Visible light switchable bR/TiO₂ nanostructured photoanodes for bio-inspired solar energy conversion

Abstract

Today, regarding the limitation and environmental side effects of fossil fuel resources, solar hydrogen production is one of the main interests in the energy research area. The development of visible light sensitized semiconductors based on non-toxic components, low cost and available bio-species is an ongoing approach for H₂ generation based on water splitting reactions. Here, two different morphologies of TiO₂ photoanodes, nanoparticulated and nanotubular, have been modified with simply extracted bacteriorhodopsin (bR) without any linker. Achieving a significant enhancement in photoconversion efficiency of TiO₂ photoanodes, η% was increased from 2.9 to 16.5 by bR addition to the TiO₂ nanoparticulate electrode, whereas it increased from 0.5 to 1.9 for the nanotubular system. This result indicated that because of the large size of the extracted purple membranes, nanoparticles showed a greater number of explosion sites for bR attachment, and they were more efficient than the nanotubes. Moreover, comparing these values with the ones reported for bR/TiO₂ photoanodes before, the simple proposed decoration method displayed a remarkable improvement in the visible response of the TiO₂ based photoanodes. This approach can be beneficial to the area of solar water splitting, photoelectrochemical sensing, and any other photoresponsive systems.

---

1. Introduction

Today, substituting fossil fuels with renewable ones is an important challenge. Considering solar energy as a free, available and clean energy resource and also considering water as a byproduct of hydrogen combustion/conversion, it seems that using solar-produced H₂ as a future energy carrier is the best solution.¹ In this regard, two separate scenarios were developed to generate solar assisted hydrogen based on water splitting reaction. In the first method, the required energy for a conventional electrolyser system is provided by photovoltaic systems,² whereas in the second one, water splits directly to H₂ and O₂ on the surface of a photoexcited semiconductor material in a photoelectrochemical (PEC) process.³ Photoelectrochemical water splitting is a more promising approach because the small current density required on the surface of the semiconducting electrode leads to lower over potentials, and the PEC system can be installed as a single monolithic device.¹

From the first report about PEC water splitting using TiO₂ photoanode by Fujishima and Honda,⁴ a huge number of researches have been published reflecting scientist's attempts to overcome the challenges in this field.^5–7 One of the important difficulties is the large band gap of the low cost, stable semiconductors, which does not match with the main part of the solar spectrum (400 nm < λ < 700 nm). The addition of visible light responsive quantum dots,⁸ dyes, and low band gap semiconductors⁹ to the main photoelectrode and also doping with various anions/cations^10–14 are ongoing procedures. In this context, finding visible light sensitized materials that are non-toxic and also highly abundant to fabricate bio-conjugated hybrid photoanodes is important. Gratzel has developed chlorophyll based solar cells in which by visible light irradiation, the photoexcited electrons were injected from chlorophyll to TiO₂ and then they entered an external circuit.^15,16 Similarly, this idea can be used for photoelectrochemical water splitting systems.

Bacteriorhodopsin (bR) is a protein in the purple membrane (PM) of Halobacteriumsalinarium, and it includes seven helices and one interior retinal chromophore.¹⁷ It can act as a visible light sensitized, activating building block in photoanodes, and it can also act as a proton pump for hydrogen generation.¹⁸ Recently, N. K. Allam et al. have attached bRs to the surface of TiO₂ nanotubes using a specific chemical linker (3-mercaptopropionic acid), and they observed a 50% increase in photocurrent densities.¹⁹ A nearly similar enhancement was reported by S. Balasubramanian for Pt/TiO₂ nanoparticles modified with bR using a long time process based on immersion of the samples in bR solution overnight.²⁰ Despite these published reports, the obtained photocurrent improvement for bR modified systems are still low and the effect of supporting semiconductor (TiO₂) morphology on the bR leading and performance is also obscure.

Herein, we have proposed a simple and fast one step method for the decoration of different TiO₂ nanostructures with bR species, which increased the photocurrent density of the systems several times. On the other hand, a comparison between TiO₂ nanotubular and TiO₂ nanoparticulate films in PEC performance of bR/TiO₂ photoanodes has been reported.

2. Experimental details

2.1. Preparation of TiO₂ photoanodes

To compare the effect of TiO₂ film morphology on the photoelectrochemical performance of bR decorated photoanodes, two different nanostructure of TiO₂ were grown: nanoparticulate and nanotubular films.

Self-organized TiO₂ nanotubes arrays (TNA) were synthesized using an anodization method which is reported before.^21,22 Briefly, titanium sheets (99.5% purity, 0.5 mm thickness) were anodized in an electrolyte solution containing deionized (DI) water and ethylene glycol with a ratio of 10 : 90 with 0.1 M NH₄F and a small amount of 1 M H₃PO₄ (for reducing the pH to 5.6). The anodization was carried out at 60 V for 120 min. After the anodization, samples were sonicated for about 30 min in pure ethanol to eliminate the depress surface layer and open the nanotube's mouth. To crystallize TNAs in the anatase phase, the prepared samples were annealed in air at 520 °C for 6 hours with a heating rate of 90 °C per hour.

TiO₂ nanoparticulate (TNP) films were fabricated using a commercial paste of well crystallized TiO₂ nanoparticles in the anatase phase with an average size of 40 nm (Sharif Solar Co.). The paste was spread on a Ti sheet using the Doctor Blade method, and then, the obtained layers were heated at 450 °C for an hour to evaporate the solvent and decompose the binders.

2.2. Decoration of TiO₂ photoanodes with bacteriorhodopsin

Purple membrane was isolated by the well-known method of Oesterhelt and Stoeckenius.²³ To attach the extracted bR on the TiO₂ surface, the freshly annealed TNA and TNP films were immersed in bR solution as soon as the heat treatment step finished. Afterwards, the bR grafted layers were dried at room temperature.

2.3. Characterizations

The optical absorption of the extracted bR solution was studied using a UV-vis spectrophotometer (StellarNet Model EPP2000) with a wavelength resolution of 1 nm. The surface morphology of the fabricated TNA and TNP films was investigated by scanning electron microscopy (SEM) (TESCAN-VEGA3-SB).

The photoelectrochemical response of the bR/TiO₂ photoanodes was studied in a three electrode glassy reactor using Ag/AgCl, Pt wire and phosphate buffer solution (pH = 8) as a reference electrode, counter electrode, and electrolyte, respectively. The reactor was equipped with a quartz window putting the photoanodes as working electrodes behind it and illuminating the layer. The light source was a xenon lamp (YUSHIO, 110 W) with a solar like spectrum. To check the photoactivity of deposited bR, the whole spectrum was eliminated except the green part (λ = 532 nm) by an optical cut off filter. Current–voltage measurements were performed using a galvanostat/potentiostat system (SAMA-500) in a linear sweep voltammetry mode with the scan rate of 100 mV s⁻¹. All reported potentials are versus the Ag/AgCl reference electrode.

3. Results and discussions

3.1. Optical measurements

The main characteristic for successfully extracted bacteriorhodopsin is its optical spectrum, which is presented in Fig. 1(a) for the freshly prepared bR solution. The observed absorption peak at λ = 568 nm revealed the activity of the extracted protein in the visible region, as also reported by others.^19,24 In this figure, the absorption spectra of bare TNP and bR decorated TNP photoanodes have been also presented. As can be seen, the characteristic peak of bR was also present for the bR/TNP photoanodes at 568 nm with a lower intensity, whereas it disappeared for the bare TNP sample.

Fig. 1 (a) Optical absorption of the extracted bR solution, bR/TNP and bare TNP samples and (b) photographs of TNP and TNA photoanodes before and after decoration with bR.

Moreover, Fig. 1(b) shows the image of the TiO₂ photoanodes before and after functionalizing with bRs (represented as the purple appearance for bR decorated TiO₂ samples) even after washing with DI water several times. This evidence confirmed the efficient attachment of photoactive species on the TiO₂ surface by a simple immersion process and in the absence of any linker.

3.2. Surface morphology

Fig. 2 represents the SEM images of the prepared TiO₂ nanostructured photoanodes. According to Fig. 2(a), the vertically oriented TiO₂ nanotube arrays were self-organized with an average wall thickness of 20 nm and diameter of 150 nm. Furthermore, as shown in on Fig. 2(b), all the nanotubes were exposed to the electrolyte, which reveals that the sonication step opened the tube's mouth effectively. As can be seen in the cross section image of the grown structure in Fig. 2(c), the thickness of the nanotubular film was about 5 μm after anodizing the Ti sheet for 2 hours. In the case of the nanoparticulate film, the SEM image in Fig. 2(d) indicates the formation of TiO₂ nanoparticles with size variation in the range of 30–50 nm. In addition, the formation of the anatase phase in such an annealing procedure has been supported in our previous reports.^21,22

Fig. 2 SEM images of (a–c) nanotubular and (d) nanoparticulate TiO₂ photoanodes before bR decoration.

3.3. Photocurrent measurements

To study the electrochemical behavior of TNA and TNP electrodes, a linear sweep technique was used in phosphate buffer electrolyte (pH = 8) under dark conditions. The recorded I–V curves, as shown in Fig. 3(a), indicated that no oxidation/reduction peak was observed for both electrodes, revealing they are chemically inert in the selected potential range. The observed increase in anodic current was related to water oxidation on the TiO₂ surface. On the other hand, it can be seen that dark anodic response of the TNP electrode was significantly higher than the one recorded for the TNA sample. At a potential of +0.5 V, the current density of the TNP layer was 0.06 mA cm⁻², which was nearly ten times more than that of the TNA electrode. This result was attributed to an enhanced film/electrolyte interface in nanoparticulate system in comparison with the nanotubular TiO₂ structure.

Fig. 3 I–V characteristics of (a) TNP and TNA systems without bR decoration in dark and irradiating green light and TNP and TNA systems decorated with bR under illumination with white (b) and green (c) light.

The photoresponse of the bare and bR decorated TiO₂ photoanodes was measured by illuminating the sample with white light (380 nm < λ < 700 nm), and the results are shown in Fig. 3(b). In the case of the TNP photoanodes, loading bR on the system caused a remarkable increase in photocurrent density from 0.09 mA cm⁻² for bare TNP to 0.64 mA cm⁻² for bR/TNP electrode in the bias of +0.5 V. The lower enhancement was also observed for nanotubes, in which the photocurrent increased from 0.07 mA cm⁻² to 0.12 mA cm⁻² by decorating TNAs with bacteriorhodopsin. It is clear that the photoactivity improvement made by the bR species is significantly higher for TNP system than the TNA one, which is attributed to a greater number of effective sites in the nanoparticulate film for grafting bR. After PM extraction, the size of the obtained bR-included membrane was several hundred nanometers. On the other hand, the tubes were 100 nm in diameter. Hence, bR attached just on the top surface of the nanotubular TiO₂ film, which led to covering the tube's mouths and lower effective bR loading in these systems. However, in both TNP and TNA based photoanodes, the addition of bR caused a negative shift in onset potential, which is beneficial for separating charge carriers and preventing photoexcited electron/hole recombination.¹ All photoelectrochemical parameters are listed in Table 1.

Table 1 Photoelectrochemical characteristics of bR modified TiO₂ photoanodes

Photoanode	J (mA cm^−2)^a	Von (mV)^a	J (mA cm^−2)^b	η (%)^b
a Illuminated with white light at fixed V = +0.5 V.b Illuminated with λ = 530 nm at fixed V = +0.5 V.
TNP	0.09	−510	0.06	2.9
bR/TNP	0.64	−590	0.34	16.5
TNA	0.07	−523	0.01	0.5
bR/TNA	0.12	−580	0.04	1.9

---

It is important to understand that this improvement is related to the visible light sensitivity of attached bR or any other changes in the TiO₂/electrolyte interface after decoration with bR. In this regard, all prepared photoanodes are also studied by illuminating the samples with a single special wavelength, λ = 532 nm, which eliminated all possible excitations in the titanium dioxide layer (Fig. 3(c)). However, as can be seen in Table 1, the significant photocurrent enhancement was still observed for bR attached photoanodes, indicating the main mechanism for the photoactivation of samples was attributed to injection of electrons from the protein-retinal complex to the semiconducting nanostructures and also attributed to H⁺ pumping by bR. In fact, the absorption of green light photo by bR caused the excitation of electrons from the HUMO to the LUMO, and because the CB of TiO₂ is located at a lower energy level, injection of the photoelectrons from bR to TiO₂ is energetically favorable.²⁴ These electrons finally transferred to the external circuit and reduced H⁺ on the Pt (counter electrode) surface, whereas water was oxidized on the photoanode surface. A schematic of this process is shown in Fig. 4. The photoelectrochemical efficiency of the prepared photoanodes was also estimated based on the photo-conversion efficiency, which is defined by the following equation:

where J, E, V_app and P are the photocurrent density of the system (mA cm⁻²), fundamental required voltage for water splitting (1.23 V), applied potential (V) and the intensity of incident photons (mW cm⁻²). These values were calculated for the systems excited with green light (λ = 532 nm) applying an anodic potential of +0.5 V that are also listed in Table 1. The efficiency was estimated at 16.6%, 3.0%, 1.6% and 0.3% for the bR/TNP, TNP, bR/TNA and TNA photoanodes, respectively. These obtained values are significantly higher than the ones reported for bR modified TiO₂ photoanodes elsewhere.

Fig. 4 Schematic of the synthesized bR decorated TNA and TNP photoanodes for the water splitting reaction.

N. Allam et al. reported the decoration of TiO₂ nanotubes with bR using 3-mercaptopropionin as a linker and observed a 50% increase in the photocurrent density of the systems.¹⁹ S. Balasubramanian et al. also modified TiO₂ nanoparticles with bR by the immersion of the particles in the bR solution overnight, leading to a ∼50% enhancement in their photoresponce.²⁰ Despite of these reports, the main advantage of the present system is the remarkable increase in the measured photocurrent, which was obtained in the absence of any chemical ligand as a linker. This may be due to the better performance of bR in buffer-based electrolyte (pH = 8) than that of citrate-based electrolyte (pH = 7) or it may be due to eliminating the sonication step that may denature the proteins. Hence, the bR species was attached to the TiO₂ surface by a simple, low cost approach, effectively leading to considerably more increase in the photoresponse of the samples (1.7 and 7 times for TNA and TNP based systems, respectively).

As mentioned before, for better mechanical and electrical connection between bR and TiO₂ nanostructures, the freshly annealed TiO₂ was immersed in prepared bR solution as soon as it cooled to a suitable temperature (∼80 °C) to avoid absorption of any type of pollution. Another method for improving these self-organized attachments is the UV treatment of the substrate and/or heating the sample after immersion in the additive solution. To compare these techniques with each other in the bR/TiO₂ nanoparticulate system, three different strategies were selected. In first sample, the hot (∼80 °C) TNP electrode just after annealing was immersed in bR solution (S1). In the second one, the surface of the TNP electrode, which was annealed a day before, was exposed to UV irradiation for an hour and then was immersed in bR solution (S2). In the third, annealed TNA was decorated with bR and then was heated in 80 °C for about 30 min (S3). Photocurrent measurements illustrated that the current densities of these bR-modified TNP photoanodes were 0.34, 0.11 and 0.28 mA cm⁻² for the S1, S2 and S3 samples, respectively. This result suggested that heating the samples after modification led to a smaller enhancement in the photoactivity of the samples due to deactivation or denaturing of some of the grafted bR on the surface during the heating step. Moreover, UV treatment was not as efficient as other two approaches.

4. Conclusions

Two different TiO₂ morphologies (nanoparticulate and nanotubular films) were modified with bacteriorhodopsin as a visible light sensitizer using a simple linker free approach. Studying I–V characteristics of the modified photoanodes revealed that the decoration of TiO₂ nanoparticles with bR led to a more significant photoresponsive enhancement (∼7 times) as compared with that of nanotubular films (∼1.7 times) due to more surface sites for the attachment of submicron bR species. Moreover, the observed photocurrent improvement made by bR decoration for the both TNP and TNA photoanodes were considerably more than the reported values for similar bR decorated photoanodes. Based on this result, tuning the cleanness of the adsorbing surface by simple methods to avoid using chemical linkers can produce efficiencies even higher than those produced using complicated linker based approaches. This is a useful strategy to develop bio-inspired systems in optobioelectronic applications. Coupling these systems with two dimensional nanostructures (such as graphene²⁵ or metal dichalcogenides), tuning the size of extracted bR and/or investigating a variety of TiO₂ morphologies can obtain even higher efficiencies in visible light induced PEC performances in the future.

Acknowledgements

The authors would like to thank Iran National Science Foundation (INSF) for financial support. Useful assistance of Dr Mohammadpour and Mr Qorbani is also greatly acknowledged.

References

1. R. Van de Krol and M. Gratzel, Photoelectrochemical Hydrogen Production, Springer, 2012.
2. J. Luo, J. H. Im, M. T. Mayer, M. T. Schreier, M. K. Nazeeruddin, N. G. Park, S. D. Tilley, H. J. Fan and M. Grätzel, Science, 2014, 345, 1593.
3. X. Chen, C. Li, M. Gratzel, R. Kosteckid and S. S. Mao, Chem. Soc. Rev., 2012, 41, 7909.
4. A. Fujishima and K. Honda, Nature, 1972, 238, 37.
5. N. Naseri, H. Kim, W. Choi and A. Z. Moshfegh, Int. J. Hydrogen Energy, 2012, 37, 3056.
6. N. Naseri, H. Kim, W. Choi and A. Z. Moshfegh, Int. J. Hydrogen Energy, 2013, 37, 2117.
7. N. Naseri, P. Sangpour and S. H. Mousavi, RSC Adv., 2014, 4, 46697.
8. S. Xie, H. Su, W. Wei, M. Li, Y. Tong and Z. Mao, J. Mater. Chem. A, 2014, 2, 16365.
9. L. A. King, W. Zhao, M. Chhowalla, D. J. Rileya and G. Eda, J. Mater. Chem. A, 2013, 1, 8935.
10. C. Wang, Z. Chen, H. Jin, C. Cao, J. Lia and Z. Mib, J. Mater. Chem. A, 2014, 2, 17820.
11. Q. Kang, J. Cao, Y. Zhang, L. Liu, H. Xuabc and J. Ye, J. Mater. Chem. A, 2013, 1, 5766.
12. H. Cui, W. Zhao, C. Yang, H. Yin, T. Lin, Y. Shan, Y. Xie, H. Gua and F. Huang, J. Mater. Chem. A, 2014, 2, 8612.
13. M. Mansoob Khan, S. A. Ansari, D. Pradhan, M. Omaish Ansari, D. Hung Han, J. Lee and M. H. Cho, J. Mater. Chem. A, 2014, 2, 637.
14. Q. Zhang, E. Uchaker, S. L. Candelariaza and G. Cao, Chem. Soc. Rev., 2013, 42, 3127.
15. B. Oregan and M. Grtazel, Nature, 1991, 353, 737.
16. M. Gratzel, Acc. Chem. Res., 2009, 42, 1788.
17. N. L. Wagner, J. A. Greco, M. J. Ranaghan and R. R. Birge, J. R. Soc., Interface, 2013, 84, 20130197.
18. L. K. Chu, C. W. Yen and M. A. El-Sayed, Biosens. Bioelectron., 2010, 26, 620.
19. N. K. Allam, C. W. Yen, R. D. Near and M. A. El-Sayed, Energy Environ. Sci., 2011, 4, 2909.
20. S. Balasubramanian, P. Wang, R. D. Schaller, T. Rajh and E. A. Rozhkova, Nano Lett., 2013, 13, 3365.
21. M. Gholami, M. Qorbani, O. Moradlou, N. Naseri and A. Z. Moshfegh, RSC Adv., 2014, 4, 7838.
22. M. Qorbani, N. Naseri, O. Moradlou, R. Azimirad and A. Z. Moshfegh, Appl. Catal., B, 2014, 162, 210.
23. D. Oesterhelt and W. Stoeckenius, Methods Enzymol., 1974, 31, 667.
24. L. S. Li, T. Xu, Y. J. Zhang, J. Jin, T. J. Li, B. S. Zou and J. P. Wang, J. Vac. Sci. Technol., A, 2001, 19, 1037.
25. P. Wang, N. Dimitrijevic, A. Y. Chang, R. D. Schaller, Y. Liu, T. Rajh and E. A. Rozhkova, ACS Nano, 2014, 8, 7995.

---