Sensor Surface via Inspiration from Nature: The Specific Case of Electron Trapping in TiO2/WO3(0.33H2O) and Reaction Center/WO3(0.33H2O) Systems

In this work, reaction centers (RCs) isolated from Rhodobacter sphaeroides purple bacteria was coupled with WO3(∙0.33H2O) via physical adsorption, where vectorial electron transfer (from RCs towards the inorganic carrier) was demonstrated using flash kinetics and photoluminescence measurements. The efficiency of the interaction between RCs and WO3(∙0.33H2O) was correlated to the components’ surface charge at the working pH and the structural/morphological and surface properties of the inorganic carrier (e.g., anchoring capacity assured by H-O, W=O bonds). The role of WO3(∙0.33H2O) as final electron trap and charge separator was proven not only in RCs/WO3(∙0.33H2O) biohybrid systems but also in TiO2/WO3(∙0.33H2O) composites. The charge transfer in the inorganic composites was evaluated by monitoring the reverse process of the color reaction (W5+ → W6+) via diffuse reflectance spectroscopy (DRS) after a previous UV-A (320400 nm) exposure. The efficiency of the charge transfer process in inorganic systems was correlated to the initial W5+ content of WO3(∙0.33H2O), followed by the photocatalytic efficiency evaluation of these inorganic composites under UV-A irradiation.


Introduction
The integration of biological components exploiting unique peculiarities and properly engineered structures has led to creating specific structures such as bio-hybrid actuators and biohybrid sensors, contributing significantly to a sustainable development [1][2]. Among the great variety of biological systems, photosynthetic reaction center proteins (RCs) have received considerable attention from the scientific community due to their broad applicability in nanobionics, optoelectronics, photonics, photovoltaics given by their high sensibility, selectivity, and high quantum yield [3]. RCs are transmembrane proteins with significant pigments and cofactors situated in the photosynthetic membrane of autotroph organisms (plants, some bacteria, and algae) [4]. Even if solar cell technology is a considerable application field of RCs because of their high quantum yield [3] (thus providing a solution for exhausting energy resources), their 'natural role' is related to the conversion of light into chemical energy, consequently contributing to food production [5].
Bacterial RCs are characterized with significantly lower structural complexity than the corresponding redox proteins from plants; thus, integrating bacterial RCs into different devices has remained a more attractive option. The unique peculiarities of bacterial RCs are assigned to the cofactors (bacteriochlorophyll -BChl; bacteriopheophytine -BPheo; ubiquinone 10 -Q A , Q B , non-heme type Fe 2+ ) within the protein. Light absorption induces vectorial electron transfer from the primary donor (P, dimer bacteriochlorophyll) to the secondary quinone (Q B ) through the formation of charge-separated species (e.g., ), followed by charge P + BPheo -, P + Q -A , P + Q -B stabilization [6].
Even if these Nature's creations manifest unique properties, their stability issues can be considered a bottleneck in their application. To overcome this limitation, RCs were coupled with several inorganic carriers, namely TiO 2 /WO 3 [7], ITO [8], Si [6], CNT [9], either by electrostatic adsorption or covalent binding [10], thus improving the stability of the whole biohybrid system.
Among the coupling methods, adsorption is considered the simplest one [11]. However, its efficiency is doubtful since the protein structure could change its initial conformation [12]. Even if the possibility of conformation change exists, several cases were reported when the functionality of RCs remained unaffected [12] after bringing it in contact with Au [13], ITO [8]. Beyond conformation change, another vital issue in the case of adsorption is related to the surface charge of components at the working pH, so the assessment of the isoelectric point (IEP) of components (e.g., , IEP RC -R26 =6.1 [14]) is essential. . WO 3 can be considered one of the most promising inorganic carriers for RCs due to its charge separation capacity and efficient electron acceptor in different composite systems. The conditions regarding the charge transfer from the RCs to WO 3 (·0.33H 2 O) are also fulfilled: (i) its relatively high midpoint potential ensures the spontaneity of electron transfer from the redox protein to WO 3 [15][16], (ii) the energy of the generated holes on its valence band is relatively low, so there is no possibility for oxidizing the protein by the semiconductor [17]. Considering the aforementioned issues, the interaction between RCs and WO 3 (·0. using the corresponding method of preparation -either physical mixing [18] or pH adjustment corresponding to the surface charge of the components of the composite [19] is considered as an efficient way to prepare composites with great sensorial and photocatalytic efficiency (in certain cases, such as phenol, methyl-orange) [20][21]. According to Lee and coworkers, the excitation of TiO 2 /WO 3 composites triggers the following steps [22]: (i) the excitation of semiconductors and the formation of charge carriers, (ii) charge-transfer (WO 3 (h + , e -) / TiO 2 (h + , e -) → WO 3 (e -) / TiO 2 (h + )), (iii) re-oxidation of W 5+ to W 6+ due to the involvement of molecular O 2 . The degradation of pollutants could be realized either by indirect mechanism (involving reactive intermediates, such as ·OH, ·O 2 -) or direct hole oxidation [22].
The joint key component of the previously described biohybrid (RCs-based) and inorganic systems (TiO 2 -based) is WO 3 because of its high affinity towards electrons (expressed by the terms of positive reduction potential) [23] and efficient charge separation capacity [18]. Tungsten trioxide has four well-known crystal phases: monoclinic, tetragonal, triclinic, and orthorhombic [20]. While the most common obtained crystal phase is monoclinic, the formation of thermodynamically unfavored crystal phases can be facilitated via shaping agents (e.g., NaCl [24], K 2 SO 4 [25]). The synthesis of partially hydrated crystals (WO 3 ·nH 2 O, where n is either 0.33 or 0.50) with enhanced gas-sensing properties has also been reported, where the adjustment of hydration level is correlated to the precursor' concentration [26].
Although some RCs based biohybrid systems were investigated previously as described before), no relevant study focused on RCs/WO 3 (·0.33H 2 O) composites; thus, this paper will offer a significant insight into this topic. Inspired by RCs/WO 3 (·0.33H 2 O) biohybrid systems, the development of TiO 2 /WO 3 (·0.33H 2 O) inorganic systems with excellent sensorial properties will be presented within this paper.

Materials
The synthesis of differently shaped WO 3 was carried out using: sodium chloride (NaCl,

Preparation of TiO 2 /WO 3 (·0.33 H 2 O) composites
The aqueous suspension of 760 mg TiO 2 (P25) and 240 mg WO 3 (WO 3 , respectively WO 3 ·0.33H 2 O) (with a final volume of 50 mL) was stirred 4 hours at 500 rpm in order to achieve a stable pH and a homogeneous suspension, followed by drying at 80 °C, 12 hours [27].

Isolation of RC
The isolation of RC from Rhodobacter sphaeroides R26 -as described previously by P.
Maróti and C. A. Wraight [28] -was performed by a classical biochemical procedure that involves ultrasonic cell disruption, separation by (ultra)centrifugation, solubilization in LDAO, precipitation with ammonium sulfate, dialysis, fine purification by ion-exchange chromatography (DEAE Sephacell). The solubilized and purified protein complex has been stored at -20 °C. Polysucrose 400 was added to the final complexes (at pH=4) to reduce the particles' sedimentation rate. The main steps of biocomposites' preparation are also presented in Fig. 3.a.

Apparatus
Scanning Electron Microscopy (SEM) was employed for the morphological and dimensional analysis of the studied sample. The measurements were carried out on a FEI Quanta 3D FEG microscope with 25 kV accelerating voltage.
The identification of the crystalline structure was accomplished by X-ray powder diffraction method (XRD) at a fixed wavelength (λ fixed =1.5406 Å insured by Cu-Kα source) and in the range of the incident angle values between 10-50° (2θ°). The diffractometer (Rigaku Miniflex) was operated at 40 kV (tube voltage) and 30 mA (tube current) [29].
A Jasco-V650 spectrophotometer recorded the reflectance spectra (DRS) of the semiconductor oxides in the UV-Visible (250-1000 nm) region. The optical signal transformation into an electrical one was realized by an integration sphere (ILV-724). The device was applied in the determination of the band-gap values and monitorization of the W 5+ / W 6+ species.
FT-IR and Raman spectroscopy were employed to determine the specific vibrational bands present in the structure of semiconductor oxides. The IR spectra of the studied samples (prepared previously in the form of pellets with KBr) were recorded at room temperature, in the range of 400-4000 cm -1 , by Jasco 6000 (Tokyo, Japan) spectrometer. Raman spectra of the samples were recorded with a multilaser confocal Renishaw inVia Reflex Raman spectrometer equipped with a RenCam CCD detector. The 532 nm laser was applied as an excitation source. The Raman spectra were collected using a 0.9 NA objective of 100× magnification. The integration time was 30 seconds, 1800 lines/mm grating for all spectra, and 10% of the maximum laser intensity -laser power of 20 mW. The spectral resolution was 4 cm -1 .
Flash-induced absorption change measurements were performed by a locally designed single beam kinetics spectrophotometer described previously by J. Tandori [30]. The absorption change could not be measured by continuous light due to the reversibility of charge recombination, so the studied sample was exposed to successive saturation flashes. The time interval between two successive flashes was set at 60 s. The specific measurement parameters were: I 0 = 100 mV, amplification: 100).
The absorption spectra of RC (in an equilibrium state and after excitation at λ excitation = 808 nm) in the range of 480-900 nm was recorded via a home-made Vis-NIR experimental setup equipped with a white light of DHL2000 light source as measuring light and CCD detector of a spectrophotometer (both from Avantes) and with a laser diode (2 W, 808 nm, Roithner). The details regarding this experimental setup were already described [31].
Fluorescence emission measurements were carried out by employing a Jasco FP-8600 High Sensitivity NIR spectrofluorometer with 1 nm spectral resolution and 10 nm excitation and

Morphological/structural and optical characterization of WO 3 crystals used for the sensor
Analyzing the SEM micrographs, hierarchical structures of fiber-like ( Fig. 1.a-  The primary crystallite size was ≈45 nm; however, a high aspect ratio difference (fiber-like structure) induces a significant error margin to the obtained value. The XRD pattern of While the thermodynamically favored crystal phase -monoclinic -was identified in the case of star-like crystals [32], the arrangement of atoms into hexagonal crystal lattice [34] was facilitated by the shaping agent [20] in the case of fiber-like crystals. The Kubelka-Munk function was applied in order to determine the necessary energy for the translocation of an electron from the VB to CB (WO 3 2.52 eV→ λ = 550 nm, WO 3 ·0.33H 2 O: 2.69 eV → λ = 461 nm) ( Fig. 1.g).
The lower band gap value was observed in the case of the WO 3 sample [20].
There were some well-defined differences between the two spectra. First of all, in the case of partially hydrated crystals, a relatively narrow band at 244 cm -1 corresponding to W-O-W bending vibrations of the bridging oxygen [36] and a low-intensity shoulder at 324 cm -1 could be identified as the specific band of W-OH 2 stretching vibrations mode [37]. In the case of WO 3 , a characteristic band identified in the neighborhood of the monoclinic lattice vibration (at 256 cm -1 ) was assigned to O-W-O bending vibrations [27,38].
In the case of WO 3 ·0.33H 2 O, two relatively broad bands were identified at 682 cm -1 and 806 cm -1 that are characteristic signals for W-O-W and O-W-O stretching vibrations [39], the band situated at 682 cm -1 being also significant evidence for hexagonal semiconductor-oxide [40]. Two successive bands could be identified in star-like crystals: the highest intensity band (at 807 cm -1 ) corresponded to the O-W-O stretching vibrations [37], while the characteristic signal at 701 cm -1 was assigned to W-O-W stretching modes. The characteristic band situated at 941 cm -1 was attributed to the symmetric stretching vibration of the terminal W=O bonds [41,42], which assures the anchoring capacity of WO 3 ·0.33H 2 O crystals.
Considering that the band situated at 695-700 cm -1 was also assigned to specific stretching vibrations in W 2 O 6, and W 3 O 8 [35,43], the presence of W 2 O 6 and W 3 O 8 on the surface of the studied semiconductor-oxide could be demonstrated by analyzing this band. In order to highlight the differences given by the content of W x O y (x 1 =2, y 1 =6; x 2 =3, y 2 =8) on the surface of

IR spectroscopy
While active vibrations in Raman spectroscopy are related to the polarizability changes, the active vibrations are associated with the dipole moment in IR spectroscopy. Analyzing the FT-IR spectra of the semiconductor oxides ( Fig. 2.b) 120 220 320 420 520 620 720 820 920 1020 1120

Inspiring from RCs coupled with WO 3 (·0.33H 2 O)
The excitation of RC (in fact of the primary donor -P) can be realized through the absorption of photons whose energy is equal to the excitation energy of either cofactor in the visible range (BPheo: 525 and 760 nm; P and BChl: 600 nm, BChl: 800 nm, P:860 nm) because the characteristic absorption band of P is situated at 860 nm [48]. According to Fig. 3.b, the functionality of P was tested through its excitation via a monochromatic diode (λ emission =808 nm).
As a result of excitation, P + is formed, which induces the disappearance of the characteristic absorption band of P (λ=860 nm).
As mentioned in the introduction, coupling RC with different inorganic carriers is a promising way to enhance this redox protein's stability. WO 3 (·0.33H 2 O) could be considered as an efficient inorganic carrier in RC/WO 3 (·0.33H 2 O) biohybrid systems, due to: (i) its relatively high midpoint potential, which ensures the spontaneity of electron transfer from the redox protein to WO 3 (E m (WO 3 ) = 1.25 eV [49]) > E m (Q A ,Q B ) = -0.17 eV [50]), (ii) the energy of the generated holes on its valence band is relatively low, so there is no possibility for oxidizing the protein complex by the holes of the semiconductor [17], (iii) its absorption threshold is located near the visible range [20].
Hexagonal partially hydrated and monoclinic star-like WO 3 semiconductors were coupled with RC (as described previously in the 'Materials and methods' part and presented in Fig. 3.a) to investigate the as-obtained binary systems. While the partially hydrated WO 3 crystal was chosen due to its anchoring capacity (adsorption) assigned to the presence of W=O and O-H bonds on the surface of semiconductors (IR spectra, [51][52]), the star-like shaped WO 3 crystal was investigated for the reason that it has contributed to the enhancement of photocatalytic efficiency of P25 in some cases [21].

Flash kinetics measurements
Relevant information about the redox state of the primary electron donor is obtained from flash-induced absorption kinetics measurements performed at its specific absorption band (860 nm) in the case of different fractions prepared at pH=4 and pH=8 [53]. Because of the pH dependence of free energy difference between and states (ΔG AB ), the P + Q A Q -B P + Q -A Q B interpretation of the kinetic parameters (in terms of time constants and amplitudes) will be carried out separately at different pH values.
Furthermore, the amount of Q B bounded to the RC can be influenced by varying the surfactant concentration, which is also highlighted by the value of the kinetic parameters [54].
Taking into consideration the fact that the solubility of quinone in a hydrophobic medium is favorable, it is evident that: (1) when the concentration of the detergent has been increased, the kinetics describing the attachment/detachment it shifts to the direction of detachment; thus, the contribution of a slow component in the absorption decay after flash excitation is decreasing [55]; (2) without the corresponding concentration of LDAO, the quinone 'is stuck' on the binding site of Q B , resulting in a high contribution and diminishing rate of recombination [55]. In addition, underneath the so-called critical micelle concentration (CMC LDAO =1 mM≈0.02% [56]), the functionality of Q B remains constant [55].
The time constant (τ 1 ) corresponding to the recombination of the primary component (P + ) was fixed on 120 ms (=0.12 s) in both cases (at pH=4 and pH=8), which value allowed Q -A → PQ A to evaluate the decay kinetics properly Because of the relatively significant differences in terms of concentrations, the supernatant fraction was diluted 10times at pH=8.

Case 1: The absorption kinetics measurements at pH=8
According to previous studies, considering the effect of relatively reduced surfactant concentration, the kinetics of charge recombination (measured at 860 nm) after a single Xe flash in the case of the reference sample at pH=8 (RC in Tris buffer solution) was described by two exponential terms representing the contribution and lifetime of fast (circa 9-10% and 120 ms) and slow (circa 88% and 2-2.1 s) components [57].  (Table 1, Fig. 3.c).
Besides, taking into consideration the fact that most of the LDAO was removed during the dialysis, but also the fact that only non-ionic detergent (Triton X-100) was added during the washing steps in preparation to remove the unbounded RC, it was expected to obtain high contribution and decreasing rate of recombination.

Case 2: The absorption kinetics measurements at pH=4
Although the free energy difference has a relatively large absolute value at pH=4 (ΔG AB ≈ -130 meV), its value is relatively low (ΔG AB ≈ -70 meV) and constant when the pH is situated between 6 and 9 [30]. According to this, the reference sample (RCs in citrate/citric acid buffer) at pH=4 was characterized via the corresponding kinetic parameters (  Fig. 4.a). The final complex fractions were diluted during the measurements; thus, the concentrations presented in Fig. 4.c-d are diluted concentration values because of highlighting the differences in kinetic parameters in the case of different fractions ( Table 2). Beyond the different surface charges of the components at his pH, pheophytization of a certain amount of P could also be observed in the structure of RC (Fig. 4.b).

Emission fluorescence measurements
According to the previous luminescence studies of RCs, emission was observed at 918 nm given by the dimer bacteriochlorophyll (P), as a result of the excitation of RCs at λ = 532 nm [58]. Although the corresponding band of photoluminescence appeared in both cases (RC vs.  composites not only as an efficient photocatalyst, but also as a qualitative sensor for some pollutants, either in the liquid phase (e. g. oxalic acid), or in the gaseous phase. As described before, the reduction of W 6+ was induced by a quantum of energy, determined by the electron transfer from TiO 2 to WO 3 (·0.33H 2 O). This charge transfer is reversible in the presence of species with high affinity to electrons (e.g., molecular oxygen).  The next question would be: how can we monitor spectroscopically the charge transfer in the latter mentioned inorganic systems (in other words the formation of W 5+ centers) in the most convenient way? In this work, we propose the recording of the reflectance spectra (DRS) in the UV-Vis-NIR range at certain well-defined moments, because while the absorption maximum of W 6+ centers is situated at 340 nm, this value is 970 nm in the case of W 5+ centers [59].
First, the DRS spectra of the composites before UV irradiation were recorded (Fig. 5.cdark blue spectra-TiO 2 /WO 3 ·0. The differences in reduction and re-oxidation ( were given by W 6 + hν W 5 + , W 5 + →W 6 + ) the following issues: (i) the ratio of W 5+ /W 6+ is higher in the case of WO 3• 0.33H 2 O than in the case of WO 3 [60] in the fundamental state, without any excitation, (ii) structural and morphological differences of WO 3 crystals.

Testing the functionality of the TiO 2 /WO 3 composites via photocatalysis
The electron transfer process from TiO 2 towards the different tungsten oxide -based components is one of the crucial steps in the degradation of model pollutants with these composites. The photocatalytic efficiency of TiO 2 /WO 3 (•0.33H 2 O) composites for the removal of oxalic acid (OA) and methyl orange (MO) from aqueous solutions was tested and analysed [18].
The proposed mechanism of photocatalytic processes in ), (ii) translocation of ( ) into the conduction band, thus = 2.69 eV, ΔE g,WO 3 = 2.52 eV evb generating holes on the valence band ( ) and electrons on the conduction band ( ) (Fig. 5.b). h + vb ecb In Fig. 5 Simultaneous degradation and detection of OA was achieved during the photocatalytic process due to the formation of W 5+ species, which was visible by the bluish colour for the suspension.
Due to the advantageous redox potential of OA and the valence band redox level (vs standard electrode) of investigated semiconductors (TiO 2 , WO 3 (·0.33H 2 O)), direct hole oxidation is also a possible pathway for OA degradation via TiO 2 /WO 3 (·0.33H 2 O). Considering the pK a values of OA (pK a1 =1.27 and pK a2 =4.28), the deprotonation of OA (thus resulting oxalate) is obvious in an aqueous solution [65].
Direct hole oxidation involved the oxidation of oxalate to oxalate radical, followed by the formation of formate radicals via decarboxylation as it was presented in Fig. 7.a [66]. The next step is either the formation of formic acid (which can be decomposed into CO 2 and H 2 O) or directly CO 2 [66]. Considering the role of OA in the presented mechanism, it can be considered as 'sacrificial electron donor' or 'hole scavenger' [67]. The indirect mechanism through radicals is also a possible scenario for its degradation, which involves the formation of the same intermediates as in the first case.
The degradation of MO most probably is accomplished through an indirect mechanism involving free radicals [68]. The mechanism includes the breaking of the azo bond and the involvement of the reactive oxygen species (radicals)in the generation of hydroxylated species  The photocatalytic efficiency of the composites (in both cases ≈77 %) after 2 hours was lower than that of the reference (commercial TiO 2 , X ≈ 96%) in the case of OA degradation (Table 3, Fig.   6.b). In contrast, TiO 2 /WO 3 ·0.33H 2 O composite has been considered the most efficient catalyst for MO degradation with a conversion of ≈97%, after 2 hours, under UV exposure and continuous stirring ( Fig. 6.a). The highest reaction rate is attributed to TiO 2 /WO 3 ·0.33H 2 O during MO degradation (Table 3). Even if a higher reaction rate could be observed in composite systems during OA degradation, the conversion rate in these cases was lower. The higher photocatalytic efficiency of WO 3 •0.33H 2 O could be explained by the presence of W x O y on its surface, which ensures enhanced mobility for electrons, but also by the higher percentage of W 5+ (WO 3 ·0.33H 2 O: 4.05%, WO 3 : 2.35%). In order to answer this question, the following simple experiment was carried out, in which on a 5 (cm) × 5 (cm) plain household tin foil 2 mL WO 3 and, 2 mL TiO 2 aqueous suspension (C = 10 g·L -1 ) was deposited on different sides of the tin foil followed by drying for 5 hours at room temperature. This procedure was repeated twice.
Furthermore, 1 mL of 5 mM oxalic acid solution was dropped onto the titania-coated side.
The coated tin foil was placed under sunlight, and immediately, the foil side which contained WO 3 turned blue (Fig. 7.b). To explain this phenomenon a charge transfer mechanism was conceived and detailed as follows. Charge carriers were generated in the VB (h + ) and CB (e -) of the semiconductors due to solar light irradiation. The photogenerated holes from the VB of the TiO 2 oxidized OA into oxalate radical at the moment when the solvated OA reached the TiO 2 particles ( Fig. 7.a, c). The oxidation product was also an electron, which was transferred from TiO 2 to a W 6+ center through the tin foil (conductor), thus reducing W 6+ into W 5+ . The blueish coloration of the WO 3 coated tin foil was due to the formation of W 5+ centers, as it was demonstrated previously via DRS measurements.
Thus, the question related to the applicability of the inorganic systems was answered,