Ozone-loaded bacterial cellulose hydrogel: a sustainable antimicrobial solution for stone cleaning

The bacterial cellulose (BC) pellicle derived from kombucha was generously supplied in the wet state from an Italian brewer located at Meda (MB). This material will be referred to as hydrogel, since it is constituted by a polymeric network swollen with water. Indeed, this material differs from cryogel that is obtained by freeze-thawing cycles and from aerogel which retains a gaseous phase in the polymeric network. Samples of approximately 10 cm × 10 cm × 0.5 cm were treated three times with an aqueous solution of 0.5 M sodium hydroxide (NaOH, Merck, Darmstadt, Germany) at 85 °C for 1 h. Afterwards the BC was washed with H₂O_dd until neutral pH and stored in H₂O_dd at 4 °C until further use.

Samples of BC hydrogel were freeze-dried using a Lyovapor L-200 instrument (Buchi, Flawil, Switzerland) and the microstructure was studied using a Zeiss Auriga 405 Field Emission-Scanning Electron Microscope (FE-SEM) (Zeiss, Oberkochen, Germany). Samples were analysed at different accelerating voltages (varying between 1.0 and 1.2 keV). Fibril and aggregate diameter distributions were evaluated using ImageJ 1.53 k Java 1.8.0_172 software for 100 measurements each on SEM images of sample surfaces.

The BC crystallinity was studied on freeze-dried samples of approximately 2 cm × 2 cm by X-ray diffraction analysis using a D8 Advance diffractometer (Bruker AXS Gmbh, Germany). The diffractometer operates with Mo–Ka radiation (λ = 0.7093 Å, 40 keV, 35 mA) in Bragg–Brentano geometry. Angular scanning was carried out between 3.3° and 29.98° with a step of 0.01°. To minimise the background, a home-made "fork type" sample stage which uses a metal support with two rods and a Mylar film as the base was employed. X-ray diffraction patterns were converted in the Cu_Kα1 (λ = 1.5418 Å) wavelength. The crystallinity index (CrI) (%) was calculated according to Segal et al. (1959) (Eq. 1), where I₂₀₀ is the maximum intensity of the (2 0 0) lattice diffraction and I_am is the diffraction intensity related to the amorphous region at 2θ ∼ 18°.

The interplanar spacing (d) was calculated by means of Bragg's law (Eq. 2), where n (integer) is the order of reflection, λ is the wavelength of the incident X-rays and θ is the angle of incidence.

The average crystallite size (L) (nm) along the cell plane related to the main peak was calculated using the Scherrer equation (Eq. 3), where K_s is a shape factor constant in the range 0.8–1.2 (typically equal to 0.89), λ is the X-ray wavelength, τ is the peak full width in radians at half maximum (FWHM), θ is the Bragg angle.

The abundance of cellulose I_α and I_β allomorphs was estimated as reported by Wada and co-workers (Wada and Okano 2001) (Eq. 4), where d₁ (nm) is the d-spacing of \((100)\) _Iα and \((1\underline{1}0)\) _Iβ planes; d₂ (nm) is the d-spacing of \((010)\) _Iα and \((110)\) _Iβ planes. Z value is > 0 for I_α-rich type cellulose and < 0 for I_β-rich type cellulose.

To study the viscoelastic properties of BC, rheological measurements were carried out using a MCR302 Rheometer (Anton Paar GmbH, Graz, Austria). The analysis was conducted according to Numata et al. (2019) (Numata et al. 2019). Briefly, BC samples were cut into circular pieces with a diameter of 25 mm and held on the 25-mm flat parallel plates of the rotational rheometer. The gap was controlled by a normal force of 0.2 N. Every specimen underwent a dynamic strain sweep, ranging from 0.001 to 100% and frequency of 1 Hz, to determine the linear viscoelastic range. An optimal strain of 0.01% was identified and consequently employed for the frequency sweeps. The measurements were carried out at room temperature, with an angular frequency in the range of 0.1–100 rad/s.

Uniaxial elongation studies were performed on the BC hydrogel according to Pacelli et al. (2018) (Pacelli et al. 2018). Briefly, the hydrogels having a rectangular shape of 45 mm × 15 mm have been fixed in an Instron 5982 (Instron, Milan, Italy). The samples were elongated at room temperature at the rate of 1 mm/min until the breaking point. From the stress–strain curve, the maximum percentage of elongation ε, the maximum yield stress σ, and the elastic modulus E were derived.

BC hydrogels were loaded with ozone by using an OZONIS STERIL 250 series ozone generator (Sepra S.R.L., Italy) with a capacity of 5.2 mmol/h of ozone with a constant gas flow rate of 1.5 l/min. Ozonated BC (OBC) hydrogels with 0.223 mmol/g (OBC-L, low), 0.381 mmol/g (OBC-M, medium), and 0.531 mmol/g (OBC-H, high) ozone concentrations were obtained by varying ozonation times and water volumes. Specifically, OBC-L was obtained after 6 h of ozonation in 60 mL of H₂O_dd while the OBC-M was produced after 24 h in the same volume of H₂O_dd. The same time was instead applied in 20 mL of H₂O_dd to generate the OBC-H samples. As an alternative control the OBC-H samples were washed 2 times with H₂O_dd, until the ozone was discharged (OBC-W). Ozone concentration in the OBC hydrogels was estimated for each batch by calculating the mmol of ozone loaded in one sample (dry weight) as follows. After ozonation, the samples were immersed in 10 ml of H₂O_dd and left at 4 °C in a rotating wheel. After 20 min, an aliquot was collected and ozone concentration in water (mmol/L) was determined photometrically using an Hach DR1900 spectrophotometer with LCK310 kit (Hach Lange, Spain), based on the colorimetric method with N,N-diethyl-p-phenylenediamine (DPD). Water was subsequently exchanged, and the procedure was repeated until no ozone was further detected (ozone < 0.001 mmol/L). The moles of total ozone loaded in the BC hydrogel samples were estimated based on its concentrations in all the exchanged water aliquots and the total volume of water used. The OBC hydrogels were stored at 4 °C until further analysis.

To evaluate the potential interaction of ozone with cellulose, chemical structures of OBC-L, OBC-M, OBC-H and the pristine BC hydrogels were analyzed by Attenuated Total Reflectance-Fourier Transform Infrared Spectroscopy (ATR-FTIR). Spectra were acquired on freeze-dried samples using a Vertex 70 spectrometer (Bruker Optics, Gmbh, Ettlingen, Germany) equipped with a single reflection Diamond ATR cell, a standard MIR source (HeNe) and a room temperature DTGS detector. The spectra were recorded with 64 scans in the mid-infrared range (400–4000 cm⁻¹) at a resolution of 4 cm⁻¹.

The water holding capacity (WHC) of OBC-M, OBC-H and the pristine BC hydrogels was calculated according to the formula reported by Ul-Islam et al. (2012) (Ul-Islam et al. 2012) (Eq. 5), where M_wet is the mass of the water-swollen sample, and M_dry is the mass of the dried sample at 60 °C until no weight change is detected. Prior to weighing the wet sample, the excess of surface water was removed by tapping with a tissue. Measurements were conducted in quadruplicate.

The microstructure analysis, the XRD analysis as well as the rheological characterization of OBC-L, OBC-M and OBC-H, were performed as described before. To assess the effect of ozonation on cellulose mechanical properties, uniaxial elongation studies were performed on oven-dried OBC and pristine BC hydrogels at 60 °C. Samples having a rectangular shape of 45 mm × 15 mm have been fixed in an Instron 5982 (Instron, Milan, Italy) with a 500N load cell and elongated at room temperature at the rate of 2 mm/min until the breaking point.

The bacterial strains utilized in this study included the gram-positive Arthrobacter aurescens, Arthrobacter agilis, Kocuria rosea, and the gram-negative Achromobacter xylosoxidans species. They were cultivated in Luria–Bertani (LB) broth under aerobic conditions at 30 °C. The fungal strains employed belonged to Aspergillus jensenii, Cladosporium cladosporioides, and Alternaria alternata species, and were cultured at 30 °C in Sabouraud medium. All bacterial and fungal species had been previously isolated in Motya (Sicily, Italy).

To assess the antibacterial activity, 0.2 mL of cells with a concentration of 2 × 10⁶ cells/mL derived from overnight cultures of A. xylosoxidans, A. aurescens, A. agilis or K. rosea was incubated at room temperature in the presence of the three ozonated BC hydrogels. At specified intervals of 1, 10, 20, and 60 min of treatment, aliquots from the various samples were subjected to a series of dilutions to ensure an appropriate concentration for plating. The dilutions were prepared using sterile H₂O_dd to maintain the integrity of the samples. Subsequently, 100 µL of each diluted sample was spread onto LB (Luria–Bertani) agar plates using a sterile spreader to ensure uniform distribution. The plates were incubated at 28 °C for 18–24 h to allow for bacterial growth. Following incubation, the number of colony-forming units (CFU) on each plate was determined by counting the distinct colonies. Each colony represents a viable bacterial cell from the original sample. The CFU count was then used to calculate the bacterial concentration in the original sample, considering the dilution factor. All experiments were performed in triplicate. Bacterial samples incubated with pristine BC hydrogel served as control. The experiments were conducted in triplicate and repeated a minimum of three times. To evaluate possible membrane damages of bacteria treated with OBC-H hydrogel, the LIVE/DEAD BacLight (Thermo Fisher Scientific) was utilized according to the manufacturer's instructions. The kit is composed of two nucleic acid binding fluorescent dyes: SYTO 9 and propidium iodide (PI). With an appropriate mixture of SYTO 9 and PI stains, bacteria with intact cell membranes stain fluorescent green, scored as viable ones; and bacteria with injured membranes stain fluorescent red. Images were acquired with a Zeiss Axiovert 25 fluorescent microscope equipped with a 100X oil-immersion objective. The anti-sporogenic activity was assessed using 0.03 mL of a spores’ suspension with a concentration of 1 × 10⁴ spores/mL of C. cladosporioides, A. jensenii, or A. alternata. Each spore suspension was incubated in the presence of ozonated BC hydrogels. At intervals of 6 and 24 h, aliquots from each sample were diluted and subsequently plated on Sabouraud agar plates. After incubation at 30 °C, the spores’ ability to germinate was determined by CFU method. Spore samples incubated with a pristine BC hydrogel served as control. The experiments were conducted in triplicate and repeated a minimum of three times.

For these experiments, specimens of marble, brick, and biocalcarenitic stone measuring 5 × 5 × 2 cm were sterilized under UV light. For antibacterial assessment, 1 cm × 1 cm areas were chosen on each specimen, and 8 × 10⁸ cells suspension of A. aurescens was applied to each area. After overnight incubation at room temperature, the contaminated areas were treated with OBC-M or OBC-H hydrogels, while the pristine BC hydrogel was taken as a control. The hydrogels were covered with sterile plastic film to minimize water evaporation and were peeled off using tweezers after 2, 6, and 8 h. To assess the antisporogenic activity, marble and biocalcarenitic stone specimens were used. For each lithotype, a 1 × 10⁴ spores’ suspension of A. jensenii, A. alternata or C. cladosporoides in YPD medium was spotted on the selected area. After a three-day incubation at 30 °C, the mycelium that originated on the stone surfaces was treated by applying OBC-M or OBC-H hydrogels, while the pristine BC hydrogel was taken as a control. The hydrogels were covered with sterile plastic film to minimize water evaporation and were peeled off by tweezers after 6 and 24 h. The determination of bacterial and fungal survival was performed as reported in Santo et al. (2023) (Santo et al. 2023). The experiments were conducted in triplicate and repeated at least three times.

The possible presence of cellulosic residues on stone surface after the cleaning treatment was evaluated by µ-Raman analysis using a Senterra spectrometer (Bruker Optics, Gmbh, Ettlingen, Germany), equipped with a CW diode-pumped solid-state laser emitting at 532 nm. Analysis was conducted on a selected area of a marble specimen before and after a 24 h-treatment with the OBC-H hydrogel. The spectral resolution was 9 cm⁻¹ and time of acquisition and number of scans were selected for reducing the fluorescence interferences. Five spectra were acquired on random spots of the selected area. A Raman spectrum of the sample dried at 60 °C was acquired as a reference.

To assess the potential reuse of the ozonated BC hydrogel, a marble specimen was used in an antimicrobial test. The specimen was sterilized under UV light and an 8 × 10⁷ cells suspension of A. aurescens species was applied to selected areas. After overnight incubation at room temperature, one area was treated with OBC-H hydrogel, which was then covered with sterile plastic film to minimize water evaporation. After 2 h, the OBC hydrogel was peeled off by tweezers and immediately placed on another area for an additional 2 h. After each treatment, treated and control areas were carefully scraped and the CFU determined as reported in Santo et al. (2023) (Santo et al. 2023). The experiment was conducted in triplicate and repeated at least three times.

The antimicrobial activity of OBC-H samples stored in their ozonated batch water at 4 °C were assessed after 1, 5, 15 and 30 days. Antimicrobial activity was evaluated on A. xylosoxidans and A. aurescens bacterial species and on A. jensenii spores as previously described. Results were expressed as a percentage of activity relative to the fresh ozonated hydrogel. The experiments were conducted in triplicate and repeated a minimum of three times.

The microstructure of the BC sample used in this study was determined by FE-SEM analysis. A network of cellulose fibrils, which are individual structural units forming aggregates, was identified. Specifically, the analysis of the sample surface (Fig. 1 a, b) and cross-sectional area (Fig. 1c) show an open structure with high and randomly oriented porosity. The percent distributions of diameters show the most frequent values are in the range 40–45 nm and 80–90 nm for fibrils and aggregates, respectively (Fig. 1 d, e). The results are in accordance with BC description reported in previous studies (Hu et al. 2014; Gao et al. 2016).

Bild vergrößern

The BC crystallinity was investigated by XRD analysis. The diffractogram showed three peaks at approximately 2θ of 14, 17 and 23 (Fig. 2a). The diffraction pattern can be assigned to a type I cellulose found in native celluloses (Atalla and VanderHart 1984; French 2014). The peaks are assigned to the Iα and Iβ allomorphs, having Miller indices of \((100)\), \((010)\) and \((110)\), and \((1\underline{1}0)\), \((110)\) and \((200)\), respectively. Indeed, the I_α and I_β crystalline forms, respectively assigned to triclinic and monoclinic systems, are both present in the type I cellulose (Atalla and VanderHart 1984). The Z index introduced by Wada and Okano (2001) estimated the abundance of the two crystalline forms, as reported in the materials and methods section (Wada and Okano 2001). The Z positive value (19.5) indicates the prevalence of the Iα allomorph. The Segal crystallinity index (Cr.I.)(Segal et al. 1959) was evaluated equal to 94.0%, showing a high ratio between crystalline and amorphous phases, and the crystallite size calculated for the main peak at 2θ of 23 was 5.81 nm. The results are in accordance with BC diffraction data obtained by other studies (Wada and Okano 2001; Fawcett et al. 2013; Tsouko et al. 2015; Vasconcelos et al. 2017). Frequency sweeps, performed by rotational rheometer MCR302, on BC samples describe the behavior as a function of time, where for high frequencies fast motion is simulated over short times, while at low frequencies slow motion is simulated over long times or at rest. In this measurement, the storage modulus G', which represents the elastic part of viscoelastic behaviour, and the loss modulus G'', which characterizes the viscous part of viscoelastic behaviour, are measured. Figure 2b shows the trends of G' and G'' as a function of angular frequency, throughout the frequency range G' > G''. This behaviour indicates that the elastic response is predominant over the viscous response as also demonstrated in (Numata et al. 2019). This behaviour classifies BC as a self-supporting hydrogel (Orlando et al. 2020). In addition, the modulus G' exhibits a linear and nearly constant trend throughout the angular frequency range. The highlighted behaviour indicates that the cellulose network is not destabilized as the angular frequency increases, presenting high stability as the value of G' is strongly influenced by the intertwining of the fibrils (Chen et al. 2018). The BC mechanical properties were investigated by tensile testing. The stress (MPa)–strain (%) plot (Fig. 2c) displays a viscoelastic behaviour. The maximum percentage of elongation ε resulted in 15,5 ± 5,6%, the maximum yield stress σ is 18,4 ± 1,1 MPa and the elastic modulus E is 1,5 ± 0,6 MPa. The prevalent elastic behaviour, in accordance with the rheological results, is associated with the stretching of fibrils. The yield point and minor plastic deformation are related to the subsequent tearing and breaking of fibrils. The results are in accordance with other studies (Chen et al. 2018; Skvortsova et al. 2019).

Bild vergrößern

Ozonated BC (OBC) hydrogels with various ozone concentrations were produced by employing different ozonation durations and volumes of water. The resulting samples are designated as OBC-L, OBC-M, and OBC-H, with ozone concentrations of 0.223 mmol/g, 0.381 mmol/g, and 0.531 mmol/g, respectively. Initially, to assess the OBCs antibacterial efficacy, a colony-forming unit (CFU) counting analysis was conducted in a liquid assay using various microbial species. Following a 10-min treatment with OBC-M and OBC-H hydrogels, the mortality of A. aurescens exceeded 98%. Notably, OBC-H treatment resulted in a 60% cell mortality after just one minute. In contrast, OBC-L caused a 40% cell mortality after a 20-min treatment (Fig. 3a). Similar trends were observed in tests involving A. agilis and K. rosea species (data not shown). For A. xylosoxidans, treatment with OBC-H produced a 90% mortality after only one minute, reaching complete suppression of viability after 10 min (Fig. 3b). The same level of suppression was achieved with OBC-M after 20 min, while a 5% survival rate was observed after 1 h of treatment with OBC-L.

Bild vergrößern

To evaluate a plausible mode of action of OBC hydrogels, the Live/Dead BacLight viability assay was used for analyzing if a bacterial cell membrane disruption occurred during the treatment. Indeed, while SYTO-9 penetrates and fluoresces green within all cells, the propidium iodide (red fluorescence) only penetrates within those with a compromised cell membrane. As shown in Fig. 3c, A.aurescens cells treated with OBC-H are stained with SYTO-9 and only few cells are stained with propidium iodide Instead, when the cells were treated with OBC-H the staining with propidium iodide was observed indicating that the ozonization affect the envelope of the cells. Indeed, the differences in antimicrobial efficacy observed among various bacterial species can be attributed to variations in bacterial cell wall composition, as A. aurescens is representative of gram-positives (alongside A. agilis and K. rosea), while A. xylosoxidans belongs to gram-negatives. However, the bacterial and fungal strains utilized in this study are established biodeteriogenic microorganisms, isolated from stone materials in various studies (Qi-Wang et al. 2011; Ettenauer et al. 2014; Tescari et al. 2018; Trovão et al. 2019).

Next, the anti-sporogenic effect against three fungal species was evaluated by assessing the germination ability of spores by the CFU counting method after treatment with various OBC hydrogels in a liquid assay. Specifically, for A. jensenii, a mortality rate exceeding 95% was evident as early as 6 h post-treatment with OCB-H hydrogel (Fig. 3d), reaching approximately 90% after 24 h. In the case of OCB-M treatment, a modest reduction in A. jensenii viability was observed at 6 h, decreasing to approximately 50% after 24 h. Conversely, treatment with the OBC-L sample showed no discernible effect on spore germination. Similar trends were observed for A. alternata and C. cladosporioides samples (Fig. 3e and f). In particular, the antibacterial effects of OCB-H are even more efficient, reaching approximately 100% after 24 h for the two fungal species.

Among the hydrogels tested, OBC-H has emerged as the most promising candidate due to its antimicrobial properties. Literature reports various antimicrobial strategies, encompassing both conventional and innovative biocides, aimed at effectively reducing microbial presence on surfaces. Among these methods are the application of preventive antimicrobial agents, such as ZnO-nanorods (Zn-NRs), graphene nanoplatelets decorated with Zn-NRs and silver nanoparticles (AgNP) in superficial coatings of stone (Carrillo-González et al. 2016; Schifano et al. 2020).

Recently, essential oils (EOs) have gained significant attention as natural biocides for the restoration of cultural heritage, serving as an alternative to chemical agents (Ranaldi et al. 2022). Nevertheless, their utilization on a larger scale is disadvantaged by factors such as production costs and a need of information regarding efficacy at low doses, treatment durability, and potential interactions with material substrates (Fidanza and Caneva 2019; Lo Schiavo et al. 2020; Romani et al. 2022).

The microstructure of OBC hydrogels was investigated by FE-SEM analysis. The OBC-H microstructure is shown in Fig. 4, where similar characteristics to the pristine BC properties are identified (Fig. 4 a, b, c). Indeed, the dimensional analysis does not show variation in the dimensions of fibrils and fibril aggregates, presenting the most frequent diameters in the range of 40–45 nm for fibrils and 80–90 nm for fibril aggregates (Fig. 4 d, e), as in pristine BC. Similar results were obtained for OBC-L and OBC-M samples (data not shown). Accordingly, the ozonation did not significantly affect the microstructure of the OBC hydrogels.

Bild vergrößern

Frequency sweeps on OBC hydrogels were used to determine their viscoelastic behavior (Fig. 5a). Like pristine BC, all samples showed a storage modulus (G') greater than the loss modulus (G'') across the frequency range, indicating that the elastic response is dominant. Additionally, the storage modulus (G') remained linear and nearly constant throughout the angular frequency range for all samples, demonstrating that the ozonated cellulose network remains stable even at higher frequencies. Furthermore, an increase in the elastic modulus of the systems can be observed with the higher content of ozone (Fig. 5b).

Bild vergrößern

The ATR-FTIR spectroscopy was employed to assess the effect of ozonation on the chemical structures of OBC-L, OBC-M and OBC-H samples. The infrared spectra of OBC-L, OBC-M, OBC-H and the pristine BC exhibit the characteristic absorption of pure cellulose as compared in Fig. 6. A new signal at 1706 cm⁻¹ in the spectra of all ozonated BC hydrogels has been observed, not present in pristine BC. Specifically, this signal is associated with the presence of carbonyl groups, more evident in OBC-H, due a partly weak oxidation of the BC cellulose, in agreement with literature reports (Chirat and Lachenal 1994; Costa et al. 2016; Tripathi et al. 2020).

Bild vergrößern

The estimated WHC of the pristine BC hydrogel is 100.8 ± 3.9, comparable with the value for OBC-M (100.9 ± 6.2), instead the WHC is higher for OBC-H (152.6 ± 13.5). Considering that the WHC is related to BC network porosity, fibril aggregation, and superficial area (Ul-Islam et al. 2012; Paximada et al. 2016), the ozonation does not significantly affect the OBC-M hydrogel. Instead, an increase in WHC of OBC-H is observed in accordance with the more evident presence of carbonyl groups observed by the FTIR spectra, attested also for other oxidized bacterial celluloses (Lai et al. 2013; Mohammadkazemi et al. 2019). This behavior is also in agreement with the higher elastic modulus of the OBC-H, which results in a greater swelling of the network structure.

The XRD analysis was performed on OBC-L, OBC-M, and OBC-H samples to evaluate possible changes in BC crystallinity due to ozonation. Similarly to pristine BC, the diffractograms showed the peaks typical of type I cellulose (Fig. 7a). Also in these samples, the presence of the Iα allomorph is indicated by the positive Z index values (28.0, 10.4, and 6.9 for OBC-L, OBC-M, and OBC-H, respectively). A slight reduction in the Cr. I. was assessed with respect to the BC (91.1%, 90.2%, and 89.1% for OBC-L, OBC-M, and OBC-H, respectively). This behavior is also observed in other studies of oxidized bacterial cellulose (Vasconcelos et al. 2020; Yang et al. 2021). Furthermore, the analysis revealed a variation in the height of the 100 peak relative to the 110 and 010 peaks in the samples. The pristine BC (pattern d in Fig. 7a or Fig. 2a) exhibited minimal preferred orientation when compared to the ideal pattern for randomly oriented Iα reported in French (2014). In contrast, both OBC-L and OBC-H samples, as well as the OBC-M, displayed a preferred orientation involving a rotation about the c-axis, favoring the 100 plane as also mentioned in French (2014). The ozonation process and/or lyophilization of OBC samples may promote the observed alignment along the fibril axis. This effect has been identified by various authors as 'uniplanar orientation' in bacterial cellulose (Takai et al. 1975; Cai and Kim 2010; Chen et al. 2011). The changes in orientation highlighted by the XRD analysis do not appear to largely impact the microstructure of the hydrogel, as revealed by SEM analysis, although slight modifications could not be excluded. To evaluate if ozonated BC has a different tensile strength with respect to pristine BC, dried OBC samples were subjected to tensile testing. As reported in Fig. 7b the profile of the two samples is almost identical. Similar results were also obtained for OBC-L and OBC-M (data not shown).

Bild vergrößern

To determine if the antimicrobial action of OBC hydrogel is due to the ozonated water that swelled the hydrogel or it is due to the ozonated part of the cellulose, the OBC-H has been dried and its antimicrobial properties were tested. Indeed, the activity of the dried OBC-H hydrogel was studied through inhibition halo measurement against A. aurescens, showing any activities (Fig. S1). According to the results, BC hydrogel can be thus identified as a carrier for ozone-dissolved water.

The shelf-life of the OBC-H hydrogel was evaluated by studying the antimicrobial activity against A. aurescens, A. xylosoxidans and A. jensenii spores in liquid assays up to 30 days from production (Fig. 8). For all the investigated period, the sample exhibited a 100% of activity against the bacterial species compared to the freshly produced OBC-H sample, demonstrating that the antibacterial activity is maintained. Instead, the activity against the fungal spores shows a 30% reduction after a 30-day storage period. This behaviour is attributed to the higher ozone concentration required against fungal spores compared to bacteria.

Bild vergrößern

After the evaluation in a liquid assay, the antibacterial activities of ozonated BC hydrogels were assessed on different lithotypes, employing the representative bacterium A. aurescens isolate (Fig. 9). In all the specimens tested, OBC-H treatment successfully resulted in almost 100% mortality within just 2 h. The OBC-M sample on marble and brick specimens showed a high degree of mortality similarly to OBC-H application, for all the time analysed. On the contrary, the OBC-M on biocalcarenitic stone resulted in a significant reduction in cell viability, surpassing 98% only after 8 h of treatment. Promising results were also achieved when testing the spore germination capability on biocalcarenite stone and marble, highlighting an antifungal activity of the OBC-H membrane exceeding 95% compared to the untreated control (Fig. 10).

Bild vergrößern

Bild vergrößern

In conclusion, after testing the antimicrobial efficacy on various lithotypes, OBC-H emerged as the most effective and promising candidate. This could potentially serve as a viable alternative to current biodeteriogen removal techniques in use, which, however, come with several disadvantages. Initially, mechanical removal methods were widely employed. Despite some advantages, these techniques may cause damage to the substrate. High-pressure water blasting, for instance, can push microorganisms deep into the heritage material (Cappitelli et al. 2020). Moreover, when microorganisms grow endolithically, penetrating pores, fissures, and cracks, mechanical removal becomes challenging. Residues in the form of single viable cells or entire colonies then become a source for rapid substrate recolonization. While natural biocides have shown success, there are drawbacks to using chemicals of natural origin, such as oils and plant extracts. The composition of these substances can vary significantly based on factors like harvesting season, geographical location, and other agronomic considerations (Nezhadali et al. 2014). Additionally, extraction methods influence the chemical composition of the extracts, making it essential to thoroughly characterize all-natural mixtures for experiment repeatability (Dhifi et al. 2016).

The potential presence of residual solvent or polymeric matter on cultural heritage surfaces after the treatment with gels generally raises concern. Indeed, the residues can impact the integrity and the appearance of treated materials, and they can be a source of degradation, particularly biodegradation (Domingues et al. 2013; Schifano et al. 2022).

In this study, the OBC hydrogels at all ozone concentrations have been easily peeled off by tweezers from the surface of the stone specimens, leaving no visible residues. To deep investigate the potential presence of cellulose residues on treated surfaces, a detailed analysis was conducted, as described in the materials and methods section, by μ-Raman spectroscopy on marble specimens (Fig. 11). After a 24-h treatment with OBC-H the Raman spectrum was collected (Fig. 11a) showing the only signals of calcite in the marble without the presence of cellulose signals (Fig. 11b). As already demonstrated for other self-standing peelable gels (Baglioni et al. 2021), the OBC-H hydrogel does not leave residues on treated surfaces.

Bild vergrößern