Fabrication and characterization of sodium alginate-silicon nitride-PVA composite biomaterials with damping properties

Silicon nitride is utilized clinically as a bioceramic for spinal fusion cages, owing to its high strength, osteo-conductivity, and antibacterial effects. Nevertheless, silicon nitride exhibits suboptimal damping properties, a critical factor in mitigating traumatic bone injuries and fractures. In fact, there is a scarcity of spinal implants that simultaneously demonstrate proficient damping performance and support osteogenesis. In our study


Introduction
Many people experience common afflictions such as low back pain, as well as discomfort in the neck and arms in their daily lives.These symptoms may serve as indicators of lumbar and cervical degenerative disc diseases.As non-invasive treatments, both medication and physical therapy are commonly employed for alleviating aches and discomforts.Nevertheless, it is postulated that up to 30% of those affected with cervical degenerative disc disease, and 10%-20% with lumbar degenerative disc disease, may eventually require surgical intervention (Secretariat, 2006).In the context of surgical techniques, spinal fusion stands as the gold standard.Typically, patients undergo the implantation of an intervertebral fusion cage, which not only directly bears body weight but also maintains the height of the intervertebral and foraminal spaces, thus preventing nerve compression and aiding in the fusion of adjacent vertebrae (Du et al., 2017;Schleicher et al., 2008).
Bioceramics have recently gained significant attention in orthopedics (Du et al., 2018a;Shekhawat et al., 2021).Among these, silicon nitride stands out as a notable contender, and it is clinically utilized in the fabrication of spinal fusion cages (Arts et al., 2017;Sorrell et al., 2004).Its advantages include high strength, osteoconductivity, antibacterial effectiveness, and excellent imaging capabilities in computed tomography (CT) and magnetic resonance imaging (MRI) (McEntire and Lakshminarayanan, 2016;Pezzotti, 2019).A range of in vivo studies have substantiated its osteoconductive and osteogenic potential through diverse models, including rabbit tibiae (Silva et al., 2008), goat interbody fusion (Kersten et al., 2019), and frontal bone defects in minipigs (Neumann et al., 2006).Moreover, both CE and FDA regulatory bodies have granted approval for its use in interbody cages, drawing on a foundation of animal studies and adherence to established compliance standards (Kersten et al., 2014).However, it is worth noting that silicon nitride's elevated Young's modulus may induce stress shielding, potentially leading to bone atrophy (Shanjani et al., 2008).In our previous study (Du et al., 2021), the dynamic Young's modulus of dense silicon nitride was measured at 298.45 ± 1.08 GPa.Increasing the porosity could significantly lower the stiffness of the silicon nitride scaffold, for example, 70% porous silicon nitride had less than 9% of the stiffness of dense silicon nitride.However, its value (26.26 ± 1.23 GPa) still far exceeds the range for human cancellous bone (typically spanning from a few hundred MPa to 2-3 GPa), raising concerns about potential complications such as cage migration or subsidence (Mobbs et al., 2018).Additionally, the pure porous silicon nitride scaffold is brittle.It can withstand high compressive loads but would suddenly break upon failure, as observed during testing.Furthermore, damping plays a critical role in averting traumatic bone injuries and fractures (Dodge et al., 2012).If a high damping capacity can be achieved for spinal implants effectively, the implants can absorb vibration and dynamic stress applied to the implants or surrounding bone (Motoyama et al., 2017).Damping helps minimize the impact forces, reducing wear on implant materials, preventing implant failure, and potentially extending the lifespan of the implant.Our previous investigations have highlighted the inadequate damping properties of silicon nitride, as demonstrated by free decay vibration tests and dynamic loading assessments (Du et al., 2021), potentially affecting its efficacy in supporting bone formation and posing concerns regarding vibration and impact-related issues.
As bone tissue itself is a natural composite material, combining inorganic and organic materials to create composite scaffolds that mimic the mechanical properties and the dynamic response of natural bone appears to be a beneficial strategy, as evidenced by many studies in the bone tissue engineering field (Du et al., 2018b;Kavitha Sri et al., 2023;Jakus et al., 2016).In our previous study (Du et al., 2023a), we 3D printed PEEK/Silicon nitride scaffolds with a triply periodic minimal surface structure for spinal fusion implants.The scaffold with 30% porosity exhibited a compressive strength (34.56 ± 1.91 MPa) and elastic modulus (734 ± 64 MPa) similar to those of trabecular bone.However, during cyclic testing, we observed that while the scaffold exhibited a large hysteresis loop, it was permanently compressed and could not recover to its original height.Therefore, the damping properties of this scaffold are mainly derived from plastic deformation, which is not ideal for a spinal fusion cage as its main goal is to restore the disc height.Therefore, it is crucial to create a silicon nitride-based scaffold with the capability for effective energy absorption and rapid elastic recovery under quasi-static, dynamic and impact loading scenarios.
Hydrogels are three-dimensional (3D) cross-linked hydrophilic polymer networks with remarkable capacity to absorb and retain significant volumes of water (Parhi, 2017).In their complex structure, hydrogels display a balanced interplay of viscoelastic and poroelastic behaviour.This is due to the friction between polymer chains during reconfiguration and the interaction between polymers and water as the water moves through the polymer network (Wang et al., 2020).Therefore, they are known for their exceptional damping properties.Yet, the use of hydrogels as a scaffold is not without its challenges.Unlike living tissues, conventional synthetic hydrogels, constructed primarily of hydrogen bonds, often tend to deteriorate when exposed to surrounding water molecules (Hu et al., 2015;Fuchs et al., 2020).As a result of their inherent softness and brittleness, relying solely on hydrogels as load bearing scaffolds proves to be extremely challenging.
Several innovative strategies have emerged in recent years to achieve strong, tough and fatigue-resistant hydrogels (Zhao et al., 2021).These strategies include double-network, nanocomposite, and polyampholyte hydrogels.The introduction of robust sacrificial bonds has led to a substantial advancement in hydrogels, significantly enhancing both their strength and deformability (Hu et al., 2015;Xiang et al., 2020;Liu et al., 2023;Kurokawa et al., 2010).Sodium alginate (SA), a polysaccharide derived from brown algae, is of special interest.The mechanical strength of SA can be elevated through cross-linking with divalent cations such as Ca 2+ and Ba 2+ (Fu et al., 2019), and SA holds a diverse spectrum of medical applications, from the gentle encapsulation of cells to safeguarding delicate bioactive compounds (Hasani-Sadrabadi et al., 2020).Polyvinyl alcohol (PVA) is a semi-crystalline polymer produced through the hydrolysis of polyvinyl acetate (Wang and Hsieh, 2010).It is widely recognized that aqueous PVA solutions can transform into hydrogels through a process of crystallization during repeated freezing and thawing cycles, all achieved without the need for noxious chemical crosslinkers.As the PVA solution undergoes freezing, ice forms within the amorphous regions, catalysing the emergence of polymer crystallites.These crystallites serve as physical crosslinking points between PVA chains, ultimately leading to the formation of a water-insoluble hydrogel (Kim et al., 2015).
Today, we still lack spinal implants exhibiting exceptional damping performance, osteogenic and antibacterial properties.Therefore, the primary objective of this study was to craft composite scaffolds, specifically sodium alginate-silicon nitride/poly(vinyl alcohol) (SA-SiN/PVA), designed for applications in spinal implants and more broadly in the realm of bone tissue engineering.The scaffold integrates the strength of silicon nitride with a stiff and tough double network hydrogel system.This biomimetic scaffold successfully overcomes the limitations of traditional ceramics and hydrogels, offering a blend of osteogenic properties and shock-absorbing functionality.This paper presents a comprehensive exploration of the mechanical characteristics, dynamic responses, osteogenic effects, and antibacterial properties of the scaffolds.
(Salt Lake City, USA).All solutions were prepared with deionized water (Biopak Polisher, Milli Q).All products were used as received without further purification.

Preparation of SiN/PVA and SA-SiN/PVA scaffolds with different crosslinking
Firstly, sodium alginate-silicon nitride (SA-SiN) was produced.Sodium alginate (SA) was mixed with deionized water at a weight ratio of 5 wt%.Subsequently, silicon nitride (SiN) powder was added with a SA: SiN ratio of 1:4, and the solution was filled into molds.The samples were frozen at − 80 • C and then freeze-dried at − 50 • C and 0.57 mbar (Freeze Drying System VACO2, Faust Laborbedarf AG, Switzerland).
Poly(vinyl alcohol) (PVA) powder was dissolved in deionized water, and the mixture was placed into a 90 • C water bath (JB Aqua 5 Plus, Grant, UK) for 3 h until the PVA was completely dissolved, resulting in a 15 wt% PVA solution.SA-SiN was slowly added to the PVA solution under constant stirring until the solution became homogeneous.The mass ratio of SA-SiN to pure PVA was 60%:40%.Subsequently, the mixed solution was filled into cylindrical molds with a diameter of 7 mm and a height of 4 mm.The samples were frozen at − 20 • C (freezer from Liebherr Medline, Switzerland) for at least 2 h and thawed at room temperature for at least 2 h.This freeze-thaw cycle was repeated three times for the physical crosslinking of the PVA hydrogel.Following this process, the samples were placed in a − 80 • C freezer for 4 h and then transferred to the freeze dryer for 3 days.The dried samples were soaked in a 4 wt% CaCl 2 solution to chemically crosslink the hydrogel.Subsequently, the samples were soaked in deionized water and named as the SA-SiN/PVA scaffold.The fabrication process and crosslinking mechanism are illustrated in Fig. 1.
To compare scaffolds with different crosslinking systems, SiN/PVA scaffolds were prepared.In short, instead of using SA-SiN, only SiN powder was added to the PVA solution.The mass ratio of SiN to PVA was 60%:40%.The same freeze-thaw and freeze-drying procedures were then applied to achieve physical crosslinking.

Preparation of SA-SiN/PVA scaffold with different mass ratio
SA-SiN/PVA samples were prepared in different ratios using the same method as described above.The mass ratios of SA-SiN to PVA were 40%:60%, 50%:50%, 60%:40%, and 70%:30%, and they were named 40% SA-SiN/PVA, 50% SA-SiN/PVA, 60% SA-SiN/PVA, and 70% SA-SiN/PVA, respectively.For example, in the case of the 60% SA-SiN samples, 3 g of SA-SiN and 13.32 g of a 15 wt% PVA solution were used.Amounts and fractions of each material powder in the final scaffold are listed in Table 1.
For the control group (SA/PVA), a 25 wt% SA solution was prepared.This solution was mixed with a 15 wt% PVA solution and stirred until the mixture was homogeneous.After this, the mixture was filled into cylinder molds, and the freeze-thaw and freeze-dry processes were performed as described above.The dried control samples were also soaked in a 4 wt% CaCl 2 solution and later stored in deionized water.

Characterization of SA-SiN/PVA scaffold
Firstly, the SA-SiN/PVA scaffolds and control group samples were sectioned and analyzed with a stereo microscope (Olympus, Japan) to observe the porous inner structure of the samples.For the swelling ratio and water content, disc-shaped samples with a diameter of 7 mm and height of 4 mm were used.The weight of fully swollen scaffolds (Ws) and dry weight (Wd) were measured.The following formulas were used to calculate the values:

Mechanical evaluation
To compare the mechanical properties with different crosslinking mechanisms, compression tests of the SA-SiN/PVA and SiN/PVA scaffolds with dimensions of approximately Ø7 mm x 4 mm were performed on an Instron E10000 (Electroplus, United Kingdom).The compression was gradually increased to 67% strain at a displacement ramp rate of 1 mm/min, followed by an unloading process.Before the tests, the height and diameter of each sample were measured using a caliper.A 1 kN load cell was used, and the samples were not preloaded.Data from these compression tests were used to calculate stress and strain.The effective elastic modulus was determined between 0% and 20% strain.
For the mechanical properties of SA-SiN/PVA with different mass ratios of SA-SiN, Ø14 mm × 7 mm scaffolds were compressed up to 67% strain and unloaded with a loading rate of 1 mm/min.A stress-strain curve was plotted, and the area underneath the curve was used to calculate the dissipated energy.The height recovery ratio was assessed by measuring each sample before and one day after compression testing using a digital micrometer.The height recovery ratio was obtained by dividing the height after testing by the initial height.
In addition, a drop-tower rig (Lewin et al., 2020) was used to measure the mechanical response of the scaffold at different impact loading rates.A moving carriage (2.5 kg, adjustable with extra weights) was  mounted on supporting columns.A hemispherical indenter (ø ¼ 40 mm) was attached to the bottom of the carriage, and a piezoelectric force sensor (208C04, PCB Piezotronics, Inc., USA) recorded forces with a measurement range of 4.448 kN, sampled at 6000 frame per second (FPS).A high-speed camera (IDT Y8-S2, Integrated Design Tools Inc., USA) was employed to capture the displacement at 3000 FPS.Subsequently, the images were used to track the displacement of the circular markers on the carriage.Image processing was conducted with MATLAB 2022b (The MathWorks, USA), where the central positions of the markers were obtained for each frame.From the primary impact force-displacement data, the actual impact velocity and the corresponding impact energy were obtained.Furthermore, the peak load and energy absorbed during impact were calculated.

Biocompatibility and cellular response test
SA-SiN/PVA and SA/PVA scaffolds, with dimensions of Ø7 mm x 4 mm, were produced and sterilized by soaking in 70% ethanol overnight.The scaffolds underwent washing with phosphate-buffered saline (PBS) and were exposed to UV light (254 nm) for 4 h.Mouse pre-osteoblast cells (MC3T3-E1) were obtained from the University of Zurich, Switzerland, and were seeded onto the scaffolds (5 × 10 3 cells per scaffold).
For cell proliferation assessment, a PrestoBlue assay kit (P50200, ThermoFisher) was employed following the manufacturer's protocol.After cell attachment, the extraction was replaced with an assay medium containing 10% PrestoBlue solution and 90% growth medium (GM) composed of Minimum Essential Medium α (MEM α) without ascorbic acid, 10% fetal bovine serum, and 1% antibiotic-antimycotic.After 30 min of incubation, the medium was collected and replaced with GM.The collected assay medium (100 μL) was analyzed by fluorescence spectroscopy at an excitation wavelength of 560 nm and emission of 590 nm on days 1, 3, and 7.The percentage of reduction was calculated.
Actin/DAPI staining and quantification were performed to observe cell attachment and distribution.Cells (1 × 10 4 ) were seeded on the scaffolds and cultured for 3 days with GM.After washing with PBS, cells were fixed in 4% formaldehyde in PBS for 15 min.Subsequently, cells were permeabilized with 0.1% Triton X-100 in PBS for 15 min and blocked with a 0.1% Triton X-100 in PBS solution with 1% BSA for 45 min.For actin immunofluorescence staining, cells on the scaffolds were stained with Alexa Fluor 568 Phalloidin for 60 min.DAPI staining was performed by staining with DAPI working solution for 10 min.Samples were analyzed using a confocal laser scanning microscope (CLSM) (Zeiss, LSM 780 upright).
The expression of osteogenic-related genes (alkaline phosphatase (ALP), osteocalcin (OCN), collagen type I (COL1), and runt-related transcription factor 2 (RUNX2)) was measured by quantitative reverse transcription polymerase chain reaction (RT-qPCR), as described previously (Du et al., 2023a).MC3T3-E1 cells (2 × 10 4 ) were seeded on the scaffolds and cultured with osteogenic medium.Total RNA was extracted after 14 days using the RNeasy Plus Mini Kit (Qiagen Inc., USA).RNA purity and concentration were determined using a Nanodrop spectrophotometer.Extracted RNA was transcribed into cDNA, and RT-qPCR was performed using TaqMan gene expression assays.The gene expression levels were normalized against the housekeeping gene GAPDH, and the results were calculated using formula P = 2 -(normalized average Ct) × 100.

Anti-bacterial test
The antibacterial efficacy of the SA-SiN/PVA and SA/PVA scaffolds was assessed against Escherichia coli (E.coli) (Serra et al., 2013) and Staphylococcus aureus (S. aureus) (Stecher et al., 2010).Bacterial overnight cultures were cultivated in Lysogeny Broth (LB) medium at 37 • C with agitation.Subsequently, the cultures were diluted 1:100 in 1 ml fresh LB and incubated under static conditions in 24-well microtiter plates (TPP, Switzerland, 92024) for 1 and 2 days at 37 • C. Following the culture period, the samples underwent washing in PBS to eliminate loosely adherent bacteria.The quantity of remaining attached bacteria was determined using the PrestoBlue assay (ThermoFisher, P50200) by adhering to the manufacturer's protocol, as specified in section 2.6.

Statistical analysis
Each testing group had at least 3 replicates.The data are expressed as mean values ± standard deviations and were analyzed using one-way ANOVA followed by Tukey's post hoc test with GraphPad Prism 8.2.0 software (GraphPad Software Inc., USA).The level of statistical significance was set at p ≤ 0.05.

Characterization of SA-SiN/PVA scaffolds with different crosslinks
Scaffolds composed of SiN/PVA and SA-SiN/PVA with varying crosslinking densities were prepared.As depicted in Fig. 2, a single crosslink refers to the execution of only the freeze-thaw and freeze-dry processes to achieve physical crosslinking within the scaffold.In contrast, double crosslinking involves the placement of scaffolds in a calcium ion solution after the physical crosslinking process, resulting in the formation of chemical crosslinks within the scaffolds.The optical image of the scaffolds in Fig. 2(C) reveals variations in ultimate size despite using the same size mold, attributable to different swelling ratios.Additionally, SEM images (Fig. 2(D and E)) illustrate distinct microstructural morphologies, with SiN/PVA scaffolds exhibiting higher porosity and more homogeneous pores compared to the SA-SiN/PVA scaffold, which features heterogeneous pores.
The effective elastic modulus of the SA-SiN/PVA scaffold was 2.46 ± 0.24 MPa, significantly higher than that of the SiN/PVA scaffold (0.63 ± 0.12 MPa).Notably, SA-SiN/PVA with double cross-linking displayed a larger hysteresis loop, indicating enhanced energy dissipation capability.Furthermore, the SA-SiN/PVA scaffold exhibited a height recovery rate of over 95% after compression to 67% strain.

Characterization of SA-SiN/PVA with varying contents of SA-SiN
Because double-crosslinked SA-SiN/PVA scaffolds exhibited favorable mechanical properties, we further investigated whether the contents of SA-SiN influence the mechanical properties of the scaffold.The inner structures of SA-SiN/PVA with varying SA-SiN concentrations and SA/PVA scaffolds are presented in Fig. 3. Clearly, the silicon nitride powders were successfully incorporated into the SA/PVA hydrogel matrix, as evidenced by the homogeneous color difference between SA-SiN/PVA and SA/PVA scaffolds due to the grey SiN powder addition in Fig. 3 (F).All SA-SiN/PVA and control group samples have cylindrical dimensions ranging from Ø14 mm to Ø15 mm × 7 mm-9 mm.The samples demonstrated 2-3 times swelling capacity, with water content ranging between 50 and 70%.Additionally, the water content of SA-SiN/PVA samples decreased with increasing SA-SiN concentration.Notably, the control group exhibited a higher swelling ratio and water content compared to the PVA/SA-SiN samples.Table 2 provides the average values for each group with standard deviations.The porous architecture of all SA-SiN/PVA with different ratios and the control group is depicted in Fig. 3. Following the freeze-drying process and crosslinking with CaCl 2 , the pores of the 40-60% SA-SiN samples appeared larger in the core and smaller in the outer region.Moreover, the 70% SA-SiN and control group samples exhibited minimal pore presence.
The results of the compression tests conducted on SA-SiN/PVA and control group samples are depicted in the stress-strain curves shown in Fig. 4. The presence of hysteresis loops enabled the determination of dissipated energy by calculating the area within the loop.In Fig. 4(C),   X.Du et al. the dissipated energy for each ratio and the control group is presented.The 60% SA-SiN samples exhibiting the highest value, approximately 0.26 MJ/m 3 .Notably, all ratios, except the 70% SA-SiN samples, were significantly different from the control group.The effective elastic modulus, as illustrated in Fig. 4(B), reveals the highest elastic modulus was obtained for the 60% SA-SiN samples.The elastic modulus increased with an increasing amount of SA-SiN (from 40% to 60%) and subsequently decreased from 60% to 70%.Significant differences were observed between the control group and the 40% SA-SiN samples, as well as between the control group and the 60% SA-SiN samples.The height recovery ratio of various SA-SiN/PVA and control group samples, as represented in Fig. 4(D), exhibited a decreasing trend with an increasing amount of SA-SiN.For instance, the 60% SA-SiN samples demonstrated a recovery ratio of 96.45%.However, the height recovery ratio of PVA/SA-SiN samples was significantly higher than that of the control group.
The construction of a drop tower setup was detailed in a prior study (Lewin et al., 2020).This setup was subsequently utilized to conduct impact tests on the SA-SiN/PVA scaffolds and SA/PVA scaffolds.We tested the scaffolds under two levels of impact energy by adjusting the weight of the crosshead.The results of these tests (Fig. 5) demonstrated that the SA-SiN/PVA scaffold exhibited exceptional elasticity.At an impact velocity of 1.61 ± 0.02 m/s and an impact energy of 3.24 ± 0.07 J, the SA-SiN/PVA scaffold fully recovered to its original shape after undergoing approximately 75 ± 8% deformation during the first impact and experiencing a maximum force of 774 ± 59 N (5.03 ± 0.38 MPa).Furthermore, when the weight increased to 5 kg and the impact energy increased to around 7 ± 0.25 J, the sample started to fracture after experiencing a maximum force of 1834 ± 384 N (11.91 ± 0.25 MPa).In contrast, the SA/PVA group exhibited significantly lower impact resistance compared to the SA-SiN/PVA scaffold.At an impact velocity of 1.58 ± 0.01 m/s and an impact energy of 3.13 ± 0.04 J, the samples had already started to fracture with a maximum force of 484 ± 57 N (3.14 ± 0.37 MPa).In terms of energy absorption capacity, the SA-SiN/PVA scaffold absorbed 3.06 ± 0.06 J of energy (94.5%) during the initial impact with a smaller weight and 6.34 ± 1.27 J of energy (90.2%) during the initial impact with a higher load.Conversely, the SA/PVA scaffold absorbed 3.11 ± 0.06 J (99.2%) of energy during the initial impact with a smaller load, demonstrating a higher energy absorption rate during the initial impact, which, however, led to the rupture of the scaffold.

Cellular response of SA-SiN/PVA scaffolds
Mouse pre-osteoblast cells (MC3T3-E1) were used to investigate the cellular response to the PVA/SA-SiN and PVA/SA scaffolds.The PVA/ SA-SiN scaffold with 60% SA-SiN was chosen as the representative group as it showed the optimal mechanical properties among the PVA/ SA-SiN groups while a PVA/SA scaffold was chosen as the negative control group.The cells on the scaffolds were stained with phalloidin and DAPI to analyze the cellular morphology.The cells were well attached and spread on PVA/SA-SiN scaffolds including the pore area, suggesting strong focal adhesion and good biocompatibility (Fig. 6 (A)).PrestoBlue staining results showed that both PVA/SA-SiN and PVA/SA scaffolds supported steady cell proliferation (Fig. 6 (B)).However, the PVA/SA-SiN group demonstrated significantly higher osteogenic differentiation (assessed by ALP, RUNX2 and COL1 gene expression    analysis) compared to the PVA/SA group (Fig. 6 (C)).

Bacterial response of SA-SiN/PVA scaffolds
In the following experiments, we evaluated the antibacterial properties of SA-SiN/PVA scaffolds.The scaffolds were placed into 24-well plates filled with statically growing E. coli or S. aureus cultures in LB medium.After 24 and 48 h of incubation at 37 • C, we measured and compared the bacterial proliferation of E. coli and S. aureus in each group using the PrestoBlue assay.For E. coli, although bacterial attachment was slightly lower on the SA-SiN/PVA scaffold compared to the SA/PVA scaffold, no significant difference between these two groups was observed (Fig. 7).Interestingly, a significant inhibition of S. aureus growth was detected on the SA-SiN/PVA scaffold compared to the SA/ PVA scaffold at both 24 and 48 h.

Discussion
As a natural composite material, bone tissue is predominantly characterized by an organic phase, primarily oriented type I collagen fibrils, functioning as the matrix, and a mineral phase, hydroxyapatite crystal, serving as reinforcement.According to the principles of composite materials, the mineral phase contributes to the strength and stiffness of bone, while the organic phase influences its toughness and viscoelasticity (Wang and Feng, 2005).The remarkable mechanical properties and dynamic response of bones extend beyond their primary constituents, encompassing the mesostructure and/or nanostructure, pivotal for conferring superior stiffness, strength, and damping properties (Lambri et al., 2014).Notably, bone fractures frequently occur under dynamic loading conditions, wherein the damping properties of bone can exert a significant influence (Dodge et al., 2012;Du et al., 2023b).Presently, conventional orthopedic implants are fabricated from materials such as titanium, PEEK, or bioceramics like silicon nitride (Heimann, 2021;Kurtz and Devine, 2007;Quinn et al., 2020).These materials exhibit hardness and rigidity but lack adequate damping properties.Consequently, it is essential for orthopedic implants and bone substitutes to possess favorable damping characteristics to ensure their safety and efficacy in vivo (Du et al., 2023b).In this study, building upon the components and structural attributes of natural bone, we developed SA-SiN/PVA composite porous scaffolds with both physical and chemical crosslinks, aiming to achieve silicon nitride-based scaffolds with good damping properties.
Hydrogels have been widely utilized in the biomedical field for several decades and are known as shock-absorbing candidate materials (Wang et al., 2020;Park et al., 2022).We explored the influence of the hydrogel crosslinking system on the mechanical properties of SiN/PVA scaffolds and SA-SiN/PVA scaffolds.Physical crosslinking included hydrogen bonding in the PVA solution and PVA crystallite formation during the freeze-thaw process.In contrast, chemical crosslinking occurs during the soaking of samples in a CaCl 2 solution, where the substitution of two sodium ions by one calcium ion occurs with the SA component.Our findings reveal that SA-SiN/PVA scaffolds, toughened and stiffened through dual crosslinking mechanisms, exhibit remarkable hyperelasticity and a substantial hysteresis loop, surpassing the characteristics of the SiN/PVA group.Notably, even SA/PVA hydrogels display an elastic modulus reaching 0.53 ± 0.03 MPa, exceeding that of many hydrogels, such as the elastic modulus of polyacrylamide hydrogels (~100 kPa) (Muniz and Geuskens, 2001) and gelatin/chitosan cryogel (2.74 ± 0.52 kPa) (Lee et al., 2021a).Additionally, the addition of silicon nitride (60 wt%) elevates the elastic modulus to 1.20 ± 0.07 MPa.Considering the relationship between the amounts of SiN added and the hydrogel matrix, the results revealed three phases.In the first phase, from SA/PVA to 40 wt% SA-SiN/PVA, the elastic modulus decreased.The likely explanation is that the primary resistance to compression in the scaffold remains the hydrogel matrix.With the introduction of a small quantity of ceramic powder, the particles are likely spaced widely apart.In this scenario, they fail to contribute significantly to load-bearing and, concurrently, disrupt the intrinsic stiffness of the hydrogel matrix.Moving into the second phase, spanning from 40 wt% SA-SiN/PVA to 60 wt% SA-SiN/PVA, the modulus exhibited a quadratic increase.This is attributed to a significant accumulation of SiN powder within the hydrogel matrix.The particles come into contact under compression, and the ceramic powder, having a much higher modulus than the hydrogel matrix, enhances the overall modulus of the scaffold.However, surpassing a specific threshold marks the onset of the third phase.This excess disrupts the crosslinking sites within the hydrogel matrix, leading to a reduction in mechanical properties.In the samples with 70% SA-SiN, the silicon nitride powder constitutes 56% by weight in the scaffold.However, there is a concern regarding potential brittleness due to the substantial amount of silicon nitride powder.This stems from the possibility that if the hydrogel matrix does not effectively encapsulate these powders, there is a risk of powder leakage from the hydrogel matrix.Such leakage is undesirable, particularly in clinical applications, where maintaining structural integrity is crucial.Furthermore, a significant reduction in pores is observed in the 70% SA-SiN samples, negatively impacting energy dissipation.The variability in the pores of the 60% SA-SiN samples may stem from differing degrees of crystallinity during freezing or CaCl 2 ingress into the outer region initially.In both scenarios, crosslinking first occurs on the exterior before progressing to the core of the sample.The water content values of PVA/SA-SiN, only slightly lower than those in a healthy intervertebral disc (Newell et al., 2017), are noteworthy.Among all groups, the 60% SA-SiN samples demonstrated the optimal stress-strain curve, exhibiting the highest dissipated energy, indicative of excellent damping properties.Moreover, they achieved the highest effective elastic modulus and exhibited robust recovery even after undergoing compressive strains up to 67%.The test results also illustrated the outstanding performance of the SA-SiN/PVA scaffolds in impact tests.We attribute this success to the presence of the double-crosslinked hydrogel system and the enhancement of bioceramic submicron-powder.Additionally, the porous structures of freeze-cast scaffolds and ceramic/hydrogel composites facilitated very large recoverable compressive deformations with varying levels of frictional energy dissipation.As the scaffolds were compressed, energy was reversibly stored, allowing for significant elastic deformations.However, friction between cell walls and between ceramic particles and the hydrogel matrix led to the dissipation of energy and damping.
Moreover, this scaffold possesses additional advantageous properties beyond its damping characteristics, specifically osteogenic ability and antibacterial properties attributed to the presence of silicon nitride.In our study, the SA-SiN/PVA scaffolds exhibited significantly heightened osteogenic expression of ALP, COL-1, and RUNX2 compared to the PVA/ SA, aligning with other investigations and validating the osteogenic potential of silicon nitride (Du et al., 2022;Lee et al., 2021b).For example, Cappi et al. (2010) reported biocompatibility between silicon nitride, fabricated through both hot-pressing and pressureless sintering, and human mesenchymal stem cells (hMSC), resulting in osteogenic differentiation of hMSC.Additionally, Howlett et al. (1989) explored the in vitro and in vivo effects of silicon nitride on rabbit skeletal cells and tissues.Their in vitro findings revealed attachment and differentiation of bone marrow stromal cells (MSC) near silicon nitride samples, although not within pores.However, their in vivo assessments demonstrated that silicon nitride implants became infused with new mature bone three months post-implantation in femoral bone marrow cavities.Beyond the inherent osteoinductive and osteoconductive properties of silicon nitride, another factor influencing osteogenic differentiation is the increased stiffness of the SA-SiN/PVA scaffold.Scaffold stiffness plays a crucial role in cell signalling and focal adhesions, guiding cells to differentiate into a tissue which has a similar stiffness to the scaffold.Soft scaffolds, such as hydrogels, lack the ideal stiffness for promoting osteogenic differentiation (Lee et al., 2021a).
Implant-associated infections present a significant complication in orthopaedic surgery (Zimmerli, 2014).To address this concern, we conducted a comparative assessment of the antibacterial properties between SA-SiN/PVA scaffolds and SA/PVA scaffolds.Our findings revealed a pronounced inhibitory effect of SA-SiN/PVA on S. aureus, a Gram-positive bacterium frequently implicated in periprosthetic joint infections (Chang et al., 2023), underscoring the clinical significance of our observations.However, no significant difference was observed for E. coli between the SA-SiN/PVA and SA/PVA scaffolds.Some studies have highlighted the inhibitory effects of silicon nitride on gram-negative bacteria, including E. coli (Akin et al., 2021;Boschetto et al., 2019).These effects are attributed to factors such as its negative charging for repulsion, hydrophilicity, and chemical interactions such as peroxynitride anion formation (Bock et al., 2017).However, sodium alginate is known for its biocompatibility with E. coli and is frequently employed in E. coli encapsulation to maintain the viability of the E. coli (Papi et al., 2005;Bassani et al., 2019).This characteristic may be a contributing factor influencing the antibacterial properties of SA-SiN/PVA scaffolds against E. coli.Furthermore, another possible explanation is that the gels have effectively encapsulated silicon nitride, leading to inadequate exposure of the silicon nitride, thus hindering its ability to inhibit the growth of E. coli.Within the spectrum of clinically employed implants, such as titanium alloys, PEEK, and stainless steel, silicon nitride has consistently demonstrated highly effective antibacterial properties.Webster et al. (2012) evaluated silicon nitride, PEEK, and titanium implants in an in vivo rat calvariae model, showcasing the excellent antimicrobial efficacy against Staphylococcus epidermidis of silicon nitride.Their study reported histological bacterial counts on implant surfaces of 0%, 21%, and 88% for silicon nitride, titanium, and PEEK, respectively, three months post-surgery.Similarly, Ishikawa et al. (2017) investigated the antimicrobial properties of stainless steel, titanium alloy, PEEK, and silicon nitride against S. aureus, with silicon nitride exhibiting superior antibacterial performance among these materials.The antibacterial efficacy of SA-SiN/PVA scaffolds not only safeguards against bacterial infection as an implant but also proposes an alternative for healthcare professionals to reduce antibiotic usage in patients, potentially mitigating the global rise in antibiotic resistance.
Despite the widely recognized osteogenic and antibacterial effectiveness of silicon nitride, understanding the underlying mechanism remains a formidable challenge.This paradoxical phenomenon, where silicon nitride both promotes bone tissue healing and inhibits bacterial proliferation, poses a conundrum.The accepted explanation revolves around the distinctive elution kinetics of nitrogen and silicon, creating an environment on the surface of silicon nitride that is toxic to bacteria and conducive to eukaryotic cells, contingent upon pH.Specifically, the simultaneous availability of Si ions scavenged from surface silanols by osteoblasts provides crucial building blocks for the synthesis of new bone tissue, while nitrogen-radical interfacial chemistry elicits seemingly opposing effects on eukaryotic and prokaryotic cells.In a comprehensive investigation into the specific surface chemistry of silicon nitride in aqueous environments by Pezzotti et al. (Pezzotti, 2019), the results revealed that a cascade of key chemical reactions leads to direct RNA/DNA damage upon the penetration of ammonia in S. epidermidis.
While the SA-SiN/PVA scaffold has shown commendable damping properties, as well as positive attributes in terms of osteogenesis and antibacterial properties, its mechanical stiffness and strength are lower than that of vertebral bone, making it challenging for use as a standalone spinal fusion cage.The proposed SA-SiN/PVA composite holds potential as a bone scaffold, when coupled with dynamic rod and screw fixation.In this configuration, the dynamic rod and screw could contribute to restoring disc height while facilitating motion preservation in the spine.Consequently, the damping scaffold will undergo compression and recovery below its yield strength to absorb energy and simultaneously induce bone formation, through controlled strain, facilitating the fusion of the two vertebrae.Nevertheless, the degradability of the SA-SiN/PVA scaffold, beneficial for creating space for bone growth, poses a challenge as its mechanical properties would change with degradation.Investigating the relationship between degradation rate and bone formation rate would be crucial and beneficial.
More generally, a promising avenue for future research involves injecting SA-SiN/PVA into a 3D porous frame, enhancing its mechanical support.Conversely, there is a pressing need to design novel hydrogels with: (i) high elasticity, capable of complete and rapid recovery from large strains under ambient conditions; (ii) superior damping properties, efficiently dissipating a significant amount of energy during impact or dynamic loading and (iii) adequate strength, able to be used as a loadbearing scaffold.

Conclusions
As a bioceramic, silicon nitride lacks notable damping properties.In our study, a novel SA-SiN/PVA composite scaffold was fabricated, which allowed for improved energy absorption efficiency and rapid elastic recovery.In addition, our study showed that the stiffness and recoverable energy dissipation were significantly improved by combining physical and chemical cross-linking.Regarding cell-material interaction, our findings indicate that the introduction of silicon nitride promotes osteoblast differentiation while inhibiting specific types of bacteria.Overall, the combination of ceramics and tough hydrogels facilitates the development of advanced composites for spinal implants with good damping and osteogenetic properties.

Fig. 2 .
Fig. 2. (A) Stress-strain curve of SA-SiN/PVA and SiN/PVA scaffold; (B) Effective elastic modulus of SA-SiN/PVA and SiN/PVA scaffold; (C) Optical image of rehydrated scaffolds with different crosslinks; (D) SEM image of SiN/PVA scaffold and (E) SEM image of SA-SiN/PVA scaffold.Error bars show standard deviation; sample size n = 3.

Fig. 5 .
Fig. 5. (A) Representative optical images during the impact test of SA-SiN/PVA scaffold, (B) Representative optical images during the impact test of SA/PVA scaffold (yellow arrow indicates the moving direction of the crosshead), (C) Representative force-displacement curve of SA-SiN/PVA scaffold under impact with 2.5 kg crosshead (blue color curve indicates the first cycle, green color indicates the second cycle, and orange color indicates the third cycle during the impact), (D) Forcetime curve of SA-SiN/PVA and SA/PVA scaffold under impact (2.5 kg and 5 kg).

Fig. 6 .
Fig. 6. (A) Representative CLSM images that show the staining of actin microfilament cytoskeletal protein (red) and nuclei counterstained with DAPI (blue) of the cells after 3 days of culturing SA-SiN/PVA scaffolds.(B) PrestoBlue results of SA/PVA and SA-SiN/PVA scaffolds on days 1, 3, and 7. (C) Osteogenic gene expression of ALP, COL 1, RUNX 2 and OCN for MC3T3-E1 cultured on SA/PVA and SA-SiN/PVA scaffolds after 14 days.Error bars show standard deviation; sample size n = 3.

Fig. 7 .
Fig. 7. Quantitative analysis of bacterial proliferation rate of E. coli and S. aureus on SA-SiN/PVA and SA/PVA scaffolds after 24 h (A) and 48 h (B) of culture.Error bars show standard deviation; sample size n = 4.

Table 1
Amounts and fractions of each material powder in the final scaffold.
X.Du et al.

Table 2
Swelling ratio and water content of SA-SiN/PVA scaffolds.