Introduction

Global climate and environmental changes affect the biodiversity and growth/metabolism of microbes1,2. Microorganisms use light energy, carbon dioxide, and minerals from stone objects for growth, releasing metabolites (e.g., organic acids and alkaloids) then can cause acidification, alkalization, and/or redox stress while promoting the participation of metal ions in complex reactions and the destruction of stone cultural relics3,4. Scenedesmus obliquus (S. obliquus) and Cyanobacteria, serving as pioneer colonizers, generate organic carbon via photosynthesis5,6. The acids produced can dissolve calcium carbonate within stone cultural relics, leading to surface corrosion and peeling, which subsequently exacerbates the weathering disease of these artifacts7.Therefore, new strategies for the protection of stone cultural relics against microbial weathering are highly sought after8.

The abovementioned microbial weathering is mainly mitigated using physical, biological, and chemical methods. Physical methods9 can accelerate the spreading of microorganisms, fail to completely remove them, or even destroy cultural relics. The harsh conditions, safety, and risks of biological methods4 cannot be assessed. Chemical methods10 mainly rely on chemical reagents or functional nanomaterials, as exemplified by bactericidal agents and essential oils. Pesticides are costly and difficult to degrade, thus having the capacity to harm the environment, human health, and cultural heritage itself, while essential oils are costly to apply and have not been evaluated in terms of longevity and safety. Nanomaterials enable sterilization and exhibit advantageous physicochemical properties, thus finding numerous applications as antifouling, photocatalytic, and antimicrobial agents11,12.

Solar radiation–responsive photocatalytic nanomaterials can rapidly kill microorganisms through advanced oxidative processes because of their ability to generate electron–hole pairs upon irradiation. The photogenerated electrons and holes interact with ambient oxygen and water, respectively, to produce reactive oxygen species (ROS), such as •O2−, •OH, H2O2, and singlet oxygen (1O2)13,14,15. These species react with the cellulose and pectin molecules comprising the bacterial cell wall to cause its fragmentation and then attack the peptidoglycans, phospholipids, and proteins of the exposed cell membrane to cause oxidative damage. Further attack on the intracellular components (nucleic acids, lipids, and proteins) ultimately results in cell death16,17. Wang et al.18 found that under ambient conditions, Escherichia coli was rapidly and efficiently killed by ultraviolet radiation upon contact with dispersed spindle filament–structured TiO2 photocatalysts. Baniamerian et al.19 showed that Fe2O3–TiO2 nanoparticles generated ROS under visible light irradiation and thus hindered the proliferation of green microalgae (Chlorella vulgaris) in freshwater and seawater. Serrà et al.20 designed a magnetically recyclable hybrid ZnO-based photocatalytic nanomaterial and evaluated its ability to kill two microalgae with different morphologies and toxicities under sunlight. Although dispersed photocatalytic nanomaterials have been widely used to kill suspended microorganisms, they suffer from inefficient light energy utilization, low killing and reuse efficiencies, and recycling difficulty. More importantly, the microbial killing mechanisms and performances of immobilized photocatalytic semiconductor films, especially in the context of stone cultural relic protection, are underexplored21,22. Therefore, the fabrication of photocatalytic films efficiently producing ROS and analysis of their action mechanisms are important for effectively preventing the microbial weathering of stone cultural relics.

To address the above knowledge gap, we herein hybridized BiVO4 (a narrow-bandgap semiconductor) with two-dimensional layered Ti2NTx-doped TiO2 and formulated the resulting BiVO4/Ti2NTx-TiO2 photocatalyst into a gel that was used to coat the surface of stone cultural relics to form photocatalytic films thereon. The thus treated relics were colonized with representative microalgae contributing to microbial weathering (Scenedesmus obliquus) to obtain an immobilized microorganism–photocatalytic film system. The photocatalyst morphology, composition, photoelectric conversion properties, and ROS production ability were examined using instrumental techniques, while the electron transitions and corresponding energy band characteristics were probed using density functional theory (DFT) calculations. The ability of the photocatalytic films to eradicate the microalgae was probed using absorbance (680 nm) and chlorophyll content measurements. In addition, we analyzed the action mechanism of the photocatalytic films and their effects on the structure of the microbial community on the relic surface.

Methods

Materials

TiO2 nanoparticles (60 nm), NH4VO3 (99%), Bi(NO3)3·5H2O ( ≥ 99%), citric acid monohydrate (≥99.9%), aqueous ammonia (25–28 wt%), anhydrous KF (99%), isopropanol (≥99.5%), ethylene glycol (>99%), and Ti(IV) isopropoxide (95%) were purchased from Aladdin. TiAlN (200 mesh) was purchased from Jilin Yiyi Technology Co. Aqueous HCl (36–38%) was purchased from Chengdu Cologne Chemical Co. Aqueous HNO3 (65–68%) was purchased from Chongqing Chuandong Chemical Group Co.

Photocatalytic material and gel preparation

Preparation of BiVO4/Ti2NTx-TiO2

TiO2 powder was calcined at 450 °C for 3 h. Ti2NTx (0.1 g; see Supplementary Method 1 for preparation details) and calcined TiO2 (1.9 g) were dispersed in deionized water, and the mixture was stirred for 6 h and centrifuged at 4000 rpm for 5 min. The precipitate was isolated and vacuum-dried at 80 °C for 12 h. The resulting composite (Ti2NTx-TiO2; 1.8 g) and BiVO4 (0.2 g; see Supplementary Method 2 for preparation details) were dispersed in deionized water, and the mixture was stirred for 8 h and centrifuged for 5 min. The precipitate was isolated and vacuum-dried at 100 °C for 12 h to afford BiVO4/Ti2NTx-TiO2. Herein, BiVO4 was selected as the dopant due to its bandgap of approximately 2.4–2.5 eV, which enables it to absorb visible light within the range of 400–500 nm and offers a high photoelectric conversion efficiency23.

Preparation of BiVO4/Ti2NTx-TiO2 gels

Ethylene glycol (2.225 mL) was heated at 60 °C in a water bath–immersed beaker upon constant stirring at 300 rpm and supplemented with Ti(IV) isopropoxide (0.5 mL). The mixture was heated to 90 °C in the water bath, treated with citric acid monohydrate (2.1 g), and stirred at 300 rpm until a clear solution was obtained. The solution was supplemented with BiVO4/Ti2NTx-TiO2 (2.1 g), stirred for 1 h, and stored in the dark to obtain the BiVO4/Ti2NTx-TiO2 gel.

Microalgae culturing and photocatalytic reaction system

S. obliquus (FACHB-13) was purchased from the Freshwater Species Bank of Wuhan Institute of Aquatic Biology, Chinese Academy of Sciences, and cultured using the BG11 medium in an intelligent artificial climate incubator (incubation temperature: 25 ± 0.1 °C, relative humidity: 70%, light cycle: 12 h light/12 h dark, light intensity: 2400 lx).

Sandstone blocks with dimensions of 12 cm × 4 cm were used to model stone cultural relics. The samples were polished with sandpaper, cleaned with deionized water, oven-dried at 150 °C for 30 min, wiped clean with anhydrous ethanol, naturally dried at room temperature, uniformly coated with the BiVO4/Ti2NTx-TiO2 gel (1.5 mL), and dried at 100 °C for 4.5 h. The resulting photocatalytic films had a thickness of ~120 μm. The film-coated and pristine samples were treated with the microbial solution (5 mL, OD680 nm = 0.1168 g/L), naturally dried, and set aside.

The setup used for the preventive conservation of stone artifacts (Fig. S1) comprised a light source, an acrylic rectangle (40 cm × 20 cm × 15 cm), a peristaltic pump, and a bottle of culture fluid. The acrylic rectangle contained an inclined support plate and a water flow conduit. The colonized sandstone samples were placed on the support plate, and the culture solution from the glass bottle was dropwise applied onto the sample surface using the peristaltic pump (flow rate = 25 mL/min).

Material characterization

Field emission scanning electron microscopy, energy-dispersive X-ray spectroscopy, X-ray photoelectron spectroscopy, X-ray diffraction, photoelectric response testing, ESR spectroscopy, and other characterization methods are described in Supplementary Method 3. DFT calculations are described in Supplementary Method 4.

Polymerase chain reaction (PCR) amplification and sequencing

DNA from microbial membranes stripped from stone was PCR-amplified from the V3-V4 variable region of the 16S rRNA gene using upstream primer 338 F (ACTCCTACGGGGAGGCAGCAG) and downstream primer 806 R (GGACTACHVGGGGTWTCTAAT) carrying barcode sequences. The amplification procedure was as follows: Predenaturation at 95 °C for 3 min, 27 cycles (denaturation at 95 °C for 30 s, annealing at 55 °C for 30 s, extension at 72 °C for 30 s) followed by stabilized extension at 72 °C for 10 min and storage at 4 °C (T100 Thermal Cycler, BIO-RAD, USA). The PCR system comprised a mixture of 5 × TransStart FastPfu buffer (4 μL), 2.5 mM dNTPs (2 μL), upstream primer (5 µM, 0.8 μL), downstream primer (5 µM, 0.8 μL), TransStart FastPfu DNA polymerase (0.4 μL), and template DNA (10 ng) made up to 20 μL. PCR products were recovered on a 2% agarose gel, purified using a DNA gel recovery and purification kit (YH-soil Soil DNA Extraction Kit, China), and quantified using Qubit 4.0 (Thermo Fisher Scientific, USA).

The library construction of purified PCR products (NEXTFLEX Rapid DNA-Seq Kit) involved (1) linkage splicing, (2) removal of splice self-associated fragments via magnetic bead screening, (3) enrichment of library templates via PCR amplification, and (4) magnetic bead recycling of PCR products. Sequencing was performed using the Illumina Nextseq2000 platform (Shanghai Meiji Biomedical Technology Co., Ltd.).

Mechanisms of photocatalytic microbe killing

Photocatalysts eradicate microorganisms through two main mechanisms, namely physical inhibition and oxidative stress24. The activation of the photocatalyst surface by light generates electron–hole pairs25, and the electrons and holes produced after separation react with ambient oxygen and water, respectively, to generate ROS, such as •O2−, •OH−, and H2O2, which attack microbial cells to cause oxidative damage and, ultimately, death26.

Figure 1 illustrates the abovementioned killing mechanism applied to S. obliquus. The protective structure of the S. obliquus cell is mainly composed of the cell wall (containing cellulose, pectin, etc.) and organic compounds on the cell membrane (e.g., peptidoglycans and phospholipids), which are susceptible to oxidation by strong oxidants and adhesion to nanoparticles to form heteroagglomerates27. The photocatalytically generated ROS (e.g., •O2− and •OH) react with the cellulose and pectin molecules of the cell wall to break their carbon chains and oxygen bridges and thus induce cell wall rupture28. The subsequent ROS attack on the phospholipids of the cell membrane cleaves the tail carboxyl groups and thus triggers rearrangements and morphological changes in the phospholipid bilayer to damage the cell membrane, increase permeability, and expose intracellular compartments to the external environment29. The cell wall and membrane damage facilitates the entry of ROS into the cell interior, which results in further cellular component destruction30. In addition, the chloroplast and mitochondrial membranes of algal cells are rich in polyunsaturated fatty acids and are particularly vulnerable to ROS attack31, mainly undergoing lipid peroxidation32. At the same time, hydroxyl radicals enter the cell and damage the DNA nucleobases and sugars to cleave the sugar-phosphate backbone and single/double DNA strands33. Superoxide radicals mainly attack guanine, which leads to base deletion/modification and the destruction of the structural integrity of DNA. These phenomena affect protein synthesis and cause abnormal physiological processes, such as membrane damage and photosynthetic protein denaturation, greatly impacting the growth and metabolism of the whole organism34. In summary, the photocatalytic production of ROS causes multiple-level damage to microalgal cells, ultimately killing them.

Fig. 1
figure 1

Mechanism of the photocatalytic eradication of microorganisms.

Full size image

Results and discussion

Photocatalyst morphology and composition

Ti2NTx had a typical two-dimensional layered structure (Fig. 2a)35, while BiVO4 comprised spherical particles with uniform size of about 220 nm (Fig. 2b)36, and Ti2NTx-TiO2 featured spherical TiO2 particles tightly dispersed on the surface of the two-dimensional layered Ti2NTx (Fig. 2c). BiVO4/Ti2NTx-TiO2 comprised spherical BiVO4 and TiO2 particles uniformly dispersed on the surface of Ti2NTx (Fig. 2d) and contained Ti, Bi, V, O, and N (Fig. S2).

Fig. 2: Morphological and compositional characteristics of photocatalysts (Ti2NTx, BiVO4, Ti2NTx-TiO2, and BiVO4/Ti2NTx-TiO2).
figure 2

(a–d) Scanning electron microscopy images of Ti2NTx (a), BiVO4 (b), Ti2NTx-TiO2 (c), and BiVO4/Ti2NTx-TiO2 (d); (e) X-ray diffraction patterns; (f) Survey X-ray photoelectron spectra; (g–j) Ti 2p (g), O 1 s (h), Bi 4 f (i), and V 2p (j) spectra.

Full size image

The X-ray diffraction pattern of BiVO4/Ti2NTx-TiO2 featured peaks at 25.273°, 48.205°, 55.186°, 62.789°, and 70.453°, attributed to the (101), (200), (211), (204), and (220) crystallographic planes of anatase TiO2, and a peak at 39.841°, corresponding to the (103) crystallographic plane of Ti2NTx37 (Fig. 2e). The peaks at 30.559°, 34.617°, 35.223°, and 42.655° corresponded to the (004), (200), (002), and (015) crystal planes of BiVO438.

The survey X-ray photoelectron spectrum of BiVO4/Ti2NTx-TiO2 revealed the presence of only Ti, O, Bi, V, and N (Fig. 2f). The Ti 2p spectrum of TiO2 featured symmetric Ti 2p3/2 and Ti 2p1/2 peaks (458.1 and 463.77 eV, respectively), and that of Ti2NTx featured Ti–O (Ti 2p3/2), Ti–O (Ti 2p1/2), and Ti–N peaks at 457.67, 463.3, and 453.28 eV, respectively (Fig. 2g)39. The Ti 2p3/2 and Ti 2p1/2 signals of BiVO4/Ti2NTx-TiO2 were observed at 457.87 and 463.55 eV, respectively (Fig. 2g). Thus, Ti in all three samples was present as Ti4+, and the Ti 2p signals shifted to lower binding energies when TiO2 was composited with Ti2NTx and BiVO4, which suggested the occurrence of electron transfer40. The O 1 s spectrum of TiO2 featured peaks at 529.25 and 531.01 eV, indicating the presence of lattice oxygen and adsorbed oxygen, respectively (Fig. 2h). In the case of BiVO4/Ti2NTx-TiO2 (Fig. 2h), these peaks were shifted to lower binding energies, indicating that the complexation of BiVO4 and Ti2NTx markedly affected the electronic structure of oxygen species41. The Bi 4 f spectrum of BiVO4 featured Bi 4f5/2 and Bi 4f7/2 signals at 163.15 and 157.89 eV, respectively42, which shifted to lower energies in the spectrum of BiVO4/Ti2NTx-TiO2, indicating the presence of Bi as free Bi3+ (Fig. 2i). The V 2p spectrum of BiVO4 featured V 2p1/2 and V 2p3/2 peaks at 523.23 eV and 515.34 eV, respectively, indicating the presence of V5+, whereas these peaks were shifted to lower energies in the case of BiVO4/Ti2NTx-TiO2, indicating the presence of V4+ (Fig. 2j)43. The N 1 s peaks of Ti2NTx at 395.02, 396.06, and 398.53 eV were ascribed to Ti–O–N, Ti–N, and satellite peak, respectively (Fig. S3)44. For Ti2NTx-TiO2 and BiVO4/Ti2NTx-TiO2, the Ti–O–N and Ti–N signals were not observed, while the satellite peak signals were shifted to higher energies by 0.71 and 0.64 eV, respectively, which confirmed the existence of a strong interaction between Ti2NTx and TiO2. The results presented in Fig. 2 and Fig. S3 confirmed the successful synthesis of BiVO4/Ti2NTx-TiO2.

Given that ROS play a key role in the photocatalytic eradication of microalgae45, we probed the production of these species on the catalyst surface by electron spin resonance (ESR) spectroscopy. Figure 3a shows that a set of six ESR peaks characteristic of DMPO-‧O2− appeared in the spectra of all samples upon light exposure (but not in the dark), which suggested the photocatalytic production of •O2−. The above signals considerably gained intensity upon the incorporation of Ti2NTx and BiVO4 and were strongest for BiVO4/Ti2NTx-TiO2. This finding suggested that among the examined samples, BiVO4/Ti2NTx-TiO2 generated the largest number of electrons, with the increase in the number of trapped photogenerated electrons leading to an increase in the levels of •O2− under visible light irradiation. Figure 3b shows that ESR peaks (peak intensity ratio = 1:2:2:1) characteristic of DMPO-•OH emerged upon the illumination of all samples (but not in the dark), indicating the production of •OH under irradiation. With the incorporation of Ti2NTx and BiVO4, these peaks gained intensity and were strongest for BiVO4/Ti2NTx-TiO2, which confirmed the high redox activity of this photocatalyst. The results of ESR analysis showed that upon irradiation, BiVO4/Ti2NTx-TiO2 produced •OH and •O2−, which could attack the microalgal cell wall to damage it and then attack the cell membrane to cause oxidative damage and thus kill the cell.

Fig. 3: Radical production characteristics and photoelectrochemical properties of photocatalysts.
figure 3

(a) Electron spin resonance (ESR) signals of •O2−; (b) ESR signals of •OH; (c) Photocurrent responses; (d) Electrochemical impedance spectra; (e) Ultraviolet–visible diffuse reflectance absorption spectra; (f) Tauc plots.

Full size image

The separation and transfer of photogenerated electrons and holes were investigated using photoelectrochemical measurements. The responses of the photocatalysts under periodic illumination on/off conditions demonstrated that the photocurrent densities of BiVO4/Ti2NTx-TiO2 markedly exceeded those of Ti2NTx-TiO2 and TiO2 (Fig. 3c). Electrochemical impedance spectroscopy measurements (Fig. 3d) showed that BiVO4/Ti2NTx-TiO2 featured the smallest charge transfer resistance (represented by semicircle diameter). Thus, BiVO4/Ti2NTx-TiO2 exhibited high charge transfer and separation efficiencies, which were conducive to the enhancement of photocatalytic activity and ability to eradicate microalgae. The absorption band edges of TiO2 and Ti2NTx-TiO2 were located at 390 nm, while that of BiVO4/Ti2NTx-TiO2 was located at 550 nm, which suggested that the incorporation of BiVO4 substantially increased light absorption ability and, hence, photocatalytic activity (Fig. 3e). The direct bandgaps of the photocatalysts were extracted from their Tauc [(αhv)2 vs. hv] plots (Fig. 3e)46 and followed the order of TiO2 (3.32 eV) > Ti2NTx-TiO2 (3.21 eV) > BiVO4/Ti2NTx-TiO2 (2.46 eV) (Fig. 3f). The hybridization of BiVO4 and Ti2NTx-TiO2 effectively reduced the bandgap, thus promoting the separation of photogenerated electrons and holes in BiVO4/Ti2NTx-TiO2 and inhibiting their recombination.

DFT calculations

DFT was used to calculate the energy band structures, density of states distributions, and work functions of the photocatalysts. The bandgaps of TiO2, Ti2NTx-TiO2, and BiVO4/Ti2NTx-TiO2 were calculated as 2.23, 0.8, and 0.73 eV, respectively (Fig. 4a–c) and were lower than the experimental values (Fig. 3f) because of the limitations of the generalized degree of approximation47. Compared with other photocatalysts, BiVO4/Ti2NTx-TiO2 featured denser energy bands and a greater overlap-based band coupling, which resulted in stronger interactions. The conduction band of TiO2 mainly consisted of Ti orbitals, with O orbitals dominating the valence band, whereas in the case of Ti2NTx-TiO2, the energy level distribution in the valence and conduction bands was mainly due to the creation of new bonds at the interfaces of the composite structure (Fig. 4d–f). In addition, Ti orbitals were the main contributors to the conduction band of Ti2NTx-TiO2, while the valence band was dominated by O and N orbitals. In the case of BiVO4/Ti2NTx-TiO2, the conduction band was dominated by the orbitals of V and Ti, and the valence band was dominated by those of the structural layer of the BiVO4 composite. The Bi and O orbitals of the BiVO4 composite had a major influence on the valence band, while V had a weak influence on the valence and conduction bands, bringing the conduction band close to the Fermi energy level. The valence band produced a certain distribution of energy levels, which increased the possibility of electron conduction.

Fig. 4: Density functional theory calculations.
figure 4

a–c Energy band structures of TiO2 (a), Ti2NTx-TiO2 (b), and BiVO4/Ti2NTx-TiO2 (c); (d–f) Density of states distributions of TiO2 (d), Ti2NTx-TiO2 (e), and BiVO4/Ti2NTx-TiO2 (f); (g, h) Surface work functions of Ti2NTx-TiO2 (g) and BiVO4 (h); (i) Charge density difference maps of BiVO4/Ti2NTx-TiO2 (BiVO4 at the top, Ti2NTx in the middle, and TiO2 at the bottom), with electron accumulation and depletion regions indicated by yellow and blue colors, respectively.

Full size image

The work functions of TiO2, Ti2NTx-TiO2, BiVO4, and BiVO4/Ti2NTx-TiO2 were calculated as 7, 5.25, 7.27, and 7.09 eV, respectively (Fig. 4g, h and Fig. S4). The hybridization of TiO2 and Ti2NTx decreased the work function and induced electron transfer from the latter to the former; the work function of BiVO4/Ti2NTx-TiO2 exceeded that of Ti2NTx-TiO2, but was lower than that of BiVO4. Therefore, the electrons were transferred from Ti2NTx/TiO2 to BiVO4. These findings indicated that Ti2NTx was mainly responsible for electron transfer in the composite photocatalyst and thereby inhibited the complexation of photogenerated carriers to increase the charge transfer and separation efficiencies of BiVO4/Ti2NTx-TiO2. The differential charge density analysis of BiVO4/Ti2NTx-TiO2 revealed that electron transfer from Ti2NTx to BiVO4 and TiO2 led to electron accumulation on BiVO4 and TiO2 and the depletion of Ti2NTx, in agreement with the results of the work function. (Fig. 4i).

Antimicrobial activities of photocatalysts

The absorbance of microalgae on the sandstone surface at the characteristic wavelength of 680 nm (Fig. 5a–c and Fig. S5) markedly changed after 28 days of incubation under growth-favoring conditions (Fig. 5d–f).

Fig. 5: Performances of different photocatalytic films (BiVO4/Ti2NTx-TiO2, Ti2NTx-TiO2, TiO2, blank) for the eradication of microalgae under different light conditions (0 h light, 12 h light/12 h dark, 24 h light) at different growth times (0, 7, 14, 21, and 28 days).
figure 5

(a–c) Absorption spectra of microalgae on the surface of sandstone coated with different photocatalytic films on the 28th day of growth under 0 h light (a), 12 h light/12 h dark (b), and 24 h light (c) conditions; (d–f) Time-dependent absorbance (680 nm) of microalgae on the surface of sandstone coated with different photocatalytic films under 0 h light (d), 12 h light/12 h dark (e), and 24 h light (f) conditions; (g) Photographs of microalgae on sandstone surfaces under different light conditions at different growth times (B: Blank, C: TiO2, D: Ti2NTx-TiO2, E: BiVO4/Ti2NTx-TiO2).

Full size image

In the absence of light (0 h light), the absorbance increased with the increasing incubation time, albeit slowly (Fig. 5d). This behavior was ascribed to the heterotrophic growth of S. obliquus under these conditions, which resulted in a low growth rate, slow cell division, and low biomass content because of the low content of organic matter in the culture medium. The biomass of S. obliquus followed the order of BiVO4/Ti2NTx-TiO2 < Ti2NTx-TiO2 < TiO2 ≈ control (blank). Thus, although the photocatalysts did not exert their photocatalytic effects in the absence of light (Fig. 3a, b), their self-cleaning ability inhibited the growth of S. obliquus, with BiVO4/Ti2NTx-TiO2 having the strongest effect.

The absorbance of the uncoated and coated samples under intermittent light conditions (12 h light/12 h dark) increased with the increase in the incubation time from 0 to 21 days (Fig. 5e) and subsequently stabilized. This finding suggested that with the prolongation of the colonization time, the microorganisms continued to grow and reproduce, and the biomass and biofilm thickness increased. Concomitantly, the amount of microbial extracellular polymers increased, and biofilm porosity decreased, which increased the mass transfer resistances of the substrate and product. Hence, microbial cell growth inside the biofilm was restricted by substrate and product inhibition48,49, which resulted in the abovementioned absorption evolution pattern. Under irradiation, the photocatalytic films produced ROS (•OH, •O2−), which attacked the microalgae to induce their death. The absorbance and absorbance increase rate followed the order of BiVO4/Ti2NTx-TiO2 < Ti2NTx-TiO2 < TiO2 < control. This finding was ascribed to the higher photocurrent density of the Ti2NTx-TiO2 film compared with that of the TiO2 film. The BiVO4/Ti2NTx-TiO2 film had an extended light absorption range and thus featured the highest efficiency of light energy utilization, photocatalytic activity, and •OH and •O2− yields, effectively eradicating microorganisms at its interface with the biofilm.

The trend observed under continuous light conditions (24 h light) resembled that under intermittent light conditions (Fig. 5f). In all cases, the absorbance decreased with time and ceased to change after the 14th day in the case of BiVO4/Ti2NTx-TiO2. This finding may be due to the fact that continuous irradiation triggered photoinhibition, which reduced the efficiency of microbial photosynthesis and possibly damaged photosynthetic pigments and reaction center proteins50. At the same time, all continuously irradiated photocatalytic films could produce more ROS to attack the microalgae and cause their death.

Comparison with Fig. 5d–f shows that in the case of BiVO4/Ti2NTx-TiO2, the absorbance under illumination was lower than that in the dark. This finding confirmed that the illumination of the BiVO4/Ti2NTx-TiO2 film produced ROS, which effectively attacked and destroyed the microbial cells of the bottom layer of Scenedesmus obliquus and inhibited its growth. The amount of ROS and, hence, strength of the antimicrobial effect, increased with the increasing illumination time, and the biomass therefore concomitantly decreased.

In the absence of light (0 h light, Fig. 5g), all samples showed surface darkening, indicating a biomass increase, in line with the results in Fig. 5d. Under light conditions, microalgae on the sample coated with Ti2NTx-TiO2 turned yellow from day 21, and the area of microalgae death was markedly larger under continuous light than under intermittent light and exceeded that observed for TiO2 at day 28. In the case of BiVO4/Ti2NTx-TiO2, the microalgae turned yellow and died by the 28th day under both continuous and intermittent light. Thus, the BiVO4/Ti2NTx-TiO2 film had a strong organic pollutant cleaning effect and photocatalytic activity, effectively eradicating S. obliquus.

Analysis of microbial community structure and chlorophyll content

Figure 6a demonstrates the effects of photocatalytic films on the chlorophyll content of sandstone surface microalgae under different light conditions. After 28 days of incubation, the chlorophyll content of sandstone surface microalgae without photocatalytic films increased to 2.8, 29.33, and 28.75 mg/L under 0 h light, 12 h light/12 h dark, and 24 h light conditions, respectively. The corresponding values for TiO2 and Ti2NTx-TiO2 were 2.64, 28.25, and 27.6 mg/L and 2.45, 26.82, and 24.89 mg/L, respectively, and those for BiVO4/Ti2NTx-TiO2 were 0.57, 1.68, and 1.37 mg/L.These results further demonstrated that the BiVO4/Ti2NTx-TiO2 photocatalytic film had the best eradication effect, especially under sustained light. This behavior was ascribed to the ability of continuous illumination to destroy the photosynthetic center of microalgae, reduce the photosynthetic efficiency, and promote ROS production51. As microalgae are photosynthetic organisms, the absence of photosynthesis and limitation of heterotrophic growth by the organic substrate in the absence of light resulted in growth deceleration. In addition, the photocatalytic materials were unable to produce ROS in the dark, which resulted in low chlorophyll content for all photocatalyst-coated samples.

Fig. 6: Chlorophyll and Biodiversity after 28 days of microbial culture.
figure 6

(a) Chlorophyll contents of sandstone with different photocatalytic films (BiVO4/Ti2NTx-TiO2, Ti2NTx-TiO2, TiO2, blank) after 28 days of incubation under different light conditions (0 h light, 12 h light, 12 h dark, 24 h light); (b) Relative abundances of different bacteria at the phylum level after different treatments; (c) Relative abundances of different bacteria at the phylum level after treatment with BiVO4/Ti2NTx-TiO2 under 24 h light conditions; (d) Principal component analysis of microbial community composition at the phylum level under different treatments. (A: initial amount; B1, B2, B3: no treatment; C1, C2, C3: TiO2; D1, D2, D3: Ti2NTx-TiO2; E1, E2, E3: BiVO4/Ti2NTx-TiO2).

Full size image

Figure 6b demonstrates the phylum-level microbial community changes for samples subjected to different photocatalyst treatments after 28 days (no photocatalytic film, TiO2, Ti2NTx-TiO2, and BiVO4/Ti2NTx-TiO2). After 28 days of incubation, Proteobacteria were the most abundant phylum in all groups, followed by Cyanobacteria, Actinobacteria, Bacteroidota, and Patescibacteria, which collectively accounted for >80% of all bacterial sequences. Notably, Bacteroidota, Proteobacteria, and Patescibacteria were most abundant in the TiO2, Ti2NTx-TiO2, and BiVO4/Ti2NTx-TiO2 groups in the absence of light, respectively, with the lowest relative abundance of Cyanobacteria observed in the TiO2 group. The higher relative abundance of Cyanobacteria in the BiVO4/Ti2NTx-TiO2 group may be due to the absence of photocatalytic activity under dark conditions.

Figure 6b also shows that the appearance of Patescibacteria on photocatalyst-treated sandstone was particularly pronounced in the presence of light. Under intermittent light conditions (12 h light, 12 h dark), the relative abundance of Bacteroidota in the photocatalyst-treated group markedly exceeded that in the control group. In addition, among the most abundant phyla, the relative abundances of Proteobacteria, Actinobacteria, and Bacteroidota in the photocatalyst-treated group under continuous light (24 h light) conditions exceeded those in the control group. The relative abundance of Patescibacteria increased on the sandstone surfaces treated with TiO2, Ti2NTx-TiO2, and BiVO4/Ti2NTx-TiO2 under light conditions, and the relative abundance of Cyanobacteria was lowest in the BiVO4/Ti2NTx-TiO2 group. Under continuous light conditions, the relative abundance of Cyanobacteria in the BiVO4/Ti2NTx-TiO2 group (22.58%) was notably lower than that in the control group (Fig. 6c). This finding was ascribed to the fact that BiVO4/Ti2NTx-TiO2 exhibited a high efficiency of light energy utilization and, hence, high activity.

In the experiments, the phylum-level microbial community (Fig. 6b, c) can be divided into beneficial and harmful species. Proteobacteria play an important role in the C, N, and S cycles and may accelerate N removal from microbial communities52. Actinobacteria and Bacteroidota are the key contributors to the C cycle, with their abundance correlating with C availability in microbial communities53. Patescibacteria, often parasitized by Actinobacteria, positively influence the utilization of soluble N and are closely associated with S-cycling microorganisms54. Notably, Patescibacteria can ferment pyruvate produced by microorganisms into acetate and lactate, thus enhancing metabolic and catabolic activities in microorganisms55. When the microorganisms enter the heterotrophic growth stage (Microorganisms that lack light or are far from the surface of the biofilm), where their growth mainly depends on nutrients such as C, N and P in the culture medium56. The higher amount of species (Proteobacteria, Actinobacteria, Bacteroidota, and Patescibacteria), which leads to a higher consumption of nutrients (C, N, and P), thereby reducing the nutrient supply needed for Cyanobacteria and S. obliquus growth, resulting in a lower abundance of Cyanobacteria and low biomass content of S. obliquus57. The metabolic activity of Cyanobacteria and S. obliquus produce organic acids that can react with minerals to cause the chemical weathering of stone. Prolonged exposure to acidic environments can weaken the structure of stone and promote its degradation58. Therefore, Proteobacteria with 2–3 µm length and 0.5–0.7 µm width, Actinobacteria with 0.5–0.8 µm diameter, Bacteroidota with 1.5–4.5 µm length and 0.5–0.8 µm width, and Patescibacteria with diameter appropriately 1 µm are beneficial species for protecting stone cultural relics against microbial weathering. In contrast, S. obliquus (length and width are respectively of 10–21 µm and 3–9 µm) and Cyanobacteria (diameter is 3–10 µm) are harmful species.

Principal coordinate analysis based on the Bray–Curtis algorithm demonstrated the changes in the composition of microbial communities in control, TiO2, Ti2NTx-TiO2, and BiVO4/Ti2NTx-TiO2 groups after 28 days of incubation under different light conditions (Fig. 6d). The initial group was clustered separately from the bacterial community after 28 days of incubation. After 28 days of incubation, the bacterial communities in the no-light and light groups also showed a clear clustering separation. The microbial communities in the 12 h and 24 h light groups were very similar in composition, which confirmed that the ROS produced by the photocatalytic film effectively killed harmful microorganisms and promoted the proliferation of beneficial microorganisms.

Together with the Figs. 5, 6, and Fig. S5, one can see that the experimental results indicated that the ROS produced by BiVO4/Ti2NTx-TiO2 films were highly prone to attack the harmful S. obliquus with large width and smooth cell walls and Cyanobacteria with large diameter and cyst-like structures composed of phospholipids and proteins, and cause their death. The intracellular nutrients released from the broken walls of S. obliquus and Cyanobacteria cells attacked by ROS provided additional energy for the growth of beneficial species (Proteobacteria, Actinobacteria, Bacteroidota, and Patescibacteria).

Conclusion

BiVO4/Ti2NTx-TiO2 was fabricated as a photocatalytic nanocomposite for the protection of stone cultural relics against microbial weathering. The hybridization of BiVO4 (narrow-bandgap semiconductor) with two-dimensional layered Ti2NTx-doped TiO2 reduced the bandgap to 2.46 eV and extended the visible light absorption edge to 550 nm, thus markedly increasing the separation efficiency of photogenerated charge carriers, efficiency of light energy utilization by TiO2, and ROS (•OH, •O2−) yield. Under the examined light conditions (24 h light and 12 h light/12 h dark), BiVO4/Ti2NTx-TiO2 photocatalytic films fully eradicated S. obliquus on sandstone surfaces after 28 days in a suitable growth environment. After 28 days of cultivation, the chlorophyll content in the BiVO4/Ti2NTx-TiO2 group increased by only 1.37 mg/L compared with that in the control group, and microbial community structure analysis showed that BiVO4/Ti2NTx-TiO2 effectively eradicated harmful bacteria (Cyanobacteria) while promoting the growth of beneficial species (Proteobacteria, Actinobacteria, Bacteroidota, and Patescibacteria). The adopted strategy can help extend the lifetime of cultural relics and inspires the development and engineering applications of semiconductor photocatalysts.