Acta Geotechnica (2023) 18:3193–3211

https://doi.org/10.1007/s11440-022-01689-0 (0123456789().,-volV)(0123456789().
,- volV)

RESEARCH PAPER

BCV bentonite hydromechanical behaviour and modelling

Jiřı́ Svoboda1 • David Mašı́n2 • Jan Najser2 • Radek Vašı́ček1 • Irena Hanusová3 • Lucie Hausmannová3

Received: 16 December 2021 / Accepted: 24 August 2022 / Published online: 5 December 2022
 The Author(s) 2022

Abstract
It is assumed that clay materials such as bentonites will be used as the buffer and backfill of high-level radioactive waste
disposal. Due to the very long timescale, the design can be based only on mathematical models that have been thoroughly
validated via the use of high-quality laboratory data. The presentation of an extensive laboratory dataset, accompanied by
the validation of a bentonite constitutive model, comprises the main topic of this paper. The objectives of the laboratory
research were to determine the basic hydromechanical properties of BCV bentonite and to study the stress path behaviour
of this bentonite. The laboratory results revealed that the hydromechanical behaviour is highly dependent on the stress path
assumed, as well as the method of water introduction into the sample. The results further indicate that the micro- and
macrostructures are constantly evolving, which exerts a significant impact on the determination of the material properties
and the development of mathematical models. Moreover, it was demonstrated that swelling pressure tests, although they
are typically considered to be element tests (with constant fields of state variables), when adopted by researchers for the
calibration of constitutive models, do not in fact constitute element tests. A hypoplastic model of the behaviour of bentonite
was implemented and validated using the experimental data obtained. It was demonstrated that the hypoplastic model
accurately predicted the main behavioural features of bentonite that are important in terms of the prediction of the
behaviour of the buffer in nuclear waste repositories.

Keywords BCV  Bentonite  Element test  Hydraulic conductivity  Hypoplastic model  Mechanical properties 
Mathematical model  Swelling

1 Introduction favourable sealing behaviour, which is based on their low

levels of permeability and high swelling ability.
It is planned that high-level radioactive waste (especially Since the radioactive waste emplaced in the DGR must
spent nuclear fuel) will be disposed of in a deep geological be isolated from the biosphere for up to 1 million years, it
repository (DGR), the safety of which is based on a multi- must be proved that the multibarrier system is able to
barrier system (see Fig. 1). The waste will be placed in waste function over such a long period of time. This is possible
disposal packages (WDP) which will be sealed with a buffer only via the creation of mathematical models that are
in disposal boreholes; the other underground workings will validated using high-quality experimental data (Wigger
subsequently be backfilled. It is assumed that clay materials [23] Laloui [13]). This paper presents both the extensive
such as bentonites and montmorillonitic clays will be used as experimental database compiled on the behaviour of BCV
the buffer and backfill materials due to their highly bentonite and the validation of a constitutive model.
The objectives of the laboratory research were to
determine the basic hydromechanical properties of BCV
& Jiřı́ Svoboda bentonite in both its powder and compacted powder forms
[email protected]; [email protected]
and to study the stress path behaviour of this bentonite, as
1
Czech Technical University, Thakurova 7, 16629 Prague 6, well as to compile a complex dataset for this material with
Czech Republic respect to its use as the reference research material in the
2
Charles University, Albertov 6, 12843 Prague 2, Czech ongoing development of the Czech DGR concept. The aim
Republic of the research team, which is composed of several insti-
3
SÚRAO, Dlážděná 6, 11000 Prague 1, Czech Republic tutions and coordinated by the Czech nuclear waste agency

123

Content courtesy of Springer Nature, terms of use apply. Rights reserved.

3194 Acta Geotechnica (2023) 18:3193–3211

Fig. 1 Scheme of the deep geological repository planned in the Czech Republic SÚRAO

SURAO, was to obtain similarly comprehensive data to a hypoplastic bentonite material model for FEM analysis
that contained in datasets available on other widely used purposes [15, 16]. The model was subsequently employed
bentonites such as MX-80 bentonite (Cui et al. [5] and in the interpretation of the behaviour of the material
FEBEX bentonite [21]). All the tests were performed by observed during the laboratory experimentation phase.
the various laboratories involved on the same batch of
bentonite in order to ensure the consistency of the tested
material, which is rather unusual for other bentonite data- 2 BCV bentonite
sets available in the literature. In order to obtain similar
datasets for the development and calibration of models, The Czech DGR concept is based on the use of bentonites
experimental results are usually obtained from various extracted from Czech bentonite deposits. BCV (Bentonite
sources that consider differing batches of material over Černý Vrch) is a Ca/Mg bentonite extracted from the Černý
several years ([6], Thatcher [20]). Since bentonites are vrch bentonite deposit located in the České Středohořı́ Mts.
natural materials, their composition and properties vary in the NW of the Czech Republic. BCV bentonite was chosen
between both sources and batches (Dixon [6]), which may as a suitable representative of Czech bentonites [9] and was
lead to biased results. A single batch-based complex first subjected to testing in 2017. Following the pilot char-
dataset provides high-quality data for the development and acterisation of BCV (Červinka et al. [3]), this bentonite was
validation of constitutive models. used in a number of laboratory [22] and Baryla et al. [1] and
The investigation of the stress path behaviour is particu- in situ projects [11] and Svoboda et al. [19].
larly important since it is known that bentonite materials BCV is produced industrially by Keramost a.s. at their
evince complex (micro)structures and that the state of the Obrnice plant. The bentonite treatment process commences
material leads to changes in its various properties and beha- with the sieving of the coarse material employing a sieve
viour. This may have important consequences in terms of the with a 20 9 20 cm mesh. The bentonite is then homo-
design of the engineered barrier system (EBS) since, for genised in a roller mill from where it is transferred to a
example, the swelling pressure may differ significantly rotary oven where the material is dried (average tempera-
depending on whether, and to what extent, the deformation of ture of 110 C) for 45 min. The drying process ensures a
the material is allowed. It is important to note that the water content of around 10% and results in the preparation
experimental dataset presented in this paper does not include of the material for final milling and air sieving. The pro-
the consideration of experiments conducted at elevated tem- cessing of the BCV material leads to a mixture where 1%
peratures. The effect of temperature is also being studied of the total weight of the grains can be larger than
extensively and will be described in forthcoming publications. 0.315 mm and at least 70% of the grains are smaller than
The data obtained from the laboratory research was used 0.063 mm. The material is supplied in 48 kg bags, larger
as a tool for the development, calibration and validation of bags or tanks [22].

123

Content courtesy of Springer Nature, terms of use apply. Rights reserved.

Acta Geotechnica (2023) 18:3193–3211 3195

BCV bentonite (see Tables 1 and 2) exhibits a high Table 2 Results of the silicate analysis of BCV bentonite [14]
proportion of minerals from the smectite group (Ca/Mg Wt % BCV_2017
montmorillonite of over 60%), while other clay minerals
(kaolinite, illite) are present to a lesser extent. With respect SiO2 51.86
to the accessory phases, the presence of genesis-related Fe- TiO2 2.34
carbonates (siderite, ankerite) and Fe oxides (mainly goe- Al2O3 15.56
thite) is of particular interest [9] and [14]. Fe2O3 11.41
It is the proportion of iron-rich phases and the iron- FeO 0.14
enrichment of the smectite in octahedral positions [7] that MgO 2.82
differentiates Czech bentonites from international stan- MnO 0.20
dards and the Ca-bentonites extracted from the majority of CaO 2.83
European deposits. Na2O 0.37
The cation exchange capacity of BCV bentonite is K2O 1.02
63.7 ± 2.2 meq/100 g and the major exchangeable cations P2O5 0.51
consist of magnesium (42.0 ± 0.6 meq/100 g) and calcium F 0.12
(15.1 ± 1.8 meq/100 g). The density of the solid particles is CO2 1.68
2.76 ± 0.02 g/cm3. The total specific surface area of BCV C 0.17
bentonite (employing EGME) is 438 ± 6 m2/g (Červinka et al. S \ 0.010
[3]). H2O(?) 9.06
The liquid limit according to the ČSN EN ISO Total 100.09
17892–12 standard [4] is 130% (according to Laufek et. al. H2O(-) 9.23
[14]) and 139% (according to Červinka et. al. [3], Haus-
mannová et. al. [9]).
This study involved the investigation of a BCV ben-
tonite powder and a uniaxially compacted powder. The
investigation of the water permeability and the total pres-
batch of the material that was subjected to experimentation
sure. The cells allowed for bottom-up saturation and the
was named BCV_2017; for more detailed information, see
continuous monitoring of total pressure at the top or both at
Laufek et al. [14].
the top and bottom, depending on the type of test. Gener-
ally, only the top total measurement approach is used for
the determination of the swelling pressure. Measurement
3 Laboratory investigation on both sides is applied only in selected tests aimed at
of the hydromechanical properties obtaining more accurate data for model calibration
purposes.
3.1 Swelling pressure (constant volume)
The top and bottom of the samples were fitted with
sintered steel permeable plates so as to prevent the leaching
The swelling pressure was measured using a constant
(‘‘mobilisation’’) of the material. The piston and the force
volume cell apparatus that allowed for the combined
sensor for the measurement of the total (or swelling)
pressure of the bentonite were positioned on the top side
Table 1 Mineralogical composition of BCV bentonite (determined between the flange of the chamber and the surface of the
using X-ray diffraction, calculated via the Rietveld method) [14] sample. A force sensor was connected to a central data
Wt% Original BCV logger.
The bentonite was uniaxially compacted into the hollow
Anatase 2.3
steel cylinder that formed the central part of the cell. Two
Quartz 11.4
laboratories were responsible for the investigation of the
Montmorillonite 69.7
swelling pressure and permeability—the Czech Technical
Mg-calcite 3.7
University (CTU) and Charles University (CU) laborato-
Goethite 3.1
ries. This allowed for the crosschecking of the results and
Hematite –
procedures.
Kaolinite 5.0 The setup used at the Charles University is shown in
Ankerite 0.6 Fig. 2. Bentonite samples of 50 mm in diameter were
Siderite 0.5 compacted directly into the cells using 36 g of bentonite
Illite 3.7 powder (hygroscopic water content of 11%). The height of

123

Content courtesy of Springer Nature, terms of use apply. Rights reserved.

3196 Acta Geotechnica (2023) 18:3193–3211

the samples varied in the range 9–13 mm depending on the dismantling, the samples were cut into slices of a thickness
required initial dry density. The samples were saturated of approximately 2 mm and the water content of each slice
with distilled water through the bottom base. The water was determined by means of the gravimetric method. This
was not pressurised during the saturation process (Fig. 2a). allowed for the determination of the water saturation dis-
The increase in the swelling pressure was monitored until a tribution in the samples. Note that the final sample (dis-
constant value was attained. Pressure and volume con- mantled at 333.78 h) was found to be fully saturated—the
trollers were connected to both bases for the measurement change in the water content along the top of the sample was
of the hydraulic conductivity (Fig. 2b), and a constant caused by the change in the dry density which, in turn, was
pressure gradient was applied. The hydraulic conductivity caused by swelling during the saturation process (which led
was measured once a constant flow rate had been attained. to density changes).
The setup used by the Czech Technical University team is
shown in Fig. 3. Bentonite samples of 30 mm in diameter and 3.2 Swelling under a constant load (oedometric
heights of 20 mm were compacted directly into the cell rings. tests)
Distilled water was used as the saturation medium, which was
propelled by compressed argon, typically at a pressure of The hydromechanical behaviour of the BCV 2017 was
1 MPa. The saturation pressure applied does not influence the investigated via the conducting of oedometric tests. A
measured values provided the saturation pressure is lower sample with a diameter of 120 mm and a height of 30 mm
than the swelling pressure of the sample [8]. The water flow was placed in a cylindrical brass cell fitted with porous
was determined manually using a graded capillary tube-based stone at the top and the bottom to allow for the saturation/
flow meter in the inflow part. Testing continued until the flow desaturation of the sample. The sample was allowed to
and the total pressure stabilised (see the example in Fig. 4). deform uniaxially (compaction/swelling) via the movement
The water flow over at least the four latest consecutive of a piston positioned on the top of the sample. The test
intervals was used for the determination of the coefficient of was pressure controlled and with free deformation. The
permeability (average of the four results). The water pressure selected uniaxial load was transferred to the sample via the
source was subsequently disconnected so as to allow for the piston, while the sample deformation was monitored via
determination of the swelling pressure (Laufek et al. [14]). the movement of the piston.
Figure 5 illustrates the dependency of the swelling This setup enabled the investigation of the stress path
pressure on the dry density as recorded by the two labo- dependence and hysteresis, as well as the influence of
ratories; the data is taken from Laufek et al. [14]. Figure 6 saturation via the alteration of the load scenario and the
shows the dependency of the hydraulic conductivity on the water regime.
dry density, again, as determined by the CTU and the CU. Three main scenarios were investigated:
Figure 7 illustrates the temporal saturation evolution for
• Initial saturation without loading (Scenario I)—com-
a dry density of approximately 1600 kg/m3. A series of
pacted samples were saturated at minimum loading (the
samples with a dry density of * 1600 kg/m3 was saturated
weight of the piston of the apparatus) then loaded
followed by dismantling at predetermined intervals. During
incrementally up to 700 kPa with free access to water,

Fig. 2 Configuration of the constant volume cells for the measurement of the swelling pressure (a) and the hydraulic conductivity (b) at the
Charles University

123

Content courtesy of Springer Nature, terms of use apply. Rights reserved.

Acta Geotechnica (2023) 18:3193–3211 3197

Fig. 3 Configuration of the constant volume cells for the measurement of the swelling pressure and the hydraulic conductivity at the Czech
Technical University

Fig. 4 Measurement of the permeability and total pressure (BCV qd = 1597 kg/m3)

123

Content courtesy of Springer Nature, terms of use apply. Rights reserved.

3198 Acta Geotechnica (2023) 18:3193–3211

Fig. 5 Swelling pressure of the BCV 2017 (Laufek et al. [14])

Fig. 6 Hydraulic conductivity of the BCV 2017

and subsequently incrementally unloaded with free clearly indicate the presence of path dependence especially
access to water. in the initial stage. However, once the samples have
• Saturation without loading after a dry cycle (Scenario become fully saturated and fully loaded, they converge to
II)—samples with natural water contents were loaded form very similar curves.
incrementally up to 700 kPa, unloaded incrementally,
saturated at minimum loading (the weight of the piston 3.3 Swelling pressure at a constant volume
of the apparatus) and loaded incrementally up to versus free swelling at constant loading
700 kPa with free access to water.
• Saturation under loading (Scenario III)—samples with The incrementally loaded oedometric test revealed the
natural water contents were loaded incrementally up to importance of the stress path on the evolution of pressure.
700 kPa, saturated at 700 kPa, unloaded incrementally Further research was subsequently conducted aimed at the
with free access to water, and loaded incrementally up investigation of the influence of the stress path on the
to 700 kPa with free access to water. swelling pressure.
A series of single-stage free swelling under constant
Figure 8 illustrates the results of Scenario I, Fig. 9 those
load tests were also conducted. Compacted samples (three
of Scenario II and Fig. 10 those of Scenario III. The results

123

Content courtesy of Springer Nature, terms of use apply. Rights reserved.

Acta Geotechnica (2023) 18:3193–3211 3199

Fig. 7 Saturation development of the BCV 2017 over time (BEACON deliverable D4.1)

initial densities were considered) were positioned in the 4 THM hypoplastic double-structure model
oedometer, loading was applied (up to 7 different loads) for bentonite
and saturation initiated. The final density was determined
once the deformation process had stabilised. The BCV dataset discussed in the previous section was
Figure 11 presents a comparison of the results of the adopted for the evaluation of a THM hypoplastic double-
constant volume cell testing (shown in red in the figure; structure model for bentonite. The model was developed by
data taken from Fig. 5) and the measurements obtained for Mašı́n [16] via the hierarchical enhancement of an earlier
free swelling under loading (shown in green in the figure). model by Mašı́n [15] that did not consider the effects of
The arrows commence at the initial dry density and end at temperature. The earlier model comprised, in turn, the
the final dry density (the blue arrows represent the relevant double-structure enhancement of earlier hypoplastic mod-
saturation phase from the oedometric tests described in the els for unsaturated and saturated soils. Interested readers
previous subchapter). are urged to refer to these references for details of the
The data clearly indicates a significant systematic dif- creation of the models. In contrast to the model version of
ference between the values of the attained swelling pres- Mašı́n [16], an additional enhancement that replaced the
sure. If deformation is allowed during saturation and original bi-linear form of the water retention curve by a
swelling, the swelling pressure is lower than the constant smooth formulation was introduced to the latest model. For
volume for the same final dry density. Mathematical reasons of conciseness, the reader is referred to Mašı́n [16]
material models need to take this factor into account. for the complete model formulation. Here, only the chan-
ges in the formulation of the model from Mašı́n [16] are

123

Content courtesy of Springer Nature, terms of use apply. Rights reserved.

3200 Acta Geotechnica (2023) 18:3193–3211

Fig. 8 Scenario I - immediate saturation at zero loading

Fig. 9 Scenario II - saturation following a dry loading cycle

123

Content courtesy of Springer Nature, terms of use apply. Rights reserved.

Acta Geotechnica (2023) 18:3193–3211 3201

Fig. 10 Scenario III - saturation at full loading (after dry loading cycle)

Fig. 11 Comparison of the swelling pressure - Urestricted swelling under constant load vs Swelling pressure at constant volume

123

Content courtesy of Springer Nature, terms of use apply. Rights reserved.

3202 Acta Geotechnica (2023) 18:3193–3211

introduced. For the sake of clarity, however, Table 3 pre-

sents the model parameters and their physical meanings.
An improvement to the WRC equation concerned the
reformulation of the macrostructural water retention model.
The basic model by Mašı́n [16], which follows on from
Mašı́n [15] and Wong and Mašı́n [24], uses the bi-linear
formulation of the water retention curve according to
Brooks and Corey [2]. The resulting water retention curve
is shown in Fig. 12. It is clear that the water retention curve

Table 3 Hypoplastic model parameters and their physical meanings

Parameter Description

uc Critical state friction angle of the macrostructure in a

standard soil-mechanics context
k Slope of the isotropic normal compression line in
lnðpM =pr Þ versus lnð1 þ eÞ space
Fig. 12 Hysteretic water retention macrostructure model adapted
j Macrostructural volume strain in pM unloading from a hypoplastic THM model for expansive soils
N Position of the isotropic normal compression line in
lnðpM =pr Þ versus lnð1 þ eÞ space is C-0 continuous only, which means that the derivatives
m Stiffness in shear oSMr =os change abruptly at the intersection of the main
ns Dependency of the position of the isotropic normal wetting/drying curves with the scanning curve and at the
compression line on suction
air entry/expulsion suction value. These discontinuities
ls Dependency of the slope of the isotropic normal
comprise by-products of a simplification adopted while
compression line on suction
defining the model equations, since real soil does indeed
nT Dependency of the position of the isotropic normal
compression line on temperature evince the non-linear dependency of SM r on suction. In
lT Dependency of the slope of the isotropic normal addition, they lead to problems in terms of the numerical
compression line on temperature performance of the model. Finally, this bi-linear formula-
m (1) Control of f u and thus the dependency of wetting-/ tion, which resembles the response of the elasto-plastic
heating-induced compaction on the distance from the constitutive model, contrasts with the general formulation
state boundary surface;
of the THM constitutive model, which is hypoplastic and,
(2) control of the double-structure coupling function and
thus, inherently non-linear.
thus the response to wetting–drying and heating–
cooling cycles (Mašı́n [15]) When redefining the water retention curve to be C-1
as Dependency of microstructural volume strains on continuous, we adopted with advantage the r k formulation
temperature proposed by Wong and Mašı́n [15]. Using this approach, it
jm Dependency of microstructural volume strains on pm was possible to define the complete bi-linear water reten-
em
r0 Reference microstructural void ratio for a reference tion curve (not only the main wetting and drying curves)
temperature T r , reference suction sr , and zero total applying:
stress (
csh Value of fm for compression M s 1kp for s  ae sen
Sr ¼ e ð1Þ
se0 Air-entry value of suction for the reference for s [ ae sen
s
macrostructural void ratio eM0
a Dependency of the macrostructural air-entry value of where ae is the parameter and the meaning of variable se is
suction on temperature clear from Fig. 12. It is calculated using:
b Dependency of the macrostructural air-entry value of
suction on temperature
se ¼ sen ðae þ ascan þ ae ascan Þ ð2Þ
ae Ratio of air entry and air expulsion values of suction for a where ascan defines the position of the current state along
macrostructure water retention model the scanning curve such that ascan ¼ 1 on the main drying
sr Reference suction for em curve and ascan ¼ 0 on the main wetting curve:
eM Reference macrostructural void ratio for the air-entry
0
value of the suction of the macrostructure
s  sW
ascan ¼ ð3Þ
Tr Reference temperature sD  sW
For the meanings of sW and sD , see Fig. 12. The rate of
the variable ascan is given by:

123

Content courtesy of Springer Nature, terms of use apply. Rights reserved.

Acta Geotechnica (2023) 18:3193–3211 3203

1  rk
a_scan ¼ s_ ð4Þ
sD ð1  ae Þ
where r k represents the ratio of the scanning curve slope (in
the lns vs. lnSr plane) and the slope of the main drying and
wetting curves. For the original model, this ratio was given
by:
8
>
> 1 for s ¼ sD and s_ [ 0
<
1 for s ¼ ae sD and s\0
_
rk ¼ k ð5Þ
>
> pscan
¼ 0:1 otherwise
:
kp

The slopes kp and kpscan are model parameters and they

are indicated in Fig. 12.
The discrete value of r k from Eq. (5) was modified in
the improved model so that the slope of the scanning curve
varied smoothly between the main drying and wetting
curves. The modification employs three parameters that are
intended to be internal (i.e. hidden from the user and hard-
coded in the implementation of the model) so that the
parameters of the complete THM model, in terms of user-
calibration, are identical for the two formulations; thus, no
recalibration is required of the already calibrated original
model for the new model. These internal parameters are
denoted pscan ¼ 3, Slim ¼ 0:75 and pwett ¼ 1:1. A new
factor f scan is used in the formulation, defined as:

ascan for s_ [ 0
f scan ¼ ð6Þ
1  ascan for s\0
_
Fig. 13 Comparison of the predicted water retention curves with the
A smoothed WRC formulation is then attained by original and modified macrostructural WRC formulations (a), the
redefining r k from Eq. (5) to effect of internal model constants on the shape of water retention
8 curves (b)
>
> 0 for s\ae sen and s_ [ 0
<  1  SM pwett
rk ¼ r
for SM
r [ Slim and s\0
_ ð7Þ 5 Simulations of laboratory experiments
>
> 1  Slim on BCV bentonite
: pscan
ðf scan Þ otherwise
A comparison of the predicted water retention curves The THM hypoplastic model was calibrated and validated
with those of the original and redefined models is provided according to the comprehensive set of experimental data on
in Fig. 13 a (note that the complete WRC is shown in terms BCV bentonite presented in Sect. 3.
of the global Sr , thus including the microstructural contri- When quantifying the degree of saturation from the
bution). The effect of the internal parameters that control experimentally measured mass water content, and when
the shape of the WRC smoothing function is shown in calculating the mass water content from the degree of
Fig. 13b. Clearly, the updated model produces a smooth saturation and the void ratio calculated by the model, a
WRC curve while maintaining the main reference wetting solid particle density of 2760 kg/m3 was adopted with a
and drying curves as asymptotic targets. Subsequently, water density of 1150 kg/m3. The latter value was deter-
Fig. 14 shows the results of the constant volume cyclic mined via the reverse-calculation of the experimental data
wetting–drying test expressed in terms of the degree of from the constant volume water retention tests, via which
saturation (in this case, the global Sr is also shown). Fig- the known mass water content, solid mass and specimen
ure 14 demonstrates that apart from acting to smooth the volume values can be used for the quantification of the
formulation of the WRC, the other properties of the model degree of saturation. With a dry density of less than
(e.g. an irreversible change in Sr during the cycles) are not 1150 kg/m3, the calculated water content exceeded the
compromised. value of 1 at low values of suction. A water density of more
than 1000 kg/m3 is consistent with Jacinto et al. [10].

123

Content courtesy of Springer Nature, terms of use apply. Rights reserved.

3204 Acta Geotechnica (2023) 18:3193–3211

Fig. 14 Evolution of the degree of saturation in the THM hypoplastic model during the cyclic water retention test. Comparison of the original bi-
linear logarithmic formulation (right) with the new smooth formulation (left)

Figure 15 (left) presents predictions of the oedometric for BCV bentonite, the experimental results for B75 ben-
loading–unloading experiments conducted on saturated tonite tested at the CTU were simulated. It was assumed
BCV bentonite performed at the CU; they demonstrate the that the unconfined compression strength of the two types
good level of representation of the loading–unloading of bentonite (extracted from the same deposit) would not
behaviour. The samples were prepared by means of the differ substantially. Cylindrical samples with aspect ratios
compaction of the bentonite powder to a dry density of of 1:1 were tested at various dry densities and an initial
1270 kg/m3, which swelled under oedometric conditions at water content of 6%, corresponding to a suction of
a constant load of 10 kPa and was subsequently subjected 100 MPa for the B75 bentonite. The experiments were
to oedometric loading–unloading. In this case, only the simulated using the BCV parameters and an initial suction
saturated loading–unloading phase was simulated (a com- value of 100 MPa. It should be stressed that while the
parison of the predictions of the swelling-loading–unload- model considers the strength parameter critical state fric-
ing oedometric tests with various path combinations is tion angle uc , the assumed effective stress for the
presented later in Fig. 23; the figure shows the void ratio macrostructure, calculated using the following Eq. (8), is
e with respect to the vertical stress rv). Although initially key to the uniaxial compressive strength calculation:
highly pre-consolidated, the samples almost immediately rM ¼ rnet  1SM ð8Þ
r s
attained normal compression lines, thus demonstrating that
the swelling process led to a decrease in the apparent pre- The unconfined strength evinced a good level of repre-
consolidation. The model provided a good representation sentation for the lower dry density sample, whereas it was
of both the loading and unloading stiffness. slightly underestimated for the higher dry density samples.
Figure 15 (right) shows the predictions of the uniaxial The reasonably good uniaxial compressive strength pre-
compression experiments conducted on unsaturated com- dictions thus indicate the suitability of the adopted effec-
pacted bentonite. In this case, since no data was available tive stress formulation.

Fig. 15 Predictions of the oedometric loading–unloading testing of saturated BCV bentonite (left) and the uniaxial loading experiments on
unsaturated compacted B75 bentonite at a water content of 6% (right)

123

Content courtesy of Springer Nature, terms of use apply. Rights reserved.

Acta Geotechnica (2023) 18:3193–3211 3205

The accurate prediction of the high uniaxial compressive demonstrated in Fig. 19, which shows the value of the state
strength can be explained with the assistance of the pre- variable relative void ratio re, where a value of 1 indicates
dictions of the water retention curves. Compacted bentonite that the normal compression line state has been attained. It
with three differing initial dry densities (1200, 1600 and is important to note that the prediction of irreversible
1900 kg/m3) was tested under constant volume (Fig. 16) microstructural deformation was a consequence of the high
and unconfined (Fig. 17) conditions. Note that the uncon- swelling properties of BCV bentonite, a factor that was not
fined WRC curves were calculated with a higher parameter so pronounced in the calibration of the model for other
N (see Table 3 for the definition thereof) value of 1.67, bentonites. However, it has not been physically substanti-
since the original value of 1.62 led to divergence. While ated and should be corrected in future versions of the
under the unconfined conditions, the model slightly model.
underpredicted the water content and the degree of satu- A further indicator considered in the model concerned
ration, the trends concerning the effect of dry density were the macrostructural effective stress. The Bishop effective
well predicted, including a lower water content at low stress equation was adopted in the model for the
suctions for the high dry density samples than for the other macrostructure with a factor v equal to the macrostructural
two tests. Less satisfactory predictions were, however, degree of saturation SM r . As already mentioned, in order to
produced by the free swelling experiments; although the predict the high uniaxial compressive strength of com-
drying curves were represented reasonably well, the water pacted bentonite, SM r must be assigned a non-negligible
content was significantly underestimated in the wetting (albeit low) value even for high suction values (see
curves. The reason for this discrepancy is as follows. Since Fig. 18). This non-negligible effective stress contributes to
the macrostructural water content was relatively low the irreversibility of the microstructural void ratio (the
(Fig. 18), the behaviour was predominantly affected by enhanced effective stress shifts the state towards the nor-
that of the microstructure and, since the microstructure was mal compression line, which subsequently limits the
considered to be fully saturated, the water retention curves microstructural swelling) and, moreover, limits the swel-
were primarily affected by the microstructural void ratio. ling potential, which is predicted reasonably well by the
Since the microstructural response was considered to be model (Fig. 20). It should be noted at this point that the
reversible in the model, the water retention curves should experimental evidence indicated a dry macrostructure for
have exhibited a low degree of hysteresis, i.e. from the higher suction values, which was clear from both the
macrostructural part only (Fig. 18). However, since BCV is ESEM images [18] and from the fact that the water content
a highly swelling bentonite, the slope of the microstructural of bentonite at higher suctions is known to be independent
swelling line is in fact close to that of the normal com- of the dry density (e.g. [17]. Thus, relating the high
pression line. Moreover, since the macrostructural effective strength of compacted bentonite with the macrostructural
stress is non-negligible so as to ensure the correct predic- degree of saturation SM r (a concept that originated from the
tion of the uniaxial compression strength (Fig. 15), the idea that the matric suction strengthens the inter-aggregate
state approaches the normal compression line during the contacts through water menisci) appears to be a simplifi-
wetting process (which is considered to be the boundary of cation of reality, since other physico-chemical phenomena
the microstructural swelling in the model), thus breaching also play a role at the inter-aggregate contact points. These
the reversibility of the microstructural deformation. This is

Fig. 16 Predictions of the water retention tests for compacted BCV bentonite under constant volume conditions at three differing initial dry
densities of 1200, 1600 and 1900 kg/m3

123

Content courtesy of Springer Nature, terms of use apply. Rights reserved.

3206 Acta Geotechnica (2023) 18:3193–3211

Fig. 17 Predictions of the water retention tests for compacted BCV bentonite under free swelling conditions at three differing initial dry densities
of 1200, 1600 and 1900 kg/m3 (calculated applying N = 1.67)

Fig. 18 Predictions of the macrostructural quantities (macrostructural water content and macrostructural degree of saturation) for the water
retention testing of compacted BCV bentonite under free swelling conditions at three differing initial dry densities of 1200, 1600 and 1900 kg/m3
(calculated applying N = 1.67)

the results of both types of experiment plotted on the same

graph. The model, in line with the experimental data,
predicted constant load swelling test results that were
consistently lower in the pressure versus dry density graph
than those of the constant volume swelling pressure tests.
The reason for the predictive capability is demonstrated
in Fig. 22 for selected tests as marked with a yellow circle
in Fig. 21. Figure 22 (left) illustrates the evolution of the
radial and axial stresses. They are obviously identical for
the swelling pressure tests since the inherent anisotropy
was not considered in the model. Conversely, with respect
to the swelling under constant load test, the radial stresses
developed substantially above the axial stress, which
Fig. 19 Predictions of the relative void ratio for the water retention remained constant, defined by the boundary conditions of
testing of compacted BCV bentonite under free swelling conditions at
three differing initial dry densities of 1200, 1600 and 1900 kg/m3
the experiment. Consequently, the swelling under constant
(calculated applying N = 1.67) load test proceeded at a high stress obliquity, i.e. the state
boundary surface was reached much earlier than in the
model components should, therefore, be carefully recon- swelling pressure tests (see Fig. 22 (right), which graphi-
sidered in future versions of the model. cally illustrates the value of the mean stress normalised by
The model swelling characteristics were very well pre- the mean stress at the state boundary surface for a given
dicted (Fig. 20), as is also clear from Fig. 21, which shows stress ratio). As discussed above, the swelling process is

123

Content courtesy of Springer Nature, terms of use apply. Rights reserved.

Acta Geotechnica (2023) 18:3193–3211 3207

Fig. 20 Predictions of the swelling pressure tests (left) and the oedometric swelling under constant load experiments at three differing initial dry
densities (right) of the unsaturated compacted BCV bentonite

plotted rather than the vertical stress; nevertheless, this

would still not have been sufficient to yield identical results
for the two types of test.
Figure 23 illustrates a particularly interesting experi-
mental data set for a model assessment conducted at the
CTU involving the performance of oedometric swelling-
loading–unloading tests applying various path combina-
tions. The graphs indicate a trend line that runs through the
experimentally obtained swelling pressure test results. It is
clear that this trend line indicates the limit for the loading
parts of the tests (thus, it coincides with the normal com-
pression line), which is reasonably well predicted by the
Fig. 21 Comparison of the predictions of the swelling pressure tests models even though the limit stress is somewhat overesti-
and oedometric swelling under constant load experiments for mated. The model also predicted other important phe-
unsaturated compacted BCV bentonite according to the model and
nomena such as the high swelling potential of the sample at
the experiment
low load compared to the much lower swelling potential of
limited once the state boundary surface is reached, which the loaded sample. Moreover, the model relatively accu-
explains the observed differences between the swelling rately predicted the significantly higher loading stiffness of
pressure and the swelling under constant load results. the unsaturated sample compared to the saturated sample.
Further, the swelling pressures in the graphs in Fig. 21 In contrast to the experiment, the model predicted high
would be higher if the radial stress or mean stress were swelling during the final unloading phase. In addition, it
failed to distinguish between the swelling potential of the

Fig. 22 Comparison of the predictions of the selected swelling pressure tests and the oedometric swelling experiments for the constant load test.
The tests selected are indicated by yellow circles in Fig. 21

123

Content courtesy of Springer Nature, terms of use apply. Rights reserved.

3208 Acta Geotechnica (2023) 18:3193–3211

Fig. 23 Swelling-loading–unloading oedometric tests for various path combinations

pre-loaded and unloaded samples when compared directly

with the swelled sample. These experimental results were
unexpected and more tests will be needed before the results
can be confirmed.
The final tests to be predicted involved the swelling
pressure testing of samples with three differing dry densi-
ties; the samples were 20 mm in height and were subjected
to an elevated bottom pore water pressure of approx.
1 MPa with free drainage at the top. The tests were
accompanied by the monitoring of the evolution of the pore
water pressure over time as the water infiltrated into the
samples. In addition, the experiment conducted at a dry
density of 1600 kg/m3 was repeated several times; the Fig. 24 Experimental and predicted evolution of the swelling pressure
samples were dismantled after various time periods and the in swelling pressure tests conducted with elevated bottom water
distribution of the water content was investigated across pressures of approx. 1 MPa
the heights of the samples. In order to be able to predict this
test, it was necessary to simulate the experiments as
boundary value problems applying the finite element
method. The models were processed in the CTU’s SIFEL
finite element code [12]. The wall friction was not con-
sidered in the simulations.
The time evolution of the swelling pressure is shown in
Fig. 24, which also shows the swelling pressure values as
obtained using element test simulations (that is, simulations
that assume constant state variable fields inside the sample,
including constant suction and degree of saturation) of the
swelling pressure tests. Consistent with Fig. 20, the swel-
ling pressures were very well predicted. However, the
model overpredicted the initial rate of the increase in the
swelling pressure, which may have been due to having Fig. 25 Water content distribution across a sample for several time
adopted a constant suction independent permeability. intervals following the commencement of the test
The evolution of the water content in the sample at
several time intervals following the commencement of the It is interesting to note that the water content was not
test is shown in Fig. 25 for the experiment with a dry homogeneous across the sample at the 334 h time point, at
density of 1597 kg/m3 (the experimental data was obtained which time the sample was practically fully saturated. In
for several samples with a target dry density of 1600 kg/ addition, Fig. 24 illustrates that the swelling pressures
m3). Although the water content was overpredicted at the predicted by the finite element model are in fact lower than
start of the test, the trends concerning the water content those predicted by the single element program (note that
distribution and evolution were predicted correctly. the bottom water pressure adds to the swelling pressure, the

123

Content courtesy of Springer Nature, terms of use apply. Rights reserved.

Acta Geotechnica (2023) 18:3193–3211 3209

swelling pressure itself is thus directly measured only if 6 Conclusion

water pressure is reduced to zero, which happened in
1306 kg/m3 and 1597 kg/m3 samples just before the end of This article provides a complex H-M characterisation of
the test). Both these observations can be explained with the BCV bentonite, which comprises the reference research
aid of Fig. 26 which shows the void ratio and the degree of material in the ongoing development of the Czech DGR
saturation distribution for the simulations from Fig. 25. It concept. A complex data set was created for mathematical
is clear that the sample had become fully saturated by the models of divalent bentonites. The BCV data set was
334 h time point; however, the void ratio is not uniform created using the same batch of material, which is often not
across the sample, which implies the non-uniform distri- the case for bentonite datasets provided in the literature.
bution of the water content. The non-uniform distribution The laboratory results showed that the hydromechanical
of the void ratio, which originated at the beginning of the behaviour of BCV bentonite is highly dependent on the
test and did not change substantially thereafter (all the void stress path taken. The state at which water is introduced is
ratio curves in Fig. 26 practically coincide), implies that also very important in this respect. These factors indicate
the sample was not homogeneous. This indicates that that the micro- and macrostructures of bentonite evolve on
adopting the single element approach to the swelling a continuous basis. This exerts a significant impact on the
pressure calculation and model calibration, which has been determination of the material properties considered in, and
adopted by most researchers in the field, is in fact not the development of mathematical models of the behaviour
exactly correct. of bentonite.
Aimed at the investigation of whether the observation It was demonstrated that the swelling pressures recorded
described above of a non-uniform state within the samples applying the constant volume test attained higher values
tested for swelling pressure was caused by the elevated than the results of the oedometric swelling test at a constant
bottom water pressure of approx. 1 MPa, the simulations load with deformation allowed. This was represented well
were repeated with a low water pressure of 1 kPa, the by the constitutive model, which allowed for the interpre-
results of which are shown in Fig. 27. The lower saturation tation of this observation. It was demonstrated that the
pressure resulted in a slower increase in the swelling main reason for the discrepancy comprised the radial
pressure (Fig. 27 left), while the final swelling pressure stresses that developed during swelling in the constant load
was still lower than that calculated applying the element test under oedometric conditions, which differed from the
test (constant fields of state variables) assumption. The stresses that developed during the constant volume swel-
distribution of the void ratio (Fig. 27 right) was practically ling pressure test. Moreover, due to the comprehensive
identical for the bottom water pressure. change in the void ratio, which is possible in the swelling
A summary of the model parameters adopted for the under constant load test, the relative proportion of
BCV simulations in this section is provided in Table 4. The microstructural and macrostructural void ratios differs
simulations were performed with dry density dependent between the swelling pressure and swelling under constant
intrinsic permeabilities of 3 9 10–20 m2 for qd = 1306 kg/ load tests. Consequently, the results of the two tests do not
m3, 1 9 10–20 m2 for qd = 1508 kg/m3 and 8 9 10–21 m2 follow an identical line when expressed in the form of a
for qd = 1597 kg/m3. final dry density versus vertical stress graph. It was also
demonstrated that the swelling pressure tests, although
typically considered to be element tests (with a constant
field of variables across the sample) when employed by
researchers for the calibration of constitutive models, do
not in fact constitute element tests. Due to the gradual
progression of the wetting front from the wetting source,
the void ratio is redistributed across the sample in a non-
uniform manner (although the average sample void ratio is,
from the test definition, constant). Consequently, the
swelling pressures determined via the simulation of the test
that applied the finite element method were lower than the
swelling pressure predicted through the assumption of
constant values of variables across the sample.
In summary, the experimental results indicate that BCV
bentonite evinces all the hydro-mechanical characteristics
Fig. 26 The void ratio and degree of saturation distribution across a that enable its use as a reference research material in the
sample for several time intervals following the start of the test

123

Content courtesy of Springer Nature, terms of use apply. Rights reserved.

3210 Acta Geotechnica (2023) 18:3193–3211

Fig. 27 Swelling pressure experiment at a dry density of 1597 kg/m3 from Fig. 24 compared with an identical experiment with a bottom
saturation pressure of 1 kPa

Table 4 Values of the parameters of the THM hypoplastic model for 745942) and research Project No. TK01010063 awarded by the
BCV bentonite Technology Agency of the Czech Republic. The second author is also
grateful for the institutional support provided by the Centre for
Parameter Unit Value Geosphere Dynamics (UNCE/SCI/006).

uc 25

k – 0.12 Open Access This article is licensed under a Creative Commons
j – 0.02 Attribution 4.0 International License, which permits use, sharing,
N – 1.62 adaptation, distribution and reproduction in any medium or format, as
long as you give appropriate credit to the original author(s) and the
t – 0.24
source, provide a link to the Creative Commons licence, and indicate
ns – 0.01 if changes were made. The images or other third party material in this
ls – 0.0 article are included in the article’s Creative Commons licence, unless
nT – -0.07 indicated otherwise in a credit line to the material. If material is not
included in the article’s Creative Commons licence and your intended
lT – 0 use is not permitted by statutory regulation or exceeds the permitted
m – 10 use, you will need to obtain permission directly from the copyright
as 1=K 0.00015 holder. To view a copy of this licence, visit http://creativecommons.
org/licenses/by/4.0/.
jm – 0.1
sr kPa -1000
em – 1.0
r0 References
csh – 0.1
se0 kPa -2,700 1. Baryla et al. (2019) BEACON Deliverable D4.1/4.2 Bentonite
eM – 0.50 mechanical evolution – experimental work for the support of
0
model development and validation
Tr K 294
2. Brooks R, Corey A (1964) Hydraulic properties of porous media.
a N=m 0.118 Hydrology paper no. 3. Colorado State University
b N=ðmK Þ -0.000154 3. Červinka R, Vašı́ček R, Večernı́k P, Kašpar V (2018) Kompletnı́
ae – 0.75 charakterizace bentonitu BCV 2017. Technical Report 419/2019.
- SÚRAO, Prague
kp0 – 1.2 4. ČSN EN ISO 17892–12 (2018) Geotechnical investigation and
testing - laboratory testing of soil. Determination of liquid and
plastic limits. Český normalizačnı́ institut, Praha, 40 p
5. Cui Y-J (2017) On the hydro-mechanical behaviour of MX80
bentonite-based materials. J Rock Mech Geotech Eng 9:565–574
development of the Czech deep geological repository 6. Dixon DA (2019) Review of the T-H-M-C Properties of MX-80
concept. These characteristics were well represented by the Bentonite, NWMO-TR-2019–07
7. Franče J (1992) Bentonity ve východnı́ části Doupovských hor.
adopted hypoplastic constitutive model, thus proving its
Sbornı́k geologických věd 30:43–90
suitability for the numerical modelling required for the 8. Hausmannová L (2017) Vliv sytı́cı́ho tlaku na hydraulickou
development of the Czech DGR concept. vodivost a bobtnacı́ tlak českých bentonitů, Disertation thesis,
CTU in Prague
Acknowledgements The team is grateful for the financial support 9. Hausmannová L, Hanusová I, Dohnálková M (2018) Summary of
provided by both the BEACON research project (Euratom research the research of Czech bentonites for use in the deep geological
and training programme 2014-2018 under Grant Agreement No. repository – up to 2018. Technical report SÚRAO 309/2018/ENG

123

Content courtesy of Springer Nature, terms of use apply. Rights reserved.

Acta Geotechnica (2023) 18:3193–3211 3211

10. Jacinto AC, Villar MV, Ledesma A (2012) Influence of water 19. Svoboda J, Vašı́ček R, Smutek J, Hausmannová L, Franěk J,
density on the water-retention curve of expansive clays. Rukavičková L, Večernı́k P (2019) Interaction experiment at the
Géotechnique 62(8):657–667 Bukov URF. Conference Proc. Underground construction Prague
11. Kober F, Vomvoris S, Finsterle S (2018) HotBENT – High 20. Thatcher KE, Newson RK, Watson SP, Norros S et al. (2017)
Temperature Buffer Experiment: Preliminary design study – Deliverable D2.2, Review of data and models on the mechanical
overview of results. Arbeitsbericht NAB17–29, NAGRA properties of bentoniteavailable at the start of Beacon, Beacon -
12. Koudelka T, Krejčı́ T, Kruis J (2018) Time integration of Bentonite Mechanical Evolution, EC grant agreement No.
hypoplastic model for expansive clays (2018). In: Proceedings of 745942
international conference on numerical analysis and applied 21. Villar MV (2002) Thermo-hydro-mechanical characterisation of
mathematics 2017, New York: AIP conference proceedings, a bentonite from Cabo de Gata. A study applied to the use of
2018. p. 1–4. vol. 1978. ISSN 0094–243X. ISBN 978–0– bentonite as sealing material in high level radioactive waste
7354–1690–1. DOI: https://doi.org/10.1063/1.5043736 repositories. Publicación Técnica ENRESA 01/2002. Madrid:
13. Laloui L, Ferrari A, Bosch JA (2020) Bentonite clay barriers in Empresa Nacional de Residuos Radioactivos
nuclear waste repositories, E3S Web Conf. 205 01003, DOI: 22. Villar MV, Armand G, Conil N, Lesquen C, Herold P, Simo E,
https://doi.org/10.1051/e3sconf/202020501003 Mayor JC, Dizier A, Li X, Chen G, Leupin O, Niskanen M,
14. Laufek F, Hanusová I, Svoboda J, Vašı́ček R, Najser J, Koubová Bailey M, Thompson S, Svensson D, Sellin P, Hausmannova L
M, Čurda M, Pticen F, Vaculı́ková L (2021) Mineralogical, (2021) Initial State-Of-The-Art on THM Behaviour of I) Buffer
geochemical and geotechnical study of BCV bentonite – initial Clay Materials and of II) Host Clay Materials, project EURAD,
state and state after thermal treatment at 200 C. Minerals. https:// EC grant agreement No. 847593
doi.org/10.3390/min11080871 23. Wigger C, Hanusová I, Hausmannová L, Heino L, Lavikainen V,
15. Mašı́n D (2013) Double structure hydromechanical coupling Leupin OX, Marshall P, Mayor JC, Meleshyn A, Pusch R, Sellin
formalism and a model for unsaturated expansive clays. Eng Geol P, Swahn J, Talandier J, Wendling J, Wieczorek K (2017)
165:73–88 Deliverable D1.1, State-of-the-Art Report, Beacon - Bentonite
16. Mašı́n D (2017) Coupled thermohydromechanical double struc- Mechanical Evolution, EC grant agreement No. 745942
ture model for expansive soils. ASCE J Eng Mech. https://doi. 24. Wong KS, Mašı́n D (2014) Coupled hydro-mechanical model for
org/10.1061/(ASCE)EM.1943-7889.0001278 partially saturated soils predicting small strain stiffness. Comput
17. Romero E, Della Vecchia G, Jommi C (2011) An insight into the Geotech 61:355–369
water retention properties of compacted clayey soils. Geotech-
nique 61(4):313–328 Publisher’s Note Springer Nature remains neutral with regard to
18. Sun H, Mašı́n D, Najser J, Neděla V, Navrátilová E (2019) jurisdictional claims in published maps and institutional affiliations.
Bentonite microstructure and saturation evolution in wetting-
drying cycles evaluated using ESEM MIP and WRC measure-
ments. Géotechnique 69(8):713–726

123

Content courtesy of Springer Nature, terms of use apply. Rights reserved.

Terms and Conditions
Springer Nature journal content, brought to you courtesy of Springer Nature Customer Service Center GmbH (“Springer Nature”).
Springer Nature supports a reasonable amount of sharing of research papers by authors, subscribers and authorised users (“Users”), for small-
scale personal, non-commercial use provided that all copyright, trade and service marks and other proprietary notices are maintained. By
accessing, sharing, receiving or otherwise using the Springer Nature journal content you agree to these terms of use (“Terms”). For these
purposes, Springer Nature considers academic use (by researchers and students) to be non-commercial.
These Terms are supplementary and will apply in addition to any applicable website terms and conditions, a relevant site licence or a personal
subscription. These Terms will prevail over any conflict or ambiguity with regards to the relevant terms, a site licence or a personal subscription
(to the extent of the conflict or ambiguity only). For Creative Commons-licensed articles, the terms of the Creative Commons license used will
apply.
We collect and use personal data to provide access to the Springer Nature journal content. We may also use these personal data internally within
ResearchGate and Springer Nature and as agreed share it, in an anonymised way, for purposes of tracking, analysis and reporting. We will not
otherwise disclose your personal data outside the ResearchGate or the Springer Nature group of companies unless we have your permission as
detailed in the Privacy Policy.
While Users may use the Springer Nature journal content for small scale, personal non-commercial use, it is important to note that Users may
not:

1. use such content for the purpose of providing other users with access on a regular or large scale basis or as a means to circumvent access
control;
2. use such content where to do so would be considered a criminal or statutory offence in any jurisdiction, or gives rise to civil liability, or is
otherwise unlawful;
3. falsely or misleadingly imply or suggest endorsement, approval , sponsorship, or association unless explicitly agreed to by Springer Nature in
writing;
4. use bots or other automated methods to access the content or redirect messages
5. override any security feature or exclusionary protocol; or
6. share the content in order to create substitute for Springer Nature products or services or a systematic database of Springer Nature journal
content.
In line with the restriction against commercial use, Springer Nature does not permit the creation of a product or service that creates revenue,
royalties, rent or income from our content or its inclusion as part of a paid for service or for other commercial gain. Springer Nature journal
content cannot be used for inter-library loans and librarians may not upload Springer Nature journal content on a large scale into their, or any
other, institutional repository.
These terms of use are reviewed regularly and may be amended at any time. Springer Nature is not obligated to publish any information or
content on this website and may remove it or features or functionality at our sole discretion, at any time with or without notice. Springer Nature
may revoke this licence to you at any time and remove access to any copies of the Springer Nature journal content which have been saved.
To the fullest extent permitted by law, Springer Nature makes no warranties, representations or guarantees to Users, either express or implied
with respect to the Springer nature journal content and all parties disclaim and waive any implied warranties or warranties imposed by law,
including merchantability or fitness for any particular purpose.
Please note that these rights do not automatically extend to content, data or other material published by Springer Nature that may be licensed
from third parties.
If you would like to use or distribute our Springer Nature journal content to a wider audience or on a regular basis or in any other manner not
expressly permitted by these Terms, please contact Springer Nature at

[email protected]