Article Cite This: Macromolecules XXXX, XXX, XXX−XXX
Effect of Freeze/Thaw Process on Mechanical Behavior of DoubleNetwork Hydrogels in Finite Tensile Deformation S. Shams Es-haghi, Morgan B. Mayfield, and R. A. Weiss* Department of Polymer Engineering, The University of Akron, 250 S. Forge St., Akron, Ohio 44325-0301, United States ABSTRACT: Tensile mechanics of double-network (DN) hydrogels synthesized from 2acrylamido-2-methylpropanesulfonic acid (AMPS) and acrylamide (AAm) with different crosslink density of the first networks was studied after applying the freeze/thaw (F/T) process. When a DN hydrogel which exhibits necking upon applying a tensile load undergoes F/T process, necking is suppressed and strain hardening is observed in a tensile deformation. The area under the stress−strain curve which measures the work of deformation and is a criterion of toughness also is reduced after the F/T process. The reduction of toughness in a DN hydrogel with more brittle first network is more pronounced. The toughness of samples reduced very markedly after the first F/T cycle; however, the toughness of the samples was very close after the second and third F/T cycles. The reduction of mechanical properties of DN hydrogels upon applying the F/T process is attributed to the damage induced by freezing of water to the internal structure of hydrogel by breaking the chemical bonds. The water content of DN hydrogels increased after the F/T process which is consistent with the hypothesis of damage induced by freezing of water. DN hydrogels immersed in liquid nitrogen for a fast freezing process shattered into small pieces, and cracks appeared on the pieces.
■
INTRODUCTION In 2003, Gong et al.1 described a new hydrogel architecture that produced extraordinarily tough materials, which they termed a double-network (DN) hydrogel. DN hydrogels are generally prepared by a two-step sequential free-radical polymerization of a soft neutral polymer network within a more highly crosslinked polyelectrolyte network, and DN hydrogels were originally believed to be interpenetrating polymer networks (IPN) if both polymers are cross-linked and semi-IPNs (SIPN) if the second polymerization was of a linear polymer. Nakajima et al.2 first questioned the presumption of a IPN or SIPN microstructure, and Shams Es-haghi et al.3,4 experimentally demonstrated that the second polymer actually grafts, at least in part, to the skeleton of the first network. Therefore, depending on whether a cross-linking agent is used in the second polymerization step or not, the actual microstructure of a tough DN hydrogel is a pseudo-IPN or pseudo-SIPN, respectively, where the use of the prefix pseudo denotes connectivity of the two networks. Shams Es-haghi et al.4 also showed that covalent connectivity between the two networks was essential for achieving toughness of a DN hydrogel. Despite their high water content (often >90 wt %), DN hydrogels possess excellent strength and toughness and undergo large tensile and compressive deformations.5 The excellent mechanical properties of DN hydrogels with large water content, however, raise the question of how a DN hydrogel is affected by a freeze/thaw (F/T) cycle. For a conventional single network hydrogel, the structure and mechanical properties may change after undergoing a F/T cycle. For example, freezing of the water within a hydrogel may change the porosity of the gel,6 produce irreversible phase separation,7 compromise the mechanical properties,8 or even fracture the gel (Figure 1). Fracture can occur because of the expansion of water when it freezes and increases the extension © XXXX American Chemical Society
of the already highly stretched network chains that may rupture covalent bonds.
Figure 1. Hydrogel produced from 2-acrylamido-2-methylpropanesulfonic acid (AMPS): (a) before F/T process; (b) after F/T process. The formulation of the sample is AMPS(1,1,2,9); see the nomenclature described in the Synthesis of Polymer Networks section.
One might reasonably expect the effect of freezing on the mechanical properties of a DN hydrogel to be different than for a conventional hydrogel because of the significantly high toughness of the DN hydrogelthat is, the ability of a DN hydrogel to dissipate energy that usually produces fracture of a conventional hydrogel. Thus, the question addressed by the study reported herein was what is the effect of freezing and thawing (F/T) cycles on the mechanical properties of DN hydrogels. Any effect of freezing on the structure or properties of the hydrogel could have important implications including Received: November 13, 2017 Revised: January 2, 2018
A
DOI: 10.1021/acs.macromol.7b02418 Macromolecules XXXX, XXX, XXX−XXX
Article
Macromolecules
photoinitiator with respect to monomer, (3) the mol % of crosslinking agent with respect to monomer, and (4) UV dose used for the reaction (i.e., intensity × time of exposure). Thus, AMPS(1,1,2,9) corresponds to the polymerization of a 1 M AMPS solution using 1 mol % OXGA, 2 mol % MBAA, and a UV dose of 9 J/cm2. Two DN hydrogels, DN1 and DN2, were synthesized in this study. The only difference between these two DN hydrogels was the concentration of cross-linking monomer used in the first polymerization step. Table 1 shows the details of recipes used for synthesizing hydrogels. Freeze/Thaw (F/T) Process. The procedure for a F/T cycle is shown schematically in Scheme 3. The DN hydrogels were cut into rectangular bars and placed in a Petri dish that was then kept in a freezer at −15 °C for 1 h to allow the water to freeze. When the water froze, the samples became opaque white (Figure 2) due to light scattering by the ice crystals formed. The frozen samples were then allowed to thaw at room temperature until the ice melted, and the sample again became transparent and soft. Following this F/T process, the samples were kept in DI water and the swelling was allowed to reequilibrate. The procedure was repeated up to three times to assess the effect of cycling the F/T process on the mechanical behavior of the DN hydrogels. The notation used to represent samples that have undergone F/T cycles is DNx-FTy, where x represents the particular DN hydrogel used and y corresponds to the number of cycles that sample had undergone. Therefore, DN1-FT2 represents an AMPS(1,1,2,9)/AAm(2,0.01,0.01,97), i.e., DN1, that experienced two F/T cycles. Rheology Measurements. Linear viscoelasticity measurements were made on the single-network AMPS hydrogels. Oscillatory shear experiments were performed with a TA Instruments ARES-G2 rheometer equipped with 25 mm parallel serrated plates covering a frequency range of 0.1−100 rad/s. The linear response region for the dynamic experiments was determined with a strain sweep. Tensile Testing. Uniaxial tensile tests were performed with an Instron 5567 universal testing machine equipped with a 100 N load cell, using a constant crosshead speed of 50 mm/min at room temperature. Rectangular samples with length = 40 mm, width = 10 mm, thickness = 2−3 mm, and a gauge length of 30 mm were used. Sandpaper was used to prevent slippage of the samples in the grips. The tensile data are reported as engineering stress versus stretch ratio, where the engineering stress, σe, was obtained by dividing the force by the original cross-sectional area of the specimen and the stretch ratio, λ ≡ L(t)/L0, where L(t) was the instantaneous sample length and L0 was the original sample length. Note that λ = εe + 1, where εe is the engineering tensile strain.
biomedical applications (e.g., drug delivery where a drug-loaded hydrogel may be freeze-dried or cryopreservation)9 or other technologies where absorbed water may freeze, such as agriculture10 or food preservation.11 In all these applications, it is important that hydrogel maintains its shape and mechanical integrity upon freezing.
■
EXPERIMENTAL SECTION
Materials. 2-Acrylamido-2-methylpropanesulfonic acid (AMPS) and acrylamide (AAm) were purchased from Sigma-Aldrich Chemical Co. and used as received. The cross-linking agent, N,N′-methylenebis(acrylamide) (MBAA), was purchased from Sigma-Aldrich Chemical Co. and was recrystallized from ethanol. A photoinitiator, 2-oxoglutaric acid (OXGA), was obtained from Fluka Chemical Co. and used as received. Synthesis of Polymeric Networks. DN hydrogels were synthesized by a two-step sequential free-radical polymerization. The first network was prepared by adding OXGA and MBAA to a 1 M solution of AMPS in deionized (DI) water. Dry nitrogen gas was bubbled through the reaction mixture for 5−10 min to remove oxygen, and the solution was injected into a glass mold made of two parallel glass slides, which was then exposed to a 365 nm ultraviolet (UV) light source. The chemistry of the AMPS polymerization to produce a water-swollen AMPS network is shown in Scheme 1. The AMPS gel
Scheme 1. Synthesis of AMPS Network
was then immersed into a 2 M solution of AAm in DI water containing the photoinitiator and cross-linking monomer, which had already been deoxygenated with N2 gas, until an equilibrium swelling was achieved. The AAm-swollen AMPS gel was then placed between two parallel glass slides and exposed to 365 nm UV light to produce the second, AAm, network (see Scheme 2). The resultant DN hydrogel sample
■
RESULTS AND DISCUSSION
Only pseudo-IPN-type DN hydrogels were used to study the effects of F/T cycles because the mechanical properties of pseudo-SIPN hydrogels change over time during storage in water due to the diffusion of physically trapped polymer chains out of the hydrogel.12 Figure 3a compares the tensile deformation of two pseudo-IPNs, AMPS(1,1,2,9)/ AAm(2,0.01,0.01,97) and AMPS(1,1,4,9)/AAm(2,0.01,0.01,97), made from different first networks. Because the first networks were brittle, it was not possible to run a static tensile experiment on them. Instead, the modulus of the AMPS networks was measured by oscillatory shear experiments (Figure 3b), which shows that the network prepared with the higher concentration of cross-linking monomer, AMPS(1,1,4,9), had a dynamic modulus, G′, about 3 times that of AMPS(1,1,2,9). That result indicates that AMPS(1,1,4,9) had the higher cross-link density and, as a consequence, was more brittle than AMPS(1,1,2,9). The initial slope of the stress− strain curve for the DN hydrogels in Figure 3a, which reflects the modulus of the first, highly cross-linked network, is consistent with the G′ data in Figure 3b. That is, the initial slope of the tensile data for the pseudo-IPN hydrogel made
Scheme 2. Synthesis of Poly(acrylamide) Gel
was immersed in DI water, which was washed a number of times with fresh water to remove any unreacted monomer. Two UV intensities, 15 and 3 mW/cm2, were used for the photopolymerization reactions, and the radiation exposure times for each intensity were 10 min (dose = 9 J/cm2) and 9 h (dose = 97 J/cm2), respectively. Each individual network is described by the name of the monomer used, followed by four numbers in parentheses that indicate (1) the molar concentration of monomer in DI water, (2) the mol % of B
DOI: 10.1021/acs.macromol.7b02418 Macromolecules XXXX, XXX, XXX−XXX
Article
Macromolecules Table 1. Formulations of the DN Hydrogels Synthesized in This Study AMPS gels (pphm)a
a
AAm gels (pphm)a 2
notation
DN hydrogel
OXGA
MBAA
UV dose (J/cm )
OXGA
MBAA
UV dose (J/cm2)
DN1 DN2
AMPS(1,1,2,9)/AAm(2,0.01,0.01,97) AMPS(1,1,4,9)/AAm(2,0.01,0.01,97)
1 1
2 4
9 9
0.01 0.01
0.01 0.01
97 97
Parts per hundred parts monomer.
from AMPS(1,1,4,9), DN2, is higher than for the DN1 made from AMPS(1,1,2,9). The brittleness of the first network, i.e., the skeleton of the DN hydrogel, was demonstrated by a simple experiment. A water-swollen DN hydrogel was locally damaged by simply squeezing the hydrogel between one’s fingers. That procedure fractures covalent bonds in the first network, which is accompanied by an increase in swelling of the damaged part of the sample due to a local decrease of the cross-link density (Figure 4). Catastrophic fracture of the DN hydrogel, however,
Scheme 3. Freeze/Thaw Cyclic Process
Figure 4. A DN hydrogel damaged locally by squeezing the sample between two fingers. The damaged part absorbs more water due to the reduced cross-link density in the damaged zone, which is identified by the arrow.
did not occur, which is attributed to the remaining structural and mechanical integrity of the second network, which can support the load following the energy dissipation by local fracture of the first network. Figure 5 shows representative uniaxial tensile data for the DN1 and DN2 hydrogels that indicate the F/T process significantly changes the mechanical properties of the DN hydrogels. The original DN1 and DN2 hydrogels exhibited necking in a tensile test. However, after a F/T process, necking was suppressed and the onset of plastic deformation occurred by strain hardening. The modulus (initial slope of stress− deformation curve) and the toughness, as measured by the work of deformation13 (i.e., the area under the stress−
Figure 2. DN1 hydrogel samples: (a) before freezing; (b) after freezing.
Figure 3. (a) Engineering tensile stress versus stretch ratio for pseudo-IPNs made from first AMPS networks with different cross-linking density (see Table 1). (b) Storage modulus versus angular frequency for the AMPS networks used to synthesize the pseudo-IPNs in (a). Symbols for AMPS network corresponds to DNx in (a). C
DOI: 10.1021/acs.macromol.7b02418 Macromolecules XXXX, XXX, XXX−XXX
Article
Macromolecules
Figure 5. Engineering tensile stress versus stretch ratio for DN1 and DN2 hydrogels: (a) DN1 hydrogel in its pristine form and after F/T cycles; (b) DN2 hydrogel in its pristine form and after F/T cycles; Black circles: pristine DN; blue circles: DN after first F/T cycle; red circles: DN after second F/T cycle; green circles: DN after third F/T cycle.
Figure 6. Work of deformation (a) and elastic modulus (b) of DN hydrogels before and after F/T cycles. Red columns: DN1 hydrogel; blue columns: DN2 hydrogel. DN1 and DN2 represent the pristine hydrogels, and the numbers indicate the number of F/T cycles. The elastic modulus values were found from the linear elastic region in stress−stretch curves.
deformation curve), decreased with each successive F/T cycle for both DN hydrogels shown in Figure 5. The work of deformation for all the samples was calculated by integration of stress−strain curves, and the results are summarized in Figure 6a. Figure 6b compares the elastic moduli of the DN hydrogels obtained from the elastic region of mechanical data in Figure 5. The first F/T cycle had the largest effect on toughness, and each successive F/T cycle had a diminishing effect. Note that for the DN2 hydrogel, which had the more brittle first network (see Figure 3b), a larger reduction of the work of deformation was observed after first F/T cycle. The decreases in the modulus and toughness of the hydrogels after F/T cycles are consistent with the breaking of covalent bonds within the chemical networks. An earlier study3 of the mechanics of DN hydrogels indicated that the initial slope of the stress−deformation curve reflected the elastic contribution of the first, brittle network. Therefore, the decrease in the modulus of the DN hydrogel by a F/T cycle is due to the reduction of the cross-link density of the brittle network. Owing to the covalent attachment of some of the polymer chains of the second network to the first network, it is also possible that some parts of the second network break by freezing. Since the second network plays the dominant role at high stretch ratios,3 the reduction in the stretch ratio at the breaking point of DN hydrogels after F/T cycles might be attributed to the breakage of the polymer chains of the second network. If this hypothesis of damaged induced by freezing of water to the structure of hydrogel is correct, it is expected that water content of samples increases in their new equilibrium states after the F/T process. Table 2 shows the equilibrium water content and elastic modulus values of DN1 and DN2
Table 2. Water Content and Elastic Modulus of DN Hydrogels before and after Successive F/T Cycles hydrogel DN1 DN1-FT1 DN1-FT2 DN1-FT3 DN2 DN2-FT1 DN2-FT2 DN2-FT3
water content (wt %)
elastic modulus (kPa)
± ± ± ± ± ± ± ±
97.20 66.70 50.90 47.00 358.20 268.70 206.70 205.40
91.63 92.80 93.11 93.33 89.20 90.23 90.53 91.10
0.18 0.01 0.02 0.04 0.13 0.29 0.14 0.20
samples before and after each F/T cycle. Note that each sample was re-equilibrated with water before the next F/T cycle. Increases in the water concentration after each F/T cycle were small but significant based on measurements of several samples of each gel. The increasing water absorption is consistent with the conclusion that F/T cycles resulted in breaking chemical bonds and decreased the cross-link density of the first network of the DN hydrogels. The disappearance of necking after the first F/T cycle (Figure 5) was also consistent with failure of the first, brittle network and stress transfer from the first network to the second one.3 Reference 3 showed that any predamaging of a DN hydrogel prior to tensile testing reduced the severity of strain localization and prevented necking. In the present study, the freezing of water provided the predamaging of the gel. That is, the expansion of the water upon freezing stressed the brittle network and broke bonds, which reduced the cross-link density and the intensity of strain localization. The effect of increasing the rate at which the water was frozen was evaluated with rapid freezing process achieved by D
DOI: 10.1021/acs.macromol.7b02418 Macromolecules XXXX, XXX, XXX−XXX
Article
Macromolecules immersing the DN hydrogel in liquid nitrogen for 1 min. In that case, both hydrogels shattered into to small pieces, and during the subsequent process of thawing the sample to room temperature, the small pieces cracked (Figure 7). The more
Figure 7. Result of fast freezing of the DN1 hydrogel: (a) frozen DN1 hydrogel after immersing in liquid nitrogen for 1 min; (b) liquid nitrogen frozen DN1 hydrogel during thawing process. Arrows show cracks in the sample.
rapid expansion of the water increased the rate of stress application to the sample, which produced catastrophic failure of both of the hydrogels. Therefore, it was not possible to perform tensile test on the samples after freezing in liquid nitrogen. An important distinction between the frozen DN hydrogels and frozen conventional single-network hydrogels is that the former can continue to sustain a tensile load during the thawing process. A single network hydrogel failed catastrophically even when the freezing step was slow, and as a result the measurements of the mechanical properties of the singlenetwork hydrogel after a single F/T cycle were not possible. Although the DN hydrogels were brittle in the frozen state, even when the thawing process was incomplete (i.e., when the interior of the gel was still frozen), they were able to sustain a substantial tensile load and undergo large deformation (Figure 8). The frozen parts are the white regions in the pictures for varying stretch ratio in Figure 8. The sample sustained a tensile load and underwent large deformations. Although the sample was not soft and some parts of the sample were frozen, the sample did not fail at the stretch ratio of 11 shown in the far right picture of Figure 8. By increasing the stretch ratio, the frozen parts became transparent, indicating that the ice melts during deformation.
Figure 8. Stretching at room temperature of an incompletely thawed DN1 hydrogel following freezing. Each picture shows what the sample looked like at the indicated stretch ratio (λ). Note that the white portions of the sample are frozen parts.
single-network hydrogel. A DN hydrogel that exhibits necking upon applying a tensile load does not exhibit necking after the first F/T cycle, and the onset of plastic deformation occurs by strain hardening. Suppression of necking after a F/T process is consistent with the previously reported3 observation that predamaging the first network of a DN hydrogel can reduce the intensity of strain localization and consequently suppress necking. Applying a single F/T cycle significantly reduces the work of deformation of a DN hydrogel. However, the damage to the DN hydrogel upon subsequent F/T cycles diminishes with each cycle, and for the DN hydrogels used in this investigation, little if any changes in the mechanical properties occurred after the second F/T cycle. The F/T process did not completely break the first networkthe elastic region in mechanical behavior of DN hydrogels originates from the contribution of the first network; thus, once the mechanical properties become nearly constant by increasing the number of F/T cycles, there are contributions from both first and second networks. By increasing the cross-link density of the first polyelectrolyte network in a DN hydrogel, the reduction of mechanical properties due to freezing becomes more pronounced. Although freezing a single-network polyelectrolyte hydrogel results in extensive internal fracture and catastrophic failure, a frozen DN hydrogel can still sustain a substantial tensile load and extend to large stretch ratios when it is only partially thawed, i.e., when the interior of the sample is still frozen. The rate of freezing of the DN hydrogel is also important with regard to the mechanical response. Rapid freezing by immersing the gel in liquid nitrogen produced catastrophic failure of the gel and shattered it into small cracked pieces.
■
CONCLUSIONS Double-network (DN) hydrogels are an important class of tough chemically cross-linked hydrogels that are synthesized by polymerizing a water-soluble monomer inside a highly crosslinked polyelectrolyte network. The first polyelectrolyte network is a mechanically weak, brittle sacrificial network that fractures during tensile deformation. As the first network fails, the stress on the gel is transferred to a second, more ductile network, which can undergo large deformation. Freezing the water inside a DN hydrogel also produces stress on the material, since water expands when the liquid transforms to ice. As with direct mechanical loading of a DN hydrogel, the expansion of water upon ice formation damages the internal structure of the gel and breaks chemical bonds. The results are similar to those of tensile loading, in that the damage is restricted to the first, brittle network and the second, ductile network picks up the load. As a result, a F/T cycle reduces the mechanical properties of a DN hydrogel but does not produce catastrophic failure as occurs with freezing a conventional, E
DOI: 10.1021/acs.macromol.7b02418 Macromolecules XXXX, XXX, XXX−XXX
Article
Macromolecules
■
AUTHOR INFORMATION
Corresponding Author
*E-mail
[email protected] (R.A.W.). ORCID
S. Shams Es-haghi: 0000-0001-6659-2828 R. A. Weiss: 0000-0002-5700-6871 Present Addresses
S.S.: Birck Nanotechnology Center, Purdue University, West Lafayette, IN 47907. M.B.M.: Thermo Fisher Scientific Inc., Bedford, MA 01730. Notes
The authors declare no competing financial interest.
■
ACKNOWLEDGMENTS This work was funded by a grant from the Civil, Mechanical and Manufacturing Innovation (CMMI) Division within the Engineering Directorate of the National Science Foundation, CMI-1300212. M.B.M. was supported by an REU supplement from CMMI.
■
REFERENCES
(1) Gong, J. P.; Katsuyama, Y.; Kurokawa, T.; Osada, Y. DoubleNetwork Hydrogels with Extremely High Mechanical Strength. Adv. Mater. 2003, 15, 1155−1158. (2) Nakajima, T.; Furukawa, H.; Tanaka, Y.; Kurokawa, T.; Osada, Y.; Gong, J. P. True Chemical Structure of Double Network Hydrogels. Macromolecules 2009, 42 (6), 2184−2189. (3) Shams Es-haghi, S.; Leonov, A. I.; Weiss, R. A. On the Necking Phenomenon in Pseudo- Semi-Interpenetrating Double-Network Hydrogels. Macromolecules 2013, 46, 6203−6208. (4) Shams Es-haghi, S.; Leonov, A. I.; Weiss, R. A. Deconstructing the Double-Network Hydrogels: The Importance of Grafted Chains for Achieving Toughness. Macromolecules 2014, 47, 4769−4777. (5) Gong, J. P. Why Are Double Network Hydrogels so Tough? Soft Matter 2010, 6, 2583−2590. (6) Ricciardi, R.; D’Errico, G.; Auriemma, F.; Ducouret, G.; Tedeschi, A. M.; De Rosa, C.; Lauprêtre, F.; Lafuma, F. Short Time Dynamics of Solvent Molecules and Supramolecular Organization of Poly (Vinyl Alcohol) Hydrogels Obtained by Freeze/thaw Techniques. Macromolecules 2005, 38 (15), 6629−6639. (7) Gupta, S.; Pramanik, A. K.; Kailath, A.; Mishra, T.; Guha, A.; Nayar, S.; Sinha, A. Composition Dependent Structural Modulations in Transparent Poly(vinyl Alcohol) Hydrogels. Colloids Surf., B 2009, 74 (1), 186−190. (8) Lozinsky, V. I.; Damshkaln, L. G.; Kurochkin, I. N.; Kurochkin, I. I. Study of Cryostructuring of Polymer Systems. 33. Effect of Rate of Chilling Aqueous Poly(vinyl Alcohol) Solutions during Their Freezing on Physicochemical Properties and Porous Structure of Resulting Cryogels. Colloid J. 2012, 74 (3), 319−327. (9) Fowler, A.; Toner, M. Cryo-Injury and Biopreservation. Ann. N. Y. Acad. Sci. 2005, 1066, 119−135. (10) Rumich-Bayer, S.; Krause, G. H. Freezing Damage and Frost Tolerance of the Photosynthetic Apparatus Studied with Isolated Mesophyll Protoplasts of Valerianella Locusta L. Photosynth. Res. 1986, 8 (2), 161−174. (11) Ustun, N. S.; Turhan, S. Antifreeze Proteins: Characteristics, Function, Mechanism of Action, Sources and Application to Foods. J. Food Process. Preserv. 2015, 39 (6), 3189−3197. (12) Shams Es-haghi, S.; Weiss, R. A. Do Physically Trapped Polymer Chains Contribute to the Mechanical Response of a Host DoubleNetwork Hydrogel under Finite Tensile Deformation? Macromolecules 2017, 50, 8267−8273. (13) Tamarin, Y. Atlas of Stress-Strain Curves, 2nd ed.; ASM International: Materials Park, OH, 2002.
F
DOI: 10.1021/acs.macromol.7b02418 Macromolecules XXXX, XXX, XXX−XXX
