Solar-driven photocatalytic water splitting is regarded as one of the most promising clean and sustainable methods for hydrogen production,often referred to as "solar fuel generation from water". Approximately 43% of solar energy lies within the visible spectrum,driving the urgent need for efficient,durable,and cost-effective visible-light-responsive catalysts for hydrogen production^[1-2]. Two-dimensional boron nitride nanosheets (BNNS) exhibit a unique combination of polar bonding,high specific surface area,thermal and oxidative stability,and wide availability. Frequently termed "white graphene",BNNS has demonstrated remarkable ultraviolet photocatalytic hydrogen evolution activity^[3-4]. However,their wide bandgap (5.5 eV) and chemical inertness pose substantial challenges for tuning their band structure via conventional functionalization strategies,thus limiting their application in visible-light-driven photocatalysis.

Previous studies have explored element doping and defect engineering to address these challenges. Element doping involves incorporating elements such as carbon,oxygen,sulfur,or fluorine into BNNS to modify its band structure and enhance photocatalytic performance^[5-8]. Nonetheless,the stringent conditions required (>800 ℃) and the low yield (<10%) hinder their scalable application. In contrast,defect engineering has emerged as a promising alternative for bandgap modulation^[9]. However,due to the strong polarity of B-N bonds,precise control over defect formation via conventional chemical methods remains difficult,with current techniques such as precursor pyrolysis^[10],ball milling^[11],and plasma treatment^[12] proving inefficient. Lu et al. introduced defects into BNNS using four types of plasma (N₂,O₂,H₂,and Ar) to regulate its oxidative dehydrogenation performance for propane. They found that O₂ treatment led to the formation of BO_x,which exhibited sintering effects that reduced catalytic performance,while N₂ treatment lowered the bandgap to around 4 eV^[12]. Although this bandgap approaches the visible region,it still fails to enable visible-light-responsive catalytic activity.

High-energy ionizing radiation has been identified as an effective tool for catalyst modifications. As early as the 1960s,researchers such as Clarke^[13],Barry^[14],and Cropper^[15] observed significant effects of radiation on the activity of solid catalysts during studies on solid materials in the nuclear reactor. Baru^[16] and Spinks^[17] also noted rapid advancements in this field. On one hand,electrons generated by high-energy radiation directly participate in catalytic reactions; in some cases,radiation enhances catalytic performance by promoting processes such as partial reduction or iron carbide formation^[13]. Furthermore,radiation effects on various catalytic activities have been found to vary significantly,ranging from no effect to activity enhancements of up to a thousand-fold. The radiation effects include catalyst poisoning,formation of electrons and holes,defect generation,and changes in adsorption capacity^[17]. These studies indicate that,compared with conventional chemical and physical methods,high-energy radiation can directly and effectively modulate catalyst structures through ionization,excitation,or interfacial radical reactions. This process,which occurs at ambient temperature and pressure,offers advantages such as high efficiency,convenience,and low carbon emissions,making it promising for large-scale preparation. Since 2004,the Radiation Chemistry Laboratory at University of Science and Technology of China has been investigating the gamma-ray effect on different catalysts (e.g.,CdS^[18],graphene^[19-20],Bi₂WO₆^[21],and BNNS^[22]),introducing defects or other active sites in catalysts to precisely modulate their surface oxidation states,crystal structures,and band structures^[18-22].

Inspired by previous work on radiation-induced exfoliation for BNNS preparation^[22],this study proposes introducing defects in BNNS by gamma-ray radiation. By varying the absorbed dose,the defect density within BNNS can be precisely controlled,resulting in BNNS with different band structures while maintaining the sheet-like morphology. This enables visible light response and enhanced photocatalytic performance for BNNS.

Hexagonal boron nitride (h-BN,analytical grade) was purchased from Aladdin Biochemical Technology Co.,Ltd.,Shanghai,China. Isopropanol,methanol,sodium sulfite,sodium sulfate,and triethanolamine were obtained from Sinopharm Chemical Reagent Co.,Ltd. and were all of the analytical grade. Deionized water was used throughout the experiments.

Preparation of BNNS: BNNS was prepared via a ball-milling method. One gram of h-BN powder was placed in a zirconia ball milling jar,and the milling parameters are shown in Table 1. After milling,the jar was washed with deionized water,and the product was centrifuged and washed three times with deionized water,followed by vacuum freeze-drying (-50 ℃,6 Pa) to obtain BNNS.

Gamma-ray irradiation experiments: One hundred milligrams of BNNS were dispersed ultrasonically in a mixed solvent of deionized water (15 mL) and isopropanol (5 mL),purged with nitrogen and sealed. The sample was irradiated with a Co-60 gamma-ray source (5.5×10¹⁴ Bq,located at the University of Science and Technology of China) at a dose rate of 62 Gy/min for different times. After irradiation,the sample was centrifuged and washed three times with deionized water,followed by vacuum freeze-drying (-50 ℃,6 Pa) to obtain the product.

The morphology of the samples was characterized by transmission electron microscopy (TEM,Hitachi H7650,operating voltage: 100 kV). Atomic force microscopy (AFM,Veeco,MultiMode V) was used to determine the accurate thickness and lateral dimensions.

The defect structure of the samples was analyzed using electron paramagnetic resonance (EPR) spectroscopy,measured with a JEOL JES-FA200 EPR spectrometer (9.081 GHz).

The crystal structure of the samples was characterized by X-ray diffraction (XRD),using a high-resolution X-ray diffractometer (Rigaku TTRIII,Cu Kα, λ=0. 154 18 nm,8 (°)/min,20°-70°).

Fourier transform infrared (FTIR) spectra were obtained using a Thermo Fisher Nicolet 6700 spectrometer (4 000~400 cm^-1) with the KBr pellet method. Ultraviolet-visible (UV-vis) absorption and diffuse reflectance spectra (DRS) were measured using a Shimadzu UV-3600 spectrophotometer. Steady-state fluorescence spectra and fluorescence decay curves were measured with a steady-state-transient fluorescence spectrometer (JY Fluorolog-3-Tou) with an excitation wavelength of 400 nm.

Photoelectrochemical tests,including photocurrent curves,Mott-Schottky plots,and electrochemical impedance spectroscopy (EIS),were conducted using an electrochemical workstation (CHI660E). A three-electrode configuration was used: platinum foil (counter electrode),Ag/AgCl electrode (reference electrode),and BNNS-coated ITO conductive glass (working electrode,1 cm×1 cm). A xenon lamp (Perfect Light PLS-SXE 300,≥420 nm,300 W) was used as the light source. The electrolyte was 0.5 mol/L saturated sodium sulfate solution.

Five milligrams of irradiated BNNS were dispersed in 40 mL of deionized water,and 5 mL of methanol was added as a sacrificial agent. The system was purged with argon and sealed. A xenon lamp (Perfect Light PLS-SXE 300,≥420 nm,300 W) was used as the light source,and the gaseous products were analyzed using gas chromatography (SHIMADZU GC-2014,Ar).

The morphology and thickness of h-BN before and after ball milling are shown in Fig. 1(a)~(e). Commercial h-BN appeared as layered agglomerated particles with a size of approximately 300 nm and a thickness greater than 50 nm. After ball milling,the product exhibited a sheet-like structure with a size of 200~300 nm,and the contrast of TEM images was significantly lower than that of commercial h-BN,indicating a substantial reduction in thickness. AFM further confirmed the thickness and size distribution of BNNS. The lateral dimensions of BNNS were mainly distributed between 180 nm and 240 nm,consistent with TEM images. About 42.3% of BNNS had a thickness between 3.0 nm and 4.5 nm,and more than 92.3% had a thickness below 7.5 nm,which was significantly different from commercial h-BN. According to the pore size distribution (Figs. 1(f)~(g)),the ball-milling process significantly damaged the h-BN crystal structure,resulting in the formation of enlarged macropores from the mesoporous structure. Correspondingly,the specific surface area of BNNS increased from 53.79 m²/g for commercial h-BN to 138.45 m²/g (Figs. 1(f)~(g)),indicating that ball-milling process exfoliated h-BN into thinner BNNS.

The FTIR spectra (Fig. 2(a)) showed two main peaks corresponding to BNNS before irradiation: a peak at 1 393 cm^-1 for the B-N in-plane stretching vibration and a peak at 807 cm^-1 for the B-N-B out-of-plane bending vibration^[23]. After gamma-ray radiation,these two peaks remained unchanged,and a peak at 1 635 cm^-1 corresponds to adsorbed water,indicating that no new functional groups were introduced into BNNS during irradiation. The XRD patterns of BNNS after ball milling confirmed that BNNS maintained its hexagonal crystal structure,with diffraction peaks corresponding to specific crystal planes shown in Fig. 2(b) (JCPDS card number 34-0421)^[24]. As the absorbed dose increased,the intensity of the (100),(101),and (102) peaks gradually decreased,almost disappearing at 177 kGy,suggesting partial crystal structure changes while the overall crystal phase remained identical.

The EPR spectra indicated the formation of defects in the irradiated samples (Fig. 2(c)). According to theoretical calculations^[25-26],high-energy radiation of boron nitride primarily generates one boron center (OBC) and three boron center (TBC) defects. Neutron and heavy ion irradiation experiments also verified the presence of OBC and TBC in irradiated BNNS^[27-29]. The EPR signal for OBC defects had a broader linewidth (ΔB>110 G). In comparison,TBC exhibited a narrower linewidth with ΔB between 20 G and 30 G. After gamma-ray irradiation,BNNS showed prominent EPR signals with a relatively narrow linewidth (ΔB≈24.4 G),consistent with the presence of TBC defects. As the absorbed dose increased,the EPR signal gradually intensified,indicating the continuous introduction of TBC defects. These defects significantly altered the light absorption properties of BNNS. As shown in Fig. 3(a),BNNS exhibited negligible visible light absorption before irradiation,with a bandgap calculated from the Kubelka-Munk function of about 5.2~5.3 eV,in line with typical boron nitride characteristics. After irradiation,the intrinsic wide bandgap decreased from 5.29 eV to 4.89 eV as the absorbed dose increased. More importantly,a broad absorption band formed in the UV-visible range,which raised with increasing dose,shifting from 3.18 eV to 2.74 eV. This was consistent with previous theoretical calculations that suggested the introduction of TBC defects altered the local electronic structure of surrounding boron atoms,generating new defect energy levels^[25-27],leading to visible and near-infrared light absorption by BNNS. Electrochemical tests further verified modifications in the band structure. The conduction band bottom (CBB) potential of irradiated BNNS decreased,enhancing the reduction ability of photoexcited electrons (Fig. 3(b)). Combining the DRS results,Fig. 3(c) summarizes the band structure of BNNS after irradiation. In addition to retaining the intrinsic wide bandgap of boron nitride,irradiated BNNS generated new intermediate energy levels,providing the prerequisites for visible-light-driven catalytic reactions,with the CBB potential shifting from -0.77 eV to -1.37 eV (vs. Ag/AgCl).

Furthermore,apart from the alternations in band structures,radiation-induced defects also impacted the separation dynamics of photogenerated carriers (Fig. 4). The steady-state photoluminescence (PL) spectra of BNNS revealed a narrow emission band centered around 430 nm and a broad band extending beyond 460 nm. Upon irradiation,a progressive decrease in fluorescence intensity was observed with increasing absorbed dose (Fig. 4(a)),suggesting that TBC defects facilitated the separation of photogenerated carriers. Moreover,as illustrated in Fig. 4(b),the carrier lifetime in BNNS remained within the nanosecond timescale before and after irradiation. Notably,a slight extension of the carrier lifetime post-irradiation could be attributed to the high polarity associated with localized π bonds between B-N atoms,highlighting the nuanced interplay between structural defects and electronic behavior.

The visible-light photocatalytic performance of irradiated BNNS was evaluated through hydrogen evolution from water splitting (Figs. 5(a),(b)). A sacrificial agent,methanol,was added to the photocatalytic system. Pristine BNNS exhibited negligible hydrogen evolution performance due to its wide bandgap (~5.3 eV),which inhibited the generation of photo-excited electrons under visible light. In contrast,irradiated BNNS demonstrated remarkable hydrogen evolution performance,with production rates increasing proportionally to the absorbed dose. At maximum efficiency,irradiated BNNS achieved a hydrogen production of 7 753.3 μmol/g,with a rate of 1 033.7 μmol/(g·h),nearly two orders of magnitude higher than BNNS synthesized by pyrolysis (Fig. 5(a)). Comparative analysis with other defect-engineered BN-based and classic photocatalysts revealed that irradiated BNNS outperformed most reported boron nitride-based catalysts and even surpassed commercial TiO₂ (P25) and CdS under visible light irradiation (Table 2). The ambient radiolytic methodology demonstrated significant advantages over high-temperature pyrolysis,offering moderate operation conditions and superior energy efficiency.

Beyond hydrogen,methane and carbon monoxide were also detected among the gaseous products,originating from methanol decomposition. This demonstrates that irradiated BNNS can simultaneously decompose methanol and water,broadening its potential applications in clean energy fields. Furthermore,as shown in Fig. 5(c),the photocatalytic activity of BNNS remained stable over five cycles,with defect structures demonstrating resilience to visible-light irradiation and surface redox reactions. Structural analysis revealed that high-energy radiation effectively transformed wide-bandgap BNNS into a highly efficient visible-light-responsive catalyst. Its superior hydrogen production performance can be attributed to the introduction of defect energy levels,enhanced visible light absorption,improved reduction potential of photoexcited electrons,and increased carrier separation efficiency,establishing irradiated BNNS as a promising candidate for advanced photocatalytic applications.

Electrochemical analyses investigated the interfacial charge transport behavior of BNNS before and after visible-light irradiation. Transient photocurrent characterization revealed that irradiated BNNS exhibited a sensitive response during light on-off cycles,with photocurrent intensity increasing alongside the absorbed dose (Fig. 5(d)). This trend mirrored the observed enhancement in hydrogen production,suggesting that radiation-induced defects improved visible light absorption and photogenerated carrier generation. Although the photocurrent signal of BNNS remained lower than that of conventional semiconductor photocatalysts due to its inherently poor conductivity,electrochemical impedance spectra (EIS) (Fig. 5(e)) showed a marked decrease in charge transfer resistance with increasing absorbed dose. This reduction in resistance indicated improved interfacial carrier mobility,further validating the role of high-energy ionizing radiation in enhancing interfacial charge transport dynamics.

These findings demonstrate that gamma-ray radiation-induced TBC defects effectively overcame the inherent limitations of BNNS in visible-light-driven catalytic applications. First,the introduction of defects established intermediate energy levels,enabling visible light absorption. Second,the defects shifted the CBB potential to more negative one,thereby enhancing the reduction potential of photogenerated electrons and thermodynamically favoring photocatalytic reactions. Third,the defects facilitated the separation of photogenerated carriers,minimizing recombination losses. Finally,the partial incorporation of defects improved interfacial charge transfer efficiency,solidifying the potential of irradiated BNNS for advanced photocatalytic application under visible light.

This study presents a novel and straightforward approach utilizing high-energy ionizing radiation to introduce defects in BNNS and achieve their visible-light-responsive catalytic activity. Gamma-ray radiation-induced TBC defects introduced intermediate energy levels into BNNS,effectively narrowing the bandgap from 5.3 eV to 2.7 eV and facilitating visible light absorption. Electrochemical analyses confirmed that these TBC defects promoted efficient separation and rapid migration of photogenerated carriers,resulting in exceptional hydrogen evolution performance. The irradiated BNNS achieved a maximum hydrogen production rate of 1 033.7 μmol/(g·h),nearly two orders of magnitude greater than BNNS synthesized via conventional methods. This innovative defect engineering strategy provides a promising pathway for developing efficient photocatalysts.

More importantly,this work highlights the potential of radiation technology in designing defect-rich structures for other wide-bandgap materials,potentially catalyzing groundbreaking advances in their catalytic performance. Such advancements position this approach as a key highlight of radiation chemistry following the study of polymer radiation effects^[35-36]. Compared to conventional material modification methods,radiation-induced defect engineering offers distinct advantages. High-energy ionizing radiation directly introduces point defects,dislocations,and vacancies in materials,allowing fine control of defect types,densities,and distributions. These radiation-induced defects not only alter the microstructure of materials to optimize their performance but also generate unique high-energy state defects,offering significant research and application value in materials science.