. 6. Illustration of the effect of shearing velocity on the coefficient of friction fracturing in hard rock considering fluid viscosity and particle size distribution. Int
(a) mobilized coefficient of friction against shearing displacement curves of J Rock Mech Min Sci 2011;48:712–27.
shearing cycles under various shearing velocities; (b) coefficient of friction [11] Zhang F, Zhu H, Zhou H, Guo J, Huang B. Discrete-element-method/
computational-fluid-dynamics coupling simulation of proppant embedment and
against shearing velocity plots of BL8 (FN ¼ 3N) and BL9 (FN ¼ 1N). fracture conductivity after hydraulic fracturing. SPE J 2017;22:632–44.
[12] Zeng J, Li H, Zhang D. Numerical simulation of proppant transport in hydraulic
systematic effect of shearing velocity can be observed on the steady state fracture with the upscaling CFD-DEM method. J Nat Gas Sci Eng 2016;33:264–77.
[13] Zhang G, Gutierrez M, Li M. A coupled CFD-DEM approach to model particle-fluid
coefficient of friction of both specimens. Since each cycle of shearing mixture transport between two parallel plates to improve understanding of
was performed on a new surface of the shale, the inter-tests variation is proppant micromechanics in hydraulic fractures. Powder Technol 2017;308:
believed to be due to surface morphology effects. 235–48.
[14] Zhang H, Liu S, Xiao H. Frictional behavior of sliding shale rock-silica contacts
under guar gum aqueous solution lubrication in hydraulic fracturing. Tribol Int
4. Summary and conclusions 2018;120:159–65.
[15] Xiao H, Liu S, Wang D. Tribological properties of sliding shale rock–alumina
contact in hydraulic fracturing. Tribol Lett 2016;62:20.
In this study, the frictional behavior of a black shale with organic [16] Zhang H, Liu S, Xiao H. Tribological properties of sliding quartz sand particle and
matter was investigated by conducting micromechanical experiments shale rock contact under water and guar gum aqueous solution in hydraulic
between two types of interfaces, i.e., black shale against Leighton fracturing. Tribol Int 2019;129:416–26.
[17] He H, Senetakis K. A micromechanical study of shale rock-proppant composite
Buzzard sand (BL) and black shale against glass bead (BG) interface,
interface. J Pet Sci Eng 2020;184:106542.
with a particular focus on the effects of normal load and shearing ve­ [18] Parker MA, Ramurthy K, Sanchez PW. New proppant for hydraulic fracturing
locity on friction. Based on shearing tests under normal loads ranging improves well performance and decreases environmental impact of hydraulic
fracturing operations. In: SPE eastern regional meeting. Society of Petroleum
from 0.5 N to 10 N, the coefficient of friction (μ) of the BL interface was
Engineers; 2012.
found to be around 0.60 to 0.75, while that of the BG interface was found [19] Alary JA, Parias T. Method of manufacturing and using rod-shaped proppants and
to be lower by a magnitude of around 0.30 in comparison with the BL anti-flowback additives. United States patent and trademark office 2013;8:562.
interface. The average μ values of the BL interface was found to be 900 B2.
[20] Zhang C, Zhao L, Yu D, Liu G, Pei Y, Huang F, et al. The evaluation on physical
higher than the friction values of the Leighton Buzzard sand against a property and fracture conductivity of a new self-generating solid proppant. J Pet
white siliceous shale (WL) interface. The coefficient of friction of both Sci Eng 2019;177:841–8.
types of black shale-proppant contacts decreased as the normal load [21] Goldsby DL, Tullis TE. Flash heating leads to low frictional strength of crustal rocks
at earthquake slip rates. Science 2011;334:216–8.
increased, while the opposite trend was observed for the WL interface. [22] Zoback MD, Kohli A, Das I, Mcclure MW. The importance of slow slip on faults
The average initial tangential stiffness of the BL interface increased from during hydraulic fracturing stimulation of shale gas reservoirs. In: SPE americas
around 36 N/mm to 65 N/mm, while that of the BG interfaces increased unconventional resources conference. Society of Petroleum Engineers; 2012.
[23] Kohli AH, Zoback MD. Frictional properties of shale reservoir rocks. J Geophys Res:
from 20 N/mm to 65 N/mm as the normal load increased from 0.5 N to solid earth 2013;118:5109–25.
10 N. The initial tangential stiffness of the BL interface was around half [24] Burwell J, Rabinowicz E. The nature of the coefficient of friction. J Appl Phys
to one third of that of the WL interface at a given normal load. The 1953;24:136–9.
[25] He H, Senetakis K, Coop MR. An investigation of the effect of shearing velocity on
differences in the observed tribological behavior between the BL and WL
the inter-particle behavior of granular and composite materials with a new
interfaces were speculated to be predominantly attributed to the micromechanical dynamic testing apparatus. Tribol Int 2019;134:252–63.

7
H. He et al. Tribology International 144 (2020) 106119

[26] Sone H, Zoback MD. Mechanical properties of shale-gas reservoir rocks—Part 1: microscopy, high-pressure mercury intrusion, and gas adsorption. Energy Fuels
static and dynamic elastic properties and anisotropy. Geophysics 2013;78: 2014;28:945–55.
D381–92. [36] Dieterich JH, Conrad G. Effect of humidity on time-and velocity-dependent friction
[27] Sone H, Zoback MD. Mechanical properties of shale-gas reservoir rocks—Part 2: in rocks. J Geophys Res: Solid Earth 1984;89:4196–202.
ductile creep, brittle strength, and their relation to the elastic modulus. Geophysics [37] Dieterich JH, Kilgore BD. Direct observation of frictional contacts: new insights for
2013;78:D393–402. state-dependent properties. Pure Appl Geophys 1994;143:283–302.
[28] Nardelli V, Coop MR. The experimental contact behaviour of natural sands: normal [38] Ando Y, Ishikawa Y, Kitahara T. Friction characteristics and adhesion force under
and tangential loading. Geotechnique 2019;69:672–86. low normal load. J Tribol 1995;117:569–74.
[29] Sandeep CS, Senetakis K. Effect of Young’s modulus and surface roughness on the [39] Michalowski RL, Wang Z, Nadukuru SS. Maturing of contacts and ageing of silica
inter-particle friction of granular materials. Materials 2018;11:217. sand. Geotechnique 2017;68(2):133–45.
[30] Sandeep CS, Senetakis K. An experimental investigation of the microslip [40] Yan W, Ge H, Wang J, Wang D, Meng F, Chen J, et al. Experimental study of the
displacement of geological materials. Comput Geotech 2019;107:55–67. friction properties and compressive shear failure behaviors of gas shale under the
[31] He H, Senetakis K. An experimental study on the micromechanical behavior of influence of fluids. J Nat Gas Sci Eng 2016;33:153–61.
pumice. Acta Geotech 2019;14:1883–904. [41] Chowdhury MA, Helali M. The effect of amplitude of vibration on the coefficient of
[32] Yang L, Wang D, Guo Y, Liu S. Tribological behaviors of quartz sand particles for friction for different materials. Tribol Int 2008;41:307–14.
hydraulic fracturing. Tribol Int 2016;102:485–96. [42] Reches Ze, Lockner DA. Fault weakening and earthquake instability by powder
[33] Sandeep CS, Senetakis K. Grain-scale mechanics of quartz sand under normal and lubrication. Nature 2010;467:452.
tangential loading. Tribol Int 2018;117:261–71. [43] Sandeep CS, Senetakis K. Exploring the micromechanical sliding behavior of
[34] Krumbein WC, Sloss LL. Stratigraphy and sedimentation. second ed. W. H Freeman typical quartz grains and completely decomposed volcanic granules subjected to
and Company; 1963. repeating shearing. Energies 2017;10:370.
[35] Wang Y, Zhu Y, Chen S, Li W. Characteristics of the nanoscale pore structure in
Northwestern Hunan shale gas reservoirs using field emission scanning electron