International Journal of Rock Mechanics and Mining Sciences 103 (2018) 68–76

Contents lists available at ScienceDirect

International Journal of
Rock Mechanics and Mining Sciences
journal homepage: www.elsevier.com/locate/ijrmms

Effects of relative humidity and mineral compositions on creep deformation T

and failure of a claystone under compression
⁎
Zaobao Liua, Jianfu Shaoa, , Shouyi Xiea, Nathalie Conilb, Wenhua Zhaa
a
University of Lille, CNRS, Centrale Lille, FRE 2016 – LaMcube – Laboratoire de mécanique multiphysique multiéchelle, F-59000, Lille, France
b
Andra, CMHM, Bure, France

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

Keywords: During the last decades, clayey rocks have been widely investigated in many countries as a potential geological
Claystone barrier for underground disposal of radioactive waste. Clayey rocks are heterogeneous materials at a micrometer
Clayey rock scale. Mechanical behaviors of clayey rocks are generally influenced by mineralogical compositions and/or
Clay minerals water content. A large number of studies have been so far performed on short term behaviors of clayey rocks.
Creep
However, very limited data are available on time-dependent deformation and failure process. This paper is
Creep failure
devoted to investigating the effects of relative humidity and mineralogical compositions on creep deformation of
Relative humidity
a hard claystone and the possibility of creep failure is studied. Influences of confining pressure and structural
anisotropy are also investigated. For this purpose, four groups of creep tests are carried out under triaxial
compression conditions on claystone samples drilled from the underground research laboratory at Bure in
France. In the first group, creep tests are performed on samples with three different values of relative humidity
(RH, 59%, 85% and 98%) under a confining stress of 6 MPa. In the second group, claystone samples with
different mineralogical compositions drilled from different geological unities are investigated. In the third group,
for a selected value of relative humidity (RH 59%), triaxial creep tests are performed under different values of
confining stress (2, 6 and 12 MPa). Finally, effects of structural anisotropy of claystone at HR = 59% are in-
vestigated by considering samples respectively drilled in parallel and perpendicular directions to the bedding
planes. A series of new experimental data are obtained and analyzed. It is found that at the tested level of relative
humidity the confining stress seems to increase creep strain and rate of the claystone. The clay minerals espe-
cially the smectite can play an essential role in time-dependent deformation and increase creep strain of the
claystone at RH = 59%. Further, creep strains are more important under higher values of relative humidity than
lower ones. In very few cases, the failure due to acceleration creep was observed but for high deviatoric stress
level. The experimental data obtained in this work can be used as a background for numerical modeling of time-
dependent behavior of re/desaturated clayey rocks in the context of geological disposal of radioactive waste.

1. Introduction claystone. The present study is a part of this research program.

Clayey rocks are intensively investigated in many countries such as
France, Swiss, Belgium and Sweden as a potential geological barrier for
underground radioactive waste disposal. An underground research la-
boratory (URL) at the Meuse/Haute-Marne Center (CMHM) has been
constructed at Bure in France by the French National Agency for
Radioactive Waste Management (Andra). The URL is situated in the
Callovo-Oxfordian (COx) claystone layer.1,2 During the last decades, an
extensive research program has been launched by Andra involving both
experimental investigations and modeling studies. And a new in-
vestigation phase is undertaken for a deeper characterization of short-
and long- term thermo-hydro-mechanical behaviors of the COx
The excavation and ventilation of underground drifts can induce an
unsaturated zone in the claystone formation and a damage zone around
the galleries.3 The existence of such a damaged zone may affect both
short and long term hydromechanical behaviors of claystone. It is
shown that the mechanical behavior of COx claystone is sensitive to
water content.4–7 Therefore, it is necessary to investigate both short-
and long-term mechanical behaviors of the host claystone under dif-
ferent saturation conditions to verify the good functionality of the ex-
cavation damaged zone.
Instantaneous mechanical behaviors of the COx claystone have been
investigated by different laboratory tests, such as oedometric com-
pression, mini compression, micro indentation, uniaxial and triaxial

⁎
Corresponding author.
E-mail address: jian-fu.shao@polytech-lille.fr (J. Shao).

https://doi.org/10.1016/j.ijrmms.2018.01.015
Received 23 September 2016; Received in revised form 23 November 2017; Accepted 4 January 2018
Available online 03 February 2018
1365-1609/ © 2018 Elsevier Ltd. All rights reserved.
Z. Liu et al. International Journal of Rock Mechanics and Mining Sciences 103 (2018) 68–76

Table 1
Geometry of samples, testing program and conditions.

Rock core Sample NO. Direction D (mm) L (mm) Pc (MPa) RH Differential creep stress q (MPa) qpeak (MPa) Actual creep level q/qpeak

EST44311 B01 ⊥ 19.8 40.12 6 59% q=40.1 44.2 (Reference) One-step creep failure
B03 19.78 40.02 85% q=30.1 34.2 (Reference)
B06 19.81 40.10 98% q=26.1 29.5 (Reference)
B09 19.79 39.96 12 59% q=52.3 59.7(Reference)
EST30442 C2 ⊥ 19.72 40.50 6 59% 36.2 56.9 64%
C6 19.70 39.37 44.9 56.4 80%
C8 19.75 40.34 Triaxial reference qpeak =442
P1 // 19.77 40.15 Triaxial reference qpeak =405
P3 19.73 39.69 36.1 45.3 80%
EST50330 L1 // 19.74 39.48 6 59% Triaxial reference qpeak =324
L3 19.76 37.85 35.5 43 83%
L4 19.70 35.08 22.3 38.3 58%
L12 19.74 33.31 33.3 48.1 69%
L7 ⊥ 19.64 38.34 20.4 44.2 46%
L8 19.81 40.13 39.3 49.3 80%
L9 19.75 37.62 Triaxial reference qpeak =354 MPa
EST44300 A03 ⊥ 19.08 37.82 12 59% Triaxial reference qpeak =62.6
A04 19.15 38.06 2 qpeak =33.4
A22 19.13 39.30 6 qpeak =46.6
A02 19.22 36.76 6 39.8 49.6 80%
A07 19.16 39.14 12 45.6 57.3 80%
A16 19.10 39.08 2 27.8 35.6 78%

(Pc: Confining pressure; D: Diameter; L: Length; RH: Relative humidity; qpeak : Peak strength; Interstitial pressure is 0.95 MPa)

compression, triaxial extension.5–16 Without giving an exhaustive re-
view of all previous studies, just mentioning a number of tests that have
been devoted to investigate effects of water content, mineral compo-
sition, structural anisotropy and temperature on mechanical behaviors
of the COx claystone. In particular, these studies have shown that the
instantaneous mechanical properties of the COx claystone were clearly
affected by water saturation degree.6,7,17 For instance, the elastic
modulus and compressive strength decreases with the increase of water
content which probably due to the process of resaturation which da-
mages the samples. The behavior of claystone becomes more brittle
under lower saturation degree i.e. after desaturation of samples. Those
results give a basic understanding of the instantaneous behaviors of the
COx claystone in the excavation damaged zone.
A number of experimental studies have also been performed on
time-dependent behaviors of the COx claystone.18–25 Some results have
suggested that the creep strain rate of the claystone could reach an
order of magnitude of 10–11~10−10/s under low differential stress. The
similar magnitudes are also observed with in situ measurement after
excavation of galleries.26 The environmental factors such as water sa-
turation degree and temperature also have an important influence on
creep deformation of the claystone. The creep deformation of claystone
is found also sensitive to loading history, differential stress level as well
as structural anisotropy.22,27 In some studies, the existence of a stress
threshold below which creep strains stabilize has been evidenced by
multi-step creep tests23,28–30 and confirmed by one-step creep test.27 It
was found that creep strain rates of the COx claystone become lower
than 10−10/s and has a decreasing trend after 3–10 days of loading
depending on differential stress level and water content of sample.
However, it is very difficult to realize tertiary creep test on the COx
claystone by one step creep test, and few results have been obtained so
far very. The existence of a creep rate threshold above which the ter-
tiary creep can occur is still an open issue. Therefore, further experi-
mental studies are still in need for a better understanding of creep be-
haviors of the COx claystone in the excavation damaged zone.
The objective of the present study is to complete existing results,
and in particular, to investigate effects of resaturation and desaturation
by controlled relative humidity, mineralogical compositions and con-
fining stress on time-dependent strains of the re/desaturated COx
claystone. The experimental results can help to evaluate the creep
deformations of the COx claystone in the excavation damaged zone at
the URL at the CMHM center.
2. Material and sample preparation
The claystone samples tested in the present work are cored in the
URL at Bure. At the micrometer scale, the COx claystone is mainly
constituted of clay minerals, calcite, quartz and other secondary mi-
nerals. The average pore size is about some ten nanometers31,32 and the
majority of pores are inside the clay minerals. The average distribution
of mineral groups in the claystones throughout the whole formation is
clay fraction (phyllosilicates) ~42 ± 11% of the rock, carbonates
~30 ± 12% of minerals, ectosilicates ~25 ± 8% of minerals; and an-
cillary minerals constitute less than 4%.4 Mineralogical analyses show
the mineral content exhibits low variability in the bed plane but sig-
nificant perpendicular to the bedding. The COx clay-rich rock porosity
lies between 14% and 20% at the URL site and is close to 18% at the
URL main level,33 and natural water content ranges between 5% and
8%.26
The COx claystone geological layer can be vertically divided into
three lithostratigraphic units,34 i.e. the argillaceous rich unit (UA), the
transition unit (UT) and the silty carbonate-rich unit (USC). There are
also three subunits in the UA, i.e. UA1, UA2 and UA3, due to variation
of physical and geological properties. The rock cores tested in this paper
are drilled from different unities so that we can evaluate the effects of
mineralogical compositions.
The tested specimens have 40 mm in length and 20 mm in diameter
and are drilled from original rock cores of 300 mm in length and about
79 mm in diameter. The samples are cut by a diamond wire saw and
polished at both ends to obtain perpendicularity of sample ends to axis
for laboratory tests. During all these operations, plastic films are used to
minimize desaturation of the specimens. The list of tested samples with
their geometry is given in Table 1. Due to the difficulty to obtain a large
number of good quality samples from a single rock core, in total, four
cores were used in the present study.
The samples after cored for laboratory tests are inevitably desatu-
rated although carefully protected. At the moment, we give estimation
of the properties of claystone not with the fully saturated samples due
to the uncertainties in property change of the water-sensitive claystone

69
Z. Liu et al.
Table 2
Estimated mineral composition of tested rock cores.a
Rock core Unit Depth Mineral composition (%)
Quartz Calcite
EST50330 UA2 489.85 27 19
EST44300 UA1 467.2 25 27
EST30442 UT 520.22 22 29
a
Notes: (1) The mineralogical composition of the clay minerals is estimated by X-ray diffraction test on the rock cores from the COx argillite formation. The compositions of other
minerals are taken from the geological referential data for Cox claystone obtained by a geological model estimation (Andra 1999). (2) I: Illite; S: Smectite; R: Ratio.
induced by full resaturation. The saturated triaxial creep tests on the
water-sensitive COx claystone samples are quite different from those of
soils or high permeable rocks and involve too many uncertainties.
Firstly, a true resaturation process of the extreme low permeable
claystone samples for mechanical tests lasts too long time, which in-
volve also many other uncertainties in condition control in laboratory.
Moreover, even after fully resaturated, the mechanical loading in
triaxial creep tests can induce a very important non-homogeneity in
pore water pressure redistribution in the low permeable claystone since
pore water redistribution due to mechanical loading takes too much
time in the low permeable medium. These conditions can involve too
many uncertainties in the deformation measured, i.e. results inter-
pretation. For example, it is not easy to know whether the measured
deformation is due to mechanical loading or pore water redistribution.
It is also unclear how long the pore water redistribution will last at
different mechanical states considering that the mechanical loading can
induce matrix compaction of the claystone. Therefore, with these con-
siderations, we carry out triaxial creep tests on the COx claystone
samples but under fully saturated conditions but with controlled sa-
turation degree.
During sample preparation, the relative humidity values are kept
and assured respectively by different chemical solutions in an isolated
container in which the claystone samples are subjected to the process of
capillary re/desaturation. The samples will not be used for mechanical
tests before the stabilization of their mass in the isolated container. The
saturation degrees of the samples are related to the relative humidity
(RH) values which the samples are subjected to for mass balance.35 To
verify this point, we have estimated the saturation degrees of the
samples drilled from EST52609 (UA) under different relative humidity
values, such as 59%, 70%, 85%, 90% and 98%, which are respectively
maintained by the over-saturated chemical solutions of NaBr, KI, KCl,
BaCl2, and K2SO4. We have observed that the mass of the COx claystone
samples increase when subjected to RH=98%, decrease when subjected
to RH = 59%, 70% and 85%, and fluctuates when subjected to RH =
90%. These observations are generally the same as reported in litera-
ture.35 The estimation of saturation degree is carried out in the fol-
lowing steps:
(i) Estimate the initial specific density ρ0 of the cylindrical claystone
samples from their initial weights and volumes, which is obtained
as ρ0 = 2.342 g/cm3;
(ii) Estimate their initial water content w0 by drying method, which is
obtained as w0 = 6.88%;
(iii) Calculate the dry density ρd of the COx claystone by ρd = ρ0/(1+
w0) = 2.230 g/cm3;
(iv) Estimate the porosity ϕ of the samples given the grain density ρs =
2.69 g/cm3 for the COx claystone, which is obtained as ϕ = 1- ρd/
ρs≈17%;
(v) Calculate the initial saturation degree Sr0 of the COx claystone by
Sr0 = w0 ρd /(ϕ ρwater) ≈90%;
(vi) Estimate the stabilized water content w(RH) of the COx claystone at
the given relative humidity values by drying method;
International Journal of Rock Mechanics and Mining Sciences 103 (2018) 68–76
Clay minerals
Total Illite Kaolinite Chlorite I/S R = 0 I/S R = 1 Other
54 13.5 – 5.40 35.10 – –
48 9.60 – 1.44 36.96 – –
49 19.60 14.70 2.45 – – 12.25
(vii) Estimate the saturation degree Sr(RH) of the samples at the given
relative humidity values by Sr(RH) = w(RH) ρd /(ϕ ρwater).
Finally, we can calculate the saturation degrees of the COx claystone
samples are respectively about 98.2%, 90%, 85%, 69.1% and 57.8%
after re/desaturation under the imposed relative humidity of 98%,
90%, 85%, 70% and 59%. The initial saturation degree of the COx
claystone samples is estimated about 90% for the rock core received for
tests.
With the above estimation on the samples of EST52609, one can
know that the saturation degrees of the COx samples are closely related
to the imposed relative humidity levels before mechanical tests.
Therefore, the relative humidity level which the samples are subjected
to for mass balance can be an easy reference of the saturation degree of
the COx claystone samples before mechanical tests.
3. Experimental program and procedure
The experimental program is summarized in Table 1. The tests re-
lated to re/desaturation effects are performed on samples from the rock
core EST44311 (at 462 m depth) with three different imposed values of
relative humidity, 59%, 85% and 98%.
Effects of mean stress are studied by creep tests with three different
confining pressures on samples from the core EST44300 which was
desaturated at imposed RH 59%. On the other hand, effects of structural
anisotropy is also investigated by creep tests on claystone samples
drilled respectively in perpendicular direction and parallel direction
with respect to bedding planes in the cores EST30442 and EST50330.
Creep strains are compared between these two loading directions under
the same stress levels. Finally, effects of mineralogical compositions are
evaluated through creep tests on samples of three different clay mineral
contents as shown in Table 2. The other minerals are not measured but
only estimated by a geological model from Andra.
All the tests are carried out by using an autonomous testing system
composed of three main components, a confining cell, an upper
chamber for axial stress loading and an interstitial fluid flow circuit.
The detailed information about the testing system was presented in
literature.27,36 The tests are conducted in a thermally insulated room at
a constant temperature of 23 ± 0.5 °C maintained by a central air-
conditioner to reduce temperature fluctuations. The axial strain ε₁ and
radial strain ε₃ are measured respectively by a pair of strain gauges in
the parallel and perpendicular directions to the sample axis. The vo-
lumetric strain ε v can be thus obtained by ε v = ε1 + 2ε3. The differential
stress q acting on the sample is obtained by
q = (P − Pf ) S / D , (1)
where P is the pressure in the upper chamber; Pf is the friction of the
piston (about 3–5 bars) obtained at the beginning of deviator loading; S
and D are respectively the cross sectional area of the loading piston and
the claystone sample. Pressures in the confining cell and in the upper
chamber are all measured and monitored by pressure transducers. For
simplicity, the piston friction Pf is thought the same for deviator and
70
Z. Liu et al. International Journal of Rock Mechanics and Mining Sciences 103 (2018) 68–76

creep loading despite that there is a slight variation of 1–2 bars. All data
are recorded in an acquisition center. Each creep test is composed of the
following steps:

• Place the sample at the right position in the cell with a filter paper
and a diffusive plate at each end, and then seal the sample by a
plastic jacket from confining oil;
• Apply the confining pressure up to a selected value, say 2, 6 or
12 MPa;
• Inject the gas pressure until a same selected value is reached at the
inlet and outlet surfaces, for example 1 MPa; then one increases the
confining pressure by 1 MPa so that the effective confining pressure
is not modified;
• After the strain stabilization under the hydrostatic stress state, load
the differential stress with a constant axial strain rate of 2×10−6/s
up to a selected value and maintain it at this value and start the
creep phase;
• At the end of creep phase, if the sample is not failed, increase the
differential stress until the sample failure is reached.

4. Experimental results

The values of differential stress used in creep stages and those of

peak stress obtained after creep stages are given in Table 1. The dif-
ferential stress ratio (qc / qpeak ) of each test is also calculated and given in
Table 1. Further, before creep tests, conventional short-term triaxial
compression tests are carried out in order to obtain a reference value of
peak stress for each confining pressure and the obtained results are also
given in Table 1. In order to facilitate comparisons, the differential
stress level qc / qpeak will be used to evaluate creep deformation under
different stress states. For tests without tertiary creep, the peak strength
of the sample is taken as that tested after creep phase. For the tests with
tertiary creep, the peak strength is taken from the triaxial reference
tests.
The stress conditions and creep strains obtained on the samples
from the rock core EST44311 are shown in Fig. 1. The first three
samples are tested under three different values of initial RH, respec-
tively RH = 59% (see Fig. 1(a)), RH = 85% (see Fig. 1(b)) and RH =
98% (see Fig. 1(c)), and under the confining pressure (Pc) of 6 MPa. The
objective is to investigate effects of relative humidity on creep de-
formation and the possibility of tertiary creep. Another forth test is
carried out under RH = 59% and Pc = 12 MPa (see Fig. 1(d)) to in-
vestigate effects of the confining pressure. As a first indication, three
creep deformation stages (primary, stationary and tertiary) are ob-
served in this group of tests. The relative humidity increases creep
deformation of the claystone. More detailed discussions will be pro-
vided in the next section of the paper.
The experimental results of another group of tests under three dif-
ferent confining pressures are shown in Fig. 2 for the samples from the
rock core EST44300. This group of tests are carried out with RH = 59%
and a differential stress ratio qc / qpeak ≈ 80%, with the objective to
evaluate the creep strain rate rather than the creep failure stress under
different confining pressures. One can see that for a same differential
stress ratio, the creep deformation is clearly more important when the
confining pressure is higher.
The third series of creep tests are performed on the samples from the
rock core EST30442 at RH = 59%. The obtained results are shown in
Fig. 3 and Fig. 4. In this series, the samples are cored respectively in the
perpendicular and parallel directions to bedding planes of the claystone
in order to study effects of structural anisotropy on creep deformation.
It seems that for a same differential stress ratio (qc / qpeak ), the creep
strains are larger in the test in the perpendicular direction than that in Fig. 1. Creep stress and strain of claystone from EST44311.
the parallel direction. However, for a same absolute value of differential
stress, one obtains larger creep strains when the sample is loaded in the
parallel direction than in the perpendicular direction due to the ani-
sotropy of peak stress as shown in Table 1.

71
Z. Liu et al. International Journal of Rock Mechanics and Mining Sciences 103 (2018) 68–76

3.0 ε1 2.0 ε1
q =39.3MPa, q /q =80%, ε₁ , ⊥ q =35.5MPa, q /q =83%, ε₁ , //
qc=45.6MPa, qpeak=57.3MPa, Pc=12MPa
2.5 1.6
qc=39.8MPa, qpeak=49.6MPa, Pc=6MPa
q =33.3MPa, q /q =69%, ε₁ , //
2.0 1.2
Creep strain (%)

Creep strain (%)

1.5 qc=27.8MPa, qpeak=35.6MPa, Pc=2MPa q =22.3MPa, q /q =58%, ε₁ , //
0.8
1.0 q =20.4MPa, q /q =46%, ε₁ , ⊥
0.4
Creep time (h)
0.5 100 200 300 400 500
Creep time (h) 0.0
0.0 q =20.4MPa, ε₃ , ⊥ q =22.3MPa, ε₃ , // q =33.3MPa, ε₃ , //
0 100 200 300 400 500 -0.4
-0.5 Pc=2MPa q =35.5MPa, ε₃ , // q =39.3MPa, ε₃ , ⊥
Pc=6MPa -0.8 ε3
ε3 Pc=12MPa
-1.0

Fig. 2. Creep strains of claystone from EST44300 under three confining pressures at stress Fig. 5. Creep strain, stress and loading direction to beddings of samples from core
level qc/qpeak≈80%. EST50330.

60 q (MPa)
0.90 ε1
q =44.9MPa, q /q =80%, ε₁ , ⊥ q =39.3MPa, q /q =80%, ε₁ , ⊥
0.75 50
q =36.1MPa , q /q =80%, ε₁ , //
q =20.4MPa, q /q =46%, ⊥
q =39.3MPa, ε₃ , ⊥ 40 q =35.5MPa, q /q =83%, ε₁ , //
0.60
q =36.2MPa, q /q =64%, ε₁ , ⊥ q =35.5MPa, ε₃ , //
q =33.3MPa , q /q =69%, ε₁ , //
Creep strain (%)

0.45 q =33.3MPa, ε₃ , // 30
q =22.3MPa, ε₃ , // q =22.3MPa, q /q =58%, ε₁ , //
0.30 20

0.15 Creep time (h) 10

100 200 300 400 500 ε (%) ε (%)
0.00 0
q =36.2MPa , ε₃ , ⊥ -0.9 -0.6 -0.3 0 0.3 0.6 0.9 1.2 1.5 1.8 2.1

-0.15 q =36.1MPa , ε₃ , // Fig. 6. Strain-stress relations of samples from core EST50330 under Pc = 6 MPa, RH =
q =44.9MPa , ε₃ , ⊥
59%.
-0.30 ε3

Fig. 3. Creep stress, strain and differential stress direction to bedding of samples from εt 1 − εt 2 Δε
core EST30442.
ε̇ = =
t1 − t2 Δt (2)

60 where ε ̇ is the creep strain rate; εt1 and εt2 are the strain values (either
q (MPa)
axial or radial) respectively at the subsequent instances t1 and t2 ; Δε is
qc=36.2MPa
qc=44.9MPa the strain increment and Δt is the time interval. Using Eq. (2), both the
50 qpeak=56.9MPa
qpeak=56.4MPa axial and radial creep strain rates are calculated and presented Fig. 7.
qc /qpeak=64%
qc /qpeak=80%
⊥ As the total duration of creep is quite different for the tests with dif-
40 ⊥
ferent values of RH, the time interval used to calculate creep strain rates
30 is also different between different tests.
qc=36.1MPa The average axial and radial creep rates at the secondary stage are
20
qpeak=45.3MPa respectively 1.1×10−8/s and 1.0×10−8/s for the samples with 98%
qc/qpeak=80% RH as shown in Fig. 7(a). For the samples with 85% RH presented in
//
10
Fig. 7(b), the average axial and radial creep rates are 1.2×10−8/s and
0.5×10−8/s, respectively. As indicated in Fig. 7(c), the average axial
ε3 (%) ε1 (%) and radial creep rates are respectively 1.5×10−7/s and 3×10−8/s for
0
-0.4 -0.2 0 0.2 0.4 0.6 0.8 1 the samples with 59% RH. Therefore, it is found that the secondary
creep rate increases when the RH is lower so when the sample is more
Fig. 4. Stress-strain relations of samples from core EST30442 under Pc = 6 MPa, RH = desaturated.
59%.
A tertiary creep stage is also observed for the claystone studied as
shown in Fig. 7. Compared to the reference peak strength in Table 1, the
The last series of tests are performed on the samples from the rock applied deviator stress values indicate that the stress levels during the
core EST50330 at RH = 59%. The stress conditions and obtained re- creep tests are very high during the creep tests. The creep rate at the
sults are presented in Fig. 5 and Fig. 6. In this series, the tested samples secondary creep stage can be thought as the threshold over which the
are also cored in both perpendicular and parallel directions to bedding tertiary creep failure can occur. It is shown that the threshold values
planes. Based on comparisons with the tests in the previous series (core vary with the RH. Therefore, it is not easy to derive a unique threshold
EST30442), one intends to investigate effects of mineralogical compo- due to the strength variability with the RH. However, there would be a
sitions on creep deformation of the claystone. relationship between the stress and the strain at the onset of the tertiary
creep, which will be discussed later.
Moreover, the creep rate prior to creep failure shown in Fig. 7. in-
5. Effect of relative humidity dicates that the failure process of the claystone is closely related to the
RH. The creep failure is more brutal in the samples with a smaller RH
In order to better describe the different stages of creep process, it is value (RH 59%) than those with greater RH values (85% and 98%). This
convenient to define the creep strain rate as follows: can be explained by the fact that at RH 59% the material loses water

72
Z. Liu et al. International Journal of Rock Mechanics and Mining Sciences 103 (2018) 68–76

Tertiary creep
Primary creep and its behavior is more brittle than at RH 85% et RH 98% where the
Secondary creep
sample absorbed water and has been softened by the water taken
3.0 ε dε/dt (/s) 10 without confinement due to the presence of clay minerals especially the
smectite.
2.5 ε₁ Axial creep rate
ε₃ Radial creep rate
2.0 10
Creep strain (%)

6. Effect of confining pressure

Creep rate (/s)

1.5
Onset of tertiary creep 10
1.0
6.1. Creep failure process

0.5
10 The creep strains from the test with Pc = 12 MPa and RH = 59%
Time (h) are shown in Fig. 7(d) and they are characterized by a clear three-stage
0.0
0 1 2 3 4 creep process. The sample failure is characterized by a tertiary creep
-0.5 ε 10
(a) Pc=6 MPa, RH=59% phase. The average axial and radial creep rates are respectively about
1.6×10−8/s and 1.5×10−8/s. Compared with those obtained in the
Tertiary creep
Primary creep test with Pc = 6 MPa and RH = 59%, as shown Fig. 7(c), it is found
Secondary creep
that the creep rates under Pc = 12 MPa are much smaller than those
4.0 ε dε/dt (/s) 10 under Pc = 6 MPa. The creep failure in the test with Pc = 12 MPa
ε₁ Axial creep rate occurs in a more progressive way and the time needed to reach the
3.0 ε₃ Radial creep rate
10
failure is longer than the test with Pc = 6 MPa.
2.0 Therefore, the creep failure of the claystone with RH = 59% is more
Creep strain (%)

Onset of tertiary creep brutal under Pc = 6 MPa than under Pc = 12 MPa. At the same time,
Creep rate (/s)

1.0
10
Time (h) both the axial and radial strains obtained before creep failure are much
0.0
0 30 60 90 120 larger under Pc = 12 MPa than under Pc = 6 MPa. The confining
-1.0 10
pressure seems to increase the rock deformability during creep process.
-2.0
10
-3.0 6.2. Strain rate and strain hardening
-4.0 ε 10
(b) Pc=6 MPa, RH=85% The creep strain rates of the samples from the core EST44300 are
presented in Fig. 8. The samples are subjected to the same differential
stress level of qc / qpeak = 80% but under three different values of con-
Primary creep Tertiary creep
Secondary creep fining. Only a primary creep stage is obtained with this differential
4.0 10
stress level with a progressive decrease of creep rates. The average of
ε dε/dt (/s)
ε₁ Axial creep rate strain rate at the final stage of test is around 10−10/s. It is found that
ε₃ Radial creep rate the axial rates are much greater than the lateral ones in all tests, leading
3.0
10
to an important volumetric compaction. Importantly, the creep strain
Creep strain (%)

2.0 rate in the claystone increases when the confining pressure is higher. In
Creep rate (/s)

Onset of tertiary creep 10

order to quantify the effect of confining pressure, the following rela-
1.0
tion37 is proposed to describe the evolution of creep strain rate with
Time (h) 10 time:
0.0
0 50 100 150 200 −m
t
-1.0
10 ε ̇ = Aε0̇ ⎛ ⎞ ⎜ ⎟ ; t ≥ t0
t
⎝ 0⎠ (3)
-2.0 ε 10
In this relation, t0 = 1hour is a normalizing parameter of time t .
(c) Pc=6 MPa, RH=98%
Three coefficients are introduced for the evolution of creep strain rate.
Tertiary creep It is assumed that the scaling coefficient A is a function of confining
Primary creep
Secondary creep pressure. The coefficient ε0̇ defines the initial creep rate and is related to
the loading rate of differential stress. The exponent coefficient m
6.0 ε dε/dt (/s) 10
ε₁ Axial creep rate Creep time (h)
5.0 Radial creep rate
ε₃ 0 100 200 300 400 500
10
4.0
Axial Pc=2MPa
3.0 Onset of tertiary creep 10 Radius Pc=2MPa
Creep strain (%)

Strain hardening law: = ε (t/t )

Creep rate (/s)

Axial Pc=6MPa
Radius Pc=6MPa
2.0 Axial Pc=12MPa
10
Radius Pc=12MPa
1.0 Pc=12MPa, axial, = 6E-07(t/t )
R² = 0.9423
(/s)

0.0 10
Pc=6MPa, axial, = 5E-07(t/t )
210 Time (h) 280
Creep rate

0 70 140 10 R² = 0.9553
-1.0
Pc=2MPa, axial, = 3E-07(t/t )
-2.0 R² = 0.9921

-3.0 ε 10 Pc=6MPa, radius, = 2E-07(t/t )

10 R² = 0.9427
(d) Pc=12 MPa, RH=59%
Fig. 7. Creep strain rate and onset of tertiary creep of claystone with different RH and Pc=12MPa, radius, = 5E-07(t/t )
R² = 0.9755
confining pressures. Pc=2MPa, radius: = 1E-07(t/t )
10
R² = 0.987

Fig. 8. Strain rate of claystone subjected to qc/qpeak = 80% under different confining
pressure.

73
Z. Liu et al.
0.5 A 8. Effect of mineralogical composition on creep strain hardening
0.4
0.3
0.2
A = 0.1167(Pc /Pc0 )0.394 + 0.0935
R² = 0.9814 (Pc0 =1MPa)
0.1
Pc (MPa)
0 clay minerals; the other minerals contents were not measured but only
0 2 4 6 8 10 12 14 16 18 20 22
Fig. 9. Relationship between coefficient A and confining pressure Pc of creep tests.
controls the evolution of creep rate and its value should be affected by
differential stress level, mineral compositions, porosity, relative hu-
midity and temperature.
For the group of tests presented here, as the samples are drilled from
a same rock core and tested at room temperature with a same RH value,
it is then not possible to determine all such dependences. From the
results presented in Fig. 8, the values of the term ( Aε0̇ ) are calculated
and respectively equal to 6×10−7/s, 5×10−7/s and 3×10−7/s, for the
confining pressures of 12, 6 and 2 MPa. Considering that the value of
coefficient (ε0̇ ) is the loading rate of differential stress, say 2×10−6/s,
one can calculate the values of scale coefficient A for different values of
confining pressure. The variation of A with confining pressure is pre-
sented in Fig. 9, with a very good correlation between calculated and
experimental values.
7. Threshold of tertiary creep
As presented above, a tertiary creep stage can occur and leads to
time-dependent failure of the claystone under certain specific condi-
tions although tertiary creep has been very rarely observed so far. It is
then important to characterize the threshold of tertiary creep onset.
According to the results shown in Fig. 7, the tertiary creep onset
point in each test can be defined by a couple of critical values of stress
(qc ′) and axial strain (ε1 ′). These values are reported as shown in Fig. 10
for different values of Pc and RH.
It is found that for a given Pc, the critical axial strain increases while
the critical differential stress decreases when the RH is higher. Further,
both critical values are influenced by confining pressure. Based on this,
it is possible to define the relationship between the critical differential
stress and axial strain using the following exponential form:
ε′
−C1
qc′ = C0 ⎜⎛ 1 ⎟⎞
⎝ζ ⎠
In this relation, ζ = 10−2 is a strain normalizing parameter.C0 is a
scale coefficient that is dependent on confining pressure; C1 is a coef-
ficient that is influenced by mineralogical composition, relative hu-
midity and temperature.
For the samples from the core EST44300 and tested under 6 MPa
confining pressure at room temperature, the following values are ob-
tained for two coefficients C0 = 76.874MPa and C1 = 1.18, as shown in
Fig. 10. It is also possible to define the tertiary creep threshold for the
sample tested under 12 MPa confining pressure. As the mineralogical
compositions of this group of samples are similar as those of the sam-
ples tested in the previous group, one takes the same value of C1 = 1.18
for both groups. Then the value of scale coefficient is determined and
equal to C0 = 280.9MPa. It is found that its value increases with con-
fining pressure.
International Journal of Rock Mechanics and Mining Sciences 103 (2018) 68–76
The mineralogical compositions of claystone vary vertically in the
different geological layers. The variation in mineralogical compositions
especially in the clay minerals can have an important influence on the
creep deformation of claystone. In the present study, the claystone
samples C6, L8 and A02, drilled respectively from the rock cores
EST30442, EST50330 and EST44300, are investigated. The three cores
were taken from three different geological units respectively the UT, the
UA2 and the UA1. The main mineral compositions of these samples are
presented in Table 2. They are mainly differentiated by the content of
estimated, so the real values can be a little different than those esti-
mated, which may play a role in the result.
Three creep tests are performed on these samples under a confining
pressure of 6 MPa with a differential stress level of qc / qpeak = 80% and a
relative humidity of 59%. The creep strains obtained three tests are
compared and presented in Fig. 11. It is found that the largest creep
strain is obtained in the sample A02, which has the maximum ratio of
the mineral I/S(R = 0) among three samples. At the same time, the
smallest creep is found in the sample C6 whose I/S(R = 0) content is
nearly zero. Therefore, the creep deformation of the claystone is
strongly affected by the content of clay minerals, especially enhanced
by the ratio of I/S(R = 0).
The clay minerals generally act as a weak zone in the claystone with
the presence of water,38 especially the interlayered smectite, i.e., the I/
S(R = 0). The illite minerals have an empirical chemical formula of
K0.6(H3O)0.4Al1.3Mg0.3Fe2+0.1 Si3.5O10(OH)2·(H2O) and the smectite has an
empirical chemical formula of (Na,Ca)0.3(Al,Mg)2Si4O10(OH)2∙n
(H2O).39 The chemical formula indicates that the smectite has a greater
potential of absorbing and losing water than the illite. This property can
further effect the volume change properties of the claystone under
humid state. During the creep of wet claystone, the I/S minerals are
likely to dislocate more actively than other minerals40 due to their high
potential of water absorption. Different properties of the illite and
smectite minerals have been also reported in some previous studies.41,42
Thus, in mineral scale, the compactions of the clay minerals under
compressive stress are more important among the smectite minerals
than among the other clay minerals. Thus, at a macroscopic scale, the
claystone with a higher ratio of I/S(R = 0) exhibits a more important
creep deformation than those with a small ratio of I/S(R = 0).
Since I/S minerals are the most unstable with the presence of water
and has a much bigger potential of compaction, the sample C6 from the
core EST30442 which has the smallest content of such minerals exhibits
the greatest peak strength after creep process, while the peak strength
of A02 is greater than that of L8.
The creep rates in the three tests are calculated and shown in
Fig. 12. The creep rates in steady creep stage are in the order of 10−10/
s. Again, the evolution of creep rates with time can be described by the
(4)
relationship given in Eq. (3), with a good correlation. The creep rates of
the three samples satisfy the decreasing order εA02̇ > εL8 ̇ , due to
̇ > εC6
the difference of content in the active I/S minerals. As a conclusion, the
active I/S clay minerals are mostly responsible for the strength reduc-
tion and large creep strains of the claystone.
9. Conclusions
The creep behaviors of desaturated or resaturated COx claystone
have been studied by several groups of creep tests. The effects of re-
lative humidity, mineralogical compositions, structural anisotropy and
confining pressure have been investigated. According to the results
obtained, some conclusions can be made:
(1) Creep rates and failure process of the claystone are closely related
to the variation of relative humidity, i.e. the water saturation. The
resaturated claystone (under higher RH values) exhibits a much
74
Z. Liu et al. International Journal of Rock Mechanics and Mining Sciences 103 (2018) 68–76

Fig. 10. Evaluation of onset criterion of tertiary creep in stress/strain space of tested claystone.

2.5
smaller secondary creep rate than that with lower ones. Creep
A02: q =39.8MPa, q /q =80%, ⊥ (77% of I/S=0 smectite in clay matrix)
failure is more brutal in the desaturated claystone (with smaller R)
2.0
values than in the resaturated ones|.
L8: q =39.3MPa, q /q =80%, ⊥ (65% of I/S=0 smectite in clay matrix) (2) The creep failure of the desaturated claystone (RH59%) is more
1.5
brutal under lower values of confining pressure than under higher
Creep strain (%)

1.0 ones. The confining pressure enhances creep deformation and creep
C6: q =44.9MPa , q /q =80%, ⊥ (0% of I/S=0 smectite in clay matrix)
strain rates of the claystone, especially radial creep strains. A power
0.5 law can quantify the relationship between the critical differential
100 200 300 400 500
stress and axial strain at the onset of tertiary creep of the claystone
0.0 with different imposed RH values.
C6: q =44.9MPa , q /q =80%, ⊥ Creep time (h)
(3) The creep strains of the desaturated claystone (at RH59%) are
-0.5 A02: q =39.8MPa, q /q =80%, ⊥ strongly affected by mineralogical compositions, especially the ac-
L8: q =39.3MPa, q /q =80%, ⊥ tive clay minerals. The active clay minerals are mostly responsible
-1.0 for the strength reduction and large creep strains of the desaturated
claystone.
Fig. 11. Creep strains of claystone samples with different mineral compositions. (4) Although creep failure of the COx claystone due to acceleration
creep has been rarely reported, the tertiary creep of the tested
claystone may occur under very high deviator stress level and other
accompanied conditions such as desaturation or resaturation. These
conditions up to now barely present in field of the URL and should
be always avoided to minimize the possibility of creep failure.

It is indicated from the experimental results that the creep de-

formation of the geological unit UA is more important than that of the
UT in the excavation damaged zone of URL. These are due to the fact
that the creep deformation of the claystone in UA is much more im-
portant than that in UT which is related to the high content of clay
minerals especially the smectite. The resaturation of the COx claystone
without constraints can induce important strength decrease and prop-
erties changes. Therefore, the resaturation of the COx claystone should
be constraint with in situ stress conditions to minimize uncertainties in
Fig. 12. Creep strain rates of claystone samples with different mineral compositions.
sample property change.

Acknowledgements

This work is supported by Andra (French National Agency for

75
Z. Liu et al.
management of radioactive waste), which is acknowledged. The
Granted support (NO. Z016013) from the Key Laboratory of Ministry of
Education on Safe Mining of Deep Metal Mines, Northeastern University
to the first author (Dr. Zaobao Liu) is also acknowledged. The authors
gratefully thank the editors and the anonymous reviewers for providing
constructive comments and suggestions that improve the presentation
of the paper.
References
1. Andra. Referentiel du site Meuse-Haute Marne. Paris: Andra; 2005.
2. Andra. Recherche et développement à l′Andra pour les projets de stockage. Paris: Andra;
2012.
3. Armand G, Leveau F, Nussbaum C, et al. Geometry and Properties of the Excavation-
Induced Fractures at the Meuse/Haute-Marne URL Drifts. Rock Mech Rock Eng.
2014;47:21–41.
4. Armand G, Djizanne H, Zghondi J, de La Vaissière R, Talandier J, Conil N. Inputs
from in situ experiments to the understanding of the unsaturated behaviour of
Callovo-Oxfordian claystone. 3rd European Conference on Unsaturated Soils - “E-
UNSAT 2016”. Paris: EDP Sciences; 2016.
5. Liu ZB, Shao JF. Moisture effects on damage and failure of Bure claystone under
compression. Geotechnique Lett. 2016;6:182–186.
6. Zhang F, Hu DW, Xie SY, Shao JF. Influences of temperature and water content on
mechanical property of argillite. Eur J Environ Civ Eng. 2013;18:173–189.
7. Hu DW, Zhang F, Shao JF, Gatmiri B. Influences of mineralogy and water content on
the mechanical properties of argillite. Rock Mech Rock Eng. 2013;47:157–166.
8. Mohajerani M, Delage P, Monfared M, Tang AM, Sulem J, Gatmiri B. Oedometric
compression and swelling behaviour of the Callovo-Oxfordian argillite. Int J Rock
Mech Min Sci. 2011;48:606–615.
9. Zhang C-L. Experimental evidence for self-sealing of fractures in claystone. Phys Chem
Earth, Parts A/B/C. 2011;36:1972–1980.
10. Zhang F, Xie SY, Hu DW, Shao JF, Gatmiri B. Effect of water content and structural
anisotropy on mechanical property of claystone. Appl Clay Sci. 2012;69:79–86.
11. Huang Y, Xie SY, Shao JF. An experimental study of crack growth in claystones. Eur J
Environ Civ En. 2013;18:307–319.
12. Yang D, Chanchole S, Valli P, Chen L. Study of the anisotropic properties of argillite
under moisture and mechanical loads. Rock Mech Rock Eng. 2013;46:247–257.
13. Masri M, Sibai M, Shao JF, Mainguy M. Experimental investigation of the effect of
temperature on the mechanical behavior of Tournemire shale. Int J Rock Mech Min
Sci. 2014;70:185–191.
14. Auvray C, Arnold G, Armand G. Experimental study of elastic properties of different
constituents of partially saturated argillite using nano-indentation tests. Eng Geol.
2015;191:61–70.
15. Yang D, Billiotte J, Su K. Characterization of the hydromechanical behavior of ar-
gillaceous rocks with effective gas permeability under deviatoric stress. Eng Geol.
2010;114:116–122.
16. Armand G, Conil N, Talandier J, Seyedi DM. Fundamental aspects of the hydro-
mechanical behaviour of Callovo-Oxfordian claystone: from experimental studies to
model calibration and validation. Comput Geotech. 2017;85:277–286.
17. Liu Z, Shao J, Xu Y Water Saturation Induced Strength Degradation of Callovo-
Oxfordian Claystone. In: Papamichos E, Papanastasiou P, Pasternak E, Dyskin
A editors. Bifurcation and Degradation of Geomaterials with Engineering Applications,
Proceedings of the 11th International Workshop on Bifurcation and Degradation in
Geomaterials dedicated to Hans Muhlhaus, Limassol, Cyprus, 2017, 21-25 May 2017.
Cham: Springer International Publishing; 2017:11-17.
18. Bérest P. Viscoplasticité en mécanique des roches. In: F. D, editor. Manuel de rhéologie
des Géomatériaux. Paris: Presses de l′ENPC;1987:235-257.
19. Ghoreychi M. Comportement thermo-hydro-mécanique des argilites de l′Est et son
76
International Journal of Rock Mechanics and Mining Sciences 103 (2018) 68–76
incidence sur les ouvrages souterrains. Actes des Journées Scientifiques de l′Andra. Les
Ulis, France: EDP Sciences; 1999:397–418.
20. Vales F, Bérest P, Gharbi H. Interaction between creep, swelling and pressure-solu-
tion during indentation tests. Clays in Natural and Engineered Barriers for Radiactive
Waste Confinement. Reims Paris: ANDRA;2002:207-208.
21. Bérest P, Charpentier J-P, Vales F. Creep of argillite under small mechanical loading.
Kiwir Conference. Paris; 2001.
22. Gasc-Barbier M, Chanchole S, Bérest P. Creep behavior of Bure clayey rock. Appl Clay
Sci. 2004;26:449–458.
23. Fabre G, Pellet F. Creep and time-dependent damage in argillaceous rocks. Int J Rock
Mech Min Sci. 2006;43:950–960.
24. Zhang C-L, Rothfuchs T. Experimental study on self-sealing capacity of clay rocks.
Clays in natural & engineered barriers for radiactive waste confinement. [ANDRA].
Lille. 2007:539–540.
25. Liu ZB, Shao JF, Liu TG, Xie SY, Conil N. Gas permeability evolution mechanism
during creep of a low permeable claystone. Appl Clay Sci. 2016;129:47–53.
26. Armand G, Noiret A, Zghondi J, Seyedi DM. Short- and long-term behaviors of drifts
in the Callovo-Oxfordian claystone at the Meuse/Haute-Marne Underground
Research Laboratory. J Rock Mech Geotechnical Eng. 2013;5:221–230.
27. Liu Z, Xie S, Shao J, Conil N. Effects of deviatoric stress and structural anisotropy on
compressive creep behavior of a clayey rock. Appl Clay Sci. 2015;114:491–496.
28. Yang D, Bornert M, Chanchole S, Wang L, Valli P, Gatmiri B. Experimental in-
vestigation of the delayed behavior of unsaturated argillaceous rocks by means of
Digital Image Correlation techniques. Appl Clay Sci. 2011;54:53–62.
29. Liu ZB, Xie SY, Shao JF, Conil N. Multi-step triaxial compressive creep behaviour and
induced gas permeability change of clay-rich rock. Géotechnique. 2017;0:1–9.
30. Liu Z, Shao J, Xie S, Conil N. Triaxial Creep Induced Gas Permeability Change and
Elastic Modulus Variation in Callovo-Oxfordian Argillite. In: Vandamme M, Dangla
P, Pereira J-M, Ghabezloo S, eds. Poromechanics VI. Paris: American Society of Civil
Engineers; 2017:1037–1044.
31. Robinet J-C. Minéralogie, porosité et diffusion des solutés dans l′argilite du Callovo-
Oxfordien de Bure (Meuse, Haute-Marne, France) de l′échelle centimétrique à
micrométrique. Poitiers: Université de Poitiers; 2008.
32. Liu Z, Shao J, Xie S, Secq J. Gas permeability evolution of clayey rocks in process of
compressive creep test. Mater Lett. 2015;139:422–425.
33. Yven B, Sammartino S, Géraud Y, Homand F, Villieras F. Mineralogy, texture and
porosity of Callovo-Oxfordian argillites of the Meuse/Haute-Marne region (eastern
Paris Basin). Mémoires De la Société géologique Fr. 2007;178:73–90.
34. Armand G, Bumbieler F, Conil N, de La Vaissière R, Bosgiraud JM, Vu MN. Main
outcomes from in situ THM experiments programme to demonstrate feasibility of
radioactive HL-ILW disposal in the Callovo-Oxfordian claystone. J Rock Mech Geotech
Eng. 2017;9:415–427.
35. Zhang F. Comportement mécanique des argilites de MHM: influences de la saturation, de
la composition minéralogique et de la température [Ph.D. Thesis]. Lille: Université Lille 1;
2011.
36. Secq J. Triaxial cell for the testing of geomaterials in compression and in shear. G01N15/
00, G01N33/24, G01N3/10. ed. France: Google Patents; 2010.
37. Norton FH. The Creep of Steel at High Temperatures. New York: McGraw-Hill book
company, inc.; 1929.
38. Gratier JP, Jenatton L, Tisserand D, Guiguet R. Indenter studies of the swelling, creep
and pressure solution of Bure argillite. Appl Clay Sci. 2004;26:459–472.
39. Roberts WL. Encyclopedia of Minerals. In: Campbell TJ, Jred. GRR. 2nd ed. Chapman
& Hall, New York: Van Nostrand Reinhold; 1990:980.
40. Ferrage E, Vidal O, Mosser-Ruck R, Cathelineau M, Cuadros J. A reinvestigation of
smectite illitization in experimental hydrothermal conditions: results from X-ray
diffraction and transmission electron microscopy. Am Mineral. 2011;96:207–223.
41. Altaner SP, Ylagan RF. Comparison of structural models of mixed-layer illite/smec-
tite and reaction mechanisms of smectite illitization. Clays Clay Miner.
1997;45:517–533.
42. Stixrude L, Peacor DR. First-principles study of illite-smectite and implications for
clay mineral systems. Nature. 2002;420:165–168.