Flexible Nano-TiO2 Sheets Exhibiting Excellent Photocatalytic and Photovoltaic Properties by Controlled Silane Functionalization-Exploring the New Prospects of Wastewater Treatment and Flexible DSSCs.

TiO2 nanoparticles surface-modified with silane moieties, which can be directly coated on a flexible substrate without the requirement of any binder materials and postsintering processes, are synthesized and characterized using X-ray diffraction, Fourier transform infrared spectroscopy, scanning electron microscopy, Brunauer-Emmett-Teller, X-ray photoelectron spectroscopy, Raman spectroscopy, photoluminescence spectroscopy, time-correlated single-photon counting, and transmission electron microscopy. The viability of the prepared surface-modified TiO2 (M-TiO2) sheets as a catalyst for the photo-induced degradation of a model dye, methylene blue, was checked using UV-visible absorption spectroscopy. The data suggest that, compared to unmodified TiO2, M-TiO2 sheets facilitate better dye-degradation, which leads to a remarkable photocatalytic activity that results in more than 95% degradation of the dye in the first 10 min and more than 99% of the degradation in the first 50 min of the photocatalytic experiments. We also demonstrate that M-TiO2 can be recycled with negligible reduction in photocatalytic activity. Further, the photovoltaic properties of the developed M-TiO2 sheets were assessed using UV-visible absorption spectroscopy, electrochemical impedance spectroscopy (EIS), and photochronoamperometry. Dye-sensitized solar cells (DSSC) fabricated using M-TiO2 as the photoanode exhibited a photoconversion efficiency of 4.1% under direct sunlight. These experiments suggested that M-TiO2 sheets show enhanced photovoltaic properties compared to unmodified TiO2 sheets, and that, when N-719 dye is incorporated, the dye-TiO2 interaction is more favorable for M-TiO2 than bare TiO2. The simple solution processing method demonstrated in this paper rendered a highly flexible photoanode made of M-TiO2 with superior charge-separation efficiency to an electrode made of bare TiO2. We propose that our findings on the photovoltaic properties of M-TiO2 open up arenas of further improvement and a wide scope for the large-scale production of flexible DSSCs on plastic substrates at room temperature in a cost-effective way.

Introduction

The holy grail of research in the fields of photocatalytic and photovoltaic conversion is a means for effective utilization of photo-generated excitons, including their quantitative separation into charge carriers, which has been a key factor in deciding the efficiencies of these processes. Efforts to develop suitable candidates that can act as potential light sensitizers and/or an intermediate system that facilitates charge separation culminated in the development of multitudes of organic dye molecules and nanostructures. Of them, TiO2 nanoparticles, which could serve both the abovementioned roles, have already revolutionized the fields of photocatalysis[1] and photovoltaics;[2] as evidenced by large numbers of research articles published each year on the use of TiO2 nanoparticles in both areas. In particular, the optoelectronic properties of both anatase and rutile polymorphs of TiO2 have been intensely investigated for the prospects of using it in light-energy conversion applications.[3−5] TiO2 is widely accepted as an n-type semiconductor with a wide indirect band gap.[6] Photoexcitation of anatase TiO2 has been reported to generate a large number of strongly bound excitons that display an intermediate character between the Wannier–Mott and Frenkel excitons, that hold a two-dimensional wave-like nature,[7] and are confined in the 001 planes in the 3-dimensional lattice.[8] In addition, the critical carrier density, which governs the dissociation of strongly bound excitons into uncorrelated electron–hole pairs, for exciton Mott transition in photodoped anatase TiO2 is found to have a remarkably high magnitude. In rutile TiO2, on the other hand, photoexcitation results in weekly bound excitons that are of the Wannier–Mott character.[9] These excitonic properties, which largely govern events that lead to the formation of charge carriers, along with the excellent electron-transport[10] and electrical properties,[11] make nano-TiO2 an appealing candidate for both photocatalytic and photovoltaic applications.

If devices employed for these applications are flexible, instead of being rigid, it will be more advantageous from an operational point of view. Recently, in the optoelectronic field, flexible devices[12] surged as fascinating new candidates that offer the unique advantage of malleability which makes carriage and installation of the devices on surfaces of any morphology easy where conventional solid-state devices fail because of their rigid and brittle nature.[13] On one hand, if TiO2 nanoparticles are to be used in wastewater purification, from a practical point of view, it is desirable to make the TiO2-based photocatalytic sheet as flexible as possible, so that the water purifier can be operated under any mechanical conditions. On the other hand, to make mechanically flexible dye-sensitized solar cells (DSSCs), it is imperative to have its components, if not all, including the photoanode, in a flexible morphology.[14]

In DSSC applications, in general, TiO2 is coated onto a rigid/flexible substrate which improves the morphology, design, and hence, the performance of the device.[15] For making a TiO2-based photoanode in DSSCs,[16] the procedure of coating requires that the TiO2 be made into a paste by adding suitable organic species, such as polymer binders, which are typically required to make the photoanode films crack-free. The TiO2 film thus made is required to be calcinated at high temperatures to make the TiO2 well-adherent onto the substrate and to remove the residual polymer from the film. In addition, the calcination process eliminates the charge transfer inhibition caused by the nonconducting polymer additive and improves the optoelectrical properties of the photoanode film.[17] However, when flexible plastic sheets are to be used as substrates of the photoanode material in a DSSC scheme, calcination poses a problem as it could degrade the plastic substrates at high temperatures, which could affect the morphology, and hence, the efficiency of the device. Though annealing of the system at a lower temperature can afford the retention of the substrate, it does also retain some fraction of the binder polymer as well, which could pose a barrier to the charge transfer process. Therefore, it is highly desirable to develop a flexible TiO2-based photoanode material that can be coated directly onto flexible plastic sheets without the requirement of a binder polymer, so that a crack-free coating that adheres well to the substrate can be obtained without a calcination process. A lot of research is being carried out worldwide to develop low-temperature fabrication methods of TiO2 photoanodes for flexible DSSCs. Lin et al.(18) developed a binder-free TiO2 photoanode coated on Ti sheets at a low annealing temperature of 120 °C and achieved a photoconversion efficiency of 3.4%. In another article, Zeng et al.(19) made a TiO2 photoanode with hydroxypropyl methylcellulose as a binder material. The binder material was made to degrade by 10 h of UV light treatment followed by annealing at 150 °C to obtain entirely flexible TiO2 sheets. They achieved a photocurrent conversion efficiency of 3.25%. Holliman et al.(20) synthesized a TiO2 photoanode material with hexafluorotitanic acid as a binder and fabricated a TiO2 photoanode by sintering these materials at 119 °C followed by treating the anode with TiCl4, and by carrying out a re-sintering at 119 °C. DSSC fabricated using this photoanode showed a photoconversion efficiency of 4.2%. Lee et al.(21) fabricated a DSSC with a hierarchically structured TiO2 (HS-TiO2) on a flexible indium tin oxide (ITO)-polyethylene naphthalate substrate via an electrospray deposition method and achieved a photoconversion efficiency of 5.57%.

Most of the articles, where fabrication of TiO2-based flexible DSSC has been discussed, show efficiencies between 0.1 and 8%, where the highest efficiency reported fabrication requires many pre/post-treatments such as ball-milling process and mechanical pressing. Toward the objective of developing flexible TiO2 sheets, some articles reported the preparation of secondary mesoporous TiO2 nanoparticle aggregates with primary anatase nanocrystallites as photoelectrode through the electrophoretic deposition method.[22] However, this method was proven to be impractical for large-scale productions. Another report introduced a lift-off technique, in which a layer of presintered porous TiO2 on a glass substrate was transferred to an arbitrary second, flexible substrate at high pressure.[23] All of the methods described above are cumbersome, require sophisticated production machinery, and are not suitable for production on large scales. These difficulties demand the need for a simple and convenient solution-processable method that is manageable for large-scale production requirements. Nonvolatile functional molecular gels/liquids prepared via surface modification with long aliphatic chains or bulky aromatic functional moieties[24] and that exhibit less viscosity at room temperature were proposed as potential candidates for optoelectronic and electrochemical device preparation and applications in the energy harvesting/storage and water purification fields.[25] These materials, when being used in synthesizing a solvent-free slurry, offer the unique advantage that no solvent is required for making the active material so that solvent evaporation problems that may arise from using a solvent in a device can be eliminated. This new approach for preparing functionalized organic and inorganic nanostructures with liquid-like behavior is reported in several recent articles.[26] Bourlinos et al. synthesized a solvent-free viscous nanofluid of anatase TiO2,[27] where poly(ethylene)glycol (PEG)-functionalized sulfonate anion C9H19–C6H4–O(CH2CH2O)20SO3– was used as a canopy. In another article, Yu et al. discuss the synthesis of a solvent-free nanofluid of TiO2 nanoparticles by surface modification using trimethoxysilane.[28] TiO2 nanoparticles coated with the organic canopy showed better dispersion after modification. The same team also studied the properties of nanofluid in another article and they found that as the temperature decreases below 10 °C, the nanofluid turned into a waxy-solid form. Zheng et al. synthesized solvent-free gold nanofluids by grafting thiol 11-mercaptoundecanoic acid onto a gold-nanoparticle surface, followed by self-assembly with a PEG-substituted tertiary amine.[29] Other well-known photocatalysts like ZnO/ZrO2 nanofluids[30] were also prepared using similar surface modification strategies. A multitude of articles[31−34] report the synthesis of silane-modified TiO2 through various routes, its rheological properties, its adhesive nature on various substrates, and so forth. However, of these articles, only a handful discuss the possibility of using silane-modified TiO2 in photocatalysis. One of those research articles reports[54] the photocatalytic activity of TiO2 surface modified with 3-aminopropyltriethoxysilane (APTES) in degrading the methylene blue (MB) dye to an extent of 88% after 240 min of light irradiation. In another report, the role of adsorption and calcination in the photocatalytic degradation of the MB dye in the presence of APTES modified TiO2 was investigated, in which a maximum of 96% degradation percentage after 240 min of irradiation was reported.[36,37] Takeshita[38] has carried out computational studies at the density functional theory (DFT) and time-dependent-DFT levels on an azobenzene-based dye covalently attached to silane coupling agents (SCAs). The SCAs were coupled to a terminal oxirane group of the photosensitizing dye containing a 1-naphthylamine moiety, and it was found that the CH–O interactions can be utilized to regulate the geometry of the dye to achieve characteristics favorable for electron injection into the conduction band of TiO2. Ukaji et al.(39) investigated the effect of surface modification with APTES and n-propyltriethoxysilane on the photocatalytic activity and UV-shielding ability of fine TiO2 particles. They found that the number of surface functional groups (NR) and the density of SCAs modified on TiO2 surface affected the photocatalytic ability and UV-shielding ability of modified TiO2. An NR of 6.2 suppressed the photocatalytic activity by up to a factor of 25%. This indicates that optimizing the amount of the surface modifier positively affects the photocatalytic ability. Efforts were made to use silane-modified TiO2 in photocatalytic dye degradation,[40] and attempts are being made to improve their photocatalytic efficiency. The details of the state of the art of research on silane-modified TiO2 in photocatalytic dye degradation are summarized in Table .

Table 1

Summary of Silane-Modified TiO2 in Photocatalytic Wastewater Treatment and Environmental Applications

coupling agent used	model pollutant	remarks	refs. no
APTES	Cr(VI)	pH plays a crucial role in the absorption of chromium. Optimization of pH, contact time, and concentration leads to maximum absorption	(54)
APTES	MB	calcination plays a key role in tuning the photocatalytic ability of the silane-coupled TiO2 photocatalyst. The calcination changed the zeta potential of the modified TiO2 photocatalyst. Calcination above 300 °C increased the photocatalytic ability and stability of the material	(36)
APTES	MB, orange II	the adsorption has a significant impact on the photooxidation of dyes. The APTES/TiO2 photocatalysts calcined at 900 °C showed high degradation rates. Functionalization via APTES suppressed the anatase-to-rutile phase transition, as well as the growth of the crystallite size	(37)
APTES	MB	the presence of Si and C in the APTES-modified TiO2 contributed to effective inhibition of the anatase-to-rutile phase transformation and the growth of the crystallite size of both polymorphous forms of TiO2 during calcination at high temperature	(40)
vinyltriacetoxysilane	rhodamine B	a large amount of modifier in the silane-modified TiO2 reduced the photocatalytic ability of the TiO2	(41)
octadecyl trichlorosilane	rhodamine B	in the case of modified TiO2, a direct electron transfer to the conduction band of TiO2 upon absorption of UV light was facilitated by the organic chain with an optimum length	(42)
APTES, 3-isocyanato propyl trimethoxy silane (IPTMS)	malachite green	a high silane modifier concentration negatively affected the photocatalytic ability	(43)
propyltrimethoxysilane, triethoxy(octadecyl)silane, trimethylchlorosilane	nitric oxide	silane-modified TiO2 paint showed significant improvement in NO reduction than that of the unmodified sample and showed good self-cleaning properties	(31)
γ-aminopropyltriethoxy silane (APTES), γ-amino propyltrimethoxy silane (APTMS), N-(2-aminoethyl)-3-amino propyltriethoxysilane (AEAPTES)	brilliant red X-3B	the degradation of the dye was 96.4% after 180 min of light irradiation. Doping of Si into the TiO2 effectively delayed the anatase-to-rutile phase transition, prevented the growth of titania grains, and increased surface area and UV light-induced photocatalytic activity	(45)
3–9(trimethoxysilyl)propylmethacrylate (KH570)	heavy metals, phenol, polybrominated diphenyl ethers (PBDEs)	Degradation of the dye was observed at 93% after 4 h of irradiation. Due to the self-condensation effect of the silane modifier, a considerable amount of silicon hydroxyl was provided as the hole trapping agent, thus improving the ability of the catalyst	(46)
APTES	heavy metals like Cu2+, Cd2+, and Pb2+	Adsorption capacities of APTES-modified hollow TiO2 nanospheres for Cu2+, Cd2+, and Pb2+ ions were found to be 12.7, 17.5, and 1.8 times more than those of unmodified samples respectively	(47)

In this article, we investigate the optical and electrochemical characteristics of nano-TiO2 flexible sheets prepared by the surface functionalization of TiO2 nanoparticles with trimethyloxysilane followed by a double sintering process and its application in wastewater treatment and photovoltaics. To the best of our knowledge, our work reports the best performing silane-modified TiO2 flexible photocatalyst sheet that decomposes the MB dye to more than 95% extent within the first 10 min of irradiation, and to more than 99% within less than an hour of irradiation, which is a significantly greater improvement over bare TiO2, which is capable of decomposing the dye to a maximum of 68% (vide infra). In a nutshell, we demonstrate that the prepared nano-TiO2 sheets (M-TiO2), compared to bare TiO2, showcase superior photocatalytic activity and reasonably high photovoltaic properties at room temperature and show far better efficiency than that of other similar reported synthesis routes.[31]

Experimental Section

A detailed experimental section is provided in Supporting Information.[49−51]

Preparation of TiO2 Nanoparticles

7 mL of titanium(IV) butoxide was added dropwise into 15 mL of ethanol under continuous stirring. Following this, a mixture of 2 mL of HCl and 1 mL of deionized water in 15 mL ethanol was added to the titanium(IV) butoxide solution under continuous stirring for 6 h at room temperature. The as-obtained precipitate was washed with water and sintered at 450 °C for 5 h in the oven.

Preparation of Surface-Modified TiO2 Nanoparticles (M-TiO2)

0.25 g of the prepared TiO2 nanopowder, 0.5 g of n-methyl glycine (sarcosine), and 0.5 g of trimethoxy silane were added to 100 mL of toluene and stirred for 24 h at 70 °C. The obtained material was washed with water and sintered at 450 °C for 5 h. The resultant powder was mixed with toluene, stirred for 5 h, and sintered at 450 °C to get the final product. The intermediate product without the final sintering process was selected as a control to find out the effect of stirring with toluene and double sintering. The intermediate product obtained was labeled as I-TiO2 and its photocatalytic activity was compared with those of TiO2 and M-TiO2. When the sintering process was repeated more than 2 times, delamination of TiO2 from the substrate occurred. M-TiO2 samples with different trimethoxy silane contents were prepared. Trimethoxy silane weight percentages varying from 0, 0.25, 0.5, 1, and 5 g were taken and the M-TiO2 preparation procedure was followed to get 5 samples with different silane contents and were labeled as 0, 1:1, 1:2, 1:4, and 1:20, respectively.

Photocatalysis

The photocatalytic efficiency of the catalysts, M-TiO2, I-TiO2, and TiO2, were compared using a standard MB dye solution. A 20 ppm stock solution of MB solution was taken in three 100 mL labeled beakers and 80 milligrams of the catalysts were, respectively, added to each beaker. The MB solution without any catalyst was labeled as blank. After the addition of the catalysts, the samples were sonicated (230 V AC 50 Hz) for 5 min in dark to ensure uniform interaction of the dye and the catalysts. After sonication, each suspension of the solution was allowed to be exposed to intense sunlight (with a power of 970 W/m2, a latitude of 9.669090, a longitude of 76.539070, and an elevation of 21 m). A 5 mL solution was collected from each sample solution at a 10 min interval until the solution turned colorless. The absorbance of the collected solution was analyzed using a UV–visible spectrometer at each time interval. The absorbance was measured at an observational wavelength of λ = 668 nm against a reference sample of deionized water using a Shimadzu UV-2450 spectrophotometer. The relative concentration of the dye before and during the catalysis process, C/C0, was plotted against the time of irradiation for calculating the percentage of degradation. A free-radical scavenger study was also performed to identify the active species responsible for photocatalytic activity. For that purpose, three different scavengers such as AgNO3 (electron scavenger), ethylene diamine tetra-acetic acid (EDTA) (hole scavenger), and isopropyl alcohol (IPA) (hydroxyl radical scavenger) were used. For evaluating the applicability of flexible M-TiO2 sheets in photocatalysis, thin films of M-TiO2 were prepared by grinding the as-synthesized M-TiO2 nanopowder and ethanol with a mortar and pestle to make a fine paste, and the subsequent coating of the paste onto a polyethylene terephthalate (PET) sheet was done by using a doctor blading technique. Photocatalytic activity of the thin film of TiO2 was also checked under identical conditions for comparison purposes (Videos S1 and S2 in Supporting Information). The photographs of the M-TiO2 thin film before and after the dye degradation process are shown in Figure .

Figure 1

(A) M-TiO2 coated PET sheet bent and immersed in dye solution before and (B) after the dye degradation.

Photoanode Fabrication

M-TiO2 was made into a paste by adding ethanol, ground with a mortar and pestle, and coated onto a flexible sheet of ITO-coated PET by a simple doctor blade technique. It should be noted that no polymer binder materials were employed during the coating process and no postsintering processes were conducted afterward. The thickness of the coating was measured using a MAHR precision outside micrometer 40A with a 0.01 mm least count and a range of 0–25 mm, and the thickness of the coating was found as 33 μm.

Cross-Cut Adhesion Test

A right-angle grid was cut into the coatings penetrating through to the substrate. The cuts were made at 45° with respect to the direction of the grain. A transparent pressure-sensitive tape was then placed over the lattice and rubbed with a fingertip to ensure good contact. The tape was removed by pulling it off steadily at an angle as close as possible to 60°. The cut area of the test coating was examined and is classified from 0 (good adhesion) to 5 (bad adhesion) by visual comparison with the illustrations in the standard, depending on the amount of flaked coating.

Preparation of Electrolytes

The (I3–/I–) redox electrolyte was prepared by mixing 0.1 M potassium iodide, 0.5 M tert-butyl pyridine, 0.05 M iodine, and 0.4 M lithium perchlorate in acetonitrile. Cobalt electrolyte was prepared by adding 0.22 M tris(2-(1H-pyrazol-1-yl) pyridine) cobalt(II) tri[hexafluorophosphate] 0.05 M tris(2-(1H-pyrazol-1-yl)pyridine) cobalt(III) di[hexafluorophosphate], 0.1 M LiClO4, and 0.2 M 4-tert-butylpyridine in acetonitrile. Cobalt complexes were added into acetonitrile followed by TBP and LiClO4 and stirred well under 50 °C for 30 min in a closed vial.

DSSC Assembly

Dye anchored M-TiO2 photoanodes were cut into a 1 cm2 square and a parafilm spacer was used as an insulator to cover the parts other than the exposed active area, and the electrolyte was filled inside the cell. Then a platinum counter electrode was sandwiched against the photoanode. The cells were sealed and connected to the electrochemical workstation for IV measurement by using crocodile clips.[52]Figure shows the images of M-TiO2-coated ITO flexible sheets without the dye (A) and with (B) dye adsorption. The M-TiO2 nanoparticles when coated on ITO-PET sheets demonstrated excellent adhesion properties; which was evident from the observation that the coated film stuck on the substrate without any delamination even after repeated mechanical bending and twisting and extensive soaking in alcohol during the N-719 dye loading process.

Figure 2

(A) As prepared M-TiO2 coated ITO sheets and (B) after N-719 dye adsorption.

Results and Discussions

Morphology and Structural Analysis

X-ray Diffraction

X-ray diffraction (XRD) patterns of unmodified TiO2[53] and M-TiO2 are shown in Figure , where both the samples show the characteristic peaks of anatase TiO2. In addition to the peaks that represent the regular anatase phase, the XRD survey of M-TiO2 also shows a characteristic diffraction peak of the rutile phase of TiO2 at a 2θ value of 27.5°.

Figure 3

XRD comparison of unmodified TiO2 (red curve) and M-TiO2 (black curve).

Having some degree of rutile character along with the anatase phase could be advantageous as photo-induced electron transfer between rutile and anatase surface states in TiO2 nanoparticles endows them with high surface photovoltage response and photocatalytic activity.[54] It should be noted that the intensity of the peak at 25° 2θ, corresponding to the diffraction from the (101) plane, is markedly enhanced for M-TiO2 compared to unmodified TiO2, which can be attributed to an increased crystallinity of the former compared to the latter sample. Studies suggest that increased crystallinity enhances the photocatalytic activity of TiO2.[55−61] M-TiO2 showcases an extra peak at 38° 2θ that corresponds to the (004) plane, which indicates that crystal growth mainly occurs along the [001] direction and that the {001} facets are exposed, a characteristic feature of anatase TiO2 nanoparticles.[62] From the XRD data, an average particle size (D) of the samples was also calculated using the Debye–Scherrer formula. Using the formula, the average particle sizes of the TiO2 particles were calculated to be 23 nm and that of M-TiO2 particles was estimated to be 22.75 nm; these values are comparable to the particle sizes estimated from the transmission electron microscopy (TEM) data (vide infra). The relative contents of the anatase and rutile phases in the M-TiO2 sample were calculated from the XRD intensity of the characteristic peaks of the phases[63] and were found to be 89.21% for anatase phase and 10.78% for rutile phase.

Fourier Transform Infrared Spectroscopy

Fourier transform infrared spectroscopy (FTIR) spectra of TiO2, I-TiO2, and M-TiO2 were recorded[64] to analyze the modalities of surface modification on TiO2, and are shown in Figure S1. The FTIR spectra of TiO2, I-TiO2, and M-TiO2 appear to be largely similar with peaks that represent frequencies of stretching (3500–3200 cm–1) and bending (1700–1300 cm–1) vibrations of O–H functional groups on the surface of TiO2 nanoparticles and, possibly, of physisorbed water on TiO2 surface and stretching (1300–1200 cm–1) vibration of Ti–O–Ti bond system. The substantially weaker peak intensities of the O–H functional groups of M-TiO2 compared to that of bare TiO2, however, indicate that a significant fraction of the hydroxyl functional groups have reacted with the silane moieties. In addition, in contrast to the spectrum of unmodified TiO2, I-TiO2 and M-TiO2 reveal a spectral feature at 1061 and 1045 cm–1 which can be ascribed to the Si–O bond stretching vibration.[65] The low-frequency band of I-TiO2 at 940 cm–1 and that of M-TiO2 at 932 cm–1 can be ascribed to the formation of Ti–O–Si bonds[66] in I-TiO2 and M-TiO2 by the interaction of the surface hydroxyl groups of TiO2 with trimethyloxysilane. The intensity of these peaks is found to be reduced in M-TiO2 compared to I-TiO2; which can be attributed to the elimination of some silane groups from the surface of TiO2 due to the double sintering process. Surprisingly, the photocatalytic performance of M-TiO2 was found to be better than that of I-TiO2 (cf.Figures S11 and S12), indicating that the double sintering process has resulted in just the optimal extent of surface functionalization required for better photocatalytic activity (vide infra). FTIR spectra of samples with different weight percentages of trimethoxy silane, such as TiO2/trimethoxy silane of 0,1:1,1:2, 1:4, and 1:20 (Figure S2), were also compared to find the effect of modifier weight percentage change on the properties of the final product. Spectra showed all the peaks mentioned earlier such as frequencies of stretching (3500–3200 cm–1) and bending (1700–1300 cm–1) vibrations of O–H functional groups on the surface of TiO2 nanoparticles and water on the TiO2 surface and stretching (1300–1200 cm–1) vibration of the Ti–O–Ti bond system. All the samples except sample 0, where silane is absent, showed the spectral feature of Si–O bond stretching vibration at ∼1060 and 1040 cm–1. A trend showing an increase in the intensity of Si–O bond stretching vibration at ∼1060 cm–1 can be attributed to the change in trimethoxy silane weight percentage in the preparation of modified TiO2. To assess the nature of the N-719 dye loading onto the TiO2 nanoparticle surface, the FTIR spectra of TiO2 and M-TiO2 samples sensitized with the N-719 dye (same quantities of samples were scratched off from the photoanode films and were immersed in solutions of the N-719 dye[67] of same concentration for the same duration) were measured and are shown in Figure S3. The absorption peaks corresponding to components of N-719, such as thiocyanate group (−S–C≡N) (∼2100 cm–1), carbonyl (∼1715 cm–1), bipyridine (∼1545 cm–1), tetrabutylammonium (TBA) counter-ions (∼1370 cm–1), and carboxylic acid and carboxylate groups (∼3500–3000 cm–1), were observed. It is apparent from the data that the intensities of the IR peaks of sensitized M-TiO2 sample are greater than that of sensitized TiO2, which indicates a greater extent of dye-loading on M-TiO2 than bare TiO2, evidently through a facile interaction, probably of hydrogen bonding in nature, between the carboxylic acid or carboxylate functional groups of the N-719 dye with Ti–O–Si bonds, and with free surface hydroxyl groups of TiO2 nanoparticles.

Raman Spectroscopy

Raman spectra of TiO2 and M-TiO2 (Figure S4) were acquired to further probe how functionalization affects the properties of TiO2. TiO2 sample shows five characteristic Raman active modes of anatase TiO2 at 140.11 cm–1 (Eg), 196.4 cm–1 (Eg), 394.08 cm–1 (B1g), 518.3 cm–1 (combination of A1g and B1g that cannot be resolved at room temperature), and 640.8 cm–1 (Eg).[68] All these peaks are due to vibrations of the O–Ti–O bonds, wherein, the vibrations of Eg symmetry are of symmetric stretching character, symmetric bending has a B1g symmetry, and antisymmetric bending vibrations have an A1g character. These characteristic vibrational frequencies and their intensity ratios confirmed the presence of phase pure anatase TiO2. M-TiO2 also exhibited similar characteristic peaks at 144.8, 394, 513.7, and 640.8 cm–1, corresponding to vibrations of the O–Ti–O bonds. However, the intensity of the peak at 144.8 cm–1 is considerably diminished in the spectrum of M-TiO2 compared to TiO2; this can be attributed to the presence of rutile characteristics in it. In addition, compared to TiO2, the A1g + B1g combination band of M-TiO2 is found to be slightly shifted to lower energy; which can be ascribed to be due to the rutile characteristics[69] of M-TiO2. A minute peak observed at 559.38 cm–1 can be assigned to the presence of silane moieties in the M-TiO2 sample (magnified view of Figure S4).

Scanning Electron Microscopy

The scanning electron microscopy (SEM) images of TiO2, I-TiO2, and M-TiO2, as shown in Figure S5A–C illustrate closely packed aggregates of grains having irregular and undefined shapes in all samples, however, surface modification in M-TiO2 appears to increase aggregates size, resulting in a spongy appearance.

Transmission Electron Microscopy

TEM images of the prepared unmodified TiO2[70] and modified M-TiO2 are shown in Figure . From the TEM images, Figure A,B, it is clear that the size of the prepared TiO2 nanoparticles is almost spherical and has dimensionalities in the range of 20–30 nm, a suitable range of particle size for optimum photocatalysis.[71]Figure C implies that TiO2 shows the presence of only the anatase phase while M-TiO2 shows both anatase and rutile phases. The inset of Figure C shows the SAED pattern of a single TiO2 nanoparticle. From the high-resolution TEM (HRTEM) image (Figure C), a fringe width of 0.34 nm is deduced for the unmodified TiO2 nanoparticles which correspond to the lattice parameter of the (101) plane of the anatase TiO2. The SAED pattern further reveals that the unmodified TiO2 nanoparticles are of polycrystalline nature. Figure D–F shows the TEM images of M-TiO2 in the order of increased resolution, and these images also reveal a particle size range of 20–30 nm.

Figure 4

TEM images of TiO2 (A–C) and M-TiO2 (D–F). (A,B) TEM images of TiO2 with different magnifications; (C) HRTEM image of TiO2, and inset of (C) is the SAED pattern of TiO2. (D,E) TEM images of M-TiO2 with different magnifications, (F) HRTEM image of M-TiO2, and inset of (F) is the SAED pattern of M-TiO2.

In comparison to the unmodified TiO2 sample, M-TiO2 has a porous structure which is shown in Figure E. The SAED pattern suggests a more polycrystalline nature of M-TiO2, with a distribution of single crystals of M-TiO2 in the matrix, which agrees very well with the XRD data. The lattice fringes observed in the HRTEM image suggest a fringe width of 0.23 nm, which can be ascribed to the exposed {001} facets of anatase TiO2. The lattice fringes with 0.35 nm width suggest the (101) plane of anatase TiO2, and the lattice fringe of 0.21 nm is attributed to the presence of the (111) plane of rutile TiO2.[72] Of the several crystallographic planes exhibited by the anatase form of TiO2, the (001) surface has been suggested to be the most reactive site for photocatalytic processes.[73] The TEM data suggests that the surface modification method has exposed the (001) facet of TiO2, and the presence of a small percentage of the rutile phase makes M-TiO2 a potential candidate in photocatalytic applications. TEM images of I-TiO2 are shown in Figure S6. From the TEM images, Figure S6A,B, it is clear that the I-TiO2 nanoparticles are also spherical in shape and have sizes in the range of 20–30 nm.

From the HRTEM image (Figure S6C), a fringe width of 0.34 nm is deduced for the unmodified TiO2 nanoparticles which correspond to the lattice parameter of the (101) plane of the anatase TiO2. Figure S6D shows the SAED pattern of I-TiO2. The SAED pattern further reveals that the I-TiO2 nanoparticles are also of polycrystalline nature.

Brunauer–Emmett–Teller Analysis

Brunauer–Emmett–Teller (BET) surface area analysis[74] was performed to study the surface area and porosity of the samples. The specific surface area of TiO2 and M-TiO2 was calculated using a BET plot and the surface area of M-TiO2 was found to be 45.6454 m2/g, and that of TiO2 was found to be 30.684 m2/g. This result indicates that surface modification has slightly improved the surface area of TiO2, which is in contrast to the common trend that modification with silane moieties significantly reduces the surface area of TiO2.[75]

The adsorption–desorption isotherm of TiO2 and M-TiO2 were plotted (Figure S7). TiO2 shows a type II isotherm, typical of a nonporous or microporous adsorbent. The arrow point, the beginning of the almost linear middle section of the isotherm, is often taken to indicate the stage at which monolayer coverage is complete and multilayer adsorption is about to begin. From this data, the pore volume and pore diameter of TiO2 was estimated to be 0.040603 cm3/g and 5.2931 nm, respectively. M-TiO2, however, belongs to a type IV classification, featuring a hysteresis loop, associated with capillary condensation in the mesopores, and a limited nitrogen uptake over a range of high (P/Po). The initial part of the isotherm is attributed to a monolayer coverage. The pore volume and pore diameter of M-TiO2 was found to be 0.124887 cm3/g and 9.53764 nm, respectively. These results confirm that both the surface area and the porosity of TiO2 have increased upon surface modification.

Energy-Dispersive X-ray Spectroscopy

Energy-dispersive X-ray spectroscopy (EDX) spectra of TiO2 and M-TiO2 were taken to analyze their elemental compositions (Figure S8). The data revealed that TiO2 consists of 61.47% Ti and 27.85% O, whereas M-TiO2 is composed of 62.29% Ti and 33.22% O. A noteworthy observation from the EDX spectrum of M-TiO2 is that it contains 4.49% of Si; this, when taken together with data presented vide supra, confirms that the silane groups are functionalized onto the TiO2 surface.

X-ray Photoelectron Spectroscopy

XPS is a commonly used technique[76] to determine the valence band maximum (VBM) in a semiconductor (Figure S9). The VBM of the prepared M-TiO2 sample was determined by a linear extrapolation method and found to be 0.88 eV. Combining the band gap value obtained from the Tauc plot (vide infra) with this result, the conduction band minimum[77] can be estimated to be around −2.05 eV. The XPS spectrum (Figure A) of the M-TiO2 sample showed characteristic peaks corresponding to Si 2p, C 1s, Ti 2p, and O 1s at 102, 285, 458, and 530 eV respectively. The fitted curve for the same gives an idea about the mass concentrations of these elements. Mass concentrations are as follows: Si 2p—14.96%, C 1s—41.42%, Ti 2p—39.82%, and O 1s—3.8%. Figure B demonstrates the core level spectrum[78] of Si 2p,[79] which can be resolved into contributions of silicone/siloxane (94%) and SiO2 (6%). In Figure C, a core level spectrum of C 1s of the M-TiO2 sample reveals a peak pattern at 285 eV, which can be further resolved into two peaks: one at 284.6 eV, due to sp2-hybridized carbon atoms (82.49%), and another one at 285.8 eV, due to carbons that are bound to an −OH group (17.51%) present in the composite.[80]Figure D illustrates the core level spectrum of Ti 2p (458 eV) with peaks resolved at 458.65 and 464.25 eV, corresponding to Ti 2p3/2 (66.49%) and Ti 2p1/2 (33.51%), respectively. Figure E illustrates the core level spectrum of O 1s, wherein, peak fitting resulted in two peaks, the first one at 529.85 eV, which can be assigned to reticular O in the oxide (78.01%), and a second one at 531.95 eV, which can be attributed to oxygen in the surface water (21.99%) molecules.

Figure 5

X-ray photoelectron spectroscopy (XPS) spectrum of M-TiO2. (A) Wide spectrum, Core level spectra of (B) Si 2p, (C) C 1s, (D) Ti 2p, and (E) O 1s.

UV–Visible Spectroscopy and Band Gap Determination

UV–visible spectroscopy was carried out to compare the absorbance spectra of TiO2, I-TiO2, and M-TiO2 and is shown in Figure A. The absorption spectrum of M-TiO2 is significantly redshifted than that of TiO2 and I-TiO2, indicating a larger fraction of the rutile phase in M-TiO2 and an appreciable reduction of the bandwidth of the M-TiO2. The optical band gap of thin films of TiO2, I-TiO2, and M-TiO2 was calculated using Tauc’s equation. Figure B shows the Tauc plot obtained from the optical absorption spectra. Tauc plot reveals a band gap of 3.27 eV for TiO2, 3.15 eV for I-TiO2, and 2.93 eV for M-TiO2, confirming that the band gap decreases upon surface modification of TIO2; an effect that can be due to the presence of the rutile phase in M-TiO2.[81] The conduction band of the rutile phase of TiO2 lies higher than that of the anatase phase. A hierarchical arrangement of the conduction bands could render an effective transfer of the photo-generated electrons of the exciton pair from the rutile phase to the lattice trapping sites of the anatase phase. This could inhibit the electron–hole recombination in the rutile phase and result in the effective separation of the photogenerated electron–hole pair. In photocatalytic applications, this ensures that the electrons reach the pollutant surface, and makes the photocatalysis more effective (cf.Figure S10). In the case of DSSC applications (Figure S11), effective separation of the electron–hole pair among the rutile and anatase phases could facilitate easy transfer of the electrons from the anatase phase to a transparent conducting electrode and boost the efficiency of the device.[82] UV–visible spectroscopy was carried out to compare the absorbance spectra of samples with different relative ratios weight percentages of trimethoxy silane with respect to TiO2viz. 0, 1:1, 1:2, 1:4, and 1:20. Figure S12 shows the absorption spectra and corresponding Tauc plots of all the combinations. Figure S12A shows the absorption spectra of all the samples. The absorption spectrum of the 1:2 sample (M-TiO2) is significantly redshifted than that of other samples. The optical band gaps of the samples were calculated using Tauc’s equation. Figure S12B shows the Tauc plot obtained from the optical absorption spectra. Tauc plots revealed band gaps of 3.27, 3.11, 2.93, 3.10, and 3.20 eV for 0, 1:1, 1:2, 1:4, 1:20 samples, respectively. The 1:2 sample (M-TiO2) showed the narrowest band gap and has the maximum coefficient of absorption. These data unequivocally illustrate that surface modification has caused a redshift in the absorption wavelength till the extent of optimum modification, after which, a blueshift in the absorption maximum was observed.

Figure 6

(A) Absorption spectra and (B) Tauc plot comparison of TiO2, I-TiO2, and M-TiO2.

Photoluminescence Spectroscopy

Photoluminescence spectroscopy (PL) gives a hint about the recombination of photo-generated excitons, where the intensity of PL spectra reveals the recombination rate; indicating that the low intensity of the PL represents a low recombination rate of photo-generated carriers. The presence of defect-related excitation peaks can be analyzed effectively by observing the PL spectra. PL spectra of TiO2, I-TiO2, and M-TiO2 were obtained by exciting the samples at a wavelength of 320 nm, and are shown in Figure S13. M-TiO2 exhibits a weaker PL emission intensity than that of I-TiO2 and TiO2 indicating that exciton recombination is more retarded in M-TiO2 than in others. Nonradiative recombination mostly occurs on the grain boundary; double sintering done during the M-TiO2 sample preparation stage effectively removes these boundaries and nonradiative defects, thus retarding the nonradiative recombination processes, which makes M-TiO2 a better candidate for photocatalytic and photovoltaic applications.[70] In addition to this, different band alignments at the interface between anatase and rutile phases in M-TiO2 also separate the excitons from recombination. It can also be reasonably envisaged that surface modification with optimum silane concentration could decrease the number of oxygen defects. The low intensity of PL emission is attributed to the hopping of electrons through small neighboring defect sites until it finds a recombination center. PL spectra of samples with different ratios of weight percentages of trimethoxy silane and TiO2 (0,1:1,1:2, 1:4, and 1:20) were compared and shown in Figure S14. When silane concentration increases, PL emission was found to decrease. However, after a point, when the concentration continues to increase, PL also concomitantly increases. Reduction in recombination sites upon modification with silane reduces the PL intensity but once the optimum silane concentration has passed, with increasing concentration of silane, silane moieties hinder the charge carriers, causing faster recombination. Among the 5 graphs, sample 1:2 shows the least value for the PL emission. PL spectra were composed of mainly two peaks, one at ∼420 nm, indicating band-edge free exciton luminescence, and the other at ∼470 nm, indicating bound exciton luminescence caused by the defect of the oxygen hole on the sample surface.[84,85] The ratio of the intensities of the peaks at 470 to 420 nm is shown in Table. S1. The ratio tends to decrease as the silane concentration increases up to an optimum level and then tends to increase. The least ratio was deduced for sample 1:2 which indicates a greater number of band-edge free excitons in sample 1:2 and, hence, better photocatalytic activity.

Time-Correlated Single-Photon Counting

Time-correlated single-photon counting (TCSPC) is a well-known technique for fluorescence lifetime measurements.[86] TCSPC detects single photons and measures their arrival time with respect to a reference signal, usually, the light source used. Using TCSPC, information about energy transfer, electron/hole transfer, indirect information about trap states (both surface traps and intrinsic traps), and a plethora of other properties of electronics excited states can be deduced. TCSPC measurement of samples with different weight percentages of trimethoxy silane (TiO2/trimethoxy silane as 0,1:1,1:2, 1:4, and 1:20), and the same samples mixed with the MB dye was compared to obtain a clear picture about the charge carrier transport inside the material (Figure S15). Sample 1:2 demonstrated the maximum lifetime while sample 0 got the least value. Factors such as the number of trap sites and the band structure of the material can influence the charge carrier lifetime of the material. Vacancies tend to be more in the unmodified form of TiO2 because the carriers trapped in a defect center cannot directly jump to the valence band. Electron hopping from one defect site to another becomes difficult because the electron finds a number of radiative centers along its path. However, in modified TiO2, the electron can travel a long way before finding a radiative center because the electrons are relaxed for a longer period.

Modified TiO2 consists of an anatase-rutile mixed-phase, wherein band bending at the interface between the anatase and the rutile occurs due to the conduction band edge difference (∼0.2 eV) between the two phases. The electron transfer occurs from the conduction band of anatase to the conduction band of rutile, and thus, due to band bending, the charge carriers are effectively separated. Charge carriers will have a long life when the anatase phase and rutile phase are in contact with each other. Apart from this, the nonradiative centers at the interface also act as shallow trap centers preventing the direct transition of electrons from the conduction band of anatase to that of rutile, and therefore, the electrons take a longer time to move from the anatase conduction band to a rutile conduction band.

Cross-Cut Adhesion Test

An adhesion test of the M-TiO2 sheet was done using a cross-cut test.[87] The result observed is shown in Figure S16. The cut area of the test coating was examined by visual comparison with the illustrations in the standard (ISO 2409), depending on the amount of flaked coating. Edges of the cuts were smooth with a detachment of less than 5% flakes from the coating, and the coating was classified as 1 according to the standard.

Evaluation of Photocatalytic Activity for the Use in Wastewater Treatment

UV–Visible Spectroscopy

To explore the photocatalytic activity of synthesized M-TiO2, photocatalytic degradation of MB was attempted because MB is a notable water pollutant, carcinogenic and mutagenic, and is a well-established dye used in numerous photocatalytic studies. The standard photocatalytic mechanism[72] involves the formation of electron–hole pairs in photoexcited TiO2; the holes oxidize the surface hydroxide ions or water molecules of TiO2 into hydroxyl radicals (OH•), and electrons react with oxygen to form superoxide radical anions (O2•–) which, if abstract protons from the aqueous media, may form hydroperoxyl radicals (HOO–). The photoexcited MB, on the other hand, may inject electrons into the conduction band of TiO2 to form a radical cation, which reacts with OH• to form the organic degradation products. Figure S17 showcase the absorption spectra of TiO2, I-TiO2, and M-TiO2 taken at different time intervals. The rate of degradation was recorded with respect to the changes in the intensity of absorption peak in the visible region where monomeric MB exhibits the (0–0) and (0–1) vibronic features at ∼664 and ∼610 nm, respectively.[89] It can be visualized that as irradiation time increases, the intensity of the absorption peaks of monomeric MB decreases. Figure A shows the graph of concentration reduction of MB taken in an aqueous solution containing bare TiO2 and M-TiO2 without any photocatalyst as a function of time (labeled as blank). The inset of Figure A(a,b) shows the photographs of MB solution in the presence of the M-TiO2 catalyst before and after degradation (after 50 min of irradiation). M-TiO2 showed 99.98% of degradation after 50 min of irradiation, whereas TiO2 showed only 68.01% degradation after 50 min of irradiation. It should be mentioned that I-TiO2 showed an intermediate efficiency of dye-degradation, that is, 81.37% degradation after 50 min of irradiation. The inset of Figure B shows the degraded MB solutions with TiO2 (a) and M-TiO2 (b) after 10 min of starting the irradiation. A comparative study of the photocatalytic activity of TiO2, I-TiO2, and M-TiO2 is shown in Figure S18. (C/C0) versus time graph of TiO2 and M-TiO2 samples is shown in Figure S18A, and the percentage of degradation of TiO2 and M-TiO2 samples with respect to time is shown in Figure S18B. Our photocatalysis results indicate that M-TiO2 is better than bare TiO2 and I-TiO2 in accelerating the degradation of MB. We attribute the better performance of M-TiO2 (color of the MB solution vanishes after 50 min of degradation using M-TiO2 as in Figure S19) to the following reasons: (i) silanization could have made the surface of TiO2 more hydrophobic,[90] which could facilitate better interaction of the surface with nonpolar parts of MB, an effect that is unlikely to occur between MB and the super-hydrophilic surface (due to the decomposing adsorbed organic residues) of bare TiO2, and (ii) the most reactive {001} facets of TiO2 is exposed upon silanization, which possibly could have promoted the rate of photocatalytic degradation of MB. In order for M-TiO2 to be used as a catalyst in practical applications, its reusability after every photocatalytic application event must be checked. Figure A shows that even at the end of the fourth consecutive cycle of photocatalysis, M-TiO2 is capable of degrading 95.44% of MB in an aqueous solution after 50 min of irradiation time; a performance that is still far superior compared to the performances of TiO2 and I-TiO2. A disadvantage of using conventional recyclable photocatalysts in powder form is that after every photocatalytic cycle, the solution needs to be centrifuged to recover the catalyst. Herein, the flexible M-TiO2 sheets dipped into the dye solution can be just lifted off once the photocatalysis is completed with negligible loss of photocatalyst from the sheet; implying that M-TiO2 sheets should be able to retain their performance over a number of catalysis cycles (Figure B).

Figure 7

(A) C/C0vs Time graph of TiO2 and M-TiO2, (B) percentage of degradation of the same after 10 min of irradiation.

Figure 8

(A) C/C0vs time graph of M-TiO2 for four consecutive cycles and (B) pictorial representation of recycling using a flexible M-TiO2 sheet.

Role of the Silane Content on the Performance of the Photocatalyst

For a better understanding of the high photocatalytic dye degradation efficiency of M-TiO2, 5 samples with different silane contents with respect to the TiO2 content (labeled 0:1,1:1,1:2, 1:4, and 1:20) were analyzed, and photocatalytic dye degradation efficiencies were compared. Figure S20 shows the (C/C0) versus time graph of M-TiO2 with different silane contents. Bare TiO2 showed a dye degradation of 68.01% after 50 min of irradiation, which was the lowest among the five. The incorporation of silane increased the dye degradation drastically (96.25%) for the 1:1 sample. 1:2 sample showed a remarkable dye degradation of 99.98%. The 1:4 sample, however, showed a decrease in performance (98.50%) and 1:5 showed a further decrease (97.87%). From the graph, it can be perceived that the presence of silane plays a crucial role in the photocatalytic performance of TiO2. An optimum level of silane weight percentage in the modified photocatalyst favors the dye degradation efficiency because the presence of silane endows better adhesion between each TiO2 particle as well as between the particles and the substrate. This enables a crack-free surface and lowers the number of trap sites and recombination and improves the charge transport. The correct length of the ligand chain connecting two TiO2 particles increases the pathway of electron transport and reduces recombination. When the nonconducting silane content exceeds a limit, it affects the conductivity and acts as a barrier to electron transport, causing a reduction in dye degradation efficiency.

Scavenger Study

A scavenger study[91] was performed to find out the active radicals responsible for the photodegradation activity of TiO2 and M-TiO2 and to elucidate the mechanism of photocatalysis. Scavengers are species that undergo a fast reaction specifically with a radical, generate a stable species that does not interfere with the reaction, erase the effect of radical in the degradation process, and help examine the nature and role of oxidative species involved in the photocatalytic reaction. IPA was used as a hydroxyl radical scavenger, silver nitrate (AgNO3) was applied as an electron scavenger, and EDTA was used as a hole scavenger. Figure S21 shows the effect of the scavengers on the photocatalytic activity of TiO2 and M-TiO2 samples. In the case of the TiO2 sample, within the first 10 min of irradiation, the addition of IPA, AgNO3, and EDTA diminished the photocatalytic activity to 40.21, 58.13, and 68.26% respectively. The apparently greater reduction in the rate of photocatalytic degradation upon adding IPA indicates that •OH radicals are the prime active species responsible for the oxidation of unmodified TiO2, consistent with the previous reports.[92] When the scavenger study was done for M-TiO2, adding AgNO3 resulted in a minimum degradation rate (48.42%) than that of IPA (52.67%) and EDTA (67.06%), indicating that electron is the active species responsible for the catalytic activity.

Modification of TiO2 with silane moieties changed the active species responsible for the photocatalytic activity, presumably due to the mixed anatase/rutile nature of M-TiO2 and due to excess electrons present in modified TiO2.

Evaluation of Photovoltaic Properties for Use in DSSC

The extent of N-719 dye loading on TiO2 and M-TiO2 was evaluated by recording the UV–visible absorption spectrum of both TiO2 and M-TiO2 of the same quantities (scratched-off from the photoanode films) soaked in the dye for the same duration and under identical experimental conditions and are shown in Figure S22. The absorption profiles of both TiO2 + N-719 and M-TiO2 + N-719 show the characteristic peaks of N-719 at ca. 390 and 530 nm. However, the absorption peak of the M-TiO2 + N-719 sample at 530 nm exhibits an enhanced intensity and a slight red-shift (536.5 nm) compared to the TiO2 + N-719 sample, implying an enhanced dye-adsorption on the surface of M-TiO2; in agreement with the FT-IR data.

Photochronoamperometry

The photocurrent densities of TiO2 and M-TiO2 were compared[93] using photochronoamperometry measurements carried out at a constant voltage of 0.5 V, and under light and dark conditions, and are illustrated in Figure . From the graph, it can be observed that when illuminated with sunlight, the M-TiO2 + N-719 sample has shown more current density than the TiO2 + N-719 sample. In dark conditions, both samples give the same magnitudes of current densities.

Figure 9

Photo chronoamperometry curve comparison of N719 dye-adsorbed TiO2 and M-TiO2.

Electrochemical Impedance Spectroscopy

EIS is a powerful tool to examine the characteristic properties of the electrode material such as resistance, impedance, and charge transfer.[94] For EIS experiments, the electrodes were immersed in (I–/I3–) electrolyte–acetonitrile mixture for 1 h, and the measurements were taken under visible light. Figure S23 shows the EIS spectra of electrodes of TiO2 and M-TiO2. We attribute the enhancement in the photovoltaic properties of M-TiO2 to the enhanced hydrophobic nature of the nanoparticle surface rendered by the silanization process that offers better interaction with the nonpolar parts of the dye and retarded back electron transfer and charge recombination processes at semiconductor/electrolyte interfaces. EIS spectra of DSSCs with TiO2 and M-TiO2 as photoanodes, N719 dye as the photosensitizer, and I3–/I– redox couple as the electrolyte were compared. To have an insight into the resistance developed at the interfaces of different components of the DSSCs, impedance curves were analyzed, the impedance data were plotted as a Nyquist diagram (Figure S24), and the fit and simulation processes were achieved using the Z-fit tool provided in the EIS spectroscopy package. From the fitted curves of Nyquist plots, and the equivalent circuit, values of resistances and capacitances formed at the interfaces of DSSCs were determined. From the values of Rct and C obtained from the EIS spectra, recombination lifetimes can be calculated (Figure S25). Table S2 shows the values of Rs, RCE, Rct, CCE, and C of the DSSCs employing TiO2 and M-TiO2 as photoanodes.

The electron recombination in the photoanode–electrolyte interface is dependent on the recombination lifetime (τoc) and is calculated by eq .Where Rct is the charge transfer resistance and C is the capacitance of the photoanode–electrolyte interface. From the data, it was estimated that the recombination lifetime of the device containing M-TiO2 (6.7 ms) is greater than that of the device containing TiO2 (5.02 ms). This proves that the carriers recombine 1.33 times faster in DSSC with TiO2 than in that with M-TiO2.

I–V Measurements

In order to illustrate the applicability of the prepared M-TiO2 in photovoltaics, DSSCs with combinations of two different dyes, and electrolytes were fabricated using M-TiO2 as the anode and platinum as the counter electrode, and the I–V characteristics of four types of DSSCs were compared viz. DSSC with N749 dye–I3–/I– electrolyte (N749–I3–/I–), N749 dye–Co(II)/Co(III) electrolyte (N749–Co(II)/Co(III)), N719 dye–I3–/I– electrolyte (N719–I3–/I–), and N719–Co(II)/Co(III) electrolyte (N719–Co(II)/Co(III) under AM 1.5 light conditions (Figure A). The input power of the light source was calibrated using a polycrystalline silicon reference solar cell. The device was exposed to direct sunlight with a power of 536 W/m2 with a latitude of 9.669090, a longitude of 76.539070, and an elevation of 21 m. Under these conditions, the J–V characteristics of the device were measured using a potentiostat SP-200 (Biologic) supported by EC-Lab software and is shown in Figure A. The results are summarized in Table .

Figure 10

(A) IV characteristics of DSSCs measured with different electrolyte and dye combinations under AM 1.5 light conditions (B) I–V measurement of the best-performing device (N719–I3–/I–) measured under direct sunlight conditions (with a device structure of ITO-N719 dye-adsorbed TiO2 photoanode–I3–/I– electrolyte–Pt counter electrode).

Table 2

I–V Measurement Results Comparison of DSSCs

DSSC	Jsc (mA/cm2)	fill-factor	Voc (V)	η (%)
N749–I3–/I–	4.4	0.31	0.50	0.7
N749–Co(II)/Co(III)	0.5	0.67	0.82	0.28
N719–I3–/I–	3.0	0.58	0.63	1.1
N719–Co(II)/Co(III)	0.8	0.46	1.02	0.41

It is evident that N719–I3–/I– DSSC exhibits the highest efficiency (average value of several measurements) among all the tested devices (Figure B) under direct sunlight, and a short-circuit photocurrent density of 5.6 mA/cm2, an open-circuit voltage of 0.68 V and a fill factor of 0.58 was obtained, and the photoconversion efficiency was found to be 4.1%. The value of photoconversion efficiency is less because of the low-lying conduction band of the ITO sheet used for the DSSC fabrication. A judicious selection of components of DSSC can improve efficiency. Sewvandi et al. have reported[65] high-performance DSSCs made of the N719 dye and TiO2 electrodes having their surface modified with hydrocarbon and fluorocarbon silanes of varying chain lengths. They attributed the high efficiency of the DSSCs to the formation of a monolayer of the silane on the TiO2 surface that prevents back electron transfer from TiO2 to the I3– species; thereby, making the injected electrons available to perform the electrical work in the device. Our results indicate that surface modification of TiO2 with silanes increases its photocurrent density and enhances the charge-separation efficiency when being employed in a photoanode scheme, which underlines the scope of using M-TiO2 type materials to boost the efficiency of DSSCs.

Conclusions

Herein, we describe the synthesis, characterization, photophysical and electrochemical studies of TiO2 nanoparticles surface modified with trimethyloxy silane functional groups. The prepared M-TiO2 particles contain both rutile and anatase phases, with the {001} facets of them exposed by silanization. Flexible sheets of M-TiO2 are prepared on a plastic substrate with a simple, scalable, and low-cost method and are used in a photocatalytic scheme and as a photoanode for DSSC applications. Based on the photocatalytic experiments, M-TiO2 sheets outperform sheets made of unmodified TiO2 nanoparticles in degrading the MB dye. Also, compared to unmodified TiO2, M-TiO2 sheets exhibited greater magnitudes of photocurrent when sensitized with the N-719 dye, and DSSC fabricated using M-TiO2 as the photoanode exhibited a photoconversion efficiency of 4.1% under direct sunlight. These two effects could be attributed to (i) better interaction between the hydrophobic surface of M-TiO2 with the nonpolar parts of the structure of the dye, facilitating better charge transport between the duo; (ii) co-existence of both rutile and anatase phases of TiO2 and photo-induced electron transfer between them; (iii) formation of silane monolayer on the surface of TiO2 which retards the back electron transfer from TiO2 to the redox electrolyte or dye; and (iv) the formation of exposed {001} facets of TiO2 upon surface modification, which confine the photo-generated excitons in the case of anatase TiO2, and thus are endowed with a pronounced photo-electrochemical activity.