Bioluminescence-induced photocatalysis on semiconducting oxide nanosheets

ABSTRACT A novel semiconductor photocatalytic reaction system employing a photo-emitting enzyme as an internal light source is pro-posed in the present study. The system completely overturns common sense that conventional photocatalytic reactions must require irradiation from an external light source. A horseradish peroxidase (HRP) catalyzing oxidative bioluminescence reaction of luminol in the presence of H2O2 and manganate nanosheets (MNSs) with a narrow bandgap were utilized for an internal light source and semiconductor photocatalysts, respectively, and both of them coexisted in a same reactant solution. In other words, nano-sized light sources were highly dispersed in the solution, resulting in photo-excitation of MNSs over the entire solution. Photo-activated MNSs simultaneously caused oxidation and reduction, where platinum hexachloride anions (PtCl62-) were utilized as a model substance to be reacted photocatalytically. According to X-ray absorption near edge spectroscopy (XANES) of MNSs after the photocatalytic reaction, the anions were mainly transformed into solid phases of PtO2 and/or Pt(OH)4 by reacting with holes in MNSs. In contrast, a control experiment without HRP, i.e. a dark experiment, did not leave any evidence for photocatalytic reaction of PtCl62-. The detailed mechanism and the advantages/disadvantages of the proposed unique system are explained.


Introduction
Needless to say, photocatalysis technology has rapidly grown in the past decades from the pioneering research on water splitting using a photo-excited TiO 2 single crystal conducted by Fujishima and Honda [1].At present, photocatalytic reaction and related phenomena have been practically applied for production of renewable energy sources [2], decomposition of toxic substances or unfavorable microorganisms [3], key components of solar cells [4], super hydrophilization of solid surfaces [5] and so on.Hence, nobody doubts photocatalytic reaction is vital to modern societies.
Up to date, a variety of inorganic (inhomogeneous) and organic complex (homogeneous) photocatalysts have been reported [6,7].What we can see in common for them is that photocatalytic reaction absolutely demands an external light source with sufficiently high photon energy to activate photocatalysts.As a matter of course, photocatalysts cannot work in the dark or places where photons do not reach.With respect to wellexplored solid inorganic semiconductor photocatalysts in solution phases, photon energy supplied to a reactor is frequently reflected or scattered by solid particles located near a light source, inducing deterioration of photon absorption efficiency and apparent photocatalytic activity because particles existing inside a solution cannot receive photon energy.
The aforementioned drawbacks seem to be never avoided unless an external light source is continued being employed.Therefore, in the present study, we attempted to construct a novel photocatalytic reaction system that can overcome the inevitable problems of conventional inhomogeneous photocatalysts.We focused on a photo-emitting enzyme as an internal light source instead of excitation by an external one because most enzymes are soluble in an aqueous phase and thus behave as highly dispersed light sources over the entire solution.This suggests that the photon energy is efficiently absorbed by semiconducting particles, then the particles trigger photocatalytic reactions independent of their location.Our previous work proved that bioluminescence from photoproteins had enough quantity of energy to evolve anodic photocurrent in n-type semiconductors (Pt-doped α-Fe 2 O 3 ) [8].Although Tatsuma and coworkers have reported that hybrid photocatalysts work even in the dark after charging photon energy [9], to our best knowledge, there is no previous literature tackling photocatalysis based on bioluminescence proposed here except for a chemiluminescence-driven photocatalytic reaction system recently reported [10].Consequently, the present manuscript discusses mechanism and advantages of the novel photocatalytic reaction system in detail.

Synthesis of colloidal solutions of MNSs
In the present study, manganate nanosheets (MNSs) were utilized as semiconductor photocatalysts because MNSs are highly dispersible in an aqueous phase and have a narrow bandgap that can be stimulated by visible bioluminescence.A colloidal solution of MNSs was fabricated with an identical technique reported by Kai et al. [11] Concretely, 20 mL of a mixed solution of 0.6 M tetramethylammonium hydroxide (TMAOH) and 3 wt% H 2 O 2 was added to a 0.3 M MnCl 2 solution (10 mL).Immediately after the addition, appearance of the solution was rapidly changed from colorless and transparent to dark blown, indicating oxidation of Mn 2+ by H 2 O 2 , followed by formation of tetravalent state of Mn with a low solubility.The reactant was magnetically stirred over a night without air sealing to complete the oxidation of Mn 2+ .After the reaction, the excess reactants in the produced colloidal solution were removed by dialysis for pure water with a centrifugal membrane filter several times.The obtained colloidal solution was diluted with a 0.1 M tris-HCl buffer solution for further characterization and photocatalytic reactions.

Characterization of colloidal solutions
A particle size distribution curve of the MNSs colloidal solution was measured by dynamic light scattering (DLS) equipped with an Ar ion laser (λ = 633 nm).The colloidal solution of MNSs was opaque and hence was diluted with a 0.1 M tris-HCl solution (pH = 9) to be [Mn] = 2.5 mM in order to reduce absorption of scattered light by MNSs.To investigate crystal structure, the colloid particles were collected by centrifugation and washed with deionized water several times, and then dried at 373 K over a night.An X-ray diffraction pattern of the obtained dark blown powder was measured with a Cu-Kα radiation.Optical absorption property of the MNSs colloidal solution ([Mn] = 0.51 mM) in a quartz cuvette was measured by a UV-Vis spectrophotometer.A Mn content in the colloidal solution was determined with inductively plasma spectroscopy (ICP) after dissolving MNSs in 1 M H 2 SO 4 .

Measurements of bioluminescence intensity
Emission intensity of bioluminescence induced by H 2 O 2induced luminol oxidation catalyzed by horseradish peroxidase (HRP) was evaluated as follows: 170 μL of 0.1 mg/ mL HRP in 0.1 M tris-HCl (pH = 7 ~ 9) or borate (pH = 10) buffer solutions, and MNSs in some cases were added to a well of a white 96-well microplate with round bottoms.Solution temperature was set to 310 K. Immediately after dispensing a mixed solution (30 μL) of luminol (1 mM) and H 2 O 2 (2 mM), change in luminescence intensity based on oxidation of luminol was monitored during 3000 s with a microplate reader to estimate luminescence lifetime under the present condition.

Measurements of bioluminescence intensity
The present study selected PtCl 6 2-as a model substance to be oxidized and/or reduced photocatalytically.To initiate the photocatalytic reactions, 0.5 mL of a mixed solution of luminol (1 mM) and H 2 O 2 (2 mM) buffered with 0.1 M tris-HCl (pH = 9) was added to 3 mL of a solution composed of 0.1 M tris-HCl (pH = 9, 1.9 mL), 8.23 mM K 2 PtCl 6 (0.5 mL), methanol (0.1 mL), 1 mg/mL HRP (0.5 mL) and 0.25 M MNSs (0.5 mL) in a lightproof container, more concretely, a glass vial covered with an Al foil.After the reaction for 1 h, which is enough time to quench bioluminescence, MNSs was recovered by centrifuging and washed with deionized water several times to get rid of soluble chemical species.The photocatalytic reaction was also performed in the identical solution excluding HRP as a control experiment.To verify whether the photocatalytic reaction actually occurred or not, energy-dispersive X-ray (EDX) spectroscopy and X-ray absorption fine structure spectrum (XAFS) measurements for the MNSs was performed, where the latter was also conducted for some reference samples having different valence states (Pt 0 , Pt II (NH 3 ) 2 Cl 2 and Pt IV O 2 ).The colloidal solution was stable and retained for a long period without precipitating.Figure 1(a, b)show a particle size distribution curve of MNSs estimated by dynamic light scattering (DLS) and a powder X-ray diffraction pattern of MNSs after washing with deionized water and drying.The particle size distribution curve is composed of a single broad peak, and the average size of MNSs was expected to be ca.145 nm.Based on the principle of DLS, the estimated value could be equivalent to planner size of MNSs because they show rapid Brownian motion in the solution.According to Figure 1b, the XRD pattern includes several diffraction lines assigned to crystal planes of hexagonal K 0.45 MnO 2 [13], even though peak intensities related to crystallinity of MNSs were small because the synthesis was carried out at moderate temperature.Interlayer distance between MnO 2 δnanosheets was calculated from d-spacing of (001) plane, which is perpendicular to a planner MnO 2 δsheet.As a result, the distance was estimated to be about 1.6 nm, that almost agrees with a sum of thickness of MnO 2 δ-(0.5 nm) and dimensions of a TMA + ion (0.9 nm), which was intercalated into the interlayer space of MNSs for charge compensation.The slight difference might be due to accommodation of water molecules besides TMA + .These results indicate that the MNSs colloidal solution was successfully produced by the simple technique.In the colloidal solution, the coexistent TMA + ions were adsorbed on MNSs with a negative surface charge through electrostatic interaction, stabilizing MNS monolayers as illustrated in Figure 1c.Optical absorption property of the MNSs colloidal solution was evaluated by means of UV-Vis spectroscopy, and the absorption spectrum was displayed in Figure 1d, where the spectrum was obtained for the MNSs solution after diluting with a 0.1 M tris-HCl buffer solution of pH = 9 ([Mn] = 0.5 mM).Absorption edge was located at ca. λ = 600 nm, suggesting MNSs obtained have a small bandgap of 2.1 eV that is almost identical to the value previously reported [12] and can be excited over a wide range of wavelength in visible region.

Results and discussion
Photoelectrochemical property of MNSs was investigated with a conventional electrochemical cell.A Pt foil was covered with a MNSs thin film by dip-coating in the colloidal solution, and then the resultant MNSs/ Pt was used as a working electrode.To check photoresponse behavior of MNSs, open circuit potential (OCP) was monitored under intermittent irradiation from a blue LED light source as a mimic of bioluminescence.The change in OCP was shown in Figure S2.In the case of a bare Pt foil, no clear shift in potential was observed.On the contrary, a small negative shift of OCP was observed when exposed to photon energy, and then the OCP approached to the original by turning off the light.Judging from the facts, together with the optical absorption property, it was concluded that MNSs are n-type semiconductors and can be excited by visible light as predicted.In the present study, HRP, which is one of oxidoreductases (redox enzymes), was utilized to catalyze a bioluminescence reaction, that is, as an internal light source.It is well known that HRP promotes oxidation of luminol in the presence of H 2 O 2 accompanied by light emission of a blue ~ UV region (Figure S3a).Since the emission spectrum and the absorption spectrum of MNSs overlapped each other, it is expected that the bioluminescence based on the HRP-catalyzed luminol oxidation emits sufficient photon energy to cause photocatalytic reactions on MNSs.In fact, the bioluminescence intensity was reduced by addition of MNSs into the mixed solution containing HRP, luminol and H 2 O 2 as shown in Figure S3b.Therefore, MNSs appear to be suitable as photocatalysts to be activated by bioluminescence.
We also investigated pH dependence of integrated emission intensity of luminol, and the result can be seen in Figure S3c.The pH dependence has a distinct maximum at pH = 9, indicating the photocatalytic reaction might be promoted at the pH value.In general, oxidation of luminol is stimulated at basic conditions; however, an HRP molecule cannot retain its threedimensional conformation at a highly basic solution and hence catalytic activity is deteriorated.Our result is not conflicted with the explanation about the effect of pH on luminescence.Figure S4 displays an emission transient of luminol oxidation catalyzed by HRP.Under the present condition that was identical to that of bioluminescence-induced photocatalytic reactions described later, the luminescence intensity was rapidly decreased with increasing in reaction time, and the emission completely disappeared around 1300s.
Figure 2 illustrates a proposed mechanism of bioluminescence-induced photocatalytic reactions on MNSs.HRP and MNSs are coexistence in an identical solution.As the bioluminescence was maximized at pH = 9 stated already, the photocatalytic reactions were carried out at a solution buffed with 0.1 M tris-HCl (pH = 9).Since both HRP and MNSs in a weakly basic solution have a negative surface charge on the basis of zeta potential measurements, both of them seem to exist separately.On the other hand, due to interaction between methyl groups of TMA + adsorbing on MNSs and hydrophobic amino acid residues of HRP, we cannot ignore possibility that a small amount of HRP and MNSs might be bounded [14].Different from other hybrid nanomaterials with synergistic functions, which are composed of inorganic nanosheets and biomolecules [15,16], the proposed reaction system is not mandatory to bind them because photons originated from luminol molecules easily and effectively reach highly dispersed MNSs.As a result of excitation of irradiated MNSs, produced photocarriers, that is, electrons in the conduction band (CB) and holes in the valence band (VB) would take place reduction and oxidation of chemical species in the solution, respectively.To initiate the photocatalytic reactions of PtCl 6 2-, the mixed solution of luminol and H 2 O 2 buffered with 0.1 M tris-HCl (pH = 9) was added to a solution consisting of 0.1 M tris-HCl (pH = 9), K 2 PtCl 6 , methanol, HRP and MNSs in a lightproof glass vial.After the reaction for 1 h, MNSs were recovered by centrifuging and washing with deionized water followed by drying.The dried MNSs after the reactions in the presence and the absence of HRP are described as "MNSs with HRP" and "MNSs without HRP", respectively.
Chemical compositions of MNSs after the photocatalytic reaction were analyzed with energy-dispersive X-ray (EDX) fluorescence spectroscopy, and the results are shown in Figure 3.In addition to characteristic X-ray peaks of Mn-Kα (5.89 keV) and Mn-Kβ (6.49 keV) as shown in Figure 3a, a broad peak assigned to Pt-Lα (4.99 keV) was clearly discernible for "MNSs with HRP" in Figure 3b.In contrast, there is no detectable peak originated from Pt-Lα in the spectrum of "MNSs without HRP".These facts demonstrate that bioluminescence based on HRP-catalyzed oxidation of luminol could bring about some photocatalytic reactions of PtCl 6 2-and resultant chemical species of Pt were deposited on the surface of MNSs as similar to a conventional photodeposition process on semiconducting photocatalysts.Independent of existence of HRP, Cl was not detectable, on the one hand, a trace amount of K was present in MNSs after the reaction.K + ions generated from K 2 PtCl 6 would electrostatically adsorb MNSs together with TMA + .Calculated molar percentages of Mn and Pt of both samples based on the EDX spectra (oxygen and other trace elements were ignored.)were inserted in Figure 3a.The Pt content in "MNSs with HRP" (average: 0.040 mol%, standard deviation: 0.004 mol% for four times of the identical experiments) was smaller than a calculated percentage of Pt when all PtCl 6 2-ions are deposited (3.2 mol%).Such a small amount of Pt may be due to low integrated luminescence intensity from oxidation of luminol.Furthermore, HRP has the absorption maximum at λ = 402 nm on the basis of a Soret band.Hence, a part of photons would be absorbed by HRP, reducing quantum efficiency of the present photocatalysis.
As a result of direct observation with scanning electron microscopies (SEM and TEM), we could not distinguish the Pt species deposited on MNSs in the case of "MNSs with HRP".In addition, we could not obtain any distinct peak using X-ray photoelectron spectroscopy (XPS) probably due to the small content of Pt.Therefore, to reveal a chemical state of Pt attached to MNSs, X-ray absorption fine structure (XAFS) spectroscopy technique was adapted for "MNSs with HRP". Figure 4 shows X-ray absorption near edge (XANES) spectra (Pt L3-edge) of "MNSs with HRP" and other reference samples with several oxidation states of Pt.In the case of "MNSs with HRP", the absorption peak was closed to that of PtO 2 rather than Pt.On the other hand, an absorption maximum of Pt(NH 3 ) 2 Cl 2 was located at the same with Pt (data not shown).The peak positions of references were almost equal to reported values in the previous paper [17].Therefore, PtCl 6 2-was predicted to be deposited as PtO 2 (Pt(OH) 4 in an aqueous phase) as a result of the bioluminescence-induced photocatalytic reaction, because Cl was not detected in "MNSs with HRP" by the XRF measurement as stated already.It is well known that PtCl 6 2-is easily reduced to Pt metal on inorganic solid photocatalysts under illumination (photodeposition) especially in an acidic solution as described below [18].
Meanwhile, Cl − ions in PtCl 6 2-are partly substituted for hydroxide anions in a basic solution (ligand exchange), producing a complex anion denoted as a general formula of PtCl 6-x (OH) x 2-.The complex anions (x=4) has been reported to react with holes generated in the valence band of TiO 2 irradiated under basic conditions Equation (2) [19].
According to the literature, the conduction band and the valence band are located at 0.14 V and 2.34 V vs. NHE at pH = 0, respectively [20].Therefore, during the reaction of PtCl 6-x (OH) x 2-, electrons in the conduction band might be consumed by reduction of Mn 4+ itself.
In fact, the average oxidation number of MNSs was slightly reduced from 3.47 to 3.32 which were estimated by difference of binding energy between splitted Mn 3s XPS peaks of MNSs (83 ~ 89 eV) before and after the reaction [21].On the other hand, the  band structure of MNSs has a possibility to cause reduction of PtCl 6 2-at the same time (calculated redox potentials from the Gibbs free formation energy of substances participating in redox reactions, Eqation (1): 0.787 V and oxidation, Equation (2): 0.855 V, where the latter was calculated for oxidation of PtCl 6 2-).Even if PtCl 6 2-was reduced to Pt, the product might be re-oxidized in a basic condition.Moreover, the conduction band edge of MNSs is considerably lower (more positive) than those of other oxide nanosheets, for example, the conduction band of titanate nanosheets (TNSs) is located at −1.27 V [12].In other words, MNSs have a lower reducing performance than TNSs at the sacrifice of the narrow bandgap.To compare the bioluminescence induced photocatalytic reaction with conventional system using an external light source, photocatalytic reaction using a UV-LED light source (principal wavelength: 365 nm) was carried out in a mixed solution containing K 2 PtCl 6 and MNSs at pH = 9.Different from the bioluminescence-induced reaction, Pt was clearly detected in the XPS spectrum of MNSs and tetravalent state of Pt was still major chemical state (Pt 4+ : Pt 2+ : Pt 0 = 88 : 12 : 0 (at%) as shown in Fig. S5).Therefore, photoreduction of PtCl 6 2-to Pt seems to be difficult using MNSs independent of incident light intensity.Hence, the oxidation of PtCl 6-x (OH) x 2-had to be the primary process during the photocatalytic reaction.In fact, as the identical reaction was carried out using Fedoped TNSs (Fe 0.8 Ti 1.2 O 4 0.8-), which can be excited by irradiating UV light with shorter wavelength than 380 nm [20], the small diffraction line assigned to Pt(111) was detected in the XRD pattern of the TNSs after reaction of PtCl 6 2-(Figure S6).So, we firstly validated that photocatalytic reactions of MNSs in an aqueous phase proceed with photon energy originated from HRP-catalyzed luminol oxidation without using any external light source.
To explore possibility of the present system, we adapted the bioluminescence-induced photocatalytic reactions for other complex ions (Ag(NH 3 ) + and Cu(NH 3 ) 2+ ) at pH = 9.Unfortunately, a large amount of Ag (3 mol%) or Cu (4 mol%) was deposited both in the presence and absence of HRP.These percentages were almost consistent with molar percentages of Ag or Cu for a total metal quantity in initial reactant solutions.Since these complex ions have a positive charge (cation), most of them might be adsorbed on the negatively charged MNSs via electrostatic interaction.Then, the cations might be reacted with photocarriers in MNSs.So, we cannot judge whether these cations actually participated in photocatalytic reactions.In summary, we reached a conclusion that cationic species are not appropriate as substances to induce redox reaction unless anionic semiconducting nanosheets are employed.

Conclusions
In this article, we proposed a novel photocatalytic reaction system with manganate nanosheets (MNSs) and photoenzymes (HRP) as semiconductors and internal light sources, respectively.Bioluminescence emitted from HRP could successfully excite MNSs, the photo-reactions of PtCl 6 2-were proceeded on the MNSs.XAFS measurements revealed presence of PtO 2 (Pt(OH) 4 ) on the surface of MNSs as a result of redox reactions.Even though the present system can be adapted to aqueous reactions only, there is a significant advantage of which any external light source is not required different from well-established conventional photocatalytic reaction systems.However, the reaction amount of targeted substances was still small which might be due to the low bioluminescence intensity.Optimization of reaction conditions such as concentration of photoenzymes (HRP) and luminescent substances (luminol) will improve photocatalytic activity of MNSs.Since optical absorption by HRP overlaps to a luminescence wavelength range of luminol, we will attempt to adapt the present system to various combinations of other photoenzymes or photoproteins without any absorbance of visible light (luciferase, aequorin, etc.) and inorganic nanosheets with an excellent photocatalytic activity to improve reaction efficiency.

Figure 1 .
Figure 1.(a) a particle size distribution curve of a colloidal solution of MNSs ([mn] = 2.5 mM).(b) A powder X-ray diffraction pattern of MNSs after washing and drying the colloidal solution.(c) An expected dispersion model of MNSs adsorbing tetramethylammonium (TMA+) ions as dispersants.(d) An optical absorption spectrum of a MNS colloidal solution ([mn] = 0.5 mM).

Figure 2 .
Figure 2. A schematic model of a bioluminescence-induced photocatalytic reaction system using MNSs and HRP (photoenzymes) as n-type semiconducting photocatalysts and internal light sources, respectively.The HRP catalyzes oxidation of luminol with H 2 O 2 accompanied by blue light emission, and then the light excites MNSs followed by proceeding photocatalytic reduction and oxidation of chemical species (O or R').The present study utilized PtCl 6 [2-] reacted both with holes in the VB and electrons in the CB.

Figure 3 .
Figure 3. Energy-dispersive X-ray fluorescence spectra of "MNS with HRP" and "HRP without HRP".The calculated molar percentages of Mn and Pt based on the spectra are inserted in the left panel.

Figure 4 .
Figure 4. X-ray absorption near edge (XANES) spectra of "MNS with HRP" together with reference samples having different oxidation states of Pt.