α‑Ketoglutaric Acid as a Sustainable and Universal Photoinitiator for Hydrogels

Abstract

Hydrogels have acquired numerous attentions in biomedicine and smart bioelectronics due to similar properties to our tissues. Initiators are crucial in vinyl polymerization for hydrogel preparation which can influence mechanical properties and physical/chemical structures. We report α-ketoglutaric acid (αKG), an intermediate in the citric acid cycle, as an ecofriendly photoinitiator for vinyl polymerization. We comprehensively investigated the mechanical properties of hydrogels initiated by comparing αKG with Irgacure 2959 (I-2959) and ammonium persulfate (APS). We found that αKG successfully initiated the polymerization in both synthetic and natural hydrogels (polyethylene glycol, polyacrylamide, gelatin, alginate, and chitosan). Meanwhile, the interactions between αKG and functional groups of macromolecules make αKG a potential additive in hydrogels, which could influence the hydrogel mechanics. In contrast to the potential toxicity of unreacted APS and the limited water solubility of I-2959, αKG proves to be a more versatile and sustainable option in gelation than the other two. This work throws new light on hydrogel fabrication with the initiator alteration for potential biomedical applications.

1. Introduction

Hydrogels are three-dimensional (3D) networks of cross-linked polymer chains and have found their enormous potentials by covering almost all areas in biomedicine from tissue engineering to drug delivery, wearable electronics and bionic electronic skin. With structural and functional resemblances to the natural extracellular matrix (ECM), hydrogels and their related physical performance have become a key focus in polymer synthesis. , To reach the best therapeutic outcomes, intensive investigations have been in the aspects of the composition, topography, and mechanics, as these factors are crucial in regulating cellular function. , Wide spectrum of cross-linking methods was explored to engineer tunable physical properties. , Among them, vinyl cross-linking is an approach of this sort in rapid initiation and fast in situ polymerization , to enable the creation of 3D networks from chain polymers in a single-step process. Methacrylate or acrylate groups are the most versatile to form the robust networks. ,

As the most common technology, free radical polymerization heavily depends on the initiators when is concerned the gelation efficiency. , Initiators could be water-soluble and nonsoluble, but they are normally considered nontoxic to cells for tissue engineering (Table S1). The most frequently used initiators for hydrogels are ammonium persulfate (APS)/tetramethyl-ethylenediamine (TEMED) redox pair − and UV initiator Irgacure 2959 [2-hydroxy-4′-(2-hydroxyethoxy)-2-methyl-propiophenone]. , Many other water-soluble photoinitiators have been studied these years in hydrogel materials regarding efficient polymerization and biocompatible concern. Lithium phenyl-2,4,6-trimethylbenzoylphosphinate (LAP) is a Norrish type I photoinitiator which can be initiated using 365 or 410 nm light irradiation, revealing its potential in visible light activation, which is reported harmless to tissue when applied in an in situ formed biomaterial. − However, the initiating efficiency of LAP, besides its high cost, in the visible light region is relatively low because of its limited absorption ability at over 400 nm. Eosin-Y is also an attracting visible light initiating photoinitiator because of its low cytotoxicity. − But a co-initiator is required to get a satisfying efficient since its type I initiating mechanism needs a hydrogen transfer process. Riboflavin (vitamin B₂) has been reported a lot in hydrogel basing on its benefit to cells and visible light activation, while creation of reactive oxygen species needs to be considered which can be harmful to tissue regeneration. , Other systems that can be activated by visible light such as camphorquinone (CQ) , and two-photon induced polymerization system , are also studied a lot, but their applications in hydrogels are limited by low water-solubility and long irradiation time, respectively.

For APS/TEMED initiation, TEMED reacts with APS through a redox mechanism that facilitates rapid radical generation, enabling fast gelation across various polymer systems. However, the cytotoxicity of TEMEM limits its clinical applications, especially in situ applications. − The corporation of APS and metal phenolic network has been reported recent years, in which quinone-catechol reversible redox reaction is utilized to accelerate free radical generation. Natural-derived phenolic structures such as dopamine, tannic acid, and lignin are studied with multivalent metal ions including Fe³⁺, Al³⁺, Cu²⁺, Co³⁺, Ni³⁺ and Zr³⁺, demonstrating their promising ability on CC bond polymerization and roles as dynamic cross-linking core and conductivity component. , However, their initiation efficiency can be inhibited by the increasing concentration of low-valent metal ions in the system. Photopolymerization, on the other hand, offers benefits such as low environmental pollution (due to the absence of volatile organics), rapid curing, low energy demands, and reactivity at room or body temperature, making it a more favorable option for biomedical applications. For example, I-2959 is frequently employed in hydrogel fabrication due to its photoinitiation efficiency and moderate cytocompatibility in certain systems, yet moderate to low water solubility (up to 0.5 wt %) , limits its uses. , To improve solubility, I-2959 is chemically modified with hydrophilic quaternary ammonium and carbohydrate residue groups, although the modified photoinitiators get higher water solubility than I-2959, their applications are limited by unrestricted migration in polymer network and unclear biocompatibility. Alternatively, I-2959 is also grafted onto polymer backbones, such as poly­(ethylene glycol) (PEG), poly­(ethylenimine) (PEI), and alginate to create water-soluble polymeric photoinitiations, however, high degrees of functionalization are often required to maintain initiating efficiency.

To address these concerns, we report efficient, capable, economical, environmentally friendly initiator α-ketoglutaric acid (αKG) for a promising UV initiator for gel preparation. αKG is a naturally occurring metabolite in the Kreb’s cycle, can offer good water solubility and biocompatibility. It goes without concerns in toxicity and environmental impact. − Our research will systematically evaluate the initiating performance. We provide a comprehensive comparison of commercial initiators with αKG and investigate the impacts of their concentrations on the mechanical properties of the hydrogels. Specifically, by comparing αKG with I-2959 and APS, the mechanical properties of hydrogels are the focus. Synthetic hydrogels, such as polyacrylamide (PAm) and poly­(ethylene glycol) diacrylate (PEGDA), as well as modified natural macromolecular hydrogels like methacrylated gelatin (GelMA), methacrylated chitosan (ChMA) and methacrylated alginate (AlgMA), were chosen for this study (Scheme ). By varying the concentrations of initiators, we modulated the polymeric chain length and hydrogel network morphology, thereby influencing hydrogel mechanics. Additionally, unreacted initiators in hydrogels served as molecular additives, interacting with macromolecules through hydrogen bonding and electrostatic interaction, which further affected the hydrogels’ properties. These findings provide valuable insights for selecting suitable initiators for biomaterial use.

1. (a) α-Ketoglutaric Acid (αKG) as a Functional Initiator for Vinyl Polymered Hydrogels; (b) Synthesis of Vinyl Polymered Hydrogels through UV Exposure; the Excess αKG Has Noncovalent Interactions with Polymer Chains, Working as an Additive, Tunning Hydrogel Mechanics.

2. Results and Discussion

2.1. Photoinitiation Kinetics of αKG

The initiation mechanism of αKG and I-2959 is shown in Figure a. αKG initiation is based on the type II mechanism, which is first photoexcited and then undergoes the intramolecular hydrogen abstraction, forming free radicals. Meanwhile, I-2959 is a type I photoinitiator and undergoes homolytic cleavage into two free radicals with UV exposure.

1.

Initiation performance of αKG initiator. (a) Initiation mechanism of αKG and I-2959, polymerization of acrylamide (Am) and poly­(ethylene glycol) methyl ether acrylate (PEGA). (b,c) UV–vis spectra and photolysis of αKG and I-2959. (d) DSC curves of αKG after UV exposure. (e,f) carbon double bond polymerization conversion of Am and PEGA initiated by αKG and I-2959.

The UV–vis absorption and photolysis spectra of αKG and I-2959 in DI water are shown in Figure b,c. Maximum absorption wavelength (λ_max) of αKG at 322 nm is attributed to the n–π* transition of ketone carbonyl groups, while the maximum absorption of I-2959 is at 280 nm due to the π–π* transition of aromatic moiety. The maximum molar extinction coefficient of αKG was 12.2 M^–1cm^–1 at 322 nm, which is much lower than that of I-2959 (ε = 11838 M^–1cm^–1 at 280 nm). However, the absorption band of αKG has a better overlap with the UV light source of 365 nm, which explains the same initiating wavelength for both αKG and I-2959. , When exposed to UV, the maximum absorption of αKG and I-2959 gradually decreases with increasing exposure time, indicating their photolysis and the generation of new products. The photoexcitation of αKG results in the loss of ketone group (Figure a), thus weakening the corresponding UV absorption at 322 nm and increasing the peak at 280 nm which assigned to the carbonyl group. The photodissociation of I-2959 leads to the formation of benzoyl and ketyl radicals. The photolysis of αKG was further investigated using differential scanning calorimeter (DSC). As shown in Figure d, the melting peak of αKG at 114 °C shows a weakening tendency and a temperature shift was observed with the increasing UV exposure time, indicating the broken structure of αKG and existing of photolysis product.

To investigate the initiating performance of αKG, its photopolymerization kinetics was characterized by NMR and compared with I-2959. In terms of acrylamide (Am) polymerization (Figure e), the CC bond conversion of αKG increased rapidly and reached over 90% at 120 s, while the conversion rate of I-2959 was approximately 50% at 120 s which is much lower than αKG, revealing a high initiating efficiency of αKG for Am at the initial stage. After that, the double bond conversion of the αKG group kept increasing and reached 98% at 240 s, and the I-2959 group was 91%, a comparable level with αKG. Poly­(ethylene glycol) methyl ether acrylate (PEGA) was selected as the acrylate model to evaluate the initiating performance of αKG (Figure f). For PEGA polymerization, the double bond conversion of αKG was higher than I-2959 within 30 s. Although the conversion rates of these two groups kept close, αKG was surpassed by I-2959 after 30 s. These results exhibit the comparable initiation efficiency of αKG to I-2959 as a photoinitiator. The low light absorption of αKG does not affect its performance.

2.2. αKG Initiation on Mechanical Properties of PAm Hydrogel

The PAm hydrogel is a biocompatible hydrogel widely used in biosensors, drug delivery, and tissue egineering. − In this study, the polymer chains were cross-linked with N,N′-methylenebis­(acrylamide) (MBA), forming a 3D network structure. Mechanical properties of the PAm hydrogels were quantitatively assessed through compressive and tensile stress–strain tests. As shown in Figure a,b, the hydrogels withstood a high compressive strain of 90% without mechanical failure. At a low concentration (0.5%), PAm hydrogels initiated by αKG exhibit a lower compressive modulus than those initiated by I-2959 and APS groups, which were 36 ± 1.4, 42 ± 1.4, and 44 ± 0.8 kPa, respectively, indicating softer αKG group hydrogels than the other two groups. Notably, increasing initiator concentrations lead to a decrease in compressive modulus across all three groups (Table S2). The modulus of PAm hydrogels reduced to half of its initial value as the initiator concentration increased from 0.5% to 1%, eventually reaching a range of 4.6–7.7 KPa at 5% concentration. But for higher concentrations (5%), there was no significant difference in compressive modulus among the three initiator groups. As shown in Figure d,e, the tensile properties followed a similar trend to the compressive tests: the tensile modulus decreased as the initiator concentration increased. The αKG-initiated hydrogels present significantly lower tensile moduli than the I-2959 and APS groups, which were about 1/2 to the I-2959 group and 1/3 to the APS group around all three concentrations (Figure e and Table S3). However, the best elongation group showed up in αKG-initiated hydrogels, which reached 1328.8 ± 34.4% at 3% concentration. I-2959 and APS groups only reached 904.9 ± 39.5% and 1117.5 ± 96.5%, respectively.

2.

Mechanical tests of 20 wt % PAm hydrogels. (a) Compressive strain–stress curves of PAm hydrogels initiated by αKG, I-2959, and APS of different concentration. (b) Compressive modulus of the hydrogels. (c) Images of PAm hydrogel, (i) hydrogels as prepared, (ii) before cyclic compression test, (iii) compressed at 50% strain, (iv) after 100 cycles of cyclic compression. (d) Tensile strain–stress curves of PAm hydrogels initiated by αKG, I-2959, and APS of different concentration. (e) Tensile modulus, strain, and stress of the hydrogels. (f) SEM images of PAm hydrogels.

Cyclic compression tests at a fixed strain of 50% were also performed to evaluate the recovery performance of the PAm hydrogels (Figure S2). The hydrogels exhibited large hysteresis upon unloading in the cycles, indicating energy dissipation due to hydrogen bond rupture during deformation. Across all initiator groups, greater hysteresis and reduced shape recovery were observed as the initiator content increased, particularly in the single-network hydrogels. The molecular weight between cross-links (M _c) can also show us the initiation performance of the initiators (Table S4). For single network PAm hydrogels, their mechanical properties are influenced by the cross-linking density and the polymer chain length. As a result, the mechanical data we obtained can be the results of combined effect of these two factors. As the cross-linker concentration in all PAm hydrogel groups were same, the mechanical properties of these hydrogels could depend on the initiation efficiency and the potential interactions between −NH₂ groups on PAm chain and small molecules.

Generally, initiator concentration can influence the polymerization process, the resulting polymer chain length, and the cross-linking density of hydrogels. With an excess of initiator, shorter polymer chains are typically formed, potentially disrupting the 3D network structure (Song, 2025 #96), and leading to reduced mechanical performance. As shown in Figure f, the pore size of PAm hydrogels increased as the initiator concentration rose, which revealed the lower cross-linking density of the hydrogels due to the short polymer chains initiated by the high initiator concentration, aligning with their decreasing mechanical properties (Figure S3). At very high initiator concentrations (5%), the originally well-defined 3D networks of these hydrogels were disrupted, likely due to the formation of shorter polymer chains, which affected network construction. Moreover, excess small molecules can affect the material’s properties due to their interactions with the macromolecules. In PAm hydrogels, unreacted αKG altered the hydrogel’s properties through hydrogen bonding and electrostatic interactions, as confirmed by FTIR analysis, where a broader absorption around 3332 cm^–1 was observed with higher αKG levels (Figure S4). This introduces an additional energy dissipation mechanism, enhancing the material’s deformation capacity. This effect was further evidenced by the greater hysteresis observed in cyclic compressive tests (5% concentration) of the αKG group compared to the other groups (Figure S2).

2.3. Mechanical Properties of PEGDA Hydrogel

PEGDA hydrogels are fabricated by polymerization of the monomers, resulting in a 3D polymer network capable of retaining large amounts of water, featuring biocompatibility and wide utilization in biomedical applications. Compression test results (Figure a,b) show that among the PEGDA hydrogels the αKG group exhibited the lowest compressive strength, while the APS group showed the highest compressive modulus across all concentration ranges (Table S5). In the αKG group, an increasing in initiator content gradually increased the modulus from 65 ± 1.7 kPa and 100 ± 10 kPa. The APS group initially showed an increase in modulus with higher initiator concentrations (0.5%–2%), reaching a peak compressive modulus of 143.2 ± 2.5 kPa, before a subsequent decline to 120.1 ± 3.7 kPa (5% concentration). In the I-2959 group, the modulus initially increased from 84.5 ± 6.1 kPa (0.5%) and then stabilized in the range of 100–110 kPa at 2–5% concentration. The compression strain for PEGDA hydrogels ranged from 36% to 50%, with the APS group achieving the highest fracture strain 47.8 ± 0.8% at low concentration (0.5–1%) and the αKG group reaching the highest (around 43%) at higher concentrations (2–5%). As a result, the PEGDA hydrogel initiated by 1% APS demonstrated the highest compression fracture stress (336 ± 5.6 kPa). As shown in Figure d,e, the tensile performance of PEGDA hydrogels followed a trend similar to the compressive results. The αKG group exhibited the lowest modulus (66.4 ± 1.2 kPa) at 1% but increased, reaching values comparable to both I-2959 and APS groups at 5% (94.8 ± 3.9 kPa). The APS group exhibited the highest tensile strain and stress at 1%, which were 51.9 ± 0.9% and 52.3 ± 1.5 kPa, respectively. Although the tensile strain and stress of the αKG-initiated hydrogels were lower than those of the APS group, they exceeded those of the I-2959 group. Cyclic compression tests on PEGDA hydrogels (Figure S5) reveal minimum hysteresis, indicating the absence of significant energy dissipation mechanisms within the PEGDA hydrogels. All groups exhibited a slight degree of stress loss, suggesting that a more uniform 3D network is formed when sufficient free radicals are available in the system, but excessive initiator concentrations may disrupt the 3D network, leading to poorer mechanical performance. The cross-linking density (M _c) also showed us the mechanic differences between these groups (Table S7). In PEGDA hydrogels, a lower M _c means more PEGDA monomers (CC bonds) fully reacted and connected to each other, demonstrating the CC bond conversion in the hydrogels, thus reflecting the initiator efficiency for the PEGDA hydrogel. The αKG groups showed higher M _c values than the other two groups.

3.

Mechanical tests of 10 wt % PEGDA hydrogels. (a) Compressive strain–stress curves of PEGDA hydrogels initiated by αKG, I-2959, and APS of different concentration. (b) Compressive modulus, strain, and stress of the hydrogels. (c) Images of PEGDA hydrogel, (i) hydrogels as prepared, (ii) before cyclic compression test, (iii) compressed at 20% strain, (iv) after 100 cycles of cyclic compression. (d) Tensile strain–stress curves of PEGDA hydrogels initiated by αKG, I-2959, and APS of different concentration. (e) Tensile modulus, strain, and stress of the hydrogels. (f) SEM images of PEGDA hydrogels.

SEM images (Figure f) showed that I-2959 and APS/TEMED initiator groups produced relatively consistent porous morphologies, implying sufficient initiation and cross-linking. However, PEGDA hydrogels cross-linked using αKG exhibited a less-defined structure compared to those initiated by I-2959 and APS/TEMED, possibly due to lower polymerization efficiency under the given formulation.

2.4. Mechanical Properties of GelMA Hydrogel

Gelatin, a natural protein derived from the hydrolysis of collagen, is characterized by its good biocompatibility, solubility, degradability, and ease of availability along with lower antigenicity than collagen. GelMA was synthesized by grafting methacryloyl groups onto the amino groups on gelatin side chains, enhancing cross-linking capacity and mechanical strengthen in the hydrogel (Figure a). , The structure of synthesized GelMA was confirmed through Fourier transform infrared (FTIR) spectroscopy, showing a CO stretching vibration absorption of the ester bond of MA at 1720 cm^–1 (Figure b-i). Nuclear magnetic resonance (NMR) spectroscopy further validated GelMA synthesis with characteristic methacrylate proton peaks observed at 5.34 and 5.59 ppm (Figure b-ii).

4.

(a) Synthesis of GelMA. (b) Characterization of GelMA. (i) FTIR spectra, (ii) ¹H NMR spectra. Mechanical properties and morphology of 10 wt % GelMA hydrogels. (c) Compressive strain–stress curves of GelMA hydrogels initiated by αKG, I-2959, and APS of different concentration. (d) Compressive modulus, strain, and stress of the hydrogels. (e) Images of GelMA hydrogel, (i) hydrogels as prepared, (ii) before cyclic compression test, (iii) compressed at 50% strain, (iv) after 100 cycles of cyclic compression. (f) Tensile strain–stress curves of GelMA hydrogels initiated by αKG, I-2959, and APS of different concentration. (g) Tensile modulus, strain and stress of the hydrogels. (h) SEM images of GelMA hydrogels.

The compression test results for GelMA hydrogels initiated with three different initiators are shown in Figure c,d. Hydrogels in the αKG-initiated group were comparatively softer, exhibiting a compressive modulus lower than those of the other groups (Table S8). As the concentration of αKG increased, the modulus further decreased, from 22.3 ± 0.7 kPa (0.5% αKG) to 13.0 ± 1.2 kPa (5% αKG). Interestingly, the softening of the αKG hydrogels allow them to sustain higher fracture strains, increasing from 60.1 ± 1.2% (similar to other groups at 0.5% initiator content) to 84.7 ± 0.9%, the highest strain observed among all groups. Consequently, the αKG hydrogels demonstrated the best fracture stress performance within the 2–5% concentration range, getting the highest compressive stress of 581.2 ± 7.1 kPa with 2% content.

Cyclic compression tests further validated the good compressive performance of the αKG hydrogels. As shown in Figure S6a and S6d, after 100 cycles of compression at 50% and 70% strain, no fractures were observed, and the stress–strain curves remained consistent with the initial cycles, indicating a stable mechanical performance. The hysteresis appeared in the unloading process, resulting from energy dissipation due to the breaking of hydrogen bonds. The αKG hydrogels exhibited stress loss between 7–10% (Table S10), which is caused by the disruption of hydrogen bonds in GelMA during the loading–unloading process.

The I-2959 group hydrogels were stiffer than those in the αKG group, reaching their highest compressive modulus at 1% content (33.3 ± 2.1 kPa). However, this modulus decreased to 17.3 ± 1.6 kPa as the initiator content increased to 5%. Hydrogels in the I-2959 group also exhibited more brittle behavior than those in the αKG group, with lower compressive strain and stress across all concentrations. In contrast, the APS group hydrogels exhibited the highest compressive modulus of 50.5 ± 0.8 kPa at 4% initiator content, with strain values ranging from 60% to 70%. However, APS-initiated GelMA hydrogels showed relatively poor performance in cyclic compression test at 50% strain; only the 3% APS group retained its elasticity, while the 1% and 5% groups fractured during testing (Figure f and S6c). Overall, GelMA hydrogels initiated with αKG performed better in compression tests, even at high αKG concentrations.

The tensile test results, shown in Figure f,g, followed a trend similar to that of the compression tests. As initiator concentration increased from 1% to 5%, αKG-initiated hydrogels showed a decrease in tensile strength (75.6 ± 3.4 kPa to 14.6 ± 1.1 kPa) but maintained stretchability to over 74.3 ± 0.6% of their original length at 3% concentration, which was less than the maximum strain observed in the APS group (95.2 ± 5.9% at 1% content). The APS group exhibited the highest tensile modulus of 91.3 ± 0.8 kPa at 3%, while the I-2959 group consistently showed the weakest performance across all parameters. The cross-linking density of GelMA hydrogel also showed us the initiation performance of the initiator, since the average molecular weights of the macromolecules are in a certain range and general high, the molecular weight between the cross-links can be seen as an indicator of initiation efficiency of the initiators. As shown in Table S11, the APS/TEMED groups showed the highest cross-linking density at a higher concentration (2–5%). However, other parameters in hydrogels also improve the hydrogel performance, such as small molecule additives and nanofillers. The potential functions of αKG in GelMA hydrogels will be discussed later and provided further insight into their mechanics.

Cross-sectional SEM images of GelMA hydrogels (Figure h) revealed a porous structure across all groups, with the pore size inversely correlated to hydrogel strength. The larger pore size and looser network structure in hydrogels also demonstrate the lower cross-linking degree, which could affect hydrogels’ mechanical strengt. For the αKG group, pore structure became looser as initiator concentration increased, contributing to a decreased modulus. The most compact network structures were observed in the I-2959 and APS groups at 1% and 3% concentrations, respectively, aligning with their respective better mechanical properties (Figure S8).

The influence of unreacted αKG on the compression performance of GelMA hydrogels was further confirmed through compression tests conducted after removing residual αKG by immersing the GelMA hydrogels in deionized water (Figure a–d). The removed αKG from GelMA hydrogels was characterized by using UV–vis spectra (Figure S9). A representative absorption of αKG at 322 nm was found, revealing the exitance of excess αKG in GelMA hydrogels. The excess αKG may interact with functional groups (e.g., −NH₂ groups) of GelMA, further influencing the properties of these hydrogels. The concentration of untreated αKG was calculated using the standard curve and the abstracted αKG from 5% group was 0.6 mg/mL hydrogel.

5.

Effect of excess αKG in GelMA hydrogel. (a) GelMA hydrogels as prepared and (b) after excess αKG removal. (c) Compressive modulus and (d) compressive strain of the hydrogels. (e) FTIR of GelMA powder and GelMA hydrogels initiated by αKG of different concentrations.

After the water treatment, the hydrogels showed a significantly enhanced compressive modulus, reduced strain, and consistent mechanical properties across various αKG concentrations. At the same time, these processed GelMA hydrogels showed consistent values in compressive modulus (around 30 kPa) and strain (about 60%), revealing an independent trend to αKG concentrations. This result highlights the role of residual αKG in modulating the compressive performance of GelMA hydrogels.

To investigate the interaction between GelMA and αKG, FTIR analysis was performed and the peak 3301 cm^–1of N–H stretching, shifted to 3287 cm^–1, indicating the formation of hydrogen bonding between GelMA and αKG (Figure e). These results show that excess αKG left in the GelMA hydrogels can modulate their mechanical properties through molecular interactions, indicating a dual function of αKG as both a photoinitiator and a functional additive.

2.5. Mechanical Properties of AlgMA Hydrogel

Sodium alginate (SA), composed of mannuronic acid and guluronic acid, is commonly used to synthesize hydrogels for biomedical applications. The synthesis of AlgMA was shown in Figure a, with the successful grafting of methacrylate groups onto the SA structure confirmed by the appearance of CO stretching variation of methacrylate units at 1718 cm^–1 in FTIR spectra (Figure b-i). The methacrylate proton peaks were found at 5.72 and 6.20 ppm in the NMR spectra (Figure S6b-ii). Prior research reports that single-network AlgMA hydrogels generally exhibit a relatively low mechanical performance, with compressive moduli between 0.5 and 180 kPa, depending on their cross-linking density, graft method, and alginate types. −

6.

(a) Synthesis of AlgMA. (b) Characterization of AlgMA. (i) FTIR spectra, (ii) ¹H NMR spectra. Mechanical properties and morphology of 2 wt % AlgMA hydrogels. (c) Compressive strain–stress curves of AlgMA hydrogels initiated by αKG, I-2959 and APS of different concentration. (d) Compressive modulus, strain and stress of the hydrogels. (e) Images of GelMA hydrogels. (f) SEM images of AlgMA hydrogels.

In this work, compressive tests were performed to assess the impact of different initiators on the AlgMA hydrogels. As shown in Figure c,d, AlgMA hydrogels initiated with αKG exhibited a compressive modulus similar to those of the APS group at low concentrations of 0.5–1% (<10 kPa) but were softer than those of the I-2959 group (approximately 12 kPa). As the αKG concentration increased to 2%, the hydrogels displayed stiffer behavior, achieving the highest modulus of 22.9 ± 1.0 kPa, followed by a decrease to 8.09 ± 0.33 kPa with further increases in concentrations. While the I-2959 group demonstrated higher moduli than the αKG group at all concentrations, it followed a similar trend, peaking at 2% with a modulus of 21.88 ± 0.37 kPa, comparable to the αKG group. In contrast, the APS group displayed softer characteristics, with moduli lower than 12 kPa across all concentrations (Table S12). The molecular weights between cross-links of AlgMA hydrogels were calculated and listed in Table S13. The higher cross-linking density provided I-2959 groups with stiffer mechanical performance than the other groups.

For all initiator groups, lower modulus in AlgMA hydrogels correlated with higher fracture strains, with the highest strain of 64.4 ± 1.0% observed in the APS group at 0.5% concentration. This suggests that higher free radicals’ availability enhances double bond cross-linking, resulting in stiffer, though more brittle hydrogels due to limited energy dissipation pathways within the AlgMA hydrogels. SEM images (Figure f) show porous structures across all groups, with a more compact network at 2% αKG and 2% I-2959 groups, contributing to their greater compressive strength (Figure S10).

2.6. Mechanical Properties of ChMA Hydrogel

Chitosan, a natural cationic biopolymer known for its high biocompatibility and biodegradability, is derived from the deacetylation of chitin through chemical treatment. The synthesis of ChMA (Figure a) was confirmed by FTIR, with a peak at 1689 cm^–1 indicating the successful grafting of methacrylate groups (Figure b-i). Additionally, the structure of ChMA was verified through the ¹H NMR spectrum, where signals of vinyl protons were shown at 5.29 and 5.49 ppm (Figure b-ii).

7.

(a) Synthesis of ChMA. (b) Characterization of ChMA. (i) FTIR spectra, (ii) ¹H NMR spectra. Mechanical properties and morphology of 2 wt % ChMA hydrogels. (c) Compressive strain–stress curves of ChMA hydrogels initiated by αKG, I-2959, and APS of different concentration. (d) Compressive modulus, strain, and stress of the hydrogels. (e) Images of ChMA hydrogels. (f) SEM images of ChMA hydrogels.

Compression tests were conducted to assess the impact of different initiators on the mechanical properties of ChMA hydrogels (Figure c,d). The hydrogels initiated by αKG and I-2959 were relatively soft, with compressive moduli of approximately 20 and 45 kPa, respectively, showing minimal variation with varying concentrations. In contrast, the compressive modulus of ChMA hydrogels initiated by APS enhanced with concentration increases, reaching a maximum of 222.2 ± 0.3 kPa at 4% APS content, considerably higher than those initiated with αKG and I-2959 groups (Table S14). For ChMA hydrogels, the abundant −NH₂ groups on its backbone can form hydrogen bonds with αKG in an acid environment, which may inhibit the formation of free radicals upon UV exposure, reducing the initiation efficiency of αKG. The interactions between αKG and ChMA can also influence the mechanical properties of obtained hydrogels. Meanwhile, ChMA precursor solution/hydrogels were nontransparent (Figure e), which can decrease the light density in photopolymerization, resulting in low initiation efficiency. As a result, the compressive moduli of both αKG and I-2959 groups were lower than those of APS/TEMED groups. The cross-linking density of ChMA hydrogels were listed in Table S15. The molecular weight between crosslinks of αKG and I-2959 groups were much higher than that of APS/TEMED groups, proving the low crosslinking density in photoinitiated ChMA hydrogel, which may be caused by the low initiation efficiency because of the limited light density that can pass through the nontransparent precursor solution.

The compression strain of ChMA hydrogels was inversely proportional to their modulus, as the ultrasoft hydrogels in the αKG group could sustain strains over 60% across all concentrations, outperforming the other two groups in terms of strain. In contrast, the stiffest hydrogel, initiated with 4% APS, fractured at a much lower strain of 24.1 ± 0.6%. SEM images of ChMA hydrogels (Figure f) display well-defined porous networks across all groups, with notably more compact structures in the 3% and 5% APS groups (Figure S11), aligning with their higher compressive modules.

3. Conclusion

In this work, we systematically analyzed the mechanical properties of vinyl polymerization in hydrogels initiated by the different initiators, focusing on how they influence the polymer chain length, hydrogel network morphology, and interactions with macromolecules. The initiation kinetics of αKG was characterized and showed competitive capacity with I-2959. The biobased photoinitiator αKG demonstrated strong potential in fabricating both synthetic and natural vinyl-grafted hydrogels, proving competitive against standard commercial initiators I-2959 and APS. Interestingly, at higher concentrations, αKG acted not only as an initiator but also as an additive, enhancing the mechanical properties of the hydrogels. In contrast, due to potential toxicity, APS and I-2959 require limited concentrations in fabricating biomedical materials. This study provides valuable insights into selecting suitable initiators to tailor hydrogel properties for specific biomedical applications.

4. Materials and Methods

4.1. Materials

The α-ketoglutaric acid (αKG), ammonium persulfate (APS), gelatin, poly­(ethylene glycol) diacrylate (PEGDA, M _n = 700), N,N′-methylenebis­(acrylamide) (MBA), methacrylic anhydride (MA), N,N,N′,N′-tetramethyl-ethylenediamine (TEMED), and poly­(ethylene glycol) methyl ether acrylate (PEGA) were purchased from Sigma-Aldrich. Iragcure 2959 (I2959) was obtained from Dempsay. Chitosan was obtained from Aladdin. Acrylamide and alginic acid sodium salt were purchased from Thermo Fisher Scientific. All of the materials were used without further purification unless otherwise noted. Deionized (DI) water was utilized in all of the experiments.

4.2. Preparation of Methyl Acrylate Materials

4.2.1. Synthesis of Methacrylated Gelatin (GelMA)

GelMA macromolecules were synthesized with methacryloyl grafting according to previously established methods. Briefly, 3 g of gelatin (10 wt %) was added in phosphate-buffered saline (PBS) at 50 °C with magnetic stirring for complete dissolution. Then, 3 mL of MA was dropped into the solution and stirred at 50 °C. In 1 h, 90 mL of DI water was added to dilute the solution and stop the reaction, followed by dialysis against DI water with 8–13 kDa cutoff at 50 °C to remove unreacted MA. The final white porous foam product was obtained after lyophilization and stored at −20 °C.

4.2.2. Synthesis of Methacrylated Alginate (AlgMA)

AlgMA was prepared through esterification of the alginate hydroxyl groups according to protocols previously described. Briefly, 0.5 g of alginate was completely dissolved in 50 mL of DI water, followed by adjusting the pH of the solution to 8 using NaOH solution. Then, 10 mL of MA was dropped slowly into the alginate solution at 0–4 °C and the pH needed to be adjusted every 30 min. After 24 h reaction at 0–4 °C, the obtained solution was dialyzed against DI water for 1 week. After freeze-drying for 5 days, the final foam-like AlgMA product was stored at −20 °C.

4.2.3. Synthesis of Methyl Acrylated Chitosan (ChMA)

ChMA was prepared using a single-step chemoselective N-acylation reaction, following established methods in the literature. First, 1 g of chitosan (3 wt %) was dissolved in 2 wt % acetic acid overnight to obtain a homogeneous solution. Five mL of MA was added to the solution and reacted for 3 h. The obtained milky suspension was purified by dialyzing it against DI water for 1 week. The solution was then lyophilized.

4.3. Characterization

¹H NMR (nuclear magnetic resonance) spectra were recorded on a Bruker Advance 300 MHz NMR spectrometer (Bruker, Billerica, MA, USA). For αKG and I-2959 initiating kinetic NMR characterization, Am and PEGA (10%) were dissolved in DI water, mixed with the photoinitiators (2%), and then exposed under UV (365 nm, 60 W, incident light intensity 2.5 W/cm²) for different time. The obtained products were freeze-dried for further test. The samples were dissolved in D₂O in a tube with an outside diameter of 5 mm for the ¹H NMR spectra. Fourier transform infrared spectra (FTIR) were characterized on a Thermo Scientific Nicolet iS-10 FTIR spectrometer, and the samples were tested using attenuated total reflection (ATR) accessory. UV–vis spectra and photolysis of αKG and I-2959 were conducted using a VMR spectrophotometer. 0.1 mol/L αKG and 10^–4 mol/L I-2959 were prepared using DI water, irradiated with UV LED of 365 nm (60 W, incident light intensity 2.5 W/cm²) at different exposure time. The characterization of the unreacted αKG in GelMA hydrogels were performed using UV–vis spectra. The standard curve of αKG was analyzed using the λ_322 nm of different concentrations, 0.01, 0.1, 1, 5, and 15 mg/mL αKG. GelMA hydrogels were immersed in DI water (5 mL) for 10 min to allow the αKG immigration and prevent the obvious shape change of the hydrogels. The obtained DI water solutions were tested using UV–vis spectra, and αKG concentrations were calculated based on the standard curve. The differential scanning calorimeter (DSC) of αKG after UV exposure were tested using a PerkinElmer DSC 8500. To prepare samples, αKG was first dissolved in DI water and exposed under UV, then the solutions were freeze-dried for further test. For DSC tests, dried products were weighed (about 10 mg) in standard aluminum pans, sealed with lids, heated at the rate of 5 °C/min from 70 to 160 °C using nitrogen as a purge gas.

4.4. Hydrogels Preparation

For αKG or I-2959 initiated hydrogels, the precursor solutions were prepared by dissolving αKG or I-2959 (0.5% to 5%, w/w) in the DI water solutions of acrylamide (20 wt %), PEGDA (10 wt %), GelMA (10 wt %), AlgMA (1.5 wt %), and 2% acetic acid solution of ChMA (1.5 wt %). The MBA (0.2%, w/w) was added into acrylamide precursor solution as the cross-linker.

For the APS initiated hydrogels, the precursor solutions were prepared by dissolving APS (0.5% to 5%, w/w) and TEMED (0.3%) in the monomers/methacrylated macromolecular solutions.

The hydrogels samples for tensile tests were prepared in dog-bone shape mold (20 × 4 × 2 mm, length × width × thickness), and samples for compression tests were prepared in a cylinder mold (10 mm in height, 8 mm in diameter). Cross-linking of photoinitiated samples was exposed under 365 nm UV light (60 W, incident light intensity 2.5 W/cm²) for 5 min. The UV light source used in this study was purchased from ShenZhen HowSuper Technology Co., Ltd., with a light power of 60 W. The distance between the light source and the samples was set to 5 cm, and the resulting UV exposure area was approximately 23.8 cm². Then the incident light intensity was calculated to be 2.5 W/cm². The experimental setup images are shown in Figure S1.

4.5. Mechanical Properties Test

The strain–stress relationships of hydrogels were assessed using an MTS Criterion 43 testing machine with a 1000 N loading cell. For tensile tests, dog-bone-shaped samples were stretched at a rate of 20 mm/min. For compression tests, cylindrical hydrogels were compressed at a speed of 10 mm/min with up to 90% strain. The resulting data were converted into stress–strain curves, and the modulus was calculated from the initial slope. Measurements were taken on three samples and averaged. The cyclic compression test was conducted at a constant speed of 20 mm/min.

The shear moduli (G) of the hydrogels were obtained by a compression test. Based on an affine network model, the shear modulus can be determined from a slope of σ vs (λ – λ^–2) plot, where σ is the elastic stress and λ is the extension ratio of the hydrogels. The molecular weight between cross-links was calculated from the modulus data according to the rubber elasticity theory using the following equation:

$$M_{\mathrm{c}}=RTc/G$$

where M _c is the molecular weight between cross-links (g/mol) and c is polymer concentration (g/m³). T is the temperature (298 K), and R is the gas constant (8.3145 J mol^–1 K^–1). ,

4.6. Statistical Analysis

Each experiment was conducted with a minimum of three samples, and data were presented as means ± SD. Statistical analysis was performed using one-way ANOVA with Tukey’s test. Statistical significance was denoted as follows: *P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.0001.

The Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/acsomega.5c03346.

• Initiator systems for vinyl-polymered hydrogel; statistical analyzed data of mechanical properties for PAm, PEGDA, GelMA, AlgMA, and ChMA hydrogels; images of experimental setting for UV-initiating process; cyclic compression test curves for PAm, PEGDA, and GelMA hydrogels; pore size distribution diagram of hydrogels; FTIR spectra of PAm hydrogels initiated by different concentration αKG; and UV–vis spectra of unreacted αKG in GelMA hydrogels (PDF)

The authors declare no competing financial interest.