Influence of Polyols on the In Vitro Biodegradation and Bioactivity of 58S Bioactive Sol–Gel Coatings on AZ31B Magnesium Alloys

The mechanical qualities of AZ31B magnesium alloys make them a promising material for biodegradable metallic implants. However, rapid degradation limits the application of these alloys. In the present study, 58S bioactive glasses were synthesized using the sol-gel method and several polyols such as glycerol, ethylene glycol, and polyethylene glycol, were used to improve the sol stability and to control the degradation of AZ31B. The synthesized bioactive sols were dip-coated onto AZ31B substrates and then, characterized by various techniques such as scanning electron microscopy (SEM), X-ray diffraction (XRD) and electrochemical techniques (potentiodynamic and electrochemical impedance spectroscopy), among them. FTIR analysis confirmed the formation of a silica, calcium, and phosphate system and the XRD the amorphous nature of the 58S bioactive coatings obtained by sol-gel. The contact angle measurements confirmed that all the coatings were hydrophilic. The biodegradability response under physiological conditions (Hank’s solution) was investigated for all the 58S bioactive glass coatings, observing a different behaviour depending on the polyols incorporated. Thus, for 58S PEG coating, an efficient control of the release of H2 gas was observed, and showing a pH control between 7.6 and 7.8 during all the tests. A marked apatite precipitation was also observed on the surface of the 58S PEG coating after the immersion test. Thus, the 58S PEG sol-gel coating is considered a promising alternative for biodegradable magnesium alloy-based medical implants.


Introduction
In the year 2022, the value of the global market for bio-implants was projected to amount to USD 117 billion. When looking into the future, it is anticipated that the market will reach USD 189.0 billion by the year 2028, expanding at a compound yearly growth rate (CAGR) of 8.12% from 2023 to 2028 [1]. Countless individuals suffer bone loss annually related to aging, accidents, congenital, or lifestyle-related diseases. Without surgery, the bulk of bone loss cannot be treated. Metals such as titanium or stainless steel in the form of screws, plates, and rods are currently the most popular biomedical implants for these surgeries, and they are often removed between around 2 and 5 years after the bones have healed [2].
Biomedical implants can be made of metals, ceramics, polymers, or composites. They can also be made of hydrogels, nanofibers, thin films, and scaffolds [3,4]. Polymeric and composite materials can be made into both load-bearing and non-load bearing, nondegradable and degradable implants. On the other hand, most metallic and ceramic However, the preparation of sol-gel derived bioactive glass requires good control of the synthesis parameters together with an appropriate selection of the precursors and thermal treatment. Several reports suggested that the use of organic polymers during the silica polymerization produces changes in the network formation, affecting the structural properties and aging of silica sol [41][42][43]. In the study on sodium alginate/glycerol thin films by Giz et al. the importance of glycerol in relation to structural properties such as stress relaxation in thin films is presented. According to their report, glycerol as an additive acts as a lubricant by replacing the stress-inducing hydrogen bonds in polyelectrolyte structures with hydroxides [44]. Stefanescu et al. report the importance of the interaction of ethylene glycol with silica precursors such as TEOS in obtaining a non-porous silica matrix with a large surface area, which can be used in numerous areas of silica sol-gel application [45]. Catauro et al., reported the preparation of a PEG/SiO 2 hybrid sol-gel coating on grade 4 titanium and observed improvement in biocompatibility and 3T3 cell proliferation as the result of PEG addition. PEG in the hybrid sol was also found to influence the hydrophilicity of the coating. Crack-free coatings were achieved at a high PEG concentration in the sol [46]. Such coatings can protect metal substrates against corrosion or, in the case of biodegradable metals such as magnesium and its alloys, they can control the degradation profile of the metals. Glycerol, ethylene glycol, and polyethylene glycol are examples of polyols with excellent bioavailability, biocompatibility, and biodegradability [47,48]. In a previous report, ethylene glycol was used as a binder for sol-gel coatings on magnesium alloys [35]. However, little attention has been paid to the sol-gel synthesis of polyol-based bioactive glass sols and their use in sol-gel derived coatings. Most studies have concentrated on hybrid silica sol coatings rather than bioactive glasses. Consequently, the deposition of polyol sol-gel bioactive glass sols can provide interesting insights for future bioactive glass coating applications in tissue engineering and biodegradable implants.
The objective of the present study is the preparation of 58S bioactive glass coatings on AZ31B Mg alloys by sol-gel method to control corrosion and to obtain bioactive coatings. The addition of different polyols such as glycerol, ethylene glycol, and polyethylene glycol with an average molecular weight of 200 g·mol −1 is considered to optimize the sol stability and obtain an appropriate silica-calcium-phosphate network. The hybrid organic-inorganic (polyol-Si-Ca-P) bioactive sols and their coating deposition on AZ31B alloys were systematically studied in terms of their structural, bioactive, and biodegradation characteristics. The prepared polyol-modified bioactive sol-gel coatings promoted the biodegradability and biocompatibility of AZ31B alloys under physiological conditions without cytotoxic effects.

58S Sol Preparation
The 58S glass (60 mol % SiO 2 : 36 mol % CaO: 4 mol % P 2 O 5 ) was prepared by using respectively TEOS, TEPI, and Ca-L-Lac as the precursors of SiO 2 , P 2 O 5 , and CaO, respectively. The 58S sol was prepared in two steps. First, Ca-L-Lac was dissolved in  Simultaneously, TEOS, TEPI, and a polyol (GLY, EG, or PEG) were mixed  and stirred for 1 h. Then, acidulated water (1N HNO 3 ) was added drop by drop to the  TEOS/TEPI/polyol solution and maintained under stirring for 1 h. Subsequently, Ca-L-Lac/methanol solution was added slowly to the sol and stirred for 1 h. The molar ratio of TEOS, TEPI, Ca-L-Lac, methanol, water, and polyol was fixed to 1.0: 0.14: 0.6: 7.5: 2.4: 4.0 and the total oxide concentration was 204 g/L. A similar sol was prepared without the addition of polyol. The different sols were respectively labeled as 58S GLY, 58S EG, 58S PEG, and 58S WP for the sols obtained respectively with glycerol, ethylene glycol, polyethylene glycol 200, and without polyol, respectively.

Dip Coating Process
The 58S coatings with and without polyol were deposited on microscopic glass-slide and AZ31B substrates. Before dip coating, the glass slides were ultrasonically cleaned in ethanol and dried with air. The sheets of commercial AZ31B were cut into plates 2 × 2 cm 2 for pH and hydrogen evolution tests and 5 × 2 cm 2 for structural, contact angle, immersion, and electrochemical characterizations. The AZ31B substrates were polished with silicon carbide sheets ranging from 120 to 2500 grit, cleaned with ethanol, and then, dried with air. All samples were dip-coated with a withdrawal rate of 30 cm/min. Coated samples were initially dried at room temperature for 1 h and subsequently cured at 160 • C for 1 h.

Ellipsometry Characterization
The thickness and the refractive index of 58S coatings deposited on microscopic glass slides after curing at 160 • C/1 h were determined by spectroscopic ellipsometry (alpha SE, J.A. Woollam). The measurements were made in the wavelength range of 250-1000 nm and at the angle of incidence 65 • .

X-ray Diffraction Studies
An analytical X'Pert PRO theta/theta diffractometer (Malvern Panalytical, Madrid, Spain) was used to measure the crystal structure of the coatings. A grazing angle of 0.5 • and 2θ range of 10-80 with a step size of 0.05 • and accumulation time of 20 s per step was used. Cu-Kα radiation (λ = 0.15418 nm) was used as the excitation source was used.

FT-IR Analysis
The structural characterization of 58S sols with and without polyols was carried out by Fourier Transformed Spectroscopy (FTIR, Perkin Elmer Spectrum 100 spectrometer, PerkinElmer, INC, Madrid, Spain), using a spectrometer with an attenuated total reflectance accessory (ATR). FTIR spectra were measured with a resolution of 4 cm −1 and 8 scans for each measurement.

Contact Angle Measurements
A 'Drop Shape Analysis System' Kruss DSA 100 system (Kruss, Hamburg, Germany) was used to measure the water contact angle. 40 µL of Hank's solution was dropped over the sample surface on three randomly selected areas and the average contact angle was measured.

Hydrogen Evolution Studies
The release of hydrogen from uncoated and coated AZ31B samples was evaluated using the procedure described by Song et al. [49]. The purpose of this experiment was to measure the amount of hydrogen gas released by the samples in Hank's solution. The sample was immersed in Hank s solution and the displacement of the solution in a graduated burette was measured [49] The amount of volume displaced is directly proportional to the amount of hydrogen gas released. For the experiment, the surface of the samples exposed to Hank's solution was fixed to 1 cm 2 and 20 mL of Hank's solution was used. The device was also placed in a thermostatic bath to maintain the temperature at 37 • C ± 1 • C. The measurement was repeated three times to verify the reproducibility.

pH Evaluation
The bare and coated AZ31B samples were immersed in Hank's solution Solution at 37 • C ± 1 • C for seven days. The pH was monitored every 24 h. Similar to hydrogen evolution, samples with an exposed surface sample of 1 cm 2 and 20 mL of Hank's solution were used. Samples were analysed in triplicate to verify the reproducibility.

Immersion Studies in Hank's Solution
Immersion tests were conducted for seven days to study the formation of an apatite layer on the surface of uncoated and coated AZ31B samples. The samples were immersed in Hank's solution. On the third and seventh days, they were extracted and analyzed using Raman spectroscopy. Raman spectra were recorded using a confocal Raman microscope (WiTec, Ulm. Germany) with a spectral resolution of 0.02 cm −1 coupled with an AFM instrument (ALPHA 300RA, WiTec, Ulm, Germany) and with 532 nm excitation laser. The images were analyzed with the WiTec Project Plus 2.08 software.

Electrochemical Characterization of the 58S-Coated Samples
The degradation of AZ31B alloys and 58S coated alloys was investigated using DC and AC signals in a Gamry FAS2 electrochemical unit (Gamry, United Kingdom, Warminster). Electrochemical impedance spectroscopy (EIS) was also applied, using a saturated calomel electrode (SCE, Radiometer, Hach Lang GmbH, Germany, Düsseldorf) as the reference electrode, metal samples as the working electrode, and platinum wire as the counter electrode. Hank's Balanced Salt Solution was used as the electrolyte and an area of 0.78 cm 2 was exposed to the solution. The EIS was done in the frequency range from 105 Hz to 0.1 Hz with a sinusoidal amplitude of 10 mV rms AC voltage applied at the open circuit potential. Three measurements of each sample were performed, and the most representative measurement was plotted. Zview 2.0 programme was used to fit the impedance plots with a compatible equivalent circuit.

Characterization of 58S Sols without and with Different Polyols and Their Coating Deposition
Transparent 58S sols were obtained without the addition of polyol and with the addition of different polyols; GLY, EG, and PEG. Table 1 summarizes the aging times of 58S GLY, 58S EG, 58S PEG, and 58S WP sols, determined as the day in which the sol turns into a gel. The 58S PEG sol had the lowest stability, at 3 days. This was associated with the polymer long-chain and high molecular weight of PEG compared to the other polyols. It also contains more hydroxyl (OH) groups than other polyols, promoting the hydrolysis reaction and the crosslinking between the PEG and TEOS hydrolyzed molecules, thus increasing the viscosity. The 58S and 58S GLY sols were stable for 4 days, and the 58 EG sol remained stable for a longer period of 5 days. In general, the hydrolysis and polymerization reactions of TEOS take place in presence of water and under acidic/basic conditions to produce a crosslinked Si-O-Si network depending on the synthesis conditions. In our scenario, the presence of a polyol alters TEOS hydrolysis; as a result, hydrolyzed TEOS molecules (Si-O-) can interact with the polyol molecules (hydroxyl groups) and produce Si-O-C bridges, altering the formation of the Si-O-Si network as described by Ravaine et al. [43]. The degree of modification depends on polyol chain length and molecular weight and affects the sol stability together with the coating's structural and biological properties [45]. Figure 1 depicts the possible reaction between the TEOS precursor and polyols. For coatings, the stability of the 58 EG sol is the only factor to consider; however, for a degradable implant application, the stability of the 58 EG sol is merely one of several factors to consider.
with the coating's structural and biological properties [45]. Figure 1 depicts the possible reaction between the TEOS precursor and polyols. For coatings, the stability of the 58 EG sol is the only factor to consider; however, for a degradable implant application, the stability of the 58 EG sol is merely one of several factors to consider.  58S coatings were prepared using the 58S GLY, 58S EG, 58S PEG, and 58S WP sols by the dip-coating process at 30 cm/min and heat treatment at 160°C/ 1 h. The thickness of the coatings was measured by ellipsometry, as reported in Table 1. In general, the thickness of the coating is affected by the addition of polyols, with the 58S PEG coating being the thickest, around 1.76 µm. Polyol was added, and the result was associated with the viscosity of the 58S PEG sol. The thickness of 58S WP and 58S EG coatings was similar, 1.5 µm, and the thickness of 58S GLY coating was 1.65 µm. In a similar study by Agustín-Sáenz et al. on mesoporous silica coatings, the authors report the increase in coating thickness with respect to the different surfactants used in their study [50]. Figure 2 shows the XRD patterns of 58S WP, 58S GLY, 58S EG, and 58S PEG powder samples obtained after drying the sols at 160°C for 1 h. Only broad diffraction bands centered around 22° (2θ) were observed, suggesting that all powders were amorphous. The addition of polyols to the 58S sol did not affect the amorphous nature or the appearance of secondary silicate phases [51]. 58S coatings were prepared using the 58S GLY, 58S EG, 58S PEG, and 58S WP sols by the dip-coating process at 30 cm/min and heat treatment at 160 • C/ 1 h. The thickness of the coatings was measured by ellipsometry, as reported in Table 1. In general, the thickness of the coating is affected by the addition of polyols, with the 58S PEG coating being the thickest, around 1.76 µm. Polyol was added, and the result was associated with the viscosity of the 58S PEG sol. The thickness of 58S WP and 58S EG coatings was similar, 1.5 µm, and the thickness of 58S GLY coating was 1.65 µm. In a similar study by Agustín-Sáenz et al. on mesoporous silica coatings, the authors report the increase in coating thickness with respect to the different surfactants used in their study [50]. Figure 2 shows the XRD patterns of 58S WP, 58S GLY, 58S EG, and 58S PEG powder samples obtained after drying the sols at 160 • C for 1 h. Only broad diffraction bands centered around 22 • (2θ) were observed, suggesting that all powders were amorphous. The addition of polyols to the 58S sol did not affect the amorphous nature or the appearance of secondary silicate phases [51].

FT-IR Analysis
The powder samples used for the XRD analysis were also analyzed by FTIR to demonstrate the formation of the 58S glass structure and to study the functional groups in the glass (Figure 3). A broad band at 450 cm −1 associated with the rocking motion perpendicular to the Si-O-Si plane, together with two wider small peaks located around 770 and 850 cm −1 , related to the symmetric vibrations of PO and the bending of the Si-O-Si band were observed. The widest and most intense band around 900-1300 cm −1 is composed of the overlap of the most important bands for this composition. The P-OH (940 cm -1 ) stretching, and Si-O-Ca (960 cm −1 ) vibration modes were identified in the broad band between 920 and 990 cm −1 , together with the rest of the vibration modes of non-condensed Si-OH. The peaks identified at 1040 and 1190 cm −1 can be attributed to Si-O-Si asymmetric and symmetric stretching and P-O-P stretching. The shoulder band at 1230 cm −1 corresponds to the symmetric and antisymmetric modes of calcium atoms linked to silica as Si-O-Ca [52]. Between 1330 and 1500 cm −1 a very broad and poorly defined peak was observed, where the -CH group of the lactate group overlaps the vibration modes of nonbridging PO2 in PO4 3− (1448 cm −1 ). Their low intensity was related to the low content of phosphates in the composition of the bioactive glass. On the other hand, a band associated with the carboxylic group (1580 cm −1 ) of the lactate group appeared. Considering that lactate is a component of human metabolism, it is physiologically degradable and its presence in the coatings does not represent any problem in terms of biocompatibility. In the region of 2500-3500 cm −1 (not shown), bands associated with O-H stretching of adsorbed water and the Si-OH and P-OH groups [47] could be identified.

FT-IR Analysis
The powder samples used for the XRD analysis were also analyzed by FTIR to demonstrate the formation of the 58S glass structure and to study the functional groups in the glass (  [52]. Between 1330 and 1500 cm −1 a very broad and poorly defined peak was observed, where the -CH group of the lactate group overlaps the vibration modes of non-bridging PO 2 in PO 4 3− (1448 cm −1 ). Their low intensity was related to the low content of phosphates in the composition of the bioactive glass. On the other hand, a band associated with the carboxylic group (1580 cm −1 ) of the lactate group appeared. Considering that lactate is a component of human metabolism, it is physiologically degradable and its presence in the coatings does not represent any problem in terms of biocompatibility. In the region of 2500-3500 cm −1 (not shown), bands associated with O-H stretching of adsorbed water and the Si-OH and P-OH groups [47] could be identified.

Contact Angle Measurements
The wettability of bioactive materials is an important parameter related to bioactivity and cell adhesion. The contact angle of AZ31B substrates with and without coatings was measured by dropping 40 µL of Hank's solution on the surface (Figure 4). The contact angle on the bare substrate was 71 • : hydrogen bubble generation was also observed indicating rapid reactivity and degradation of the Mg alloy. The deposition of a 58S coating decreased the contact angle to around 50 • for 58S GLY and 58S EG, 40 • for 58S WP, and 35 • for 58S PEG. The result was associated with the presence of silanol groups on the surface. No hydrogen bubbles were observed, suggesting at least partial or even complete elimination of degradation of the Mg alloy. For biodegradable implants, a highly hydrophilic surface is preferred over a hydrophobic one because it is more capable of interacting with cells, ions in bodily fluids, and with other biological entities [53,54]. A hydrophobic surface can minimize these interactions. The highest hydrophilicity was observed for the 58S PEG coating.

300
The wettability of bioactive materials is an important parameter related to bioactivity 301 and cell adhesion. The contact angle of AZ31B substrates with and without coatings was 302 measured by dropping 40 µL of Hank's solution on the surface (Figure 4). The contact 303 angle on the bare substrate was 71°: hydrogen bubble generation was also observed indi-304 cating rapid reactivity and degradation of the Mg alloy. The deposition of a 58S coating 305 decreased the contact angle to around 50° for 58S GLY and 58S EG, 40° for 58S WP, and 306 35° for 58S PEG. The result was associated with the presence of silanol groups on the sur-307 face. No hydrogen bubbles were observed, suggesting at least partial or even complete 308 elimination of degradation of the Mg alloy. For biodegradable implants, a highly hydro-309 philic surface is preferred over a hydrophobic one because it is more capable of interacting 310 with cells, ions in bodily fluids, and with other biological entities [53,54]. A hydrophobic 311 surface can minimize these interactions. The highest hydrophilicity was observed for the 312 58S PEG coating.

318
The hydrogen gas evolution for the uncoated and coated AZ31B alloys with 58S-319 based coatings obtained with the use of various polyols was measured in Hank's solution 320 at 37°C as a function of immersion time (up to 7 days or 168 h) to evaluate the in vitro 321

300
The wettability of bioactive materials is an important parameter related to bioactivity 301 and cell adhesion. The contact angle of AZ31B substrates with and without coatings was 302 measured by dropping 40 µL of Hank's solution on the surface (Figure 4). The contact 303 angle on the bare substrate was 71°: hydrogen bubble generation was also observed indi-304 cating rapid reactivity and degradation of the Mg alloy. The deposition of a 58S coating 305 decreased the contact angle to around 50° for 58S GLY and 58S EG, 40° for 58S WP, and 306 35° for 58S PEG. The result was associated with the presence of silanol groups on the sur-307 face. No hydrogen bubbles were observed, suggesting at least partial or even complete 308 elimination of degradation of the Mg alloy. For biodegradable implants, a highly hydro-309 philic surface is preferred over a hydrophobic one because it is more capable of interacting 310 with cells, ions in bodily fluids, and with other biological entities [53,54]. A hydrophobic 311 surface can minimize these interactions. The highest hydrophilicity was observed for the 312 58S PEG coating.

318
The hydrogen gas evolution for the uncoated and coated AZ31B alloys with 58S-319 based coatings obtained with the use of various polyols was measured in Hank's solution 320 at 37°C as a function of immersion time (up to 7 days or 168 h) to evaluate the in vitro 321

Hydrogen Evolution and pH Evaluation
The hydrogen gas evolution for the uncoated and coated AZ31B alloys with 58S-based coatings obtained with the use of various polyols was measured in Hank's solution at 37 • C as a function of immersion time (up to 7 days or 168 h) to evaluate the in vitro biodegradability and to obtain information about the corrosion resistance of the coating systems (Figure 5a).
Spontaneous corrosion of Mg substrate takes place as a result of its reaction in an aqueous environment according to the reaction.
so, the amount of released H 2 gas is proportional to the degradation rate of the alloy. The higher the H 2 gas emission, the stronger the alloy degradation. The bare alloy showed the highest evolution of hydrogen, with a maximum release of 4.5 mL/cm 2 H 2 gas after 160 h of immersion (Figure 5a). Some irregularities in hydrogen evolution were observed in all AZ31B coated samples, which was associated with the formation of the corrosion products that adhere to the corroded surface and then peel off. In the case of 58S GLY, the H 2 gas release was unstable: after an initial increase a decrease until 100 h was observed, followed by a marked increase up to a maximum value of 3.9 mL/cm 2 after 168 h of immersion. The coating 58S EG showed a linear release of H 2 gas after 75 h of immersion, followed by a constant release rate, reaching a maximum of 2.5 mL/cm 2 . The coatings 58S WP and 58S PEG displayed the most favorable H 2 gas release behavior. During the first 24 h of immersion in SBF, the volume of H 2 increased slowly and then decreased until reaching a stable stationary phase, without significant variations, until the end of the test (168 h). solution for 7 days at 37 °C, as shown in Figure 5b). Initially, a quick increase of pH from 344 7.2 to 8.4 during the first 24 h was observed for all the samples. However, for bare AZ31B, 345 the pH was maintained at 8.4 for the remaining 7 days while the pH of coated samples 346 decreased, achieving a steady value of 7.8-7.6. The lowest pH was measured for the 58S 347 WP coating. This increment was related to the release of Mg 2+ , Ca 2+ , and HPO4 2+ ions from 348 the substrate and/or coating to the medium. 349 In general, the coated samples provide better control of H2 and pH evolution than 350 the bare sample. In this case, after 24 h of immersion in Hank´s solution, the 58S coatings 351 could react with the medium, forming an apatite layer that would regulate the variation 352 of pH and favor tissue regeneration. Ng et al. reported that the magnesium degradation 353 rate was considerably higher at the pH values between 7.4 and 5.5, and the alloy was far 354 more stable between 7.4 and 8 [55]. After 20 hours of immersion, the coated samples main-355 tained a pH of 7.7-8, which favored the deposition of the apatite layer. Similar to H 2 evolution, the pH change was also monitored with the immersion time. An abrupt pH increase results in a local alkalization of the area causing poisoning and cell death. In the present study, all coatings, including bare AZ31B, were immersed in Hank's solution for 7 days at 37 • C, as shown in Figure 5b. Initially, a quick increase of pH from 7.2 to 8.4 during the first 24 h was observed for all the samples. However, for bare AZ31B, the pH was maintained at 8.4 for the remaining 7 days while the pH of coated samples decreased, achieving a steady value of 7.8-7.6. The lowest pH was measured for the 58S WP coating. This increment was related to the release of Mg 2+ , Ca 2+ , and HPO 4 2+ ions from the substrate and/or coating to the medium.
In general, the coated samples provide better control of H 2 and pH evolution than the bare sample. In this case, after 24 h of immersion in Hank s solution, the 58S coatings could react with the medium, forming an apatite layer that would regulate the variation of pH and favor tissue regeneration. Ng et al. reported that the magnesium degradation rate was considerably higher at the pH values between 7.4 and 5.5, and the alloy was far more stable between 7.4 and 8 [55]. After 20 hours of immersion, the coated samples maintained a pH of 7.7-8, which favored the deposition of the apatite layer.

Immersion Studies in Hank's Solution
To analyze the formation of the apatite layer on the 58S coatings and uncoated samples, AZ31B and 58S coated with and without polyol were immersed in Hank s solution at 37 • C for 0, 3, and 7 days and the surface morphology was analyzed by Raman spectroscopy (Figure 6). In the bare sample, the formation of apatite was not confirmed, which was associated with the fast degradation rate of the Mg alloy. In HAP crystals, the characteristic vibration bands of PO 4 groups are (i) ν2 bending of P-O-P at 472 cm −1 , (ii) ν4 bending of P-O-P at 563 and 602 cm −1 , (iii) ν1 stretching of P-O at 960-962 cm −1 [56], and (iv) ν2 asymmetrical stretching of (PO 2 ) at 1200-1300 cm −1 [57]. The most noticeable change in amorphous calcium phosphate (ACP) is a 10 cm −1 shift of the ν1 stretching towards 950 cm −1 . As a result, the ACP to HAP transformation is characterized by a shift in the ν1 symmetrical band from a broad peak at about 950 cm −1 to a narrow peak around 960 cm −1 . For the 58S WP coating, the peaks attributed to apatite (ν4 P-O-P and ν1 P-O) and ν2 asymmetrical stretching of PO 2 (ν as PO 2 ) 1250-1300 cm −1 were observed after 3 and 7 days of immersion. The ν as PO 2 peaks were more intense after 7 days. In the case of 58S GLY coating, phosphate peaks were also identified, but with lower intensity. In addition, after 7 days of immersion, the intensity of the apatite peak remained unchanged, and the ν as PO 2 peak decreased. For 58S EG coating, the intensity of apatite and ν as PO 2 peaks increased gradually with the immersion time. The 58S PEG coating displayed the highest intensities of apatite and ν as PO 2 peaks of all coated samples during the whole duration of the experiment.

395
Electrochemical impedance spectroscopy (EIS) analysis was used to analyze the bio-396 degradation of the uncoated and coated samples during immersion in Hank´s solution at 397 37 °C. The study was conducted for a 58S PEG-coated sample and AZ31B bare Mg alloy 398 after 24 h, 3 and 7 days of immersion. Figure 7a,b shows the Bode and the phase angle 399 diagrams of both samples as a function of immersion time. The shape of the impedance 400 spectra and the value of the impedance modulus │Z│ at the low-frequency domain (f < 1 401 Hz) can be directly used to estimate the corrosion resistance performance of the coating 402 [59]. Figure 7a shows the │Z│ modulus (f < 1 Hz) of the bare Mg alloy after 24 h of im-403 mersion with a value of ~10 4 Ω cm 2 . After 3 days of immersion │Z│ modulus decreased 404 down to ~10 3 Ω cm 2 , one order of magnitude lower than after 24 h. However, after 7 days 405 of immersion, the impedance value increased again above 10 4 Ω cm 2 . To explain this be-406 havior, different electrochemical phenomena must be considered at the solution/substrate 407 interface. Due to a non-controlled degradation of the bare Mg alloy in Hank's solution, 408 the apatite layer was unable to form on the surface after 3 and 7 days, as documented by 409 the absence of apatite peaks in the Raman spectra (Figure 7a). The increment of impedance 410 after 7 days of immersion is thus associated with the formation and deposition of an un-411 stable layer composed of corrosion products, such as Mg·(OH)2 on the surface, which can-412 not provide a stable barrier and cannot control the hydrogen or pH evolution during the 413 immersion. Considering the H 2 gas evolution, the 58S GLY coating demonstrated the worst behavior associated with low apatite formation (Figure 6c) after being exposed to Hank's solution for a longer period. Then again, 58S PEG coating showed the most favorable properties in terms of H 2 and pH evolution as well as abundant apatite formation. This may be related to the presence of polyethylene glycol, which may have had an impact on the crosslinking of the silica network. A similar study by Li et al. reported that an increase in PEG (molecular weight 600) to silica ratio in the sol promoted the apatite formation [58]. Based on the result the 58S PEG coating was subjected to additional degradation analysis.

Degradation Analysis
Electrochemical impedance spectroscopy (EIS) analysis was used to analyze the biodegradation of the uncoated and coated samples during immersion in Hank s solution at 37 • C. The study was conducted for a 58S PEG-coated sample and AZ31B bare Mg alloy after 24 h, 3 and 7 days of immersion. Figure 7a,b shows the Bode and the phase angle diagrams of both samples as a function of immersion time. The shape of the impedance spectra and the value of the impedance modulus |Z| at the low-frequency domain (f < 1 Hz) can be directly used to estimate the corrosion resistance performance of the coating [59]. Figure 7a shows the |Z| modulus (f < 1 Hz) of the bare Mg alloy after 24 h of immersion with a value of~10 4 Ω cm 2 . After 3 days of immersion |Z| modulus decreased down to~10 3 Ω cm 2 , one order of magnitude lower than after 24 h. However, after 7 days of immersion, the impedance value increased again above 10 4 Ω cm 2 . To explain this behavior, different electrochemical phenomena must be considered at the solution/substrate interface. Due to a non-controlled degradation of the bare Mg alloy in Hank's solution, the apatite layer was unable to form on the surface after 3 and 7 days, as documented by the absence of apatite peaks in the Raman spectra (Figure 7a). The increment of impedance after 7 days of immersion is thus associated with the formation and deposition of an unstable layer composed of corrosion products, such as Mg·(OH) 2 on the surface, which cannot provide a stable barrier and cannot control the hydrogen or pH evolution during the immersion. When the 58S PEG coating was deposited on the AZ31B alloy, the Bode plot (Figure 419 7b) exhibited an impedance modulus of 10 3 Ω cm 2 with an inductive behavior at low fre-420 quency after 24 h of immersion. However, after 3 and 7 days of immersion, the low fre-421 quency inductive loop disappeared, and a capacitive behavior appeared along with an 422 increment of the impedance modulus to the values of about 10 5 Ω cm 2 . Comparing the 423 impedance behavior of AZ31B bare Mg alloy and 58S PEG coating, the increment of the 424 impedance modulus with the immersion time for the 58S PEG sample could be associated 425 with the formation of stable corrosion products such as apatite documented by Raman 426 spectroscopy (Figure 6e). 427 In order to comprehend the corrosion progress of 58S PEG coating with immersion 428 time, the EIS data was fitted using two different equivalent electrical circuit (EEC) as 429 shown in Figure 7 (c,d). The circuits are constituted of the Rs (the solution resistance), the 430 resistance and the constant phase element of the outer layer (Rout and CPEout), the charge 431 transfer resistance and constant phase element of the Faradaic reaction (Rdl and CPEdl) on 432 the substrate interface, and the inductive behavior in a low-frequency domain (f < 1 Hz) 433 (Rads and Lads). To estimate the total corrosion resistance for each system, the polarization 434 resistance (Rp) values were calculated according to the fitted equivalent circuit, and Rp 435 values are summarized in Table 2. When the 58S PEG coating was deposited on the AZ31B alloy, the Bode plot (Figure 7b) exhibited an impedance modulus of 10 3 Ω cm 2 with an inductive behavior at low frequency after 24 h of immersion. However, after 3 and 7 days of immersion, the low frequency inductive loop disappeared, and a capacitive behavior appeared along with an increment of the impedance modulus to the values of about 10 5 Ω cm 2 . Comparing the impedance behavior of AZ31B bare Mg alloy and 58S PEG coating, the increment of the impedance modulus with the immersion time for the 58S PEG sample could be associated with the formation of stable corrosion products such as apatite documented by Raman spectroscopy (Figure 6e).
In order to comprehend the corrosion progress of 58S PEG coating with immersion time, the EIS data was fitted using two different equivalent electrical circuit (EEC) as shown in Figure 7c,d. The circuits are constituted of the R s (the solution resistance), the resistance and the constant phase element of the outer layer (R out and CPE out ), the charge transfer resistance and constant phase element of the Faradaic reaction (R dl and CPE dl ) on the substrate interface, and the inductive behavior in a low-frequency domain (f < 1 Hz) (R ads and L ads ). To estimate the total corrosion resistance for each system, the polarization resistance (R p ) values were calculated according to the fitted equivalent circuit, and R p values are summarized in Table 2. The presence of an inductive loop after 24 h of immersion suggests the release of adsorbed ions [60], likely associated with the coating dissolution. The coating dissolution is dominated by the release of Ca 2+ ions and thus, by the ion exchange between the 58S PEG coating and Hank s solution [61]. After 3 days, and due to the release of Ca 2+ ions to the medium, nucleation sites are created on the surface of the sample, and the deposition of apatite on the substrate takes place. The large increment of R p value from 58.81 Ω cm 2 to 103,055.6 Ω cm 2 confirms that the corrosion resistance of the substrate was significantly enhanced due to the protective effect of the layer formed because of the precipitation of corrosion products. After 7 days, the R p value increased up to 172,157.4 Ω cm 2 . The slight increment of the Rp value was attributed to the saturation of the nucleation sites and the reduction of ion exchange activity substrate/medium. Although the precipitation rate of the corrosion products seems to be slower, the high Rp value, above 10 5 Ω cm 2 after 7 days, confirmed the effectiveness of the 58S PEG coating in terms of control of degradation of the AZ31B substrate in Hank s solution and confirmed the results of the measurements of hydrogen and pH evolution.
In conclusion, the in vitro bioactivity of the 58S PEG coating has been demonstrated. It should be noted that the amorphous bioactive materials showed higher bioactivity under physiological conditions than the crystalline ones [62]. The surface biomineralization has been induced and the degradation rate of the AZ31B Mg alloys was been reduced. The developed 58S PEG coating could be considered for potential application in degradable implants in biomedicine.

Conclusions
Three distinct polyols were incorporated into 58S sol, which was then dip-coated onto an AZ31B alloy substrate. The uniformity and particle homogeneity of the coatings were evaluated, as well as their structural and surface properties and bioactivity. The amorphous nature of 58S sols with and without added polyols was confirmed by structural analysis. Particle size, as measured by scanning electron microscopy (SEM), varies from 0.5 µm ± 0.03 for 58S WP to 0.3 µm ± 0.04 for 58S EG and from 0.1 µm ± 0.08 for 58S GLY to 4.6 µm ± 0.08 for 58S PEG, depending on the polyol used. The presence of elements linked to the composition of the corresponding 58S sol was confirmed by the chemical composition analysis (EDAX). The hydrophilicity of the contact angle decreased from 71 • for bare alloy to 35 • for 58S PEG coatings under physiological conditions. The linear H 2 evolution and pH instability of uncoated magnesium alloys with a maximum release of 4.5 mL/cm 2 and a pH of 8.4 were effectively controlled by 58S PEG coatings to an almost negligible amount of H 2 evolution and a pH of 7.8 upon 160 h of immersion. Due to the high bioactivity and high degree of crosslinking in the silica network, the PEG-based 58S coating demonstrated the best hydrophilicity, hydrogen evolution, and pH stability during immersion tests, which are essential prerequisites for biomedical applications. The observed control of the surface activity of the 58S PEG coating was attributed to the passivation of the surface through the formation of an apatite layer after 48 h, which was confirmed by the presence of various phosphate peaks in the Raman spectroscopy. The electrochemical impendence study of the degradability of the bare alloy and 58S PEG coating revealed that the 58S PEG coating could control the rapid degradation of Mg alloys under physiological conditions.
Further in vitro and in vivo studies are planned to analyze the efficiency of the 58S PEG coating on the degradability of magnesium alloys.