An Alkaline Based Method for Generating Crystalline, Strong, and Shape Memory Polyvinyl Alcohol Biomaterials

Abstract Strong, stretchable, and durable biomaterials with shape memory properties can be useful in different biomedical devices, tissue engineering, and soft robotics. However, it is challenging to combine these features. Semi‐crystalline polyvinyl alcohol (PVA) has been used to make hydrogels by conventional methods such as freeze–thaw and chemical crosslinking, but it is formidable to produce strong materials with adjustable properties. Herein, a method to induce crystallinity and produce physically crosslinked PVA hydrogels via applying high‐concentration sodium hydroxide into dense PVA polymer is introduced. Such a strategy enables the production of physically crosslinked PVA biomaterial with high mechanical properties, low water content, resistance to injury, and shape memory properties. It is also found that the developed PVA hydrogel can recover 90% of plastic deformation due to extension upon supplying water, providing a strong contraction force sufficiently to lift objects 1100 times more than their weight. Cytocompatibility, antifouling property, hemocompatibility, and biocompatibility are also demonstrated in vitro and in vivo. The fabrication methods of PVA‐based catheters, injectable electronics, and microfluidic devices are demonstrated. This gelation approach enables both layer‐by‐layer and 3D printing fabrications.


Samples sizes used for
-O. Designed PVA-Hs have the capability to be uniformly incorporated different nanomaterials. To prepare composite films, magnetic nanoparticles (MNPs), graphene or CNTs were added to the PVA solution, followed by casting and crosslinking. It is noted that the wt% of elements PVA and nanomaterials can be tuned depending on the required properties.
According to the Figure 1M, combining PVA with MNPs led to a stretchable magnetic membrane which can be used as a magnetic actuator. 0.1 g MNPs were mixed with 5 mL of PVA solution, sonicated and stirred, before casting on a petri dish to form a film. Afterward, the dried film was crosslinked using NaOH solution. Video S2 shows how magnetic membrane can dance by a moving magnet. Figure 1N shows graphene was uniformly distributed in the PVA network in a PVA/graphene membrane. Figure   1O displays a nanomembrane made with PVA/CNT solution (PVA with 10 mg/mL of CNT). The solution was poured on a dish, dried overnight, and made a homogenous film following by crosslinking. NaOH solution for 5 min, it was used for artificial muscle tests ( Figure 2B, 2C and Figure S7) and measuring the contraction force ( Figure 2C(3)). After soaking the strip in NaOH solution, it was manually stretched and was attached to a 0.5 kg weight reaching the final length of 38 cm ( Figure 2C (1,2)). Figure 2C(3). Similarly, after immersing the strip in NaOH solution, it was elongated to L1 =32 cm and positioned between tensile machine grips. After calibration, the strip was extended manually to reach 5 N preload, to mimic weight hanging. The force was recorded while the water was continuously added to the both sides of the PVA-H strip. Figure 3A-D. First solution: 10 wt% AgNP and 10 wt% PVA blended; second solution: 10 wt% PVA; third solution: 8.8 wt% CNT and 10wt% PVA blended. 0.5 g of each solution were casted on a small petri dish sequentially. First solution was casted and allow it to totally dry at room temperature, then, similarly second and third solutions were poured and air dried.

Preparation of PVA composite in
After drying, the rod with coated layers was soaked in NaOH solution for 15 min for crosslinking. The composite membrane was easily peeled off from the metal rod and washed with water.
Swelling ratio. Swelling ratio was calculated according to following formula: Swelling ratio= [Wf−W0/W0]; in which, Wf is the weight of PVA-H in fully swollen state and W0 denotes the one for dried state. Figure 3. To make small tubes, a metal rod was immersed in PVA solution and rotated in a customized motor with applied drying fan, creating a uniform layer of PVA coating. By using a high rotation speed, we can reach a uniform coating. The process of coating was repeated to reach the desired thickness. Afterwards, the dried PVA tube was crosslinked by 6M NaOH solution followed by washing with water.

Tubes and catheters made in
Measuring burst pressure of tubes and balloon diameter (Figure 3). Tubes were attached and sealed to a pipe and filled with water and connected to a compressor. With increasing the pressure, tubes were blown and the burst pressure, when they performed excessive expansion and ultimately explosion, were recorded. Figure 4A). PVA can serve as stabilizer to effectively distribute CNTs in the solution without coagulation. The amount of CNT can be tuned to obtain required conductivity. First, 100 mg PVA was dissolved in 5 mL water at 90 °C, followed by adding 50 mg CNT.

Preparation of an injectable electronic (
The mixture was further stirred and sonicated for 1 h. Afterwards, 400 mg PVA was dissolved in the mixture under stirring at 90 °C (to reach PVA concentration of 100 mg/mL), and the mixture was further shaken in moving bath at 60 °C overnight. This PVA-CNT ink was used as printing ink to print an electronic mesh on the back of a petri dish as seen in Figure 4A. When prints were dried, they were soaked in NaOH solution followed by immersion in water to physically crosslink PVA and CNT networks. Figure 4B. To make the substrate, 5 ml of 100 mg/mL PVA was poured in a petri dish and left overnight at room temperature to dry. Then, PVA-CNT was printed on the dried PVA substrate and let dry for several minutes, followed by immersion in NaOH solution and water in sequence to crosslink PVA. Figure 4C and D. 20 ml of PVA solution (100 mg/mL) was poured in a large Petri dish and dried overnight at the ambient temperature. Then, a 5 wt% alginate solution was prepared for printing on the PVA membrane as sacrificial material. After printing the alginate patterns, the dried PVA with 3D printed alginate was covered with another layer of PVA solution, followed by drying at room temperature. Then, the dried membrane was immersed in NaOH solution for 15 min. After the PVA gelation, the membrane was washed with water and the printed alginate was readily removed from edges be pressing the membrane.

Cytocompatibility tests
HaCaT and 3T3 cell lines were cultured with RPMI (Hyclone, USA) and 10% fetal bovine serum (Gibco, USA) in a 5% CO2 incubator at 37 °C. The HaCaT and 3T3 cells were seeded in 96-well plates at 5x10 3 /100 μL per well for measuring methabolic activity and in 24-well plates containing cover glasses at 1x10 4 /300 μL per well for live/dead assay. For toxicity assay, the culture medium was replaced with the medium extracted from PVA-H after 24 h of incubation, and the untreated medium was used as a blank control. The cells in the 96-well plate were assessed using a cell counting kit (CCK-8, Dojindo, Kyushu, Japan) test on days 1, 3 and 5 after seeding. At each time point, the medium was replaced with RPMI, and 10 μL of CCK-8 solution was added to each well, followed by incubation at 37 °C for 2 h. The mean optical density (O.D) value was determined at 450 nm using an enzyme-linked immunosorbent assay reader (Thermo Varioskan Flash, USA), and the cell viability of cells on cover glasses was tested by cell live/dead kit (Invitrogen, USA). The relevant reagent added to incubation at 37 ℃ for 20 min, and the resultant images were observed under laser confocal (Zeiss LSM780, Germany). were rinsed with deionized water for three times, and added 1ml deionized water to ultrasonic 10 minutes, take 100 μl plated and incubated at 37 ℃ for 14 h.

Hemocompatibility
Citrated fresh human blood (anticoagulation to blood radio, 1:9) and purified counting platelets (1x10 12 units / ml) were provided by Department of blood transfusion, Southwest Hospital, Chongqing, china. In hemolysis assay, PVA membranes (1 cm x 2.5 cm) were washed for five times with deionized water and then rinsed with normal saline. The samples were incubated at 37 ℃ for 30 min with 10 ml normal saline, and then 200 μl diluted whole blood (whole blood: saline; 8: 10) were added and incubated at 37 ℃ for 1 h. The normal saline was used as negative control and the distilled water was used as positive control.
The tubes were centrifuged for 10 min at 1500 rpm and then the optical density of the supernatant fluid was read by Microplate reader (Thermo, varioskan flash, USA) at 545 nm. The HR was calculated according to the following formula: ODs, ODn, ODp were the corresponding OD values of sample, negative control and positive control groups. Measurements were repeated three times for each group.

Subcutaneous embedding in vivo and H-E staining
The membranes were punched into 6-mm diameter discs and sterilized. After general anesthesia with 1% pento-barbital, two mice were operated on. The dorsal surface was shaved and cleaned with 75% alcohol.
Using surgical knife, three subcutaneous pockets were created through two separate neat openings of 1cm in length lateral to the spine. One PVA membrane was implanted into each of the first two pockets and anchored to surrounding connective tissue. The wounds were closed using sutures. In the third pocket, no membrane was used to serve as control. The mice were euthanized four weeks postoperatively to remove the skin tissues where the membranes were implanted. Tissue slices were embedded in paraffin, dyed and stained using H&E staining. To investigate the effect of dehydration on the mechanical properties of PVA-H, a PVA-H strip (0.3 mm thickness) was tested for five consequent times with a speed of 5 mm/min immediately after removing from water ( Figure S1B). It should be noted that the strip was slipped from the clips without rupture in every test, and thus the strip was immediately installed for the next tensile test. This film with 0.3 mm thickness is so strong that the slippage occurred in all tests ( Figure S1B) since the large tensile force exceeded the grip ability to restrain the samples. According to Figure S1B, the fourth and fifth test (the most dried) showed the highest strength (stress before slippage) and elastic modulus, while the first test (the least dried) showed the lowest strength. This observation implies that the mechanical strength of the PVA-H increases with dehydration (it becomes stiffer with loosing water). This is because the lower water content of a PVA-H leads to denser assembly of PVA chains [10] . The fact that the tensile test of the fifth test (160 min out of water) followed the same trend as the fourth test (150 min) indicates that the dehydration further than 150 min (already dried) has negligible effect on the mechanical performance of this strip and that the strip is reusable without performance change.

The effect of using different alkali metal hydroxides on mechanical properties, crystallinity and morphology of PVA-Hs
To investigate the behavior of the PVA-H during cyclic tensile test, another strip (0.1 mm thickness) was stretched-released for 10 cycles from 30% to 58% strain ( Figure S1C). During 10 cycles in 8 min, the continuous dehydration of the strip is evident the gradual increase of elastic modulus and mechanical strength. For example, the maximum stress of tenth cycle was 14% more than that of the first cycle as seen in Figure S1C. In figure S1D and E, the same strip was exposed to water vapor generated by a humidifier to prevent dehydration of the strip during the cyclic test. In contrast to the cyclic tensile test carried out in the room condition (cyclic tensile test), the strip maintained its mechanical properties when it was exposed to the humidifier as indicated by the mere difference between first and last cycles. The same sample was further tested in the presence of humidity to show its stress-strain behavior in the fully hydrated state ( Figure S1F).  characteristics of crystalline PVA. The peak at 2q » 20° is attributed to the diffraction of the (101) plane [11] . The change in the crystallinity degree is confirmed by the increase of the area of this diffraction peak shown in Figure S2, which is the result of larger crystallites in PVA-H as shown by Hong et al [12] . Figure S3. SEM images of a PVA-H film, which was freeze dried and gold coated before SEM.

Dissociation of PVA-H with heating in water
To investigate the temperature stability of the PVA-Hs, they were immersed in water with a certain temperature. A PVA-H was stable at temperatures below 65 °C, but it started to change at around 70 °C. To probe the effect of heating, a PVA-H was heated at 70 °C for 30 min, and then it was stored in water at room temperature overnight. It was observed that heating treatment substantially decreased the mechanical properties and increased the swelling ratio from 0.8 to ~9.5. The SEM images in Figure S4 show that the partial dissociation of PVA in water at 70 °C led to macro-sized porosity. Hence, the porosity and the swelling ratio of PVA-Hs can be also tuned by using heat treatment in water (~70 °C). The rate of PVA-H dissociation in water increases with raising the temperature higher than 70 °C, and it will be fully dissolved in water around 90 °C, indicating the recyclability of PVA-based devices.

Effect of the molecular weight on PVA-H formation
Our NaOH-induced crosslinking method was used for PVA polymers with different molecular weights including, 1) MW 205,000 (85% hydrolyzed), 2) MW 89,000-98,000 (99% hydrolyzed), 3) MW 31,000-50,000 (98-99% hydrolyzed), 4) MW 13,000-23,000 (98% hydrolyzed) and 5) MW 9,000-10,000 (80% hydrolyzed) (Sigma-Aldrich). We were able to obtain PVA hydrogels for PVA polymers with high molecular weights (i.e. 1, 2, 3), using our method. For the molecular weights less than 31,000, this gelation method was ineffective, i.e. PVA films were degraded in the washing steps. Accordingly, high molecular weight is an important parameter for applying this method. Figure S6A shows the optical images of PVA-Hs formed using PVA polymers with three higher molecular weights. In addition, this experiment reveals that using PVA with higher degree of hydrolysis does not inhibit the crosslinking, since groups #1 and #2 both have high molecular weights, but different hydrolysis ratios and we were able to make strong hydrogels using our approach. Figure S6B also presents the tensile behavior of PVA-Hs made with PVA polymers # 2 and 3. According to the tensile experiment, both molecular weights can result in PVA hydrogels with high mechanical strength more than 3 MPa.     1 Elements Au and Pd are the result of gold coating of the sample before conducting SEM. 2 The small amount of Na is the residual sodium ions after washing.

EDX study
According to the EDX spectra in Figure 3D, PVA-CNT layer contains highest amount of carbon element compared to other two layers. Ag peak is also appeared in third layer. Accordingly, the presented method enables synthesizing micro-and nano-scale membranes with different materials and properties.

Videos
Video S1: Demonstration of resistance of a PVA film against pointed metals.
Video S2: Demonstration of a dried PANI-coated PVA-H ribbon lifting 9 Kg. Demonstration of a PVA/MNPs hybrid hydrogel moving with a magnet. pH detection: the color of a PANI-coated PVA film changes from green to blue/purple when it is transferred between acidic and basic solutions.
Video S3: Demonstration of shape memory effect of PVA-H ribbon: water-assisted elongation recovery of the NaOH-treated and elongated ribbon.
Video S4: Demonstration of artificial muscle behavior of PVA-H ribbon: a weight was lifted by the contraction force generated in the NaOH-treated and elongated ribbon upon supplying water.
Video S5: Demonstration of PVA-H (not/treated with NaOH) performance carrying 9 Kg and recovering the permanent deformation upon adding water.
Video S6: Demonstration of injectable electronics: injection of a printed PVA/CNT hydrogel into water.
Video S7: Demonstration of stretchable electronics: PVA/CNT mesh printed on a PVA-H film.
Video S8: Demonstration of all-PVA microfluidic channels on a PVA-H membrane and tube.