Electron transfer and H2 evolution in hybrid systems based on [FeFe]- hydrogenase anchored on modified TiO2

An efficient conversion of solar energy into renewable fuels is one of the most important targets in the last decades. The hybrid system composed by [FeFe]-hydrogenase anchored to the surface of TiO2 could be a valid alternative to typical noble metal co-catalyst in the H2 production from solar light. In this paper we investigate by Electron Paramagnetic Resonance the interaction of hydrogenase with the surface of three distinct types of TiO2 anatase, namely the bare oxide, the N-doped one and a substoichiometric blue material (TiO2-x). Experiments in dark and under irradiation show that the anchored enzyme is in electronic contact with the solid. The catalytic activity under solar light of the three hybrid systems has been preliminary evaluated. The production of H2 is higher for the system containing N-TiO2 with respect to that based on the bare oxide indicating a role of the N intra band gap states in harvesting the visible components of sunlight. The system containing reduced TiO2, in spite of its strong absorption in the visible, is less active suggesting that specific centers for visible light absorption are needed to obtain an efficient photoexitation process.


Introduction
Since the discovery, in 1972, of the photosplitting of water in hydrogen and oxygen [1] based on a photo-electrochemical cell with an anode made of titanium dioxide, this compound became a system of reference in the area of photochemical and photo-electrochemical applications. Titanium dioxide, or titania, is nowadays the most important photocatalyst employed in processes for pollutants abatement, [2] as biocide, [3,4] in odour control [5] and in self-cleaning of glasses or of external surfaces. [6] As to the mentioned applications for hydrogen production from water, the advent of more efficient (but often much more expansive) materials and of complex approaches both for the preparation of photoelectrodes and for the direct photocatalytic water splitting, [7,8] eclipsed for quite a long time the role initially played by titania. This material, which has the great advantage to be cheap, abundant and non toxic, actually shows some serious drawbacks hampering its direct use in the process of water photosplitting. The first one is the large value of the TiO2 band gap (around 3.2 eV for the anatase polymorph) that implicates the use of UV photons to perform the excitation of electrons from the valence band to the conduction band limiting therefore the use of sun light (poor of UV components at the earth surface) in photochemical applications of this material. The second drawback is the electrochemical potential of the conduction band electrons which is only slightly more negative than the H + /H2 potential thus limiting the reductive capability of photoexcited electrons. For this reason the most successful photocatalytic applications of titania are in the field of oxidation (e.g. pollutants remediation) so exploiting the excellent oxidative potential of the valence band holes. [9] The need of overcoming these two serious constraints is at the origin of a long activity aiming to modify the solid in order to improve its performances. Among the methods to adjust the photochemical properties of titania extending to the visible range its optical absorption, it is worth mentioning: i) the grafting of sensitizers at the surface of the solid [10] analogous to what done for Dye Sensitized Solar Cells (DSSC, photovoltaic applications); ii) the use of supported metal nanoparticles (mainly Au) to exploit light absorption due to plasmonic effect; [11,12] iii) the doping of the solid with various dopants. In the latter case the most intriguing case is that of doping with non-metals. In particular doping with nitrogen [13][14][15][16] or with combination of elements including nitrogen [17,18] has been the object of intense research activity and of warmly debated results. As a matter of fact, nitrogen doped titania is now currently used in photocatalytic applications employing visible light. [19][20][21] On the other hand, as far as the production of hydrogen by reduction is concerned, the use of cocatalysts is necessary to overcome the described potential limitation. Quite recently hydrogenase-TiO2 nanohybryds have been prepared and tested in photocatalytic process for hydrogen production based on solar light conversion. [26][27][28][29][30][31][32] The purpose of the present work is to investigate the hybrids prepared coupling hydrogenase with modified titanias. The latter are materials based on the anatase polymorph modified in order to make them capable of absorbing visible light. In particular we employed, comparing the results with those of a sample of bare anatase, two kind of materials: i) N-doped anatase, a yellow material, and ii) a substoichiometric reduced anatase which shows a deep blue colour. In N-doped TiO2 the presence of nitrogen defects in the lattice of the oxide generates intra band gap energetic levels available to excite the electrons of the valence band to the conduction band under visible light. These states cause the yellow colour of the powder. [13][14][15]33] The substoichiometric TiO2 is prepared via a particular synthetic route that lead to a partially reduced, blue-colored oxide (TiO2-x) which is stable in air and shows the presence of stable Ti 3+ ions. The color in this case is due to the tail of a broad absorption centered in the infrared region and typical of reduced titanias. [34][35][36][37][38][39][40] Our aim is therefore to overcome the two limitations of titania described before, using doped solids, capable of harvesting some visible components of the solar spectrum and coupling them to an efficient enzyme for hydrogen reduction. Our attention is paid not only to the efficiency of the hydrogenase-TiO2 materials in hydrogen production but also to the mechanism of interaction of the protein with the solid and to the effect of irradiation on the hybrid materials. All this has been monitored by Electron Paramagnetic Resonance (EPR). This technique has been widely used in recent years to investigate the active sites of hydrogenase which are, in particular redox states, paramagnetic. [41][42][43][44][45] EPR, however, has also been used to characterise the paramagnetic centers present in titanium dioxide. In particular the process of photo-induced charge separation creates an electron and a hole which can be stabilized by the solid or transferred from its surface to an adsorbed molecule always producing paramagnetic states. [46][47][48] The EPR technique is therefore much suited to follow the process occurring in the hydrogenase-TiO2 system both in dark and under irradiation.
The EPR experiments of the present work have been performed in aqueous solutions containing thionine and triethanolamine (TEOA). TEOA molecule acts as a buffer and as sacrificial electron donor (or scavenger of the photogenerated holes) for TiO2. [28] Thionine plays the role of stabilizing agent for the oxidized form of the H-cluster (named Hox state) characterized by a diamagnetic [4Fe4S] 2+ sub-cluster and a paramagnetic [Fe(I)-Fe(II)] sub-cluster. [43] In the present paper the results will be organized as follows: firstly the effect of the irradiation on the thionine-TEOA aqueous solution and on the suspension of the various bare supports in the same solution will be monitored by EPR in order to investigate the photoinduced processes in the absence of the enzyme. Then the same approach will be followed for the suspension containing the hydrogenase/oxide nanohybrid. Finally, the hydrogen production ability under solar irradiation of the three hybrid systems in TEOA solution will be compared.

Materials and Methods
Hydrogenase preparation: CpHydA [FeFe]-hydrogenase was cloned from Clostridium perfringens SM09, recombinantly expressed and purified under strict anaerobic conditions as previously described. [49,50]  X-Ray Diffraction: Powder X-ray diffraction (XRD) patterns were recorded with a PANalytical PW3040/60 X'Pert PRO MPD diffractometer using a copper Ka radiation source (0.154056 nm). The intensities were obtained in the 2θ range between 20° and 80°. X'Pert High-Score software was used for data handling.

UV-Visible absorbance:
The UV-Visible absorption spectra were recorded using a Varian Cary 5 spectrometer, coupled with an integration sphere for diffuse reflectance studies, using a Carywin-UV/scan software. A sample of PTFE with 100% reflectance was used as the reference. Due to the high absorbance of the rd-TiO2 material, this latter was mixed with a portion of white MgO.

Surface area measurements:
The surface area measurements were carried out on a Micromeritics ASAP 2020/2010 apparatus using the Brunauer-Emmett-Teller (BET) model for N2 adsorption measurements. Prior to the adsorption run, all of the samples were outgassed at 573 K for 3 h.

UV-visible irradiation:
All the samples were irradiated with UV-visible light using a 1600 W xenon lamp (Oriel instruments) equipped with a IR water filter. The effect of irradiation on EPR spectra was investigated irradiating the sample in the EPR cavity at 77 K for 15 minutes.

EPR characterization: Continuous Wave Electron paramagnetic resonance (CW-EPR) experiments
were performed with a Bruker EMX spectrometer operating at X-band (9.5 GHz), equipped with a cylindrical cavity operating at 100 kHz field modulation. All the spectra were recorded with the following experimental parameters: microwave power 10 mW, Modulation Amplitude 0.2 mT and temperature 77 K.

H2 photoevolution test:
The test was performed in a 7 mL glass vial filled with 1 mL of O2-free sample solution containing 25 mM TEOA pH 7, 1 mg/mL TiO2 and 50 nM CpHydA.
They were irradiated for 60 minutes under natural sunlight at a flux of approximately 650 W/m 2 .
During irradiation the samples were kept on ice. Control experiments in the absence of TiO2, or the enzyme or in the dark were also performed. All the samples were tested at the same time, under the same experimental conditions. Hydrogen evolution has been measured via gas chromatography. The gaschromatographer (Agilent Technologies 7890A) was equippedwith purged packed inlet, Molesieve 5A column (30 m, ID 0.53 mm, film 25 μm) and thermal conductivity detector; argon was used as carrier gas. Quantitative separation was achieved in less than 3 min at 60°C.

Oxide characterization.
XRD analysis were performed on the three distinct polycrystalline systems namely: bare TiO2, N-TiO2 and rd-TiO2. The corresponding XRD patterns are reported in Figure 2a), b) and c) respectively.
All the diffraction pattern are characterized by peaks at the same position that correspond to those of the anatase polymorph however, while both TiO2 and N-TiO2 exhibit the same linewidth, the pattern of rd-TiO2 shows much narrower peaks since this oxide possesses significantly larger crystallites.
This result is also confirmed by the surface area measurements (BET method), indicating about 80 m 2 /g for TiO2 and N-TiO2 and about 10 m 2 /g for rd-TiO2.   The spectrum of the as prepared thionine in TEOA solution in dark (Figure 4(a)) shows a very weak, negligible signal centered at g=2.005. This signal however increases upon irradiation (Figure 4(b)) and disappears by heating the sample at room temperature in dark (Figure 4(c)). The isotropic signal at g=2.005 is due to the formation of small amounts of the thionine radical cation, [51,52] and is not observed irradiating a solution containing TEOA only. The radical species is stable in the frozen solution only (77K) and is destroyed raising the temperature.  The TiO2 powder in air ( Figure 5A(a)) is EPR silent, and also the spectrum of the as prepared TiO2 suspension in dark ( Figure 5A(b)) shows very weak signals including that at g= 2.005, due to the presence of thionine. The irradiation with UV-visible light of the suspension causes a change in the EPR spectrum ( Figure 5A(c)) which becomes dominated by the narrow isotropic signal of the irradiated thionine (marked by an asterisk). Beside this signal a broader one, centered at g= 1.93 and unambiguously due to Ti 3+ ions in anatase [46] is present. The features of this spectrum are stable after irradiation (lamp off) maintaining the sample at 77K while, after thaw-freeze (RT-77K) (data not shown for sake of brevity), they disappear and the EPR spectrum becomes the same observed ( Figure 5A(b)) before irradiation. The effect of the irradiation is therefore clear. Apart for the formation of the thionine radical ion (which is a sort of side-effect of the process and which will be no longer discussed in the following) the experiment indicates that the electrons formed upon irradiation (or, at least, a fraction of them) remain in the solid where they are trapped by Ti 4+ ions which reduce to Ti 3+ . [47,48] The holes photogenareted in parallel are consumed by the electron donor (TEOA) in the system and the classic electron-hole recombination, which would occur increasing the temperature, becomes of course impossible. This is the reason of the formation of the broad signal due to reduced trivalent titanium centers in TiO2.

Irradiation of
The same experiment was performed also for the N-doped and the reduced TiO2 materials.
In the former case the final result is the same observed for bare TiO2 as Ti 3+ centers are observed upon irradiation of the frozen suspension ( Figure 5B(c)). The spectra before irradiation are affected by the presence of nitric oxide (NO) which is a by-product of the N-TiO2 synthesis and which does not interfere with the photochemistry of the material. [15] The typical EPR signal of NO (whose amount is not constant in different samples depending on details of the preparation) is highlighted in Figure 5B(a) and Figure 5B(b) and it is no more visible after irradiation probably because the NO molecule reacts with the photoformed charge carriers.
In the case of the substoichiometric TiO2 (rd-TiO2) the spectrum of the powder ( Figure 5C(a)) reveals, beside of an isotropic signal at g= 2.003, possibly connected with electron trapped at lattice defects, signals with g values between 1.99-1.88 typical of Ti 3+ ions in anatase. [46,47] In particular a narrow signal centered at g=1.992 and a broader one at g=1.93 are visible, respectively due to isolated Ti 3+ in regular position of the lattice (bulk) and to various families of Ti 3+ formed at the surface in relatively disordered environment. [46,47] The suspension of the rd-TiO2 in TEOA-thionine solution ( Figure 5C(b)) maintain the same EPR feature of the powder (Figure 5C(a)) while the UV-visible irradiation ( Figure 5C(c)) causes few modification essentially due to the thionine radical cation. Is more difficult, in this case, to appreciate the change in intensity of the broad signal of Ti 3+ centers because the sample already contains structural Ti 3+ center prior irradiation.
Summarizing while TiO2 and N-TiO2 show electron trapping ability (as Ti 3+ ) upon illumination of suspension, the same doesn't occur in the case of rd-TiO2 (a material which is electron rich already before irradiation).

EPR experiment on TiO2-impregnated CpHydA systems in TEOA-thionine solution.
The same kind of CW-EPR experiments previously described in Figure 5  The spectra of the first and second hybrid systems and their behavior under irradiation are similar, at least in qualitative terms and will be discussed together.
The EPR spectra of the as prepared oxidized hybrid systems in dark, Figure 6A(b) and Figure 6B =2.0363 and g3 =1.9954 while for Hox-CO the main values are at g =2.075 and g=2.008. [50] The fact that the spectra observed for the surface-anchored protein strictly correspond to those of the free protein, clearly suggests that the active site is not strongly perturbed by the interaction with the oxide and that it likely preserves its local geometry and chemical environment (at least the part of the Hcluster that is monitored by EPR) during the impregnation process and the related treatments.
Irradiation of the two hybrid systems (TiO2, N-TiO2) with UV-visible light ( Figure 6A(c) and Figure   6B(c)) causes small but significant changes in the EPR spectra. The weak signal due to the Hox-CO species decreases in intensity, in agreement with previous studies, [53,54] while a new small signal, having a component clearly appreciable at g= 1.93, appears. This signal corresponds to that of the reduced [2Fe2S] + clusters, with is characterized by g values g1=2.0227 and g2= 1.936,. Such redox center is peculiar of this type of hydrogenases and it is paramagnetic in reductive conditions. [55,56] This signal is also detectable upon irradiation of the protein alone and thus it is not an evidence of electron transfer from TiO2 to the protein. The features of this signal remain visible when the lamp is turned off and maintaining the sample at 77K (data not shown). After a thaw-freeze treatment (RT-77K in dark) the signal due to the reduced [2Fe2S] + disappears and the features of the EPR spectrum of the oxidized hybrid systems before irradiation are fully recovered (spectra (d) in Figure 6A and Figure 6B).
Rather than the appearance of new weak signal upon irradiation, the most important feature of the sequence of spectra shown in Figure 6A and Figure 6B is that the broad EPR signal, typical of reduced Ti 3+ centers in irradiated-TiO2 and N-TiO2 suspensions ( Figure 5), is not observed upon irradiation of the same samples containing the anchored protein. This is a clear evidence of the occurrence of electron transfer from TiO2 to the protein under irradiation. The photoexcited electrons, in this case, are not stabilized by the matrix as Ti 3+ ions, but are most likely transferred from the oxide to the protein.
In the case of the rd-TiO2-CpHydA material things are in part different. Also in this case the main features of the spectrum of the as prepared hybrid system ( Figure 6C(b)) are those of the Hox state of the protein. It is worth to note that in the right side of this spectrum only a weak signal of Ti 3+ ions is observed. However the Ti 3+ signal is clearly visible in the spectra of rd-TiO2 both as dry powder and in suspension ( Figure 5C(a, b)). This indicates that, upon contact with the protein the reduced Ti 3+ centers have been oxidized suggesting that excess electrons have been nearly completely scavenged by the protein itself. In Figure 7 the spectrum of the rd-TiO2 suspension alone is compared to that of By the way in all hybrid systems tested upon irradiation a slight decrease of the Hox signal intensity is observed suggesting, for a fraction of the protein, a reduction of the active site to an EPR silent form. These three facts converge in indicating that the electrons generated by irradiation of the solid are scavenged by the protein even at low temperature and become potentially available for H + reduction.
With this in mind the photocatalytic activity of the various hybrids in the production of hydrogen under solar irradiation can now be compared.

Tests on H2 evolution rate.
In order to obtain a preliminary evaluation of the photocatalytic activity for the TiO2-based hybrid systems these were submitted to a test of hydrogen evolution under solar light irradiation.
The amount of photogenerated H2 was measured, via gas chromatography, for the three hybrid systems based on TiO2, N-TiO2 and rd-TiO2 respectively. The corresponding blank tests were performed in the same condition using the oxide supports only. Figure 7 reports the H2 evolution rates the three TiO2-CpHydA hybrid systems in comparison with those of the TiO2 powders alone. In both cases the powders were suspended in a TEOA solutions and exposed at direct sunlight for 60 minutes. Control experiments performed in the dark or in the absence of TiO2 did not show any detectable H2 production (data not shown). Figure 7 clearly shows that all the hybrid systems present a significantly higher activity in H2 evolution with respect to that observed for the TiO2 powders alone.

Conclusions
We have characterized for the first time by EPR spectroscopy the behavior of different TiO2-CpHydA hybrid systems under irradiation. The experiments reported in this paper demonstrate that the various treatments undergone by the protein along the anchoring on the TiO2 powder seem not affect the Hcluster geometry and the catalytic capability of the protein.
Furthermore, the experiments shown in Figure 6 show that the hydrogenase molecule in the hybrid system is indeed a scavenger of photoexcited electrons, in fact these are no more trapped by Ti 4+ centers as it occurs in the case of the oxide supports in the absence of the protein ( Figure 5).
This result is relevant as it definitely shows that hydrogenase after surface anchoring on TiO2, is in electronic contact with the solid thus allowing the electron transfer and the H + reduction.
In terms of photocatalytic activity the present work remains a proof of concept. The H2 evolution tests, however, shows that, in spite of a quite modest production of H2, the hybrid system could have a strong potential in this field since the functionalization of TiO2 with the protein sustains an H2 evolution rate at least one order of magnitude higher than that of the TiO2 powder alone.
Finally, the N-TiO2-CpHydA sample exhibits the highest activity among all hybrid systems tested in this work. This is due to the presence of intra band gap levels 2.5 eV below the conduction band that, as previously shown by some of us, [13][14][15] are available to excite the electrons of the valence band to the conduction band even under visible light. This allows the employment of a fraction of photons present in the solar light [14] which are not effective in the case of bare TiO2. To conclude, once demonstrated the potential of oxide supported hydrogenase in H2 photoproduction, the search for an optimal solid matrix to improve the photocatalytic performance under sunlight is now open.