Electron Beam Irradiation on the Production of a Si- and Zr-Based Hybrid Material: A Study by FTIR and WDXRF

Sol-gel production of hybrid materials has, to some extent, revolutionised materials’ engineering and the way science and technology perceive the creation of new materials. Despite that, the method presents some limitations that are circumvented by radiation processing. Electron beam irradiation was used to promote synthesis of hybrid structures while using silanol-terminated PDMS, TEOS and TPOZ as precursors. Evaluation of the method’s performance was executed by gel fraction determination, WDXRF and FTIR-ATR. Results showed that, although there is some pre-irradiation reactivity between precursors, radiolysis induces scission on multiple sites of precursor’s structures, which induces hybrid network formation to a greater extent. Characterisation allowed determining electron beam irradiation to be effective in the creation of Si–O–Zr bonds, resulting in the production of a Class II hybrid material.


Introduction
Hybrid materials (HMs) are an example of technology's contribution to the evolution of societies and civilisations. HMs can be found in nature (bone, nacre and wood) but humans have prepared artificial ones since time immemorial. Some examples are the pre-history's Lascaux's hybrid paints, antiquity's Maya blue dye, the Chinese rice-lime mortars and the modern age's Prussian Blue pigment. Researchers have been inspired by the widespread existence of these materials, and so research and development in this field started a few decades ago. The understanding of the chemistry involved and the control over the preparing processes have enormously extended the variety and versatility of designed HMs for a wide range of applications; however, special attention has been devoted to silicon-based HMs since the 17th century, when the first silicates' gelation experiments took place [1][2][3].
Since then, the majority of novel HMs has been produced by sol-gel methods. In 1985, in the United States of America, silicon-based precursors-polydimethylsiloxane (PDMS) and tetraethylorthosilicate (TEOS)-were used for the first time to form a class II hybrid material by sol-gel methods. This designation is attributed to HMs whose structures are composed exclusively or partially by covalent, by iono-covalent, or by Lewis acid-base bonds linking the organic and inorganic moieties. Further materials were developed with the goal of incorporating transition metals oxides in PDMS-TEOS hybrid matrices [2,3]. These offer many options, from production to applications, covering a wide range, such as • catalysts not needed to promote cross-linking; • no solvents or water required; • residues after hybrid network formation are reduced; • higher cross-linking degree due to presence of more activated sites on precursors structures generated by radiolysis.
However, the preparation of HMs requires considerably high doses. This is often a drawback when γ-radiation is used due to the very long irradiation time needed and low radioresistance of samples' containers to highly accumulated doses.
Directing this work for applications in the cultural heritage field, it is possible to determine that there are recurrent deterioration forms affecting different types of historical materials that require multiple types of interventions. Loss of cohesion of porous materials such as ceramics, stone and mortars features as one the most problematic, along with loss of adhesion related to layered materials. Typically, this is the result of interaction with deteriorating agents such as water, pollutants, biological agents or exposure to mechanical stress [21]. Therefore, conservation-restoration interventions aim many times at consolidating and fixating materials and structures, while at the same time, it is necessary to resort to biocides or biocolonisation preventers and water repellents in order to minimise recurrence of deterioration. It is thus understandable that this implies procedures involving several techniques and materials. By resorting to a multifunctional material that can address several of these issues simultaneously, conservation-restoration interventions can become less time-consuming and requiring less human resources, therefore becoming less costly.
Centring the discussion on the PDMS-TEOS system, there are specific properties of the precursors that are appropriate in a material developed for this means. PDMS shows suitable flexibility, with elastomeric properties, transparency and lack of colour. It is also hydrophobic and presents high thermal stability and low susceptibility to degradation by agents such as ultraviolet light, oxygen or ozone [22]. Nevertheless, it presents poor structural properties and its biocompatibility [23] may, ironically, be incompatible with the objective of being used in the formulation of a material that should prevent biodeterioration. TEOS, on the other hand, has been widely used as a consolidant of historical materials and more occasionally on the formulation of protective coatings also employed on cultural heritage [7,[24][25][26][27]. This alkoxide (ALK) enables the network formation of PDMS-based materials and provides it structural properties besides enhancing the consolidation effect.
Building on this, the authors aimed at developing the same type of multifunctional material, resorting to electron beam irradiation. The advantages of this process include a more expedite control of dose rates, more safety for the operators and a very significant time economy, since much higher dose rates are easily achieved than with gamma irradiation, which decreases total processing time. The use of this method for the production of PDMS-TEOS hybrid materials and taking advantage of shorter irradiation times is possible mainly because there is little to no dependence of the PDMS-TEOS system on dose rate, as shown by a previous study [17].
Given the goal of conferring biocide or biocolonisation prevention properties to the final product and the aforementioned biocompatibility of PDMS, it was necessary to introduce a third precursor in the formulation. Based on the antimicrobial properties of zirconium [28][29][30] and the group's previous experience with Zr in PDMS-TEOS prepared by γ-irradiation, TPOZ was the choice of Zr precursor to be used.

Materials
The precursors were used as received from the suppliers ( Figure 1): dures involving several techniques and materials. By resorting to a multifunctional material that can address several of these issues simultaneously, conservation-restoration interventions can become less time-consuming and requiring less human resources, therefore becoming less costly. Centring the discussion on the PDMS-TEOS system, there are specific properties of the precursors that are appropriate in a material developed for this means. PDMS shows suitable flexibility, with elastomeric properties, transparency and lack of colour. It is also hydrophobic and presents high thermal stability and low susceptibility to degradation by agents such as ultraviolet light, oxygen or ozone [22]. Nevertheless, it presents poor structural properties and its biocompatibility [23] may, ironically, be incompatible with the objective of being used in the formulation of a material that should prevent biodeterioration. TEOS, on the other hand, has been widely used as a consolidant of historical materials and more occasionally on the formulation of protective coatings also employed on cultural heritage [7,[24][25][26][27]. This alkoxide (ALK) enables the network formation of PDMS-based materials and provides it structural properties besides enhancing the consolidation effect.
Building on this, the authors aimed at developing the same type of multifunctional material, resorting to electron beam irradiation. The advantages of this process include a more expedite control of dose rates, more safety for the operators and a very significant time economy, since much higher dose rates are easily achieved than with gamma irradiation, which decreases total processing time. The use of this method for the production of PDMS-TEOS hybrid materials and taking advantage of shorter irradiation times is possible mainly because there is little to no dependence of the PDMS-TEOS system on dose rate, as shown by a previous study [17].
Given the goal of conferring biocide or biocolonisation prevention properties to the final product and the aforementioned biocompatibility of PDMS, it was necessary to introduce a third precursor in the formulation. Based on the antimicrobial properties of zirconium [28][29][30] and the group's previous experience with Zr in PDMS-TEOS prepared by γ-irradiation, TPOZ was the choice of Zr precursor to be used.

Materials
The precursors were used as received from the suppliers (

Choice of Formulation
The choice of the best formulation was based on compositions previously produced using gamma irradiation and already characterised from the point of view of morphology, composition, structure and thermal behaviour [16,31].
Given the objective of conferring biocidal properties on the material produced, a preliminary assay was carried out to identify the formulation with the best performance

Choice of Formulation
The choice of the best formulation was based on compositions previously produced using gamma irradiation and already characterised from the point of view of morphology, composition, structure and thermal behaviour [16,31].
Given the objective of conferring biocidal properties on the material produced, a preliminary assay was carried out to identify the formulation with the best performance in this scope, evaluating the intrinsic bioactivity of these materials on selected microbiological contaminants. Best results were achieved by compositions with [TPOZ] = 20%. Formulations with mass ratios PDMS/ALK < 1 result in powdery samples, without cohesion, due to the absence of connections between the inorganic regions and the polymeric network [10]. For this reason, the chosen formulation was PDMS:TEOS:TPOZ with 67:13:20 m% (Precursors ratio is expressed in mass percentage (m%) and not weight percentage (wt%), since all the precursors quantities necessary for hybrids preparation were measured using an analytical electronic balance with internal calibration, which gives direct reading of samples masses), respectively-mass ratio PDMS/ALK > 2-since this resulted in monolithic, flexible, homogeneous, transparent and colourless materials [10]. Irradiations were also performed on PDMS samples to determine a dose threshold that could guarantee gelation in irradiated mixture samples.
A sample which was previously prepared by γ-irradiation was also used for comparison in this study, specifically for the gel point determination of electron beam-irradiated PDMS. The sample was irradiated in the 60 Co irradiator facility of Instituto Superior Técnico. The attained dose was 700 kGy at a dose rate of 30 kGy·h −1 .

Samples Preparation and Conditioning
Precursors were mixed by alternating mechanical stirring with a vortex mixer and ultrasound, keeping the vessel in a water bath at normal temperature.
Aliquots of~6 mL of the mixture making up to 5 mm in height were poured into colourless transparent polystyrene (PS) boxes with square bases with~3 cm sides. These were then flooded with gaseous N 2 , closed and sealed inside low-density polyethylene (LDPE) bags also flooded with N 2 , making sure the bag would not be inflated enough to affect irradiation geometry.
To minimise pre-irradiation reactions between the precursors, the already sealed samples were maintained at a temperature < 10 • C until placement in targeted position. Please address Appendices A and B for further details about the preparation and irradiation procedures and, in Figure A1, about beam range through Mix simulated with ESTAR [32].
Samples' names are attributed according to their composition and status as described in Table 1:

Irradiation Parameters
Irradiations were performed in a linear electron accelerator (LINAC, adapted from GE Saturne 41, EuroMeV, Paris, France) located at the Ionizing Radiation Facility (IRIS) from C2TN/IST-UL. The system was set to a 10 MeV electron beam whose peak current was 50 mA. The pulse width and repetition frequency were 4 µs and 25 Hz, respectively.
The absorbed doses were estimated using calibrated radiochromic films FWT-60 Far West Technology, Goleta, CA, USA. The routine dosimeters were placed between the LDPE bag and the PS plate sample holder.
Samples were irradiated for periods up to 55 min, in partial exposure times between 2 min and 12 min 25 s of continuous irradiation, to minimise temperature increase and effect. Temperatures after partial irradiations reached values between 29 • C and 42 • C. The samples were cooled between partial irradiation exposures by subjecting them to T ∼ −10 • C, until they reached T < 10 • C.

Gel mass Fraction Determination
Gel mass fraction was determined by calculating the ratio between mass of the sample after processing and mass of the sample before processing [33,34], according to Gel mass f raction % = mass a f ter processing mass be f ore processing × 100.
Samples processing was composed of the following steps: 1.
Extraction of unreacted materials and precursors' fragments by immersion in tetrahydrofuran for 72 h; 2.
Evaporation of extraction solvent and extracted substances by drying in air for 6 days; 3.
Drying at 80 • C for 12 h in a laboratory oven, to guarantee non-matrix materials' evaporation.

Wavelength Dispersive X-ray Fluorescence
Wavelength dispersive X-ray fluorescence (WDXRF) analyses were executed in CENIMAT/FCT-NOVA with a PANalytical XRD-WDS 4 kW AXIOS sequential spectrometer equipped with a rhodium tube (Almelo, The Netherlands).

Fourier Transform Infrared Spectroscopy
Chemical bonds in e-HMs were studied by Fourier transformed infrared spectroscopy (FTIR) in attenuated total reflectance mode (ATR). Acquisition was performed in the range 4000-400 cm −1 , with resolution of 4 cm −1 , for 64 scans. These analyses were performed with a Nicolet iS50 spectrometer by Thermo Scientific (Waltham, MA, USA).

Gel Fraction and Gel Point Determination
Irradiated PDMS samples were flowing after 68 kGy of accumulated dose, although very viscous and apparently non-flowing at D = 113 kGy. Gel mass fractions for these doses corresponded to 89.5% and 99.5%, respectively, indicating a significant increase in the number of created bonds in the matrix between the two dose values. The third dose, 213 kGy, corresponds to 99.4% gel mass fraction, showing no significant difference in the influence of accumulated dose on gel mass fraction regarding the previous point and, therefore, in crosslinking promotion. These factors corroborate the estimation of gel point of PDMS to be in the 68-113 kGy range of accumulated dose. Given this, this range was considered to be the threshold dose, above which it would be possible to achieve gel point in the 67:13:20 m% mixture of PDMS:TEOS:TPOZ. Table 2 and Figure 2 that none of the e-Mix samples presents a gel mass fraction close to 100%, in opposition to what happens with e-PDMS, which means that there was no complete gelation with the used doses. This also implies the lingering presence of fragments of the precursors in all irradiated materials, regardless of accumulated dose. There is, however, a significant increase in these values, from 53.8% to 85.5%, for materials irradiated with accumulated doses of 248 kGy and 320 kGy. This could point to this range as comprising gel point for e-Mix. Nevertheless, materials with accumulated doses above 388 kGy, should be studied in the future to clarify this point.

WDXRF Spectroscopy
Si/Zr ratios in atomic fraction (χat) calculated from WDXRF data represented in Table 3 show that all samples, with exception of e-Mix215, exhibit a lower content in Si than the theoretical value calculated for the as-prepared mixture. This decrease, according to Gomes et al. [17], is in all probability the reflection of evaporation of TEOS before irradiation, causing Zr content in each sample to apparently increase. Si content decrease occurs also in air-Mix, strengthening this hypothesis.

WDXRF Spectroscopy
Si/Zr ratios in atomic fraction (χ at ) calculated from WDXRF data represented in Table 3 show that all samples, with exception of e-Mix215, exhibit a lower content in Si than the theoretical value calculated for the as-prepared mixture. This decrease, according to Gomes et al. [17], is in all probability the reflection of evaporation of TEOS before irradiation, causing Zr content in each sample to apparently increase. Si content decrease occurs also in air-Mix, strengthening this hypothesis. It is also visible that e-Mix irradiated with 70, 132, 154, 178, 320, 370 and 388 kGy present a higher decrease in Si content than air-Mix. This suggests these samples may have further lost Si after irradiation, meaning the latter may have recombined with volatile fragments produced by radiolysis and evaporated due to not being incorporated in the postirradiation matrix. Another conclusion that can be drawn is that the higher stoichiometry favouring Zr content is achieved with 320 kGy, originating a network of approximately 1 atom of Zr to 4.12 atoms of Si proportion.
To understand whether Zr was incorporated in the network or not, WDXRF was again performed on samples after processing for unreacted materials and fragments' extraction.
The results in Table 4 and Figure 3 show that when comparing the samples before processing , there is an increase in Zr content in air-Mix and irradiated samples with up to 248 kGy, except e-Mix178, after processing. Like the exception, samples irradiated with doses 320-388 kGy present lower Zr contents. Higher Zr contents after extraction processing infer that the unreacted materials and fragments that are extracted after irradiation, and therefore never formed covalent bonds with the produced network, are richer in Si than in Zr. The remaining material is, opposingly, richer in Zr and the higher stoichiometryfavouring Zr content is achieved with 132 kGy, originating a network of approximately 1 atom of Zr to 4.12 atoms of Si proportion. The seemingly trend inverts after 248 kGy. For the doses above this value, and for 178 kGy, the Zr content decreases after processing, implying the extraction removes fragments that are richer in Zr than in Si content.  If sample e-Mix178 is considered an outlier, these values suggest that extracted Si-rich fragments are the result of PDMS and TEOS pre-irradiation alterations and that doses ≥ 320 kGy cause the degradation of the produced matrix by breaking it and producing more Zr-rich fragments.
However, given the noncompliant behaviour of e-Mix178, more attention should be devoted to this matter and the tendency observed must be confirmed.

PDMS
The identification and assignment of FTIR bands, by comparison with the literature, of PDMS and irradiated PDMS are summarised in Table 5.  However, a possibly relevant difference is the presence of a very discrete shoulder at 2927-2917 cm −1 in the spectra of e-PDMS113 and e-PDMS213 Figure 5). This wavenumber is associated with an asymmetric stretching of C-H in methylene which is the result of scission of Si-C and C-H bonds and subsequent recombination to longer aliphatic chains where -CH2-is present, that is Cn≥2Hn+1, if it remains bound to the polymeric chain or Cn≥2Hn+2 if it forms a free alkane. This would mean that recombined aliphatic fragments would integrate the new material's matrix or, at the very least, that free alkanes would remain trapped in it [10]. The fact that the presence of -CH2-groups occurs for e-PDMS113, e-PDMS213 and γ-PDMS700 strengthens the hypothesis of the gel point being in the previously mentioned range of 68-113kGy, meaning that a dose lower than 113 kGy would not be sufficient to induce radiolysis of methyl groups and subsequent recombination and cross-linking.  However, a possibly relevant difference is the presence of a very discrete shoulder at 2927-2917 cm −1 in the spectra of e-PDMS113 and e-PDMS213 Figure 5). This wavenumber is associated with an asymmetric stretching of C-H in methylene which is the result of scission of Si-C and C-H bonds and subsequent recombination to longer aliphatic chains where -CH 2 -is present, that is C n≥2 H n+1 , if it remains bound to the polymeric chain or C n≥2 H n+2 if it forms a free alkane. This would mean that recombined aliphatic fragments would integrate the new material's matrix or, at the very least, that free alkanes would remain trapped in it [10]. However, a possibly relevant difference is the presence of a very discrete shoulder at 2927-2917 cm −1 in the spectra of e-PDMS113 and e-PDMS213 Figure 5). This wavenumber is associated with an asymmetric stretching of C-H in methylene which is the result of scission of Si-C and C-H bonds and subsequent recombination to longer aliphatic chains where -CH2-is present, that is Cn≥2Hn+1, if it remains bound to the polymeric chain or Cn≥2Hn+2 if it forms a free alkane. This would mean that recombined aliphatic fragments would integrate the new material's matrix or, at the very least, that free alkanes would remain trapped in it [10]. The fact that the presence of -CH2-groups occurs for e-PDMS113, e-PDMS213 and γ-PDMS700 strengthens the hypothesis of the gel point being in the previously mentioned range of 68-113kGy, meaning that a dose lower than 113 kGy would not be sufficient to induce radiolysis of methyl groups and subsequent recombination and cross-linking. The fact that the presence of -CH2-groups occurs for e-PDMS113, e-PDMS213 and γ-PDMS700 strengthens the hypothesis of the gel point being in the previously mentioned range of 68-113kGy, meaning that a dose lower than 113 kGy would not be sufficient to induce radiolysis of methyl groups and subsequent recombination and cross-linking.

Precursors vs. Non-Irradiated Mix
All typical bands present in PDMS's spectrum are also present in the non-irradiated mixture's spectrum as presented in Figure 6 and Table 6. Nevertheless, there are shifts in the latter, as well as when comparing to TEOS and TPOZ, namely at the following points:

Precursors vs. Non-Irradiated Mix
All typical bands present in PDMS's spectrum are also present in the non-irradiat mixture's spectrum as presented in Figure 6 and Table 6. Nevertheless, there are shifts the latter, as well as when comparing to TEOS and TPOZ, namely at the following poin  Comparison between the spectrum of Mix and the alkoxides' shows that the mo significant difference is the absence in Mix's spectrum of some weak bands present in t alkoxides':   Comparison between the spectrum of Mix and the alkoxides' shows that the most significant difference is the absence in Mix's spectrum of some weak bands present in the alkoxides': At the same time, a discrete shoulder at 1033 cm −1 is only present in Mix's spectrum. This is attributed to Si-O-Zr. This implies pre-irradiation reactivity of precursors, as already proposed by Gomes et al. [10] and corroborated by WDXRF results.
Bands at 466 cm −1 and 493 cm −1 are attributed to the convolution) of bands at 472 cm −1 from TEOS with 459 cm −1 from TPOZ and 500 cm −1 from PDMS with 482 cm −1 from TPOZ, respectively.
The shifts associated with methyl groups could be explained by the scission of methyl groups and recombination with other fragments or activated centres either in PDMS's or in alkoxides' structures. Rebonding with higher atomic mass elements would decrease the vibrations' frequencies, since higher atomic mass shifts the vibration frequency to lower values while rebonding with lower atomic mass elements would result in the opposite effect. Another cause is the occasional substitution of Si by Zr. The most probable provenance of these Si atoms should be the silanol terminations, given the typically high reactivity of hydroxyl. Nevertheless, it could also happen in a region farther from the chain's ending, implying that PDMS should suffer chain scission and recombination with TPOZ even before irradiation. The shifts associated with Si-O-Si bonds seem to support this possibility, since the occasional substitution of Si in Si-O-Si by Zr is compatible with the mentioned shift. The shift regarding the phonon bands follows the same type of mechanism as the substitution of Si from the PDMS chain's backbone by Zr.
Analysing the TEOS spectrum, bands at 1073 cm −1 and 1100 cm −1 may be attributed to Si-EtOx vibrations since Si-OR groups give rise to strong bands (one or more) between 1110 cm −1 and 1000 cm −1 [44]. However, based on the study of TEOS hydrolysis [40], these bands may be attributed to Si-O-Si symmetric stretching in linear structures, indicating changes in TEOS structure before Mix preparation, as mentioned above, such as hydrolysis induced by humidity from the air [9]. Furthermore, it is evident that the absorbances associated with alkyl groups (-CH 3 and -CH 2 -), besides shifting probably due to atomic mass effect related to the presence of Zr, tend to disappear in Mix. Likewise, the disappearing of the bands at 1275 cm −1 and 1252 cm −1 attributed to C-O stretching and shifts at 603 cm −1 , 534 cm −1 and 482 cm −1 assigned by Colomer et al. [48] to (Zr-O)C stretching, both ensembles in propanol, indicating that this vehicle for TPOZ begins volatilising before irradiation. This event should render TPOZ more susceptible to reacting with its surroundings, causing its alteration.
In addition, the absence of most bands related to the alkoxide bonds in Mix's spectrum is in agreement with premature modifications of the precursors, which occur previously to irradiation.

e-Mix
Comparing Mix's and e-Mix's spectra, the most significant differences are the following bands only present in the spectrum of the non-irradiated mixture: Looking with further detail at the double band at 1078/1008 cm −1 , it is visible in Figure 7a that, besides full width at half maximum of 1008 cm −1 peak is higher for e-Mix, the intensities ratio between them does not remain constant, contrarily to what approximately happens with PDMS samples. As reported by Kongwudthiti et al. [50], infrared vibrations of Si-O-Zr bonds may go from 965 cm −1 to as far as 1025 cm −1 , in mixed oxides of silica-modified zirconia, depending on the concentration of Si and Zr.
In the present study, however, the ratio between concentrations of Si and Zr is much higher, as shown by WDXRF results. This, of course, should induce a shift of the band to higher frequencies due to atomic mass effect of Si. Moreover, Si-O-Zr bonds in these materials' matrix, are part of a much larger molecule, which also influences the shift. This is in accordance with the assignment of 1033 cm −1 band to Si-O-Zr in Mix's spectrum. The incorporation of higher contents of Zr in the matrix would then shift the band to lower frequencies. Considering the presence of a band at 493 cm −1 , assigned to phonons from Si-O-Si and Si-O-Zr allied to the varying ratio of intensities of Si-O-Si asym/sym stretching double band in e-Mix, it is most likely that 1008 cm −1 band is the result of the sum between Si-O-Si symmetric stretching and Si-O-Zr vibrations whose frequency should fall in this same value, thus increasing its FWHM and intensity.
These data would, therefore, mean that there is the formation of Si-O-Zr bonds in the electron beam-produced matrix, attesting to the presence of a hybrid material.
Additionally, as seen in Figure 7, int1008/int1078 is very similar both for Mix and air-Mix samples and, therefore, although there is some cross-linking between PDMS and the alkoxides before processing, the lack of irradiation prevents further incorporation of Zr in the network. There is neither obvious dose effect in this ratio evolution until 248 kGy; however, after this point, the value evidently approaches the ones calculated for (γ/e-)PDMS samples. This is in agreement with gel point estimation and suggests that, beyond this threshold, the network starts being destroyed by incident radiation and degrades by scission into fragments forming few to no Si-O-Zr bonds. Furthermore, although the higher stoichiometry-favouring Zr content for remaining materials after post-irradiation processing is achieved with 132 kGy, there is also a significant increase in Zr content between 248 kGy and 320 kGy, pointing, once again, to this range of accumulated dose as optimal for inducing the formation of Si-O-Zr bonds that integrate the As reported by Kongwudthiti et al. [50], infrared vibrations of Si-O-Zr bonds may go from 965 cm −1 to as far as 1025 cm −1 , in mixed oxides of silica-modified zirconia, depending on the concentration of Si and Zr.
In the present study, however, the ratio between concentrations of Si and Zr is much higher, as shown by WDXRF results. This, of course, should induce a shift of the band to higher frequencies due to atomic mass effect of Si. Moreover, Si-O-Zr bonds in these materials' matrix, are part of a much larger molecule, which also influences the shift. This is in accordance with the assignment of 1033 cm −1 band to Si-O-Zr in Mix's spectrum. The incorporation of higher contents of Zr in the matrix would then shift the band to lower frequencies. Considering the presence of a band at 493 cm −1 , assigned to phonons from Si-O-Si and Si-O-Zr allied to the varying ratio of intensities of Si-O-Si asym/sym stretching double band in e-Mix, it is most likely that 1008 cm −1 band is the result of the sum between Si-O-Si symmetric stretching and Si-O-Zr vibrations whose frequency should fall in this same value, thus increasing its FWHM and intensity.
These data would, therefore, mean that there is the formation of Si-O-Zr bonds in the electron beam-produced matrix, attesting to the presence of a hybrid material.
Additionally, as seen in Figure 7, int 1008 /int 1078 is very similar both for Mix and air-Mix samples and, therefore, although there is some cross-linking between PDMS and the alkoxides before processing, the lack of irradiation prevents further incorporation of Zr in the network. There is neither obvious dose effect in this ratio evolution until 248 kGy; however, after this point, the value evidently approaches the ones calculated for (γ/e-)PDMS samples. This is in agreement with gel point estimation and suggests that, beyond this threshold, the network starts being destroyed by incident radiation and degrades by scission into fragments forming few to no Si-O-Zr bonds. Furthermore, although the higher stoichiometry-favouring Zr content for remaining materials after post-irradiation processing is achieved with 132 kGy, there is also a significant increase in Zr content between 248 kGy and 320 kGy, pointing, once again, to this range of accumulated dose as optimal for inducing the formation of Si-O-Zr bonds that integrate the produced matrix. For the mentioned range, the achieved stoichiometry would then vary between 1:7.33 and 1:5.39 (Si:Zr).
Based on the presented results, the sequence of events represented in Figure 8 is proposed as a general mechanism of production of hybrid materials by electron beam irradiation in the PDMS:TEOS:TPOZ/67:13:20 system: Based on the presented results, the sequence of events represented in Figure 8 is proposed as a general mechanism of production of hybrid materials by electron beam irradiation in the PDMS:TEOS:TPOZ/67:13:20 system:

Conclusions
Electron beam irradiation is effective in promoting cross-linking in PDMS:TEOS:TPOZ/67:13:20 system and inducing Si-O-Zr bonds between the precursors. This results in the formation of a Class II hybrid material where covalent bonds are responsible for the network formation.
While there is some pre-irradiation reactivity between precursors, radiolysis promotes scission on multiple sites of precursors' structures, which induces hybrid network formation to a greater extent than that if sol-gel method was used, since the latter would promote cross-linking mainly originating on silanol terminations of PDMS. This system's gel point is estimated to be achieved for the dose range of 248-320 kGy.
Even though in-air curing is attainable, the degree of formation of Si-O-Zr bonds is heavily reduced and, as such, Class II hybridisation between the polymer and the alkoxides cannot be proven for these conditions.

Conclusions
Electron beam irradiation is effective in promoting cross-linking in PDMS:TEOS:TPOZ/ 67:13:20 system and inducing Si-O-Zr bonds between the precursors. This results in the formation of a Class II hybrid material where covalent bonds are responsible for the network formation.
While there is some pre-irradiation reactivity between precursors, radiolysis promotes scission on multiple sites of precursors' structures, which induces hybrid network formation to a greater extent than that if sol-gel method was used, since the latter would promote cross-linking mainly originating on silanol terminations of PDMS. This system's gel point is estimated to be achieved for the dose range of 248-320 kGy.
Even though in-air curing is attainable, the degree of formation of Si-O-Zr bonds is heavily reduced and, as such, Class II hybridisation between the polymer and the alkoxides cannot be proven for these conditions. While there is an evident effect of the irradiation in the structural hybridisation mechanism, there is not an obvious dose effect on the extent of these reactions, i.e., there is not a linear-neither direct nor inverse-correlation between absorbed dose values and direction or magnitude of reactions. Stopping time between partial irradiations may prevent the linear evolution of the studied parameters. Further studies should be carried out to clarify this point, as well as to address the reproducibility of results correlated to the studied doses. It is also important to introduce the temperature variable and understand its role in the process, as well as to gather more information about the material's morphology through SEM studies. Moreover, after establishing a production procedure, extensive characterisation and research should also be carried out regarding the produced material's application, namely as a multifunctional conservation-restoration product.
Simulation with ESTAR: stopping power and range tables for electrons from the National Institute of Standards and Technology [32] showed that, assuming 0.001205 g·cm −3 as the density of dry air at 20-25 • C, and 0.99238 g·cm −3 as the PDMS:TEOS:TPOZ mixture's density, beam range is~50 m and 5.1 cm in each material, respectively, confirming the beam can reach the sample and cross it with neglectable energy loss. Simulation with ESTAR: stopping power and range tables for electrons from the National Institute of Standards and Technology [32] showed that, assuming 0.001205 g‧cm −3 as the density of dry air at 20-25 °C, and 0.99238 g‧cm −3 as the PDMS:TEOS:TPOZ mixture's density, beam range is ~50 m and 5.1 cm in each material, respectively, confirming the beam can reach the sample and cross it with neglectable energy loss.