Journal of Physics: Conference

Series

PAPER • OPEN ACCESS You may also like

- Effect of laser power on microstructure
Biaxial tension test of soft materials and properties of Ti-Al/WC composite
coatings
Hongwei Liu, Baoquan Dong, Xun Gong et
To cite this article: Minxing Li et al 2024 J. Phys.: Conf. Ser. 2808 012040 al.

- Experimental study on the flow capacity of

shale fractures with different degrees of
sand filling
Ting Chen, Chunlei Yu, Zhigang Sun et al.
View the article online for updates and enhancements.
- Preparation and performance verification
of calcium silicate composite electrical
fireproof materials
Yin Wang, Yu Wang, Dedong Yang et al.

This content was downloaded from IP address 180.149.26.122 on 02/08/2024 at 15:09

2024 3rd International Conference on Materials Engineering and Applied Mechanics IOP Publishing
Journal of Physics: Conference Series 2808 (2024) 012040 doi:10.1088/1742-6596/2808/1/012040

Biaxial tension test of soft materials

Minxing Li1, Ming Yang2, Huiming Wang1, 3 and Ronghao Bao1, 3, 4*

1
Department of Engineering Mechanics, Zhejiang University, Hangzhou, Zhejiang,
310027, China
2
Hangzhou Intellisense Technology Co., Ltd, Hangzhou, Zhejiang, 311107, China
3
Key Laboratory of Soft Machines and Smart Devices of Zhejiang Province, Hangzhou,
Zhejiang, 310027, China
4
Center for Soft Machines and Smart Devices, Huanjiang Laboratory, Zhuji 311816,
China
*
Corresponding author’s e-mail: [email protected]

Abstract. A specific loading machine is designed and fabricated to test the constitutive
properties of soft materials by biaxial tension in this study. The finite element model under
ABAQUS is launched to simulate the biaxial tension of a squared specimen, and it is verified by
a theoretical model. Using the finite elements model, the effects of clamping conditions of the
squared samples, including the clamping width, clamping depth, and clamping numbers, are
systematically studied, which helps improve future machine design. The comparison of
numerical results with experimental results is conducted by equibiaxial test and pure shear test.
Finally, a novel testing procedure is suggested with the combination of numerical simulation and
practical experiments to enhance the testing result.

1. Introduction
Due to their inherent flexibility, capacity for large deformation, and quick response to external stimuli
like mechanical, electrical, and magnetic fields, soft materials are now widely utilized to fabricate
flexible electronic devices, such as flexible sensors, actuators, and screens [1-2]. In the forthcoming
intelligence era, soft materials are anticipated to play more important roles in various areas, such as
medicine, healthcare, human-computer interaction, and energy and environment [3-4].
To fully realize the aforementioned functional designs in practical applications, it is very crucial to
understand the basic mechanical properties of such materials [5]. However, unlike traditional hard
materials, soft materials exhibit hyperelasticity and undergo significant deformation in most applications,
where the relation between strain and stress is complicated. In the realms of scientific research and
industrial design, hyperelastic constitutive models are frequently employed to characterize the
mechanical properties of soft materials, specifically applied in the analysis of electromechanical systems
in the design of flexible electronic devices. Additionally, they are instrumental in analyzing the
deformation modes of devices under external loads in operational environments.
Uniaxial tensile experiments alone are insufficient to adequately determine the hyperelastic
constitutive model of soft materials. More kinds of tests should be supplemented simultaneously to
acquire the proper constitutive model, such as the uniaxial compressive test, equibiaxial tension test,
pure shear test, and so on [6]. Currently, the equibiaxial tension test on planar specimens (such as
cruciform specimens, squared specimens, and circular specimens) is the most feasible way to be

Content from this work may be used under the terms of the Creative Commons Attribution 4.0 licence. Any further distribution
of this work must maintain attribution to the author(s) and the title of the work, journal citation and DOI.
Published under licence by IOP Publishing Ltd 1
2024 3rd International Conference on Materials Engineering and Applied Mechanics IOP Publishing
Journal of Physics: Conference Series 2808 (2024) 012040 doi:10.1088/1742-6596/2808/1/012040

conducted in mechanical laboratories, owing to its simplicity of tensile machine. For cruciform
specimens, the efficient testing area (denoted as a nearly uniform distributed stress or strain state in this
paper) is very limited during the test, even the geometry of the specimens has been optimized by
numerous researchers [7-8], which will consequently affect the accuracy of final material parameters.
Regarding to cruciform specimen, it can maximally achieve a uniformly distributed equibiaxial strain
region of no more than 30% [9] due to the size restriction of the testing machine, which is still inadequate
to fully capture the mechanical properties. Avanzini et al. [10] conducted biaxial tests using natural rubber
with various geometric shapes and demonstrated that squared specimens could attain a wider efficient
testing area in comparison with cruciform specimens, which means the more accurate constitutive model
parameters. However, limited work has been done in the mechanical testing on squared specimens [11],
so there remains a deficiency of commercial apparatus available for researchers to conduct such
experiments in laboratories.
Regarding the biaxial testing on squared specimens, the working frames and clamping system are
very crucial to the result, which should enable the clamps to move freely along the perpendicular sides
to the load direction, resulting in the free expansion of specimens to achieve a wide efficient testing area
[12]
. The finite element method (FEM) is utilized to investigate the optimal procedures for conducting
biaxial tensile tests on squared specimens. Fujikawa et al. [13] proposed to introduce the appropriate
notches near the fixture area to obtain accurate nominal stress-strain relation through experiment.
However, the effective cross-sectional area to calculate the nominal stress is not correctly suggested in
their work, leading to a significant deviation from theoretical expectations for conventional specimens.
Luo et al. [14] concentrated on the comparative analysis of two clamping methods: two-corner fixed
stretching and single-corner fixed stretching by FEM, showing that the latter could improve the overall
stress distribution effectively. However, the displacement boundary conditions in their model were
imposed on the distributed lines rather than the clamps, neglecting the effects of the clamping system,
which is very crucial to such experiments.
In this work, customized testing equipment is designed and fabricated to apply biaxial tension test
on squared planar soft material specimens or standard dog bone specimens to acquire an accurate
constitutive model with various tension conditions. A finite element model is built to simulate the biaxial
tension test as well, which is verified by the theoretical model. Furthermore, the effects of the clamping
system on the testing result are systematically investigated by the numerical model. The experimental
results and numerical analysis both validate that the proposed experimental approach could yield highly
accurate testing results. Finally, a calibrating procedure is suggested for practical experiments with our
testing apparatus.

2. Experimental testing methods and corresponding finite element models

2.1. Experimental test methods

2.1.1 Biaxial tensile machine. To achieve better testing results, a biaxial tensile machine on soft
materials, as shown in Figure 1, is developed in our laboratory at Zhejiang University with the
cooperation of Hangzhou IntelliSense technology company (Model: ZY-BiFSM). Squared specimens
(or standard dog-bone specimens) can be attached to the platform of the machine by four rows of clamps
as illustrated in Figure 1(a), where the specimens can be stretched through the travel mechanism driven
by four independent servo motors as shown in Figure 1(b), with an arbitrary tension ratio between the
two loading directions. Owing to the specific design, the clamps attached to the four guide rails can be
freely moved during the test, resulting in sufficient deformation of the whole specimen. As can be seen
from Figure 1(a), even grid lines were marked on the surface of the specimen before the test. A wide
effective deformation area could be obtained within our test since the grids progressed uniformly during
the test. A series of load cells with appropriate loading capacity is installed between the clamps and
sliding blocks on the guide rails to measure the loading force by our own developed data acquisition
system as demonstrated in Figure 1(c). The mechanism is driven by the control algorithm embedded in

2
2024 3rd International Conference on Materials Engineering and Applied Mechanics IOP Publishing
Journal of Physics: Conference Series 2808 (2024) 012040 doi:10.1088/1742-6596/2808/1/012040

the corresponding software, and all the loading and deformation data are recorded for further data
analysis.

Figure 1. (a) Experimental setup on squared specimens; (b) Schematic diagram of the main structure
of the biaxial tensile machine; (c) Details of the load cells and fixture clamping system.

2.1.2 Experimental procedure and experimental data processing. As mentioned above, when the grids
were marked uniformly on the surface of the specimen, the grids around the central region were found
to move outward evenly during the tension, which is sufficient to assume that the central part remained
uniform strain state during the test. To acquire the accurate stretch ratio, four dark points are marked on
the central area (30 mm  30 mm) of the specimen (220 mm  220 mm), as shown in Figure 2. A digital
camera with high resolution is installed on the overhead center of the specimen to capture the traces of
marked points. The distortion of the image is calibrated before the test. The digital images are recorded
and utilized to calculate the stretch ratio at each frame. Then the nominal strain can be obtained
according to the stretch ratio. Regarding the nominal stress, it can be calculated as S  Fi （
/ L0 ×d0）,
i
where L0 represents the original length of the sample (220 mm), and d0 denotes the initial thickness of
the sample (1.0 mm) before the test, respectively. Here the summation means the overall of the loads
along each side, and the load is averaged within four sides.

Figure 2. (a) Image of the four dark points before the test; (b) Image of the four dark points when the
recording time is 300 s, taken from the digital camera.

Figure 3. (a) Schematics of finite element model in ABAQUS with a quarter of specimen. (b) Contour
plot of strain distribution along y-direction by FEM.

3
2024 3rd International Conference on Materials Engineering and Applied Mechanics IOP Publishing
Journal of Physics: Conference Series 2808 (2024) 012040 doi:10.1088/1742-6596/2808/1/012040

2.2. Finite element model

Based on the experimental setup, a quarter of the specimen is modeled using ABAQUS for further
parameter analysis as depicted in Figure 3(a), owing to the symmetry of the test. Appropriate symmetric
boundary conditions are applied. The shell element, S4R, is adopted in the finite element model with
sufficient mesh elements. The displacements of the clamps are applied as boundary conditions while the
reaction forces are taken as the applied loads. The primary key parameters include W (the clamp width),
H (the clamping depth), WO (the distance between the corner edge and the neighboring clamp), and N
(the number of clamps at each side), respectively. With our primary analysis, the length (or width) L of
the FEM is selected as 90 mm (which is a little different from the testing specimen), and the thickness
is taken as 0.5 mm. N is in the range of (4,8), W is in the range of (7-17 mm), H is in the range of (4-10
mm), and WO is in the range of (10-28 mm), respectively. Since the specimen is fixed by the clamps
tightly enough, the clamping areas are modeled as rigid bodies for simplicity. A classic Ogden model
with 3rd order is adopted in FEM, with strain energy density function (SEDF) as
3

W(1 , 2 , 3 )   i  1  2  3  3
i i i (1)
i 1  i

where the parameters are [15]:

1  0.618MPa,1  1.3, 2  1.18 103 MPa, 2  5.03, 3  9.81103 MPa, 3  2.0 .
Under equibiaxial loading conditions, the theoretical formula of nominal stress can be expressed as
[13]
:
3


S ( )   i  i 1   2i 1  (2)
i 1

where λ is the stretching ratio of either side during the test.

Within the FEM simulation, the stretch ratio and nominal stress are calculated as follows:
u (t )  a0 (3)
 (t )  ，S (t )  RFtotal (t ) （
/ L0 ×d0）
a0
where a0 is the original size of the observation area with nearly uniform deformation, taken as 40 mm;
𝑢(t) is the average displacement along the edge of the observation area (red line as shown in Figure 3b)
at each frame. RFtotal(t) is the sum of the reaction forces on each side. The original length L0 and thickness
d0 in the finite element model before tension are 90 mm and 0.5 mm, respectively. With the FEM result,
the relation between nominal stress and nominal strain (stretch ratio minus 1) can be obtained and will
be compared with the theoretical one presented by Equation (2). Systematic analysis of the parameters
on the results will be further conducted.

3. Results and discussion

3.1. Finite element result analysis

3.1.1. Effects of key parameters. Figure 4(a-d) illustrates the influences of WO, W, H, and N on the
nominal stress-nominal strain relation obtained by Equation (3), respectively. All the results show that
the FEM results fit very well with the theoretical one when the nominal strain is less than 1.4,
demonstrating the validity of the FEM model of the equibiaxial test on squared specimens. It can also
be seen from the figures that WO and H exhibit relatively limited effects on the nominal stress-nominal
strain curve, while W and N have notable influences when the nominal strain is greater than 1.5. It can
be inferred that the nominal stress-strain curve is mainly related to W and N. An increase (decrease) in
N within the studying range elevates (lowers) the nominal stress-nominal strain curve, and then W needs
to be reduced within the studying range to align well with the theoretical one. Regarding experiments,
W and N need to be optimized for good results, which corresponds to an optimal specimen size for test
preparation.

4
2024 3rd International Conference on Materials Engineering and Applied Mechanics IOP Publishing
Journal of Physics: Conference Series 2808 (2024) 012040 doi:10.1088/1742-6596/2808/1/012040

Figure 4. The influence of geometric parameters on the nominal stress-strain curve, WO(a), W(b),
H(c), and N(d).

3.1.2. Experimental methods based on finite element results. Based on the discussion presented in 3.1.1,
when the clamping number N in the experiment is fixed as 5, W will emerge as the primary factor for
the result. The constitutive model of soft material is not available before the test. Therefore, the optimal
W (corresponding to specimen size) could not be rightly determined by FEM. To achieve a better
constitutive model by experiment, it is significant to select an appropriate specimen size considering
that W is almost fixed when the testing apparatus is finished. Therefore, we propose a novel
experimental testing procedure with the cooperation of FEM simulation, which is schematically
illustrated in Figure 5. When the theoretical curves in 3.1.1 were replaced by those obtained through
corresponding experiments, the optimal specimen size can be gradually improved to make the FEM one
fit well with the experimental one, so that the constitutive model obtained by the experiment can be used
by another simulation model for further analysis.

Figure 5. Proposed testing procedure with the aid of FEM.

5
2024 3rd International Conference on Materials Engineering and Applied Mechanics IOP Publishing
Journal of Physics: Conference Series 2808 (2024) 012040 doi:10.1088/1742-6596/2808/1/012040

3.2. Experimental results

3.2.1. Equibiaxial tensile results. Renowned for their remarkable flexibility and adaptability, silica gels
are leading materials in contemporary scientific research across diverse fields. Particularly in flexible
electronics, silica gels stand out as a pivotal material, offering unmatched versatility in crafting flexible
electronic devices. Its distinctive properties render it ideal for a wide array of applications, spanning
from wearable sensors to stretchable circuits. Hence, this study opts for silica gel as a prime example of
functional soft materials for conducting mechanical property testing experiments. Squared specimens
made of FC2211 silica gel (size 220 mm × 220 mm × 1 mm) were prepared for the equibiaxial test by
our testing machine presented in 2.1.1. N, W, H, and WO were set as 5 mm, 8 mm, 8 mm, and 30 mm,
respectively. The maximal tensile was 180 mm, performed evenly within 300 s, resulting in a stretch
ratio of 1.818. Meanwhile, a series of uniaxial tensile tests on the same materials with standard dog-
bone specimens were conducted both on an Instron universal testing machine and our testing machine.
The results show good agreement of these two apparatus. By combining the equibiaxial and uniaxial
experimental data, the following SEDF with 3rd reduced polynomial can be obtained with curve fitting
(R² = 0.99589)
3
W(1 , 2 , 3 )   Ci 0  12  2 2  32  3
i
(4)
i 1

where C10=0.18942, C20=-0.026151, C30=0.0066057, respectively. The fitted curves are plotted in Figure
6(a) in solid lines. A new finite element model in ABAQUS was then built with the constitutive
parameters identified by the above experiments as well as the identical clamping conditions with the
practical experimental test. Figure 6(b) plots the curves between nominal stain and nominal stress, for
FEM and experiment results, respectively. It can be affirmed that the finite element model agrees very
well with the experimental one. Therefore, it can be further utilized to improve the experimental test
result with the procedure presented in 3.1.2.

Figure 6. (a) Curve fitting by uniaxial and biaxial experimental data, (b) Comparison of FEM and
experimental results under equibiaxial tension.

3.2.2. Validation by pure shear tests. To further validate the finite element model, a pure shear test [13]
was taken by the simulation and experiment simultaneously, with identical test conditions. The
specimens were fixed on two opponent sides without any motion, while the other two opponent sides
were stretched gradually by a given displacement. The stretch ratio curves were obtained for both
directions. Figure 7 depicts the changing of forces along two directions (x represents the tensile direction,
and y represents the fixed direction) with the stretch ratio, showing the validity of constitutive parameter
identification and the finite element model as well.

6
2024 3rd International Conference on Materials Engineering and Applied Mechanics IOP Publishing
Journal of Physics: Conference Series 2808 (2024) 012040 doi:10.1088/1742-6596/2808/1/012040

Figure 7. Comparison of finite element results and experimental results of pure shear test.

4. Conclusions
Biaxial tension tests on soft materials are conducted in this paper with FEM and our own developed
experimental machine. Based on the FEM, the effects of the clamping system on the testing result are
systematically explored, and a novel testing procedure to combine the FEM with the experimental results
is presented. The following conclusions are reached:
1) The clamping width and clamp number are the key parameters to affect the testing results.
2) The finite element model can predict the constitutive model well with the theoretical one.
3) The finite element model agrees well with experiment results, and our testing machine can
produce a wide efficient deformation area.
4) A novel testing procedure to combine the finite element model and experimental testing is
presented, which can be utilized to determine an optimal specimen size during practical tests with
specified materials.

Acknowledgments
This work is financially supported by the National Natural Science Foundation of China (12072315),
funded by the specialized research projects of Huanjiang Laboratory, Zhuji, Zhejiang Province as well.

References
[1] Pilz da Cunha, M.; Ambergen, S.; Debije, M. G.; Homburg, E. F. G. A.; den Toonder, J. M. J.;
Schenning, A. P. H. J. (2020) A soft transporter robot fueled by light. Adv. Sci., 7 (5), 1902842.
[2] Ren, Z.; Hu, W.; Dong, X.; Sitti, M. (2019) Multi-functional soft-bodied jellyfish-like swimming.
Nat. Commun. 10 (1), 2703.
[3] Kight, A.; Pirozzi, I.; Liang, X.; McElhinney, D.B.; Han, A.K.; Dual, S.A.; Cutkosky, M. (2023)
Decoupling transmission and transduction for improved durability of highly stretchable, soft
strain sensing: applications in human health monitoring. Sensors, 23, 1955.
[4] Yin, R.; Wang, D.; Zhao, S.; Lou, Z.; Shen, G., (2021) Wearable sensors-enabled human-machine
interaction systems: from design to application. Adv. Funct. Mater. 31 (11), 2008936.
[5] Collins, I.; Hossain, M.; Dettmer, W.; Masters, I., (2021) Flexible membrane structures for wave
energy harvesting: a review of the developments, materials, and computational modeling
approaches. Renewable Sustainable Energy Rev. 151, 111478.
[6] Sasso, M.; Palmieri, G.; Chiappini, G.; Amodio, D., (2008) Characterization of hyperelastic rubber-
like materials by biaxial and uniaxial stretching tests based on optical methods. Polym. Test.
27 (8), 995-1004.

7
2024 3rd International Conference on Materials Engineering and Applied Mechanics IOP Publishing
Journal of Physics: Conference Series 2808 (2024) 012040 doi:10.1088/1742-6596/2808/1/012040

[7] Palacios-Pineda, L.M.; Perales-Martínez, I.A.; Moreno-Guerra, M.R.; Elías-Zúñiga, A. (2017) An

Optimum Specimen Geometry for Equibiaxial Experimental Tests of Reinforced
Magnetorheological Elastomers with Iron Micro- and Nanoparticles. Nanomaterials, 7, 254.
[8] Putra, K. B.; Tian, X.; Plott, J.; Shih, A., (2020) Biaxial test and hyperelastic material models of
silicone elastomer fabricated by extrusion-based additive manufacturing for wearable
biomedical devices. J. Mech. Behav. Biomed. Mater. 107, 103733.
[9] Keerthiwansa, R.; Javorik, J.; Kledrowetz, J.; Nekoksa, P., (2018) Elastomer testing: The risk of
using only uniaxial data for fitting the Mooney-Rivlin hyperelastic-material model. Mater.
Tehnol, 52, 3-8.
[10] Avanzini, A.; Battini, D., (2016) Integrated experimental and numerical comparison of different
approaches for planar biaxial testing of a hyperelastic material. Adv. Mater. Sci. Eng, 6014129.
[11] Esmaeili, A.; George, D.; Masters, I.; Hossain, M., (2023) Biaxial experimental characterizations
of soft polymers: A review. Polym. Test. 128, 108246.
[12] Jiang, M.; Sridhar, R. L.; Robbins, A. B.; Freed, A. D.; Moreno, M. R., (2021) A versatile biaxial
testing platform for soft tissues. J. Mech. Behav. Biomed. Mater. 114, 104144.
[13] Fujikawa, M.; Maeda, N.; Yamabe, J.; Kodama, Y.; Koishi, M., (2014) Determining stress-strain
in rubber with the in-plane biaxial tensile tester. Exp. Mech. 54 (9), 1639-1649.
[14] Luo, H.; Zhu, Y.; Zhao, H.; Ma, L.; Zhang, J. (2023) Simulation analysis of equibiaxial tension
tests for rubber-like materials. Polymers, 15, 3561.
[15]Holzapfel, G. A., (2002) Nonlinear solid mechanics: A continuum approach for engineering science.
37 (4), 489-490.