- Email: [email protected]
On the cyclic deformation behavior, fracture properties and cytotoxicity of silicone-based elastomers for biomedical applications
Polymer Testing 60 (2017) 117e123
Contents lists available at ScienceDirect
Polymer Testing journal homepage: www.elsevier.com/locate/polytest
Material Behaviour
On the cyclic deformation behavior, fracture properties and cytotoxicity of silicone-based elastomers for biomedical applications L. Bernardi a, *, R. Hopf a, D. Sibilio a, A. Ferrari c, A.E. Ehret a, b, E. Mazza a, b a
ETH Zurich, Institute for Mechanical Systems, 8092 Zürich, Switzerland Empa, Swiss Federal Laboratories for Materials Science and Technology, 8600 Dübendorf, Switzerland c ETH Zurich, Laboratory for Thermodynamics in Emerging Technologies, 8092 Zürich, Switzerland b
a r t i c l e i n f o
a b s t r a c t
Article history: Received 8 February 2017 Accepted 15 March 2017 Available online 16 March 2017
This paper provides results from a comprehensive experimental characterization on five silicone-based elastomers used as substrates for mechanobiological studies or in soft biomedical implants. A previous paper was recently published which focused on the large strain deformation behavior of these materials. This second part analyzes their reliability for biomedical applications in terms of changes of deformation behavior with the history of loading (long term cyclic behavior), ability to resist loads in the presence of defects (fracture properties), and cytotoxicity. For the latter, all materials are confirmed to be non-toxic which is a prerequisite for their use in mechanobiological studies or as part of implants and biomedical devices. The response in long term uniaxial tests over 2200 000 cycles was characterized and the results indicate general stability of the mechanical response with, for some conditions, softening mechanisms active mainly in the initial phase of the test (500 000 cycles). A critical aspect of elastomer performance and their suitability for application in biomedical devices concerns their fracture properties. The tearing energy varies in a range from brittle (with approximately 80 J/m2 for PDMS Sylgard 184) to tough (with approximately 900 J/m2 for SMI G/G 0.020). © 2017 Elsevier Ltd. All rights reserved.
Keywords: Silicone-based elastomers Biomedical applications Tearing energy Long term cyclic behavior Cytotoxicity
1. Introduction This article presents the second part of a comprehensive experimental campaign for the characterization of five siliconebased elastomers, i.e. two types of PDMS (Sylgard 184 and Sylgard 186, Dow Corning), and three RTV elastomers vulcanized at room temperature (SMI G/G 0.020”, Specialty Manufacturing Inc.; RTV 4528 and RTV 4420, Blue Stars Silicones). These materials were shown to cover a range of mechanical characteristics representative of those of the elastomers commonly used for mechanobiological studies and biomedical devices. The first part of this study [1] provided experimental data and corresponding model equations for the multiaxial monotonic large strain deformation behavior of the elastomers. This paper describes their long term cyclic response, fracture properties and cytotoxicity. Little is reported in previous studies regarding these characteristics, despite their importance for a number of applications. Our companion paper [1] provides a general overview on applications of silicone-based
* Corresponding author. E-mail address: [email protected] (L. Bernardi). http://dx.doi.org/10.1016/j.polymertesting.2017.03.018 0142-9418/© 2017 Elsevier Ltd. All rights reserved.
elastomers and the necessity to characterize their deformation response under a variety of mechanical loading conditions. The specific motivation for the analysis of their cyclic response is found in biomedical applications such as for bioreactors mimicking cardiovascular pulsatile conditions [2,3], as part of endovascular grafts and cardiovascular devices [4,5], or other implants exposed to cyclic deformation, e.g. artificial urinary sphincters [6], external shells of breast implants [7] or skin support [8]. In these applications, the mechanical response might change with the history of loading since the long term deformation behavior might differ from that in virgin conditions. This question was addressed here through experiments with >2000 000 cycles at nominal strains of up to 30%. Note that the objective of these experiments was not the characterization of the fatigue behavior of the elastomers, which would require a much higher number of cycles. As explained in Ref. [1], the range of deformations considered for the present work were originally motivated by a specific application, viz. as blood propulsion membrane in a pulsatile Ventricular Assist Device (VAD) [9,10]. Another essential feature of materials fulfilling a structural function is their toughness. The ability to resist loads despite the presence of (small) defects is important in particular in view of
118
L. Bernardi et al. / Polymer Testing 60 (2017) 117e123
possible imperfections arising from the manufacturing process. Such defects might affect the mechanical integrity and reliability of biomedical devices or implants. An example is given in the application of silicone elastomers as a joint replacement, e.g. finger joints [11,12], for which fracture initiation and crack growth were analyzed in previous studies [13,14]. The influence of temperature and wet environment (typical for physiological conditions) on the cyclic response and fracture properties of RTV 4420 is also addressed. Finally, in view of the application as implants or substrates for mechanobiological studies, all materials were subjected to cytotoxicity tests. 2. Materials The characteristics of the elastomers analyzed in this study were described in detail in Ref. [1]. PDMS Sylgard 184, a two component system of base polymer and crosslinker was used in 10:1 ratio; PDMS Sylgard 186 was also used in 10:1 ratio; Silbione® RTV 4528 A&B and Silbione® RTV 4420 A&B are Room-Temperature Vulcanized (RTV) elastomers for which the two components are mixed in 1:1 ratio; SMI G/G 0.02000 also belongs to the RTV family but it is commercially available. All the materials presented in this work are widely used as substrates in mechanobiological studies [15e17], or in soft biomedical implants such as skin-like applications or external maxillofacial prosthetics [18,19]. Samples for long term cyclic and tearing tests were prepared according to the protocol described in Ref. [1]: the elastomer components were mixed by hand for 2e3 min, then degassed and poured into petri dishes of 90 mm diameter. After additional 45 min of degassing, samples were cured in an oven at 60 C for 4 h, except for RTV 4420 which was cured at room temperature (23 C). The samples for the long term cyclic tests were rectangular, with a width to length ratio of 1:4, and samples for tearing tests were rectangular with a width to length ratio of 1:6, see Fig. 1.
Reliable gripping of the sample is provided by gluing a strip of sandpaper onto the surface of the custom-made metal clamps. The clamps are displacement controlled in that a screw is used to apply the initial compression and this remains unchanged during the test. An environmental chamber with a bath, heating coil and a temperature sensor allows testing in wet conditions and at a prescribed temperature (here: 37 C). Fig. 2 shows the set-up equipped with the environmental bath and the heating coil with a tearing test sample mounted in the clamps. 3.2. Data analysis Deformation in each sample is reported as engineering strain ε ¼ l 1 , where l is the stretch in loading direction. Stress is reported as force per unit of initial area (first Piola-Kirchoff stress P). Local strains are calculated from image-based determination of the in-plane deformation. To this end, an optical flow tracker (described in Refs. [1] and [20]) was applied in order to extract the displacement field in the central region of each test piece in order to avoid boundary or clamping effects. 3.3. Effect of the environmental conditions on deformation behavior Mechanobiological studies as well as many biomedical devices expose the elastomers to a temperature of 37 C and wet conditions. For example, in cells studies (e.g. Ref. [3]), the silicone membranes are left several days in cell culture medium inside an incubator. In order to mimic these environmental conditions, a few tests were performed in distilled water and saline solution (resembling a cell culture medium) at 37 C. Saline solutions were prepared with a known concentration of salts in distilled water (0.9 mg/mL NaCl in distilled water [22]). In order to assess the influence of environmental conditions, monotonic uniaxial tests were carried out on RTV 4420 samples left 24 h in distilled water or in saline solution at 37 and compared with results obtained at room temperature in dry conditions.
3. Experimental 3.4. Long term cyclic tests 3.1. Mechanical testing set-up The set-up used for both tearing experiments and long term cyclic tests was described in Refs. [1] and [20]. It consists of 2 horizontal hydraulic actuators, 100 N load cells (calibrated for up to 20 N, MTS System, Eden Prairie, USA), a CCD camera (Pike F-100B Allied Vision Technologies GmbH, Stadroda, Germany) with 0.25 telecentric lens (NT55-349 Edmund Optics GmbH, Karlsruhe, Germany) used for local strain analysis (see Refs. [20,21]).
Cyclic tests were performed in uniaxial configuration. Samples were cyclically stretched between 0 and 30% nominal strain at 1 Hz, up to a total of 2200 000 cycles. Thus, the total duration of each test was about 60 h and only 1 sample was tested for each material. 200 cycles of initial pre-conditioning at 1 Hz were applied before starting the cyclic experiment in order to stabilize clamping conditions. Forces and images were recorded during the test every 12 h (i.e. every 430 000 cycles) in order to monitor the stress decay associated with the applied local strain for increasing number of cycles. Forces were recorded at a frequency of 512 Hz while images were recorded at a frequency of 20 Hz. A quantitative measure, the reduction in peak stress Pdecay was calculated normalizing the stress at each time point Pi with respect to the value associated with the applied strain in virgin conditions Pinitial . RTV 4420 was tested also at 37 C in distilled water; the sample was stored in the corresponding solution and inside an incubator (37 C) 24 h prior to the test. The incubator was chosen because it is representative of cell cultures. 3.5. Tearing tests
Fig. 1. Sample dimensions in mm for the long term cyclic test (a) and tearing test (b).
Tearing tests were performed following a protocol based on the original work by Rivlin and Thomas [23]. A cut of 20 mm was created on one of the lateral edges of the test piece (Fig. 1), which was loaded up to the point of crack propagation with a nominal strain rate of 0.3%/s. Fig. 3 shows an example of a sample at the
L. Bernardi et al. / Polymer Testing 60 (2017) 117e123
119
Fig. 2. Testing set-up equipped with the environmental chamber and the temperature coil, with a tearing test sample mounted.
Fig. 3. Tearing test sample at the beginning of the test (left) and at the time point of crack propagation (right). The blue rectangle delimitates the area were the optical flow tracker is applied to extract the local deformation. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)
beginning of the test and at the time point of crack propagation. For each material, 3 cut specimens were tested. Additionally a sample without cut was tested under the same conditions, in order to provide a reference for the calculation of the strain energy at the time point of crack propagation, as described in Refs. [23] and [24]. The stretch at which the crack starts to propagate was determined based on a visual inspection of the recorded camera images. This indicates that crack propagation often occurred well before complete sample rupture or before visible changes in the loaddisplacement curves. The local state of deformation was measured from optical flow tracking in the region indicated in Fig. 3. These data were used to calculate the tearing energy by integrating the first Piola-Kirchoff stress over the local stretch of
the reference sample up to the critical stretch value (W0). The tearing energy was calculated as G ¼ L* W0 [24], where L* is the length of the sample in the unstrained state. To correct for slippage, local values of strain were used. Results are reported in terms of tearing energy and critical local stretch (mean ± standard deviation). The five materials were tested in dry conditions at room temperature, while the effect of temperature and wet environment was studied for RTV4420. For this material, tests were performed at 37 C in distilled water and the sample was stored in the corresponding liquid solution in the incubator for 24 h prior to the test.
120
L. Bernardi et al. / Polymer Testing 60 (2017) 117e123
3.6. Cytotoxicity tests Due to the envisaged application as implant materials, the cytotoxicity of the elastomers was evaluated according to [25] and [26]. Importantly, the chosen test is more sensitive than classical live/dead assays. It can detect toxicity levels which are compatible with cell survival yet significantly interfere with the cell metabolism and activity. In short, the 5 materials were coated with PolyL-Lysine (PLL) 0.01% solution (Sigma Aldrich, USA) and PC12 cells were cultured in RPMI-1640 culture medium supplemented with 10% horse serum and 5% FBS (all Sigma Aldrich, USA), 2 mM Lglutamine, 100 U/ml penicillin and 100 ml/ml streptomycin (all Life Technologies, USA). The day of the test, the cells were incubated with the different elastomer materials, previously placed inside the wells of 12 well plates, at a density of 3 104 cells/well (3 repetitions for each material) or seeded into wells of 12 well plates in direct contact with PLL-coated tissue culture Poly Styrene TCPS (6 control wells without materials). Cells were incubated at 37 C and 5% CO2 for 24 h, then washed 3 times with PBS, incubated again for 24 h with fresh culture medium, and finally stimulated with nerve growth factor at a concentration of 100 ng/ml (NGF, Sigma-Aldrich) in induction medium (RPMI medium with 5% horse medium and 1% FBS). In order to check the functionality of the nerve growth factor, cells were stimulated at the same time point in 3 wells of the control, while they were not stimulated in the last 3 control wells. Cells were incubated for 3 days before imaging. As described in Ref. [26], cell images were acquired using an inverted Nikon-Ti wide field microscope (Nikon, Japan) with an Orca R-2 CCD camera (Hamamatsu Photonics, Japan). 10 images per well were collected at random spots with a 20 magnification, 0.45 numerical aperture long distance air objective (Plan Fluor, Nikon, Japan), using the white field channel with a differential interference contrast (DIC) filter. The length of neurites generated by PC12 cells on stimulation with NGF was used to detect the presence of cytotoxic effect from the tested materials interfering with the neuronal differentiation mechanism. The length of neurites was measured using the NeuronJ plug-in of ImageJ [27e29]. Membrane protrusions with length smaller than 10 mm were excluded from this analysis. Statistical comparison of the neurite length was performed using a non-parametric Mann-Whitney test to identify variations between the tested samples. 4. Results According to [1], a different but consistent color is associated with each material, and is systematically used in each figure. 4.1. Effect of environmental conditions on deformation behavior Fig. 4 reports the results of the monotonic uniaxial tests of RTV 4420 performed at 37 C either in distilled water or in saline solution (average ± standard deviation, n ¼ 3). For comparison, the average curve previously reported in Ref. [1] for the same material is also shown in the figure. Note that, in this figure, to be consistent with the curve in Ref. [1], stress is reported as Cauchy stress (force per unit of deformed area). Two observations can be made. First, the difference between the samples stored and tested in saline solution and samples stored and tested in distilled water is insignificant. For this reason, and since the maintenance of a constant salt concentration for the long term tests is cumbersome due to water evaporation, the environmental conditions for the tests presented in this work are reproduced as distilled water at 37 C. Second, there is no difference between the material stored and tested at room temperature and dry environment and the material stored and tested in the different environmental conditions.
Fig. 4. Monotonic uniaxial stress-strain curves of RTV 4420 tested after 1 day in distilled water or saline solution at 37 C. For comparison, the nominal curve from Ref. [1] (room temperature and dry conditions) is also reported.
4.2. Long term cyclic tests Fig. 5 reports the first Piola-Kirchhoff stress decay over the measured 2200 000 cycles. The long term behavior of most materials indicates a stable response, while RTV 4528 and Sylgard 184 are characterized by an initial decrease in peak stress that stabilizes after approximately 500 000 cycles. RTV 4528 shows the largest decrease (around 25%), while for the other materials the decrease is always lower than 8%. No macroscopic failure of the samples was observed. Fig. 5 reports also the results for the sample tested in water at 37 C. This sample shows some progressive decay when compared to the same material tested in dry conditions at room temperature. Note, however, that the stress reduction is modest with a maximum of 7% and an average of about 3% at the end of the test.
4.3. Tearing tests The stretch at crack propagation and tearing energy provide an indication of the sensitivity of the elastomer to pre-existing defects. Crack-like defects might be related to the manufacturing or the mounting process, or arise due to extreme operating conditions. The tearing energy values can be used for a corresponding fracture mechanics analysis by assuming a specific crack size and position, and calculating the loading level required to cause crack propagation. Information on the critical stretch can help to classify the materials in terms of their brittleness and independently of their stiffness, whereas tearing energy is important in the case of force or pressure controlled loading conditions. Note that, while the tearing energy is a property of the material, the stretch at propagation depends on the particular geometry and the test configuration, see e.g. Ref. [24]. Here it is reported for comparison among the considered elastomers, since the same geometry was used for all the materials. Test results are reported in Fig. 6. They indicate that PDMS Sylgard 184 is significantly more brittle than the other materials, and cracks propagate at relatively small stretches (around l ¼ 1.1). RTV 4528 is the most resistant in terms of stretch at propagation, with critical values around l ¼ 1.75. For the other three materials, the behavior is similar and cracks propagate at a stretch of around l ¼ 1.3e1.4. Regarding the tearing energy, SMI performs best, with a tearing energy of around 920 J/m2, while PDMS Sylgard 184 has the lowest tearing energy of only 80 J/m2, i.e.
L. Bernardi et al. / Polymer Testing 60 (2017) 117e123
121
Fig. 5. Stress decay over 2200 000 cycles normalized with respect to the value at the beginning of the test.
Fig. 6. Fracture behavior of the tested materials, average and standard deviation for each material (n ¼ 3). Top: Stretch at propagation for samples of same dimensions. Bottom: Tearing energy.
one order of magnitude smaller. The results of RTV4420 samples tested in distilled water at 37 C are also reported in Fig. 6. Although both stretches at crack propagation and tearing energy are lower than the corresponding quantities tested in nominal conditions, the difference is small (below 2.5%) and, at least for the low number of samples considered, not statistically significant. 4.4. Cytotoxicity tests Cytotoxicity tests were performed as previously described [26] to detect potential interference with the delicate molecular
pathways mediating neuronal differentiation of PC12 cells upon NGF-stimulation [30]. In particular, the neurite length was measured as described in section 3.6, yielding a linear parameter which directly reports on the potential cytotoxicity mediated by the contact with the elastomeric substrates, as compared with standard culturing conditions on tissue culture polystyrene (TCPS). The data presented in Fig. 7 demonstrate that none of the tested substrates caused a significant reduction of neurite length as compared with the control. Altogether, these results confirm that the materials under study do not show cytotoxicity in vitro, thus supporting their further characterization for potential in vivo applications.
122
L. Bernardi et al. / Polymer Testing 60 (2017) 117e123
conditions, as shown in Fig. 4. Contrary to other materials such as PDMS, known for its variable mechanical properties due to temperature and time of curing [35], and also for its severe aging over time [20], RTV 4420 is very stable in terms of temperature of testing in the range of interest of this work. Furthermore, it does not show aging in the first two months from preparation (data not shown). The present results show that the polymer is fully crosslinked already at room temperature and, contrary to other fully crosslinked RTV materials, as in Ref. [36], it shows no temperature dependence of its stress-strain behavior in the range of temperature investigated that is representative of physiological conditions. The elastomers analyzed in the present work show no cytotoxic effects in vitro. Of course, in case of applications as a component of human implants (such as in devices that are in contact with blood), the materials would have to be specifically qualified. 6. Conclusions Fig. 7. NGF induced PC12 differentiation on different materials. Box spans from 1st to 3rd quartile, outliers are removed from the plot. Black lines indicate the medians. Sample size is indicated above each population. Note that for the control without NGF, only 5 neurites had a length greater than 10 mm.
5. Discussion In this paper, the characteristic energy for tearing, or tearing energy [23], was used as measure of the fracture properties of the elastomers among other possible quantities such as J-integral [31] or crack tip opening displacement [32]. This quantity allows prediction of the resistance to the (Mode I) propagation of existent crack-like defects for arbitrary loading conditions. It has been chosen due to its straight-forward experimental evaluation [31]. Typical values for silicone-based elastomers found in literature are in line with the results of the current work [33,34]. Our results highlight the brittleness of PDMS Sylgard 184: with 80 J/m2, it is the material with the lowest tearing energy among the tested elastomers. This material is widely applied for in vitro cell studies since it is easy to fabricate and transparent, thus allowing live imaging applications. However, its brittleness represents a severe limitation with respect to applications that involve mechanical loading to large strains since pre-existing defects might be expected to grow, possibly leading to catastrophic failure. PDMS Sylgard 186 might represent a viable alternative, since it shows elevated tearing energy as well as negligible stress decay over 2200 000 cycles. However, its viscosity might complicate the fabrication process; in fact, our experience shows that membranes of homogeneous thickness are more difficult to obtain with typical preparation methods. RTV 4528 shows the maximum stress decay in the long-term cyclic tests, continuing beyond the 200 cycles reported in Ref. [1] and reaching nearly 500 000 cycles. In this regard, considering a biomedical application for which a long term stable response is required, SMI and RTV 4420 seem more suitable candidate materials. In particular, RTV 4420 showed good toughness properties, stable long term cyclic behavior and, importantly, it can easily be fabricated and tailored to a specific shape required for a specific design of a biomedical device. The effect of wet conditions and temperature on the long term cyclic behavior and on the fracture properties is analyzed for RTV 4420 and no relevant differences can be observed with respect to the samples tested in nominal conditions. No previous data on the combined effect of these factors were found in literature. Interestingly, the material chosen for the environmental conditions study shows no dependence of the deformation properties on these
This work presents a characterization of five silicone-based elastomers in terms of cyclic response and fracture properties. Important for applications in biomedical devices, implants and bioreactors, the critical stretch and tearing energy of the elastomers were determined in corresponding tests, indicating for the latter differences of more than one order of magnitude among the elastomers. This evaluation highlighted in particular the brittleness of PDMS Sylgard 184, thus limiting its reliability in terms of mechanical integrity. Long term tests involving up to 2200 000 loading cycles were performed and the corresponding decay in peak stress was analyzed. One elastomer, RTV 4528, showed a significant decay in peak stress while all others displayed a stable response. The influence of “physiological” environmental conditions on the mechanical properties of RTV 4420 was investigated and no relevant differences could be observed. Finally, the non-cytotoxic characteristics of the elastomers were demonstrated, which are a prerequisite for their use in mechanobiological studies or as part of soft implants and biomedical devices. Acknowledgments This project is supported by the “Stiftung PROPTER HOMINES Vaduz/Fürstentum Liechtenstein”, the “Schwyzer-Winiker Stiftung” and the ETH Zurich Foundation. This work is part of the Zurich Heart project of Hochschulmedizin Zürich. References [1] L. Bernardi, R. Hopf, A. Ferrari, A.E. Ehret, E. Mazza, On the large strain deformation behavior of silicone-based elastomers for biomedical applications, Polym. Test. 58 (2017) 189e198. [2] X.Q. Brown, K. Ookawa, J.Y. Wong, Evaluation of polydimethylsiloxane scaffolds with physiologically-relevant elastic moduli: interplay of substrate mechanics and surface chemistry effects on vascular smooth muscle cell response, Biomaterials 26 (16) (2005) 3123e3129. [3] B.J. Bachmann, L. Bernardi, C. Loosli, J. Marschewski, M. Perrini, M. Ehrbar, P. Ermanni, D. Poulikakos, A. Ferrari, E. Mazza, A novel bioreactor system for the assessment of endothelialization on deformable surfaces, Sci. Rep. (6) (2016) 38861. [4] W.-M.P.F. Bosman, T.J. van der Steenhoven, D.R. Su arez, E.R. Valstar, A.C. de Vries, H.L.F. Brom, M.J. Jacobs, J.F. Hamming, The effect of injectable biocompatible elastomer (PDMS) on the strength of the proximal fixation of endovascular aneurysm repair grafts: an in vitro study, J. Vasc. Surg. 52 (1) (2010) 152e158. [5] G. Soldani, P. Losi, M. Bernabei, S. Burchielli, D. Chiappino, S. Kull, E. Briganti, D. Spiller, Long term performance of small-diameter vascular grafts made of a poly (ether) urethane-polydimethylsiloxane semi-interpenetrating polymeric network, Biomaterials 31 (9) (2010) 2592e2605. [6] P. Costa, G. Poinas, K.B. Naoum, K. Bouzoubaa, L. Wagner, L. Soustelle, M. Boukaram, S. Droupy, Long-term results of artificial urinary sphincter for women with type III stress urinary incontinence, Eur. Urol. 63 (4) (2013) 753e758.
L. Bernardi et al. / Polymer Testing 60 (2017) 117e123 [7] S. Necchi, D. Molina, S. Turri, F. Rossetto, M. Rietjens, G. Pennati, Failure of silicone gel breast implants: is the mechanical weakening due to shell swelling a significant cause of prostheses rupture? J. Mech. Behav. Biomed. Mater. 4 (8) (2011) 2002e2008. [8] B. Yu, S.-Y. Kang, A. Akthakul, N. Ramadurai, M. Pilkenton, A. Patel, A. Nashat, D.G. Anderson, F.H. Sakamoto, B.A. Gilchrest, et al., An elastic second skin, Nat. Mater. 15 (2016) 911e918. [9] “Zurich Heart,” 2015. [Online]. Available: http://www.hochschulmedizin.uzh. ch/de/projekte/zurichheart.html. [Accessed 03 May 2016]. [10] “Hybrid Membrane,” 2016. [Online]. Available: http://www.zurichheart.ethz. ch/hybridmembrane. [Accessed 03 May 2016]. [11] “Swanson Finger Joint Implants,” ©2016. [Online]. Available: http://www. wright.com/products-upper/swanson-finger-joint-implants. [Accessed 15 June 2016]. [12] “NeuFlex MCP/PIP Finger Joint Implants,” ©2015. [Online]. Available: http:// www.fingerreplacement.com/docs/Finger/Products/Info/neuflex.html. [Accessed 15 June 2016]. [13] T.J. Joyce, Currently available metacarpophalangeal prostheses: their designs and prospective considerations, Expert Rev. Med. Devices 1 (2) (2004) 193e204. [14] T.J. Joyce, A. Unsworth, A literature review of “failures” of the Swanson finger prosthesis in the metacarpophalangeal joint, Hand Surg. 7 (01) (2002) 139e146. [15] I. Schoen, W. Hu, E. Klotzsch, V. Vogel, Probing cellular traction forces by micropillar arrays: contribution of substrate warping to pillar deflection, Nano Lett. 10 (5) (2010) 1823e1830. [16] S. Fusco, V. Panzetta, V. Embrione, P.A. Netti, Crosstalk between focal adhesions and material mechanical properties governs cell mechanics and functions, Acta biomater. 23 (2015) 63e71. [17] G. Fois, M. Weimer, T. Busch, E.T. Felder, F. Oswald, G. von Wichert, T. Seufferlein, P. Dietl, E. Felder, Effects of keratin phosphorylation on the mechanical properties of keratin filaments in living cells, FASEB J. 27 (4) (2013) 1322e1329. [18] G.L. Polyzois, A. Hensten-Pettersen, A. Kullmann, An assessment of the physical properties and biocompatibility of three silicone elastomers, J. Prosthet. Dent. 71 (5) (1994) 500e504. [19] V.A. Chalian, R.W. Phillips, Materials in maxillofacial prosthetics, J. Biomed. Mater. Res. 8 (4) (1974) 349e363. [20] R. Hopf, L. Bernardi, J. Menze, M. Zündel, E. Mazza, A.E. Ehret, Experimental and theoretical analyses of the age-dependent large-strain behavior of Sylgard 184 (10: 1) silicone elastomer, J. Mech. Behav. Biomed. Mater. 60 (2016) 425e437.
123
[21] M. Hollenstein, A.E. Ehret, M. Itskov, E. Mazza, A novel experimental procedure based on pure shear testing of dermatome-cut samples applied to porcine skin, Biomech. Model. Mechanobiol. 10 (5) (2011) 651e661. [22] M. Pensalfini, A.E. Ehret, S. Stüdeli, D. Marino, A. Kaech, E. Reichmann, E. Mazza, Factors affecting the mechanical behavior of collagen hydrogels for skin tissue engineering, J. Mech. Behav. Biomed. Mater. 69 (2017) 85e97. [23] R.S. Rivlin, A.G. Thomas, Rupture of rubber. I. Characteristic energy for tearing, J. Polym. Sci. 10 (3) (1953) 291e318. [24] M. Pharr, J.-Y. Sun, Z. Suo, Rupture of a highly stretchable acrylic dielectric elastomer, J. Appl. Phys. 111 (10) (2012) 104114. [25] A. Ferrari, M. Cecchini, M. Serresi, P. Faraci, D. Pisignano, F. Beltram, Neuronal polarity selection by topography-induced focal adhesion control, Biomaterials 31 (17) (2010) 4682e4694. [26] T. Vasileiou, D. Foresti, A. Bayram, D. Poulikakos, A. Ferrari, Toward contactless biology: acoustophoretic dna transfection, Sci. Rep. 6 (2016). [27] C.A. Schneider, W.S. Rasband, K.W. Eliceiri, et al., NIH Image to ImageJ: 25 years of image analysis, Nat. methods 9 (7) (2012) 671e675. [28] J. Schindelin, I. Arganda-Carreras, E. Frise, V. Kaynig, M. Longair, T. Pietzsch, S. Preibisch, C. Rueden, S. Saalfeld, B. Schmid, et al., Fiji: an open-source platform for biological-image analysis, Nat. methods 9 (7) (2012) 676e682. [29] E. Meijering, M. Jacob, J.-C.F. Sarria, P. Steiner, H. Hirling, M. Unser, Design and validation of a tool for neurite tracing and analysis in fluorescence microscopy images, Cytom. Part A 58 (2) (2004) 167e176. [30] A. Ferrari, P. Faraci, M. Cecchini, F. Beltram, The effect of alternative neuronal differentiation pathways on PC12 cell adhesion and neurite alignment to nanogratings, Biomaterials 31 (9) (2010) 2565e2573. [31] R. Sto cek, T. Horst, K. Reincke, Tearing energy as fracture mechanical quantity for elastomers, Adv. Polym. Sci. 275 (2017) 361e398. [32] K. Reincke, W. Grellmann, R. Lach, G. Heinrich, Toughness optimization of SBR elastomerseuse of fracture mechanics methods for characterization, Macromol. Mater. Eng. 288 (2) (2003) 181e189. [33] L.C. Yanyo, F.N. Kelley, Effect of chain length distribution on the tearing energy of silicone elastomers, Rubber Chem. Technol. 60 (1) (1987) 78e88. [34] G.D. Genesky, C. Cohen, Toughness and fracture energy of PDMS bimodal and trimodal networks with widely separated precursor molar masses, Polymer 51 (18) (2010) 4152e4159. [35] M. Liu, J. Sun, Q. Chen, Influences of heating temperature on mechanical properties of polydimethylsiloxane, Sens. Actuators A Phys. 151 (1) (2009) 42e45. [36] T. Rey, G. Chagnon, J.-B. Le Cam, D. Favier, Influence of the temperature on the mechanical behaviour of filled and unfilled silicone rubbers, Polym. Test. 32 (3) (2013) 492e501.
Contents lists available at ScienceDirect
Polymer Testing journal homepage: www.elsevier.com/locate/polytest
Material Behaviour
On the cyclic deformation behavior, fracture properties and cytotoxicity of silicone-based elastomers for biomedical applications L. Bernardi a, *, R. Hopf a, D. Sibilio a, A. Ferrari c, A.E. Ehret a, b, E. Mazza a, b a
ETH Zurich, Institute for Mechanical Systems, 8092 Zürich, Switzerland Empa, Swiss Federal Laboratories for Materials Science and Technology, 8600 Dübendorf, Switzerland c ETH Zurich, Laboratory for Thermodynamics in Emerging Technologies, 8092 Zürich, Switzerland b
a r t i c l e i n f o
a b s t r a c t
Article history: Received 8 February 2017 Accepted 15 March 2017 Available online 16 March 2017
This paper provides results from a comprehensive experimental characterization on five silicone-based elastomers used as substrates for mechanobiological studies or in soft biomedical implants. A previous paper was recently published which focused on the large strain deformation behavior of these materials. This second part analyzes their reliability for biomedical applications in terms of changes of deformation behavior with the history of loading (long term cyclic behavior), ability to resist loads in the presence of defects (fracture properties), and cytotoxicity. For the latter, all materials are confirmed to be non-toxic which is a prerequisite for their use in mechanobiological studies or as part of implants and biomedical devices. The response in long term uniaxial tests over 2200 000 cycles was characterized and the results indicate general stability of the mechanical response with, for some conditions, softening mechanisms active mainly in the initial phase of the test (500 000 cycles). A critical aspect of elastomer performance and their suitability for application in biomedical devices concerns their fracture properties. The tearing energy varies in a range from brittle (with approximately 80 J/m2 for PDMS Sylgard 184) to tough (with approximately 900 J/m2 for SMI G/G 0.020). © 2017 Elsevier Ltd. All rights reserved.
Keywords: Silicone-based elastomers Biomedical applications Tearing energy Long term cyclic behavior Cytotoxicity
1. Introduction This article presents the second part of a comprehensive experimental campaign for the characterization of five siliconebased elastomers, i.e. two types of PDMS (Sylgard 184 and Sylgard 186, Dow Corning), and three RTV elastomers vulcanized at room temperature (SMI G/G 0.020”, Specialty Manufacturing Inc.; RTV 4528 and RTV 4420, Blue Stars Silicones). These materials were shown to cover a range of mechanical characteristics representative of those of the elastomers commonly used for mechanobiological studies and biomedical devices. The first part of this study [1] provided experimental data and corresponding model equations for the multiaxial monotonic large strain deformation behavior of the elastomers. This paper describes their long term cyclic response, fracture properties and cytotoxicity. Little is reported in previous studies regarding these characteristics, despite their importance for a number of applications. Our companion paper [1] provides a general overview on applications of silicone-based
* Corresponding author. E-mail address: [email protected] (L. Bernardi). http://dx.doi.org/10.1016/j.polymertesting.2017.03.018 0142-9418/© 2017 Elsevier Ltd. All rights reserved.
elastomers and the necessity to characterize their deformation response under a variety of mechanical loading conditions. The specific motivation for the analysis of their cyclic response is found in biomedical applications such as for bioreactors mimicking cardiovascular pulsatile conditions [2,3], as part of endovascular grafts and cardiovascular devices [4,5], or other implants exposed to cyclic deformation, e.g. artificial urinary sphincters [6], external shells of breast implants [7] or skin support [8]. In these applications, the mechanical response might change with the history of loading since the long term deformation behavior might differ from that in virgin conditions. This question was addressed here through experiments with >2000 000 cycles at nominal strains of up to 30%. Note that the objective of these experiments was not the characterization of the fatigue behavior of the elastomers, which would require a much higher number of cycles. As explained in Ref. [1], the range of deformations considered for the present work were originally motivated by a specific application, viz. as blood propulsion membrane in a pulsatile Ventricular Assist Device (VAD) [9,10]. Another essential feature of materials fulfilling a structural function is their toughness. The ability to resist loads despite the presence of (small) defects is important in particular in view of
118
L. Bernardi et al. / Polymer Testing 60 (2017) 117e123
possible imperfections arising from the manufacturing process. Such defects might affect the mechanical integrity and reliability of biomedical devices or implants. An example is given in the application of silicone elastomers as a joint replacement, e.g. finger joints [11,12], for which fracture initiation and crack growth were analyzed in previous studies [13,14]. The influence of temperature and wet environment (typical for physiological conditions) on the cyclic response and fracture properties of RTV 4420 is also addressed. Finally, in view of the application as implants or substrates for mechanobiological studies, all materials were subjected to cytotoxicity tests. 2. Materials The characteristics of the elastomers analyzed in this study were described in detail in Ref. [1]. PDMS Sylgard 184, a two component system of base polymer and crosslinker was used in 10:1 ratio; PDMS Sylgard 186 was also used in 10:1 ratio; Silbione® RTV 4528 A&B and Silbione® RTV 4420 A&B are Room-Temperature Vulcanized (RTV) elastomers for which the two components are mixed in 1:1 ratio; SMI G/G 0.02000 also belongs to the RTV family but it is commercially available. All the materials presented in this work are widely used as substrates in mechanobiological studies [15e17], or in soft biomedical implants such as skin-like applications or external maxillofacial prosthetics [18,19]. Samples for long term cyclic and tearing tests were prepared according to the protocol described in Ref. [1]: the elastomer components were mixed by hand for 2e3 min, then degassed and poured into petri dishes of 90 mm diameter. After additional 45 min of degassing, samples were cured in an oven at 60 C for 4 h, except for RTV 4420 which was cured at room temperature (23 C). The samples for the long term cyclic tests were rectangular, with a width to length ratio of 1:4, and samples for tearing tests were rectangular with a width to length ratio of 1:6, see Fig. 1.
Reliable gripping of the sample is provided by gluing a strip of sandpaper onto the surface of the custom-made metal clamps. The clamps are displacement controlled in that a screw is used to apply the initial compression and this remains unchanged during the test. An environmental chamber with a bath, heating coil and a temperature sensor allows testing in wet conditions and at a prescribed temperature (here: 37 C). Fig. 2 shows the set-up equipped with the environmental bath and the heating coil with a tearing test sample mounted in the clamps. 3.2. Data analysis Deformation in each sample is reported as engineering strain ε ¼ l 1 , where l is the stretch in loading direction. Stress is reported as force per unit of initial area (first Piola-Kirchoff stress P). Local strains are calculated from image-based determination of the in-plane deformation. To this end, an optical flow tracker (described in Refs. [1] and [20]) was applied in order to extract the displacement field in the central region of each test piece in order to avoid boundary or clamping effects. 3.3. Effect of the environmental conditions on deformation behavior Mechanobiological studies as well as many biomedical devices expose the elastomers to a temperature of 37 C and wet conditions. For example, in cells studies (e.g. Ref. [3]), the silicone membranes are left several days in cell culture medium inside an incubator. In order to mimic these environmental conditions, a few tests were performed in distilled water and saline solution (resembling a cell culture medium) at 37 C. Saline solutions were prepared with a known concentration of salts in distilled water (0.9 mg/mL NaCl in distilled water [22]). In order to assess the influence of environmental conditions, monotonic uniaxial tests were carried out on RTV 4420 samples left 24 h in distilled water or in saline solution at 37 and compared with results obtained at room temperature in dry conditions.
3. Experimental 3.4. Long term cyclic tests 3.1. Mechanical testing set-up The set-up used for both tearing experiments and long term cyclic tests was described in Refs. [1] and [20]. It consists of 2 horizontal hydraulic actuators, 100 N load cells (calibrated for up to 20 N, MTS System, Eden Prairie, USA), a CCD camera (Pike F-100B Allied Vision Technologies GmbH, Stadroda, Germany) with 0.25 telecentric lens (NT55-349 Edmund Optics GmbH, Karlsruhe, Germany) used for local strain analysis (see Refs. [20,21]).
Cyclic tests were performed in uniaxial configuration. Samples were cyclically stretched between 0 and 30% nominal strain at 1 Hz, up to a total of 2200 000 cycles. Thus, the total duration of each test was about 60 h and only 1 sample was tested for each material. 200 cycles of initial pre-conditioning at 1 Hz were applied before starting the cyclic experiment in order to stabilize clamping conditions. Forces and images were recorded during the test every 12 h (i.e. every 430 000 cycles) in order to monitor the stress decay associated with the applied local strain for increasing number of cycles. Forces were recorded at a frequency of 512 Hz while images were recorded at a frequency of 20 Hz. A quantitative measure, the reduction in peak stress Pdecay was calculated normalizing the stress at each time point Pi with respect to the value associated with the applied strain in virgin conditions Pinitial . RTV 4420 was tested also at 37 C in distilled water; the sample was stored in the corresponding solution and inside an incubator (37 C) 24 h prior to the test. The incubator was chosen because it is representative of cell cultures. 3.5. Tearing tests
Fig. 1. Sample dimensions in mm for the long term cyclic test (a) and tearing test (b).
Tearing tests were performed following a protocol based on the original work by Rivlin and Thomas [23]. A cut of 20 mm was created on one of the lateral edges of the test piece (Fig. 1), which was loaded up to the point of crack propagation with a nominal strain rate of 0.3%/s. Fig. 3 shows an example of a sample at the
L. Bernardi et al. / Polymer Testing 60 (2017) 117e123
119
Fig. 2. Testing set-up equipped with the environmental chamber and the temperature coil, with a tearing test sample mounted.
Fig. 3. Tearing test sample at the beginning of the test (left) and at the time point of crack propagation (right). The blue rectangle delimitates the area were the optical flow tracker is applied to extract the local deformation. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)
beginning of the test and at the time point of crack propagation. For each material, 3 cut specimens were tested. Additionally a sample without cut was tested under the same conditions, in order to provide a reference for the calculation of the strain energy at the time point of crack propagation, as described in Refs. [23] and [24]. The stretch at which the crack starts to propagate was determined based on a visual inspection of the recorded camera images. This indicates that crack propagation often occurred well before complete sample rupture or before visible changes in the loaddisplacement curves. The local state of deformation was measured from optical flow tracking in the region indicated in Fig. 3. These data were used to calculate the tearing energy by integrating the first Piola-Kirchoff stress over the local stretch of
the reference sample up to the critical stretch value (W0). The tearing energy was calculated as G ¼ L* W0 [24], where L* is the length of the sample in the unstrained state. To correct for slippage, local values of strain were used. Results are reported in terms of tearing energy and critical local stretch (mean ± standard deviation). The five materials were tested in dry conditions at room temperature, while the effect of temperature and wet environment was studied for RTV4420. For this material, tests were performed at 37 C in distilled water and the sample was stored in the corresponding liquid solution in the incubator for 24 h prior to the test.
120
L. Bernardi et al. / Polymer Testing 60 (2017) 117e123
3.6. Cytotoxicity tests Due to the envisaged application as implant materials, the cytotoxicity of the elastomers was evaluated according to [25] and [26]. Importantly, the chosen test is more sensitive than classical live/dead assays. It can detect toxicity levels which are compatible with cell survival yet significantly interfere with the cell metabolism and activity. In short, the 5 materials were coated with PolyL-Lysine (PLL) 0.01% solution (Sigma Aldrich, USA) and PC12 cells were cultured in RPMI-1640 culture medium supplemented with 10% horse serum and 5% FBS (all Sigma Aldrich, USA), 2 mM Lglutamine, 100 U/ml penicillin and 100 ml/ml streptomycin (all Life Technologies, USA). The day of the test, the cells were incubated with the different elastomer materials, previously placed inside the wells of 12 well plates, at a density of 3 104 cells/well (3 repetitions for each material) or seeded into wells of 12 well plates in direct contact with PLL-coated tissue culture Poly Styrene TCPS (6 control wells without materials). Cells were incubated at 37 C and 5% CO2 for 24 h, then washed 3 times with PBS, incubated again for 24 h with fresh culture medium, and finally stimulated with nerve growth factor at a concentration of 100 ng/ml (NGF, Sigma-Aldrich) in induction medium (RPMI medium with 5% horse medium and 1% FBS). In order to check the functionality of the nerve growth factor, cells were stimulated at the same time point in 3 wells of the control, while they were not stimulated in the last 3 control wells. Cells were incubated for 3 days before imaging. As described in Ref. [26], cell images were acquired using an inverted Nikon-Ti wide field microscope (Nikon, Japan) with an Orca R-2 CCD camera (Hamamatsu Photonics, Japan). 10 images per well were collected at random spots with a 20 magnification, 0.45 numerical aperture long distance air objective (Plan Fluor, Nikon, Japan), using the white field channel with a differential interference contrast (DIC) filter. The length of neurites generated by PC12 cells on stimulation with NGF was used to detect the presence of cytotoxic effect from the tested materials interfering with the neuronal differentiation mechanism. The length of neurites was measured using the NeuronJ plug-in of ImageJ [27e29]. Membrane protrusions with length smaller than 10 mm were excluded from this analysis. Statistical comparison of the neurite length was performed using a non-parametric Mann-Whitney test to identify variations between the tested samples. 4. Results According to [1], a different but consistent color is associated with each material, and is systematically used in each figure. 4.1. Effect of environmental conditions on deformation behavior Fig. 4 reports the results of the monotonic uniaxial tests of RTV 4420 performed at 37 C either in distilled water or in saline solution (average ± standard deviation, n ¼ 3). For comparison, the average curve previously reported in Ref. [1] for the same material is also shown in the figure. Note that, in this figure, to be consistent with the curve in Ref. [1], stress is reported as Cauchy stress (force per unit of deformed area). Two observations can be made. First, the difference between the samples stored and tested in saline solution and samples stored and tested in distilled water is insignificant. For this reason, and since the maintenance of a constant salt concentration for the long term tests is cumbersome due to water evaporation, the environmental conditions for the tests presented in this work are reproduced as distilled water at 37 C. Second, there is no difference between the material stored and tested at room temperature and dry environment and the material stored and tested in the different environmental conditions.
Fig. 4. Monotonic uniaxial stress-strain curves of RTV 4420 tested after 1 day in distilled water or saline solution at 37 C. For comparison, the nominal curve from Ref. [1] (room temperature and dry conditions) is also reported.
4.2. Long term cyclic tests Fig. 5 reports the first Piola-Kirchhoff stress decay over the measured 2200 000 cycles. The long term behavior of most materials indicates a stable response, while RTV 4528 and Sylgard 184 are characterized by an initial decrease in peak stress that stabilizes after approximately 500 000 cycles. RTV 4528 shows the largest decrease (around 25%), while for the other materials the decrease is always lower than 8%. No macroscopic failure of the samples was observed. Fig. 5 reports also the results for the sample tested in water at 37 C. This sample shows some progressive decay when compared to the same material tested in dry conditions at room temperature. Note, however, that the stress reduction is modest with a maximum of 7% and an average of about 3% at the end of the test.
4.3. Tearing tests The stretch at crack propagation and tearing energy provide an indication of the sensitivity of the elastomer to pre-existing defects. Crack-like defects might be related to the manufacturing or the mounting process, or arise due to extreme operating conditions. The tearing energy values can be used for a corresponding fracture mechanics analysis by assuming a specific crack size and position, and calculating the loading level required to cause crack propagation. Information on the critical stretch can help to classify the materials in terms of their brittleness and independently of their stiffness, whereas tearing energy is important in the case of force or pressure controlled loading conditions. Note that, while the tearing energy is a property of the material, the stretch at propagation depends on the particular geometry and the test configuration, see e.g. Ref. [24]. Here it is reported for comparison among the considered elastomers, since the same geometry was used for all the materials. Test results are reported in Fig. 6. They indicate that PDMS Sylgard 184 is significantly more brittle than the other materials, and cracks propagate at relatively small stretches (around l ¼ 1.1). RTV 4528 is the most resistant in terms of stretch at propagation, with critical values around l ¼ 1.75. For the other three materials, the behavior is similar and cracks propagate at a stretch of around l ¼ 1.3e1.4. Regarding the tearing energy, SMI performs best, with a tearing energy of around 920 J/m2, while PDMS Sylgard 184 has the lowest tearing energy of only 80 J/m2, i.e.
L. Bernardi et al. / Polymer Testing 60 (2017) 117e123
121
Fig. 5. Stress decay over 2200 000 cycles normalized with respect to the value at the beginning of the test.
Fig. 6. Fracture behavior of the tested materials, average and standard deviation for each material (n ¼ 3). Top: Stretch at propagation for samples of same dimensions. Bottom: Tearing energy.
one order of magnitude smaller. The results of RTV4420 samples tested in distilled water at 37 C are also reported in Fig. 6. Although both stretches at crack propagation and tearing energy are lower than the corresponding quantities tested in nominal conditions, the difference is small (below 2.5%) and, at least for the low number of samples considered, not statistically significant. 4.4. Cytotoxicity tests Cytotoxicity tests were performed as previously described [26] to detect potential interference with the delicate molecular
pathways mediating neuronal differentiation of PC12 cells upon NGF-stimulation [30]. In particular, the neurite length was measured as described in section 3.6, yielding a linear parameter which directly reports on the potential cytotoxicity mediated by the contact with the elastomeric substrates, as compared with standard culturing conditions on tissue culture polystyrene (TCPS). The data presented in Fig. 7 demonstrate that none of the tested substrates caused a significant reduction of neurite length as compared with the control. Altogether, these results confirm that the materials under study do not show cytotoxicity in vitro, thus supporting their further characterization for potential in vivo applications.
122
L. Bernardi et al. / Polymer Testing 60 (2017) 117e123
conditions, as shown in Fig. 4. Contrary to other materials such as PDMS, known for its variable mechanical properties due to temperature and time of curing [35], and also for its severe aging over time [20], RTV 4420 is very stable in terms of temperature of testing in the range of interest of this work. Furthermore, it does not show aging in the first two months from preparation (data not shown). The present results show that the polymer is fully crosslinked already at room temperature and, contrary to other fully crosslinked RTV materials, as in Ref. [36], it shows no temperature dependence of its stress-strain behavior in the range of temperature investigated that is representative of physiological conditions. The elastomers analyzed in the present work show no cytotoxic effects in vitro. Of course, in case of applications as a component of human implants (such as in devices that are in contact with blood), the materials would have to be specifically qualified. 6. Conclusions Fig. 7. NGF induced PC12 differentiation on different materials. Box spans from 1st to 3rd quartile, outliers are removed from the plot. Black lines indicate the medians. Sample size is indicated above each population. Note that for the control without NGF, only 5 neurites had a length greater than 10 mm.
5. Discussion In this paper, the characteristic energy for tearing, or tearing energy [23], was used as measure of the fracture properties of the elastomers among other possible quantities such as J-integral [31] or crack tip opening displacement [32]. This quantity allows prediction of the resistance to the (Mode I) propagation of existent crack-like defects for arbitrary loading conditions. It has been chosen due to its straight-forward experimental evaluation [31]. Typical values for silicone-based elastomers found in literature are in line with the results of the current work [33,34]. Our results highlight the brittleness of PDMS Sylgard 184: with 80 J/m2, it is the material with the lowest tearing energy among the tested elastomers. This material is widely applied for in vitro cell studies since it is easy to fabricate and transparent, thus allowing live imaging applications. However, its brittleness represents a severe limitation with respect to applications that involve mechanical loading to large strains since pre-existing defects might be expected to grow, possibly leading to catastrophic failure. PDMS Sylgard 186 might represent a viable alternative, since it shows elevated tearing energy as well as negligible stress decay over 2200 000 cycles. However, its viscosity might complicate the fabrication process; in fact, our experience shows that membranes of homogeneous thickness are more difficult to obtain with typical preparation methods. RTV 4528 shows the maximum stress decay in the long-term cyclic tests, continuing beyond the 200 cycles reported in Ref. [1] and reaching nearly 500 000 cycles. In this regard, considering a biomedical application for which a long term stable response is required, SMI and RTV 4420 seem more suitable candidate materials. In particular, RTV 4420 showed good toughness properties, stable long term cyclic behavior and, importantly, it can easily be fabricated and tailored to a specific shape required for a specific design of a biomedical device. The effect of wet conditions and temperature on the long term cyclic behavior and on the fracture properties is analyzed for RTV 4420 and no relevant differences can be observed with respect to the samples tested in nominal conditions. No previous data on the combined effect of these factors were found in literature. Interestingly, the material chosen for the environmental conditions study shows no dependence of the deformation properties on these
This work presents a characterization of five silicone-based elastomers in terms of cyclic response and fracture properties. Important for applications in biomedical devices, implants and bioreactors, the critical stretch and tearing energy of the elastomers were determined in corresponding tests, indicating for the latter differences of more than one order of magnitude among the elastomers. This evaluation highlighted in particular the brittleness of PDMS Sylgard 184, thus limiting its reliability in terms of mechanical integrity. Long term tests involving up to 2200 000 loading cycles were performed and the corresponding decay in peak stress was analyzed. One elastomer, RTV 4528, showed a significant decay in peak stress while all others displayed a stable response. The influence of “physiological” environmental conditions on the mechanical properties of RTV 4420 was investigated and no relevant differences could be observed. Finally, the non-cytotoxic characteristics of the elastomers were demonstrated, which are a prerequisite for their use in mechanobiological studies or as part of soft implants and biomedical devices. Acknowledgments This project is supported by the “Stiftung PROPTER HOMINES Vaduz/Fürstentum Liechtenstein”, the “Schwyzer-Winiker Stiftung” and the ETH Zurich Foundation. This work is part of the Zurich Heart project of Hochschulmedizin Zürich. References [1] L. Bernardi, R. Hopf, A. Ferrari, A.E. Ehret, E. Mazza, On the large strain deformation behavior of silicone-based elastomers for biomedical applications, Polym. Test. 58 (2017) 189e198. [2] X.Q. Brown, K. Ookawa, J.Y. Wong, Evaluation of polydimethylsiloxane scaffolds with physiologically-relevant elastic moduli: interplay of substrate mechanics and surface chemistry effects on vascular smooth muscle cell response, Biomaterials 26 (16) (2005) 3123e3129. [3] B.J. Bachmann, L. Bernardi, C. Loosli, J. Marschewski, M. Perrini, M. Ehrbar, P. Ermanni, D. Poulikakos, A. Ferrari, E. Mazza, A novel bioreactor system for the assessment of endothelialization on deformable surfaces, Sci. Rep. (6) (2016) 38861. [4] W.-M.P.F. Bosman, T.J. van der Steenhoven, D.R. Su arez, E.R. Valstar, A.C. de Vries, H.L.F. Brom, M.J. Jacobs, J.F. Hamming, The effect of injectable biocompatible elastomer (PDMS) on the strength of the proximal fixation of endovascular aneurysm repair grafts: an in vitro study, J. Vasc. Surg. 52 (1) (2010) 152e158. [5] G. Soldani, P. Losi, M. Bernabei, S. Burchielli, D. Chiappino, S. Kull, E. Briganti, D. Spiller, Long term performance of small-diameter vascular grafts made of a poly (ether) urethane-polydimethylsiloxane semi-interpenetrating polymeric network, Biomaterials 31 (9) (2010) 2592e2605. [6] P. Costa, G. Poinas, K.B. Naoum, K. Bouzoubaa, L. Wagner, L. Soustelle, M. Boukaram, S. Droupy, Long-term results of artificial urinary sphincter for women with type III stress urinary incontinence, Eur. Urol. 63 (4) (2013) 753e758.
L. Bernardi et al. / Polymer Testing 60 (2017) 117e123 [7] S. Necchi, D. Molina, S. Turri, F. Rossetto, M. Rietjens, G. Pennati, Failure of silicone gel breast implants: is the mechanical weakening due to shell swelling a significant cause of prostheses rupture? J. Mech. Behav. Biomed. Mater. 4 (8) (2011) 2002e2008. [8] B. Yu, S.-Y. Kang, A. Akthakul, N. Ramadurai, M. Pilkenton, A. Patel, A. Nashat, D.G. Anderson, F.H. Sakamoto, B.A. Gilchrest, et al., An elastic second skin, Nat. Mater. 15 (2016) 911e918. [9] “Zurich Heart,” 2015. [Online]. Available: http://www.hochschulmedizin.uzh. ch/de/projekte/zurichheart.html. [Accessed 03 May 2016]. [10] “Hybrid Membrane,” 2016. [Online]. Available: http://www.zurichheart.ethz. ch/hybridmembrane. [Accessed 03 May 2016]. [11] “Swanson Finger Joint Implants,” ©2016. [Online]. Available: http://www. wright.com/products-upper/swanson-finger-joint-implants. [Accessed 15 June 2016]. [12] “NeuFlex MCP/PIP Finger Joint Implants,” ©2015. [Online]. Available: http:// www.fingerreplacement.com/docs/Finger/Products/Info/neuflex.html. [Accessed 15 June 2016]. [13] T.J. Joyce, Currently available metacarpophalangeal prostheses: their designs and prospective considerations, Expert Rev. Med. Devices 1 (2) (2004) 193e204. [14] T.J. Joyce, A. Unsworth, A literature review of “failures” of the Swanson finger prosthesis in the metacarpophalangeal joint, Hand Surg. 7 (01) (2002) 139e146. [15] I. Schoen, W. Hu, E. Klotzsch, V. Vogel, Probing cellular traction forces by micropillar arrays: contribution of substrate warping to pillar deflection, Nano Lett. 10 (5) (2010) 1823e1830. [16] S. Fusco, V. Panzetta, V. Embrione, P.A. Netti, Crosstalk between focal adhesions and material mechanical properties governs cell mechanics and functions, Acta biomater. 23 (2015) 63e71. [17] G. Fois, M. Weimer, T. Busch, E.T. Felder, F. Oswald, G. von Wichert, T. Seufferlein, P. Dietl, E. Felder, Effects of keratin phosphorylation on the mechanical properties of keratin filaments in living cells, FASEB J. 27 (4) (2013) 1322e1329. [18] G.L. Polyzois, A. Hensten-Pettersen, A. Kullmann, An assessment of the physical properties and biocompatibility of three silicone elastomers, J. Prosthet. Dent. 71 (5) (1994) 500e504. [19] V.A. Chalian, R.W. Phillips, Materials in maxillofacial prosthetics, J. Biomed. Mater. Res. 8 (4) (1974) 349e363. [20] R. Hopf, L. Bernardi, J. Menze, M. Zündel, E. Mazza, A.E. Ehret, Experimental and theoretical analyses of the age-dependent large-strain behavior of Sylgard 184 (10: 1) silicone elastomer, J. Mech. Behav. Biomed. Mater. 60 (2016) 425e437.
123
[21] M. Hollenstein, A.E. Ehret, M. Itskov, E. Mazza, A novel experimental procedure based on pure shear testing of dermatome-cut samples applied to porcine skin, Biomech. Model. Mechanobiol. 10 (5) (2011) 651e661. [22] M. Pensalfini, A.E. Ehret, S. Stüdeli, D. Marino, A. Kaech, E. Reichmann, E. Mazza, Factors affecting the mechanical behavior of collagen hydrogels for skin tissue engineering, J. Mech. Behav. Biomed. Mater. 69 (2017) 85e97. [23] R.S. Rivlin, A.G. Thomas, Rupture of rubber. I. Characteristic energy for tearing, J. Polym. Sci. 10 (3) (1953) 291e318. [24] M. Pharr, J.-Y. Sun, Z. Suo, Rupture of a highly stretchable acrylic dielectric elastomer, J. Appl. Phys. 111 (10) (2012) 104114. [25] A. Ferrari, M. Cecchini, M. Serresi, P. Faraci, D. Pisignano, F. Beltram, Neuronal polarity selection by topography-induced focal adhesion control, Biomaterials 31 (17) (2010) 4682e4694. [26] T. Vasileiou, D. Foresti, A. Bayram, D. Poulikakos, A. Ferrari, Toward contactless biology: acoustophoretic dna transfection, Sci. Rep. 6 (2016). [27] C.A. Schneider, W.S. Rasband, K.W. Eliceiri, et al., NIH Image to ImageJ: 25 years of image analysis, Nat. methods 9 (7) (2012) 671e675. [28] J. Schindelin, I. Arganda-Carreras, E. Frise, V. Kaynig, M. Longair, T. Pietzsch, S. Preibisch, C. Rueden, S. Saalfeld, B. Schmid, et al., Fiji: an open-source platform for biological-image analysis, Nat. methods 9 (7) (2012) 676e682. [29] E. Meijering, M. Jacob, J.-C.F. Sarria, P. Steiner, H. Hirling, M. Unser, Design and validation of a tool for neurite tracing and analysis in fluorescence microscopy images, Cytom. Part A 58 (2) (2004) 167e176. [30] A. Ferrari, P. Faraci, M. Cecchini, F. Beltram, The effect of alternative neuronal differentiation pathways on PC12 cell adhesion and neurite alignment to nanogratings, Biomaterials 31 (9) (2010) 2565e2573. [31] R. Sto cek, T. Horst, K. Reincke, Tearing energy as fracture mechanical quantity for elastomers, Adv. Polym. Sci. 275 (2017) 361e398. [32] K. Reincke, W. Grellmann, R. Lach, G. Heinrich, Toughness optimization of SBR elastomerseuse of fracture mechanics methods for characterization, Macromol. Mater. Eng. 288 (2) (2003) 181e189. [33] L.C. Yanyo, F.N. Kelley, Effect of chain length distribution on the tearing energy of silicone elastomers, Rubber Chem. Technol. 60 (1) (1987) 78e88. [34] G.D. Genesky, C. Cohen, Toughness and fracture energy of PDMS bimodal and trimodal networks with widely separated precursor molar masses, Polymer 51 (18) (2010) 4152e4159. [35] M. Liu, J. Sun, Q. Chen, Influences of heating temperature on mechanical properties of polydimethylsiloxane, Sens. Actuators A Phys. 151 (1) (2009) 42e45. [36] T. Rey, G. Chagnon, J.-B. Le Cam, D. Favier, Influence of the temperature on the mechanical behaviour of filled and unfilled silicone rubbers, Polym. Test. 32 (3) (2013) 492e501.