Recent progress in one dimensional TiO2 nanomaterials as photoanodes in dye-sensitized solar cells

Exploiting the vast possibilities of crystal and electronic structural modifications in TiO2 based nanomaterials creatively attracted the scientific community to various energy applications. A dye sensitised solar cell, which converts photons into electricity, is considered a viable solution for the generation of electricity. TiO2 nanomaterials were well accepted as photoanode materials in dye-sensitized solar cells, and possess non-toxicity, high surface area, high electron transport rates, fine tuneable band gap, high resistance to photo corrosion and optimum pore size for better diffusion of dye and electrolyte. This review focuses on various aspects of TiO2 nanomaterials as photoanodes in dye-sensitized solar cells. TiO2 photoanode modification via doping and morphological variations were discussed in detail. The impact of various morphologies on the design of TiO2 photoanodes was particularly stressed.


Introduction
The excessive consumption of fossil fuels to meet the increasing energy demand has resulted in profound consequences such as global warming, environmental pollution, etc. To address the challenges raised by environmental degradation and the energy crisis, the development of a sustainable energy production approach is required. Even though various renewable energy resources such as geothermal energy, biofuels, and wind-tidal energies can substantially contribute to sustainable energy development, solar energy has to play a major role. In this era of growing energy demand and environmental issues, solar energy is the only choice le in front of us as solar radiation possesses immense potential, vast abundance and environmental friendliness. Out of the 4 million exajoules (10 18 J) of solar energy reaching the earth, around 50 000 EJ's are easily exploitable. 1 The signicance of solar energy becomes more clear from the fact that the world's annual energy consumption Mr Deepak Joshy is presently working as a Senior Research Fellow under the CSIR-UGC JRF scheme in the Department of Chemistry, University of Calicut under the guidance of Dr Pradeepan Periyat. His research mainly focuses on the design and development of nanomaterials for energy saving and environmental applications. His major research areas are photocatalysts and adsorbents for water purication, inorganic pigments for cool coatings and supercapacitors.
for the year 2017 was around 565 EJ's which is only a fraction of the harvestable solar energy (i.e. 50 000 EJ's). 2 But in the present scenario, electricity produced from the renewable energy sector constitutes only 8.4% of the total world electricity production. 2 More than 20% of the contribution to the renewable energy sector is made by solar energy. 3 But this share of solar energy is signicantly small compared to the magnitude of solar power available for harvesting. The lack of efficient and economically viable solar harvesting techniques is the main reason for this huge difference. Solar cells are one of the widely employed solar energy conversion devices. Solar cells are classied into different generations based on the material and technology present in them. We have rst generation silicon solar cells as the most widely commercialized solar cell type. 4,5 Silicon solar cells have efficiencies of around 26% which is almost near the theoretical efficiency maximum. 6 At the same time, secondgeneration solar cells are based on thin-lm technology and employ materials such as CdTe, CIGS, etc. [7][8][9][10] The efficiency of second-generation solar cells has reached up to 21.7% recently. 6 But the large-scale commercialization of rst and secondgeneration solar cells is very expensive. Also, the components used in these two generations are capable of causing environmental hazards. So, to develop an economically viable and ecofriendly way of solar energy conversion, the third generation of solar cells, known as dye-sensitized solar cells (DSSCs) was introduced. 11-14 The concept of a DSSC was introduced by Michael Gratzel in 1991. 15 Signicant attraction garnered by DSSCs is due to their large-scale and economically viable Then he worked as a CRI researcher at the Center for Bio production capability along with exibility. The efficiencies of DSSCs introduced so far are found to be within 14.3%. 16 If the issues associated with low efficiency and shorter durability are addressed properly, DSSCs are the best to play a vital role in the development of a sustainable energy culture. Also, large-scale commercialization of DSSCs can convert a major share of the available solar radiation. Performance enhancement, durability increment and production cost reduction of DSSCs can be carried out in various ways. Mainly it involves either the synthesis of novel component materials/modication of existing materials or the introduction of more rational DSSC designs. As a DSSC assembly consists of various components such as a photoanode, sensitizer(dye), counter electrode, redox electrolyte, etc., investigation into performance enhancement can be carried out on any one of these components. In this review, special emphasis is given to the DSSC photoanode and the contribution of one dimensional TiO 2 nanomaterial based photoanodes for the critical development of DSSCs.

DSSC structure & working-an overview
A photoanode forms the core part of the DSSC and it consists of a transparent conducting oxide (TCO) glass plate or plastic substrate on which nanometre-sized semiconducting metal oxide particles are deposited and sintered. Fluorine-doped tin oxide (FTO) is the most widely employed substrate. 17,18 Usually, a mesoporous TiO 2 lm having a thickness of around 10 mm is coated on the FTO substrate to facilitate electron transfer. 19 Sensitizer or dye molecules responsible for light absorption are adsorbed onto the mesoporous nanomaterial lm. Upon irradiation by solar radiation, the dye molecule will get excited and emit an electron. This emitted electron will get injected into the conduction band of the nanocrystalline semiconductor metal oxide lm. From there the electron will get transferred to the FTO plate and then to the external circuit. The dye molecule which got oxidised on solar irradiation is now regenerated by employing a redox couple. 20,21 The iodide/triiodide redox system is the most widely used one. 22 Upon reduction of the oxidised dye molecule, the iodide ion gets oxidised into triiodide. The regeneration of the redox couple is facilitated by electrons which reach the counter electrode through the external circuit. The counter electrode consists of a thin layer of Pt nanoparticles deposited on a conducting glass substrate. 23,24 Here the electron emitted from the dye molecule travels through the mesoporous nanomaterial layer deposited on the FTO plate and then ows through the external load and nally reaches the counter electrode to regenerate the redox system. The redox electrolyte system present in between the two electrodes not only facilitates dye regeneration but also prevents the recombination of conduction band electrons with oxidised dye molecules. While fabricating the DSSC, as shown in Fig. 1 the photoanode and cathode are sandwiched together with a layer of the polymer lm in between them. 25 Then the redox electrolyte system is injected into the space in between so that the triiodide ions can diffuse towards the counter electrode and get reduced to iodide ions. A number of such repeating cycles produce current in the circuit. The resultant efficiency of a DSSC depends on the electron recombination rate. The lower the electron recombination rate, the higher the efficiency will be. The lower electron recombination rate results from the high electron transport rate. 26

Important terms associated with DSSCs
The important parameters associated with dye sensitised solar cells are incident photon-to-current efficiency (IPCE), open circuit voltage (V OC ), short circuit current (J sc ), ll factor (FF) and cell efficiency.
1.2.1 Short circuit photocurrent (J sc ). Fig. 2a shows the current versus voltage curve associated with a solar cell circuit in the dark and under illumination. When the electrons travel from the anode to the cathode of the solar cell, and on the other hand when the circuit is reverse biased, the current ows will be zero. This is due to the hindrance caused by the high energy barrier of the donor. Aer the energy barrier is crossed, a very low current will ow. This feature is represented by the dark curve. When the solar radiation is absorbed by the donor, charge carriers will be generated easily. There will be a reverse current due to the electron ow from the anode to cathode. This reverse current which has resulted in the absence of an external voltage is termed as photocurrent or short circuit photocurrent (J sc ). This is represented by the hollow bubbled curve.

Open circuit voltage (V OC ).
Since the short circuit current is a result of the reverse bias current, it is possible to compensate the current by applying a forward voltage. Under this condition, there will be a point where the current becomes zero. The voltage corresponding to this point is referred to as the open circuit voltage (Voc). On the other hand, the open circuit voltage is the maximum voltage available from a solar cell when there is no external load connected and the external current owing through the cell is zero. 27 1.2.3 Fill factor (FF). The solar cell ll factor (FF) gives us an idea of the performance of the solar cell. 28 It corresponds to the ratio of the actual performance to the theoretical maximum power output of a solar cell.
A higher ll factor favours the maximum output of a solar cell. The graphical representation of the ll factor is shown in Fig. 2b. This graph shows the solar cell output current and power as a function of voltage. It is the ratio of the area of the larger rectangle (pale green) to the area of the smaller rectangle (yellow).
1.2.4 Incident photon-to-current efficiency (IPCE). The incident photon-to-current efficiency (IPCE) is a measure of the ratio of the photocurrent (converted to an electron transfer rate) vs. the rate of incident photons (converted from the calibrated power of a light source) as a function of wavelength. 29 IPCE is also known as quantum efficiency which measures the efficiency of a device to convert incident photons into electrical energy at a given wavelength.
1.2.5 Solar cell efficiency. The efficiency of a solar cell is dened as the fraction of incident solar power which is converted to electricity. Solar cell efficiency depends on the spectrum, intensity of incident sunlight and the temperature of the solar cell. Therefore, conditions under which efficiency is measured must be carefully controlled to compare the performance of one device to that of another one. 30 Terrestrial solar cells are measured under AM1.5 conditions and at a temperature of 25 C. The maximum output power is given by Then efficiency is given by the equation, is the open circuit voltage, J SC is the short circuit current and FF is the ll factor. 11

Role of the photoanode in a DSSC
Among the functional components of a DSSC, the photoanode forms the crucial part. The nature of the material used as the photoanode and its morphology are important factors determining the overall performance of a DSSC. The functions of a photoanode include dye pickup, electron injection, transportation and collection which in turn inuence the photocurrent, photovoltage and power conversion efficiency. Efforts are being made to improve the efficiency of DSSCs by introducing novel materials as photoanodes and by modifying the morphologies of existing materials. The essential requirements for an ideal photoanode are: (i) Higher surface area facilitates the adsorption of dye molecules to a greater extent. 31,32 (ii) Photoanodes should have higher electron transport rates so that the electron injection from the dye to the external circuit through the photoanode occurs smoothly. 33 (iii) Photoanodes should have suitable band gap alignment with the energy levels of the sensitizer. 34 (iv) It should possess high resistance to photo corrosion. 35 (v) Photoanode materials should possess a pore size that can be optimized to achieve better diffusion of dye and electrolyte. 36 (vi) It should possess the ability to absorb/scatter sunlight for the improved performance of the dye. 37  (vii) The photoanode material needs to be in optimum contact with the dye molecules and the conducting substrate. 38 The above-mentioned characteristics are vital in achieving a better photoconversion efficiency.
This review discusses the signicance of TiO 2 nanomaterials as photoanodes in DSSCs. The characteristic photovoltaic properties and important modications of TiO 2 for photoanode applications are surveyed. A summary of research on TiO 2 and its important one-dimensional morphologies and their modi-cations to be used as photoanode materials in DSSCs is given here.

Photoanode materials
The limitations associated with the conventional materials lead to the investigation of more effective photoanode materials, which have advanced properties with bulk and surface modications. Commonly investigated materials include metal oxides such as TiO 2 , ZnO, [39][40][41][42] [54][55][56] All these are wide band gap semiconductor materials whose structure, morphology and crystallinity decide the performance of the DSSCs. The band structures of these materials are shown in Fig. 3. 57 Among these materials, TiO 2 and ZnO are the most widely used materials for DSSC fabrication. [58][59][60][61][62] Although ZnO has better electron mobility compared to TiO 2 , its efficiency is less than those DSSCs employing TiO 2 . This lower efficiency of ZnO emerges due to decreased dye adsorption and instability in acidic environments. 63 Therefore, TiO 2 has superior photovoltaic applicability compared to ZnO. Four polymorphs of TiO 2 are known, viz, as rutile (tetragonal), anatase (tetragonal), brookite (orthorhombic) and TiO 2 (B) (monoclinic) as shown in Fig. 4. [64][65][66] Among these polymorphs of TiO 2 , anatase is preferred over rutile for photovoltaic applications irrespective of rutile's greater stability and lower band gap. This is because the anatase phase has a higher conduction band energy level, absorptive affinity and a lower electron-hole recombination rate. 67 Since the synthesis of brookite TiO 2 is difficult, its applicability as a photoanode remains less explored. 68 DSSCs employing anatase TiO 2 showed efficiencies ranging from 12-14%. 69,70 So TiO 2 is considered the best choice available as a photoanode material due to its (a) cost-effectiveness, (b) easy availability, (c) good stability along with non-toxicity and (d) suitable optical and electronic characteristics. 67 Furthermore, most of the stable sensitizers showing higher light absorption capability have their LUMO positioned favourably with the conduction band of TiO 2 . These favourable qualities of TiO 2 initiated further investigations to improve the functioning of TiO 2 photoanodes. During this process of improvement, the main challenges to deal with include the (i) large band gap (3.2 eV) of TiO 2 causing adsorption in the UV region 71 and (ii) low internal electron transport rate. 72 Furthermore, research aimed at better performance of TiO 2 by focusing on factors such as high surface area, an increased light scattering effect, enhanced interface quality, fast electron mobility and better charge collection ability. The physical, chemical and optical properties of TiO 2 depend not only on its intrinsic electronic structure but also on its shape, size, porosity, pore size distribution, organization and surface features. One approach to increase the photoconversion efficiency is by maximising the surface area of TiO 2 and thereby enhancing the reaction at the interface of the photoanode and interacting media. The extent of dye adsorption depends on the surface area available. The greater the dye pickup, the more the electron/current density that will be generated. Semiconductor mesoporous TiO 2 (Fig. 5), nanorods ( Fig. 6a and b), nanowires ( Fig. 6c and d), nanotubes (Fig. 6e), nanosheets and various other nanoarchitectures have been employed and explored for enhanced dye adsorption. [73][74][75][76] Besides increasing the surface area, enhancing electron mobility is also a crucial factor in improving the performance of the photoanode. Defects in TiO 2 act as electron traps at the grain boundaries and the absence of defects will assure better collection of injected electrons from the semiconductor. Another method involves surface modication of TiO 2 semiconductors which has remarkable inuence on charge separation, electron mobility and the recombination process. 77 Attempts are being made to minimise electron recombination losses due to grain boundaries and also to extend the absorption of TiO 2 towards the near-infrared region.

TiO 2 photoanode modication by doping
Electronic properties of TiO 2 can be effectively modied by doping i.e. by the deliberate insertion of impurities into the TiO 2 lattice. 78 Doping in the TiO 2 lattice results in an increase in free charge carriers and conductivity. 79 This is due to the defectridden nature of TiO 2 and it inuences the electronic structure and trap states present in the lattice. 80 While doping, either the Ti 4+ cation or O 2− anion can be replaced. TiO 2 has a band structure consisting of conduction band (CB) energy levels formed by the Ti 4+ orbitals and valence band (VB) energy levels comprised of O 2− 2p orbitals. 67 Thus replacing Ti 4+ and O 2− ions will alter the CB and VB structures respectively. 81 The atomic radii of the dopants should be comparable to the ions to be replaced. 82 The sensitizer dye molecules used to bind with Ti atoms in the TiO 2 lattice and the replacement of Ti with other atoms will affect the dye adsorption due to different binding strengths between the dye and the dopant. 83,84 It was also observed that the growth rate of TiO 2 nanoparticles is inhibited by dopants resulting in smaller particles having increased surface area. 85,86 As a result of an increase in surface area, dye adsorption and current densities are improved. 87 This will automatically enhance light absorption and facilitate the use of thinner lms having a lower recombination rate in DSSCs. 32 Various morphologies of TiO 2 have a major inuence on its optical and electronic properties. 88 It is found that one-dimensional nanostructures such as nanowires, nanorods and nanotubes possess better charge transport properties compared to nanoparticle assemblies. 89 However, these 1D structures have lesser surface area than nanoparticle systems and thereby have reduced dye pick-up capability. 90 Dopants having better dye adsorption capacities can improve the photovoltaic performance of 1D nanostructures. 91,92 At the same time, nanoparticle assemblies benet from dopants that cause increased charge transfer. 93 Hence, it is difficult to identify clearly whether performance improvement is caused by an increase in absorption or electronic effects in doped TiO 2 .
Based on the general electronic conguration, dopants can be classied into alkaline earth metals, 94-96 metalloids, 97-99 nonmetals, 100-102 transition metals, 103 post-transition metals 104,105 and lanthanides. [106][107][108] Signicant attempts were made for codoping in which more than one dopant is introduced into the lattice for enhanced device performance. [109][110][111] Doping in a TiO 2 lattice alters its phase, at band potential, recombination rate, electron transport rate and dye adsorption capability. 79 Anatase to rutile phase transformation is inhibited by doping, 112 thereby reducing charge carrier recombination. The at band potential will experience either a positive or a negative shi i.e. a positive shi involves the downward shi of the CB and Fermi level (E F ) whereas a negative shi involves the upward shi of the CB and E F . A decrease in the number of defect states upon doping increases the lifetime of photogenerated electrons and reduces recombination losses. 113 At the same time, a decrease in the number of trap states may lead to enhanced electron mobility. 114 Dopants inuence the growth rate of nanoparticles, thereby affecting their size, surface area and the number of grain boundaries. Also, dopants decide the binding strength between the doped surface and dye molecules. 115 Recently, Mndoped TiO 2 with IR absorbing capability was reported to be an excellent photoanode for dye-sensitized solar cells and it possesses 79% higher efficiency compared to commercial P25. 116

Nanostructured TiO 2 photoanode
A revolutionary breakthrough in the eld of photoanode fabrication with the introduction of a mesoporous TiO 2 nanoparticle photoanode was pioneered by Gratzel and co-workers in 1991. 117 They replaced the bulk TiO 2 photoelectrode with a nanostructured architecture and obtained an efficiency of 7.1-7.9% by employing a trimeric Ru complex. 118 Detailed guidelines for the fabrication of a TiO 2 nanoelectrode for DSSCs with efficiency >10% are given by Gratzel and team. 119 They introduced a double-layered TiO 2 lm comprising of a light-absorbing layer of anatase with 20 nm thickness and a light scattering overlayer of 200-400 nm sized anatase particles. 119 Soon several multilayered TiO 2 photoanodes were introduced to achieve broadband light connement without affecting dye adsorption capacity. 120 Several modications and novel nanoarchitectures of TiO 2 have been introduced to exploit the nanoscale properties for  better DSSC performance. 121 Zero dimensional TiO 2 nanoparticles serve as the growth centres for advanced nanoarchitectures with enhanced performance parameters. When the particle size approaches the nanometre range, the nanostructure band gap increases as a result of quantum size effects and it is possible to adjust the valence and conduction band energy levels of the nanosemiconductor with respect to the redox potentials of the redox couple used.

One-dimensional nanostructures as photoanodes
One-dimensional nanostructures employed in DSSCs consist of nanotubes, nanowires, nanobres, nanobelts and nanoribbons. One-dimensional nanomaterials gained attention through the pioneering work by Iijima. 122 A well-ordered arrangement of one-dimensional nanostructures can provide a direct electron transport pathway from the semiconductor lm to the conducting substrate. This results in a reduction in the electron recombination rate and an increase in PCE. 123

Nanowires
TiO 2 nanowire structures serve as conned conducting channels with their long charge diffusion lengths preventing charge recombination and thus facilitating better charge transport. 66 It is this fast charge transport and better charge collection ability which made them possible candidates for DSSC fabrication. By using a dense array of long, thin nanowires as dye scaffolds, it is possible to increase both dye pickup and carrier collection efficiency. Also, nanowire photoanodes are found to be more suitable for non-standard electrolytes such as solid inorganic phases or polymer gels having higher recombination rates. 123 Transient photocurrent and photovoltage measurements were carried out on TiO 2 nanowires and it was found that the electron transport time and its dependence on illumination intensity are similar to that of TiO 2 nanoparticles. 124 However, the ratio of electron-hole recombination time and electron collection time of TiO 2 nanowire-based DSSCs are about 150 times that of nanoparticle-based solar cells and it indicates the improved collection efficiency of nanowire arrays. Scheme 1 gives a brief idea of various attempts made to improve the key factors affecting the overall performance of TiO 2 nanowire based DSSCs. From the large collection of literature on TiO 2 nanowire based DSSCs, here we are selectively discussing major research outcomes involving efficient synthesis methods for nanowires, novel architectures and signicantly enhanced efficiencies. Many of the nanowire synthesis methods were based on hydrothermal treatment with slight modications.
Feng et al. presented a straightforward low-temperature method to fabricate single-crystalline rutile TiO 2 nanowires through a nonpolar solvent/hydrophilic solid substrate interfacial reaction under hydrothermal conditions. 75 Nanowires having lengths up to 5 mm can be grown vertically from the TCO glass substrate along the preferred direction by this approach. This arrangement has given an efficiency of 5.02% under AM1.5 irradiation by employing N719 dye on 2-3 mm long TiO 2 nanowire arrays. An added advantage of this technique is the low temperature employed, which favours the use of polymers for cell fabrication. Thus, low-temperature methods of photoanode fabrication are found to be compatible with polymer substrates and can result in exibility. As part of developing exible and lightweight DSSCs, Liao et al. introduced hierarchial TiO 2 nanowire (HNW) arrays grown on a Ti foil substrate instead of FTO. 125 These HNW arrays contain long TiO 2 nanowire trunks and short TiO 2 nanorod branches and are prepared by a twostep hydrothermal process. Liao et al. are the rst to report such HNW arrays fabricated on Ti foil and they also replaced the Pt counter electrode with another electrode in which PEDOT is electrodeposited on the ITO-PET substrate. Even though HNW array DSSCs showed increased efficiency (4.32%) compared to that of NW-based DSSCs, the former possesses a comparatively reduced electron lifetime and transport time. As a further step toward realising lightweight and exible DSSCs, cells were constructed with a vertical TiO 2 nanowire array grown in situ on carbon bre substrates as shown in Fig. 7. In 2015, Liu et al. reported the controllable formation of TiO 2 nanowire arrays on a titanium mesh. This method is a hydrothermal one capable of forming nanowire arrays (NWAs) with an average diameter of 80 nm. 127 Along with studying the inuence of NWA preparation conditions on DSSC parameters, Liu et al. also focused on the role of the sensitization temperature and time on DSSC performance. 127 They found that a higher sensitization temperature would benet dye molecule inltration to the internal areas of NWA lms, and the complete covering of monolayer dye molecules on the surface of TiO 2 NWAs would enhance the photovoltaic properties of the DSSC. By maintaining optimum conditions, Liu et al. obtained an efficiency of 3.42% for a exible DSSC. Apart from maintaining optimum dye pickup, they were found to have depressed charge recombination also. Later several groups tried enhancing the available surface area of the photoanode by growing additional structures on TiO 2 nanowires. In 2014, single-crystal-like 3D TiO 2 branched nanowire arrays were fabricated by Sheng et al. with the 1D branch epitaxially grown from the primary trunk. 128 This attempt made by Sheng et al. increased the available surface area by 71% and they also exhibited a fast charge transport property compared to one-dimensional TiO 2 nanowires. Sheng obtained an efficiency of 4.61% for branched nanowire-based DSSCs. 128 It is found that the presence of other nanostructures as branches on nanowires creates additional boundaries in electron transport. Later, DSSCs with multilayered photoanodes, in which different functions were assigned to specic layers, were fabricated. For example, in 2013, Bakshayesh et al. introduced a new TiO 2 structure with corn-like nanowire morphology having high surface area and crystallinity synthesised by the surface tension stress mechanism. 129 They have adopted a double-layer DSSC design consisting of an underlayer of anatase TiO 2 nanoparticles and an overlayer of corn-like TiO 2 nanowires as shown in Fig. 8a and Fig. 8b shows the photocurrent against voltage curves. By adopting a triple function mechanism with effective management of light scattering, dye sensitization and photogeneration of charge carriers, Bakshayesh et al. attained an efficiency of 7.11%. Here the increased surface area of corn-like nanowires resulted in improved dye sensitization and short circuit current density. At the same time, the presence of NPs on corn-like nanowires increased light scattering. Most of the multilayer photoanodes follow such division of labour among various layers. Another new morphology reported in 2013 consists of thornbush like TiO 2 nanowires (TBWs) prepared by a facile single-step hydrothermal method using potassium titanium oxide oxalate dehydrate, diethylene glycol (DEG) and water at 200 C. 130 These TBWs consist of a large number of nanoplates and nanorods. Depending on the change in the DEG/water composition, the diameter, as well as the morphology of TBWs, varies. TBWs having a diameter of 200 nm shows higher efficiency (5.2%) than those having a diameter of 400 nm (4.5%) and 600 nm (3.4%). Further treatment of TBW200 with gra-copolymerdirected, organized mesoporous TiO 2 helps to increase the surface area and interconnectivity of TBWs leading to an enhanced efficiency of 6.7%. As the electron transport in TiO 2 nanowires depends on their length, efforts were made to modify the nanowire length. In 2011, a multicycle hydrothermal synthetic process to produce vertically oriented, single-crystalline rutile TiO 2 nanowires with lengths in between 1 and 8 mm was reported by Zhou and co-workers. 131 It is observed that a further increase in the nanowire length does not give an expected increase in efficiency. This can be attributed to a decrease in surface area by fusion and widening at the base of nanowires.
In the same year, a double-sided brush-shaped (DSBS) TiO 2 nano architecture consisting of highly ordered TiO 2 nanowires aligned around an annealed TiO 2 nanoparticle layer was prepared by C. Zha and team via the hydrothermal method. 132 Here nanowire growth is seeded by the annealed nanoparticle layer and it supports the DSBS structure. It was found that by varying the hydrothermal reaction time, structural properties  like the crystalline phase, phase composition, length of the nanowires and thickness of the nanoparticle layer can be tuned. They have obtained an efficiency of 5.61% from an 8 hour reaction time and nanowires of length 6 mm. Thus, the microstructure has a signicant inuence on solar cell performance and the DSBS structure is a greatly promising one. Later Qiang Wu and co-workers fabricated a DSSC based on a self-assembled, vertically aligned TiO 2 nanowire photoelectrode sensitized with N719 (ref. 133) sensitizer and it could achieve an efficiency of 9.40% (Fig. 9). 134 Here TiO 2 nanowires on an FTO glass plate have a tuneable length in the range of 15-55 mm and are suitable for multi-layered photoanode conguration. 134 136 By controlling the ethanol content in the reaction mixture, the length and separation between the NWs can be controlled. A reaction mixture containing 20 mL ethanol facilitated the formation of TiO 2 NWAs with an efficiency of 8.9%. This can be further enhanced to 9.6-10.2% by incorporating a light scattering layer into the TiO 2 NWA-based DSSC.
Another mode of enhancement of light absorption and efficiency of DSSCs utilizes the anti-reecting (AR) property of photoanode materials. Another approach is by the exploitation of the surface plasmonic effect of metal nanoparticles. Yen et al. fabricated a DSSC utilizing both the antireecting character and plasmonic effect. Their DSSC was equipped with a 3D TNW-AuNP plasmonic electrode having antireective (AR) TiO 2 nanowires (TNWs) serving as light-harvesting antennae coupled with Au nanoparticles (NPs). 137 These plasmonic functionalized electrodes (PFEs) exhibited a remarkable plasmonic red shi from 520 nm to 575 nm. Such PFEs were developed to overcome the narrow absorption range and low absorption coefficient of dyes. It was found that TiCl 4 treatment can increase the efficiency of TNW-AuNP hybrid DSSCs from 6.25% to 9.73%. In the same year, Lee et al. investigated the photovoltaic performances  of back-illuminated DSSCs employing TiO 2 NP/NW composite lms of various weight percentages. 138 These DSSCs used a cobalt-based electrolyte system. In NP/NW composite lms NPs help to increase the surface area whereas NWs facilitate efficient charge transfer. The highest efficiency of 7.37% was shown by DSSCs with 10 wt% of NWs in the NP/NW composite lm which is 20% improved compared to that of a pure NPbased DSSC. Apart from the widely used hydrothermal method of nanowire synthesis, Jin et al. proposed an approach for fabricating TiO 2 nanowire networks on Ti foil using a Ti corrosion reaction in KOH aqueous solutions at different temperatures, followed by a further ion-exchange process. 139 These photoanodes were suitable for bendable DSSCs and have exhibited an efficiency of 1.11% for back illumination. The same team also tried growing TiO 2 nanowire networks on FTO substrates by wet corrosion and obtained an efficiency of 1.0% under AM 1.5 illumination. 140 In 2017 Li and co-workers reported a simple single-step hydrothermal method to prepare single-crystalline self-branched anatase TiO 2 NWs by using TBAH and CTAB as co-surfactants. 141 These single crystalline self-branched TiO 2 NW-based DSSCs exhibited an efficiency of 6.37% which is due to an enhanced percentage of exposed (010) facets having high dye adsorption capacity. Again in 2018, another design was introduced into the class of FDSSCs (ber shaped dye-sensitized solar cells) which consists of a stretchable spring-like Ti@TiO 2 nanowire array as the photoanode by Liu and co-workers. This design was intended to achieve higher exibility and elasticity of DSSCs and the spring-like Ti@TiO 2 array was prepared by the hydrothermal technique at a controlled NaOH concentration. Liu's team was the rst to achieve a 100% stretching degree in FDSSCs with a PCE retention rate of 95.95% upon bending accompanied by 100% stretching strain. 142 Another major nding was made by Ni et al., that the enhanced surface area of nanowire arrays achieved on prolonged etching treatment can improve the dye intake but it reduces the PCE by increased electron recombination. Even then, Ni and co-workers succeeded in attaining an improved PCE of 9.39% by employing an additional scattering layer of TiO 2 particles along with rough surface rutile TiO 2 nanowire arrays. 143 Thus, light scattering layers can make up for the loss encountered by extended rates of photo electron recombination. As a way of minimising charge recombination and improving electron transport across TiO 2 nanowires, developing composites with materials having high electrical conductivity and charge carrier mobility was found exciting. One such approach was carried out by Makal and Das in 2021 by fabricating a reduced graphene oxide laminated one-dimensional TiO 2 -B nanowire composite based photoanode. 144 The effect of different reduced graphene oxide loadings on the PCE of the DSSCs was evaluated and it was found that an 8 wt% loading showed a PCE of 4.95%. Makal and Das used the same design to develop another photoanode material where they encapsulated TiO 2 -B nanowires with graphitic carbon nitride (g-C 3 N 4 ) and the PCE for the developed cell was found to be 5.12%. 145 This is one of the highest efficiencies reported for a DSSC with a TiO 2 -B based photoanode. Signicant efforts carried out for increasing the performance of nanowire-based DSSCs are discussed above and are tabulated in Table 1.

Nanorods
Nanorods (NRs) are introduced into DSSC photoanode fabrication ( Fig. 10) with the objective of increasing DSSC performance by exploiting their one-dimensional nanoscale properties. TiO 2 nanorods facilitate easy electron transfer by utilising their specic geometry (Fig. 11) and reducing the ohmic loss during the electron transfer through the mesoporous titania layer. 147 Various methods have been used for the synthesis of TiO 2 nanorods for DSSC applications. Along with the introduction of novel synthesis strategies, efforts were made to modify the morphology and surface properties of TiO 2 nanorods as given in Scheme 2.
A widely used fabrication method for TiO 2 nanorods was the solvothermal method. Several studies on TiO 2 nanorods were reported based on solvothermal techniques. [147][148][149][150][151][152][153][154][155][156][157][158][159][160] Most of the solvothermal methods used for TiO 2 nanorod fabrication employ water as a solvent i.e. the hydrothermal method. Sometimes the solvothermal method may be a multi-step process or a combination of more than one synthetic method. In 2005 Jiu et al. reported the synthesis of single crystalline anatase TiO 2 nanorods by a surfactant-assisted hydrothermal method. These nanorods were found to exhibit an efficiency of 7.06%. Higher ll factor (FF) values of TiO 2 NRs were observed compared to those of P25 nanoparticles, indicating the easier electron transport (lower resistance) through nanorods. In 2006, Jiu et. al. prepared single-crystalline anatase TiO 2 nanorods by a surfactant-assisted hydrothermal method with control over the size and diameter. Here nanorods having a length of 100-300 nm and diameter of 20-30 nm were synthesized and they showed an efficiency of 7.29%. 147 Jiu et al. achieved shape conservation and size regulation of nanorods with the help of a copolymer. Later Liu et al. found that the diameter, length and density of the nanorods prepared by the solvothermal method could be varied by changing the growth conditions such as growth time, growth temperature, initial reactant concentration, acidity and additives. 161 Efforts of Liu et al. resulted in the formation of single-crystalline rutile TiO 2 nanorods having an efficiency of 3%. In the very next year, a signicant increase in conversion efficiency (7.9%) was achieved by De Marco et al. using TiO 2 nanorods prepared by a single-step solvothermal technique. 149 The obtained anatase TiO 2 nanorod crystals were converted into screen printable paste for easy application into DSSCs. Here the solvothermal method helped to avoid coarse aggregation and shape loss of nanorods. Using the alkali hydrothermal technique, a novel morphology i.e. fan-shaped rectangular parallelepiped TiO 2 rods were synthesized by Shao et al. with a conversion efficiency of 6% obtained for 1 M NaOH utilizing DSSCs and this efficiency is found to be 66.7% higher than that of P25-DSSCs. 153,162 Investigations carried out at different NaOH concentrations showed that the morphology and crystal phase of the nanorods are affected by the alkali concentration. Kathirvel [158][159][160] Oriented rutile TiO 2 nanorod arrays were grown in situ on an FTO (uorine doped tin oxide) coated glass substrate by using a mixed acid medium composed of titanium tetraisopropoxide (TTIP), hydrochloric acid (HCl), acetic acid (AcOH) and water (H 2 O) as the solvent. 152 An acid mixture with HCl : AcOH in a volume ratio of 4 : 8 produced oriented, uniform, thin rutile nanorods which were found to be superior to single acid-grown nanorods. They have shown an efficiency of 4.03% using 2.3 mm long nanorods. Zhang and co-workers introduced an oriented attachment mechanism for the fabrication of size and shape tuneable anatase TiO 2 NRs. 158 These single crystalline long TiO 2 NRs have reduced grain boundaries which lead to enhanced charge collection. Thus, long thin NRbased DSSCs exhibited an efficiency of 8.87%. In 2015 Liu et al. prepared single-crystalline anatase TiO 2 nanorods by a solvothermal method in which tetrabutylammonium hydroxide (TBAH) is used as the morphology controlling agent. 159 DSSCs fabricated using these NRs have achieved an efficiency of 8.66%. Recently, several single and multi-step solvothermal fabrications of TiO 2 NRs were reported. 160,[163][164][165] Later it is found that employing a composite structure of nanorods (NR) and nanoparticles (NP) in the photoanode can complement the advantages of each other. The rst bilayer NR/NP composite structured photoanode was fabricated by Hafez et al. 150 Here the

Review
Nanoscale Advances sol-gel derived NP layer was coated with hydrothermally prepared NRs and this DSSC assembly exhibited an efficiency of 7.1%. Several NR/NP bilayered photoanode structures were fabricated later. 151,155,156,[166][167][168][169][170] These bilayer structures provided enhanced light scattering as well as increased surface area, which is nally reected in their increased efficiencies (Fig. 12). The bilayer design proposed by Rui and co-workers enabled the synthesis of size tuneable rutile TiO 2 nanorod microspheres by controlled hydrolysis during solvothermal synthesis and can be employed as a scattering over-layer in bilayer DSSCs. 156 Rui et al. achieved an efficiency of 8.22% whereas the efficiency of the single layer reference cell was 7.00%. Better performance of the TiO 2 NP/NR composite was obtained by Chatterjee and coworkers by employing a 1 : 1 wt% composite obtained from hydrothermally derived NRs and commercially available TiO 2 powder. 171 The obtained PCE of 8.61% was the highest among the photoanodes fabricated out of TiO 2 NR/NP composites. Shao et al. introduced low-temperature fabrication of photoanodes for exible DSSCs by electrophoretic deposition (EPD). 167 The obtained NR/NP structures (NRPs) have high surface charges and wide size distribution. In addition, a multiple EPD process was adopted to form a better quality microstructured photoanode. In the same total time, the efficiencies of the multiple devices are more than 2.2 times that of one-step devices. Without any calcination or compression, the best device fabricated with multiple EPD and a thin layer of nanoparticles gives a conversion efficiency of 4.35%. Using a two-step hydrothermal method a novel bilayer structure composed of one-dimensional nanorods, which can serve as direct electron transport pathways and a three-dimensional hierarchical structure acting as a light scattering as well as large surface area layer for dye loading was introduced by Li et al. 168 The dependence of the TiO 2 NR/NP structured composite's photoanode efficiency on the amount of NRs was studied by researchers and it was found that efficiency is the maximum for composites having 10% NR content (4.89%). 172 Another method employed for nanorod synthesis was electrospinning. Fujihara et al. employed electrospun TiO 2 nanobres as precursors for nanorods and these rods were spray dried onto an FTO plate. 173 These spray deposited nanorods overcame the adhesion difficulties of nanobres and exhibited an efficiency of 5.8%. Another electrospun synthesis was done by Jose et al. and they yielded 5.1% efficiency. 190 In 2009 DSSCs having an efficiency of 9.52% were fabricated by Lee et al. using a combination of solgel and electrospinning techniques. Here nanorods were electrospun from a solution of titanium n-propoxide and polyvinyl acetate in dimethyl formamide. 174 They have conducted a comparative study of nanorod and nanoparticle based DSSCs. These studies revealed that nanorod DSSCs have a pore volume double that of nanoparticle cells. Obviously the surface area available for sensitizers in nanorods is $2.5 times that of the nanoparticle based DSSC at equal TiO 2 weights. Also, the electron-hole recombination time for nanorods is found to be more than eight times that of nanoparticle DSSCs. Later in 2013 MWCNTs were introduced as electron transporting superhighways into TiO 2 nanorods by the electrospinning technique (Fig. 13). This incorporation was done by Yang et al. and they obtained an efficiency of 10.24%. 175 In the same year Chen et al. employed a microemulsion electrospinning technique to synthesize TiO 2 nanorods as a composite of mesopores and macropores by using paraffin oil droplets as the template. 166 This microemulsion electrospinning approach achieved 8.53% efficiency for bilayer DSSC architecture with NRs as the scattering layer.

Review Nanoscale Advances
In 2011 it has been reported that nanorods prepared by the hydrothermal method were used as seeds for the synthesis of a TiO 2 nano-branched photoanode. 176 This nano-branched array has achieved an efficiency of 3.75%.
Efforts were made to make NR-based DSSCs exible by growing NRs on substrates other than the FTO plate. Such an innovative electrode architecture was introduced by Guo and their team, in which bunched TiO 2 nanorods were grown on carbon bres by a 'dissolve and grow' method. 177 By using these carbon bres coated with bunched TiO 2 nanorods, tube-shaped exible 3D DSSCs have been prepared with an efficiency of 1.28%. Wang et al. made a similar attempt for enhancing exibility by preparing single crystal nanorod assembled TiO 2 cloth from carbon cloth templates. Wang et al. employed a rapid microwave heating process for TiO 2 cloth synthesis and this DSSC assembly yielded an efficiency of 2.21%. Wang et al. investigated the inuence of sensitizers such as N719, C218 and D205 on TiO 2 nanorod array-based photoanodes and found out that the best efficiency is obtained from C218. 178 Also, the performance of TiO 2 NR array DSSCs depends strongly on the annealing temperature. 179,180 Researchers found that annealed nanorod arrays show an improvement of more than 400% in efficiency compared to unannealed NRs. This increase can be attributed to reduced recombination and better electric contact between the NRs and FTO substrate. Efficiency improvement of TiO 2 nanorod-based DSSCs can be made by surface modications such as nanoparticle decoration, 181 chemical etching, 182 doping, 183 TiCl 4 treatment 184 etc. Ghaffari et al. showed that the generation of photoelectrons by Au NPs can help electron-hole separation and thereby result in enhanced ll factor (FF) and short circuit current values (J sc ). 181 The photochemical reduction process under ultraviolet radiation was employed for loading Au NPs on TiO 2 NRs. Yao et al. prepared Nd doped TiO 2 NRs by the solvothermal method. 183 Here the doped Nd ions enhanced the injection of excited electrons and thereby decreased the electron-hole recombination. The obtained efficiency was 4.4% which is 33.3% greater than that of the undoped analogue. Later, a combination of TiO 2 nanomorphologies with other nanomaterials gained considerable interest. One such photoanode material consists of an overlayer of TiO 2 nanorods and an underlayer of TiO 2 embedded ZnO nanoowers. 185 The doublelayered design was introduced by Chen and co-workers and they achieved a photo conversion efficiency of 8.01%. A similar composite design consists of rutile TiO 2 nanorods incorporated with an a alumina thin lm. 186 Here a alumina facilitated an enhanced electron lifetime and better charge transport. The a alumina incorporated TiO 2 nanorod arrays exhibited a PCE of 6.5%.
To boost the charge collection efficiency of TiO 2 nanorods, a layer of reduced graphene oxide (rGO) was wrapped over the nanorod surface by Subramaniam and his team. 187 2 wt% rGO loaded nanocomposite was found to show a superior photo conversion efficiency of 4.54%. Another effort was made by Roy and co-workers to improve the electron transport which involved the wrapping of rutile TiO 2 nanorod arrays with polyaniline (PANI) on their surface. 188 Being a conducting polymer, PANI enhanced the electron transport and the conjugation helped to capture more photoelectrons which in turn diminished the recombination between photo excitons. The association of a TiO 2 NR photoanode with PANI made more visible light absorption possible. Later Tang and co-workers reported that the higher the aspect ratios of TiO 2 NRs, the higher the dye loading capacity and efficiency will be. 189 Here NRs having different aspect ratios were prepared by controlling the reaction time. Also, improved performance of branched TiO 2 NRs was investigated by Hu et al. 164 . Prominent research on TiO 2 nanorod-based DSSC photoanodes is tabulated in Table 2.

Nanotubes
TiO 2 nanotubes were introduced with a hollow cavity structure, possessing a higher active surface area. Their enhanced absorption capacity and fast electron transport ability were examined by researchers 194,195 and they were found suitable for DSSC applications as shown in Fig. 14. Gaps in TiO 2 mesostructures, which act as electron traps, can be avoided by using nanotube arrays having small boundaries in between. This results in an increased diffusion length i.e. the distance travelled by an electron in a tube before recombination. 196 It was estimated that the diffusion length of a nanotube cell is approximately 100 mm. 197 So up to this limit, the tube size can be increased which in turn increases the surface area without promoting recombination. The oxidised species from the electrolyte can easily escape from nanotubes because of their 'open' structure as compared to from the mesoporous layers. Once it is oxidised, it will diffuse towards the cathode and get reduced back rapidly to minimise recombination losses. 198 The drawbacks of TiO 2 nanotubes are their high manufacturing cost and time-consuming preparation techniques. 199 The efficiency of DSSCs employing TiO 2 nanotubes is dictated by factors such as morphology and the crystalline structure of the tubes. It is observed that reduction in the tube diameter has a superior role in efficiency compared to the increase in the tube length. 200 Also, dye loading depends on the annealing temperature of the nanotubes. 200 The next factor is collection efficiency, which can be improved by minimising recombination losses by introducing modications on the tube surfaces. 201,202 Scheme 3 provides an overall idea of the efforts made to improve TiO 2 nanotube based photoanode performance in DSSCs. Some of the signicant nanotube-based DSSC architectures are discussed in this section. One of the earliest attempts involved the use of nanotube powders formed by a surfactant template-assisted technique. 203 These disordered nanotube array has achieved an efficiency of 4.88%. 203 The most widely employed technique for nanotube synthesis was electrochemical anodization. [202][203][204][205][206][207][208] These electrochemical anodizations were carried out in different electrolytes by different research groups. Yi et al. investigated the inuence of parameters like water content, anodization time and post-treatment of nanotubes on the photoconversion efficiency of the photoanodes. 209 It is found that efficient systems can be developed out of a combination of anodization with other synthetic methods for the fabrication of hybrid and multi-layered photoanodes. [210][211][212][213][214][215][216] Another simpler route of TiO 2 nanotube synthesis proposed by Ramakrishnan and co-workers involved the hydrothermal treatment of quasi-crystalline TiO 2 nanoparticles. 217 They were reported to have achieved hydrothermal conditions even without using an autoclave. Apart from using individual nanotubes, designs involving nanotube arrays were made by certain research groups. 194,204,[206][207][208]210,[212][213][214][215][218][219][220] For example, a highly ordered array of TiO 2 nanotubes with a length of 360 nm was introduced in 2006 by Mor and co-workers with a PCE of 2.9%. 221 This reduction in PCE is due to the comparatively smaller thickness of the photoelectrode. Mor et al. claimed that the efficiency can be further improved towards the ideal limit of $31% by increasing the length of the nanotube array up to a few micrometres. Wang et al. showed that efficiency can also be increased by treatment of TiO 2 nanotube arrays with TiCl 4 and ozone. 202 Such a surface engineered TiO 2 nanotube array attained an efficiency of 7.37%. Lei et al. brought the efficiency to 8.07%, employing one dimensional TiO 2 nanotube arrays ( Fig. 15a and b) prepared by anodization. 206 The I-V plots obtained by Lei and team for the TNT/FTO lm in comparison with the P25/FTO lm are shown in Fig. 15c. Another report by Hun Park and co-workers shows that TiO 2 nanotubes longer than 15 mm, transplanted on an FTO plate achieved an efficiency of 5.36% aer TiCl 4 treatment. 207 TiO 2 nanotube arrays were also subjected to doping with other elements. [222][223][224] The intention of adding dopants is to perfectly match the LUMO of dye molecules to the conduction band of TiO 2 for efficient electron transfer. This kind of perfect alignment will result in enhanced electron injection and reduced electron recombination. Yang et al. introduced TiO 2 nanotubes doped with Nb by anodizing Ti-Nb alloys (h ¼ 3.21%). 222 Using the same technique, So et al. prepared Ru-doped TiO 2 nanotubes from Ti-Ru alloys(h ¼ 5.16%). 223 While adding metal content it is found that the efficiency is the maximum for an optimum value of metal concentration. Also, metal doping will have an adverse effect if the length of the tubes is not correctly controlled. Subramanian et al. fabricated DSSCs with borondoped TiO 2 nanotube arrays. 224 Here B doped nanotube arrays   were prepared by anodization and they achieved an efficiency of 3.44% with an increased electron lifetime and reduced interface resistance. Besides doping, surface decoration with NPs and microspheres was also tried by different groups. 205 227 The multi-hierarchical design, which was obtained via a novel vacuum-assisted colloid lling route has 4 times better charge transport and 3.2 times better dye intake than that of the conventional TiO 2 nanotube array-based photoanodes. Cirak and co-workers proposed a hybrid photoanode design comprising ZnO nanorods and TiO 2 nanotubes. 228 The hybrid photoanode fabrication was carried out by a two-step synthetic process, anodic oxidation of TiO 2 nanotubes followed by hydrothermal deposition of ZnO nanorods over the TiO 2 nanotubes. They found the hydrothermal treatment temperature to be a determining factor of PCE. The hybrid design attained a PCE almost double that of the normal TiO 2 nanotube photoanode. As part of achieving increased dye pick up and enhanced light scattering, multilayered photoanode architectures involving TiO 2 nanotubes were also introduced. 208  Several other associations of TiO 2 nanotube arrays with other electron-conducting materials such as graphene and g-C 3 N 4 were later introduced. Wang et al. fabricated a single-layer graphene/TiO 2 nanotube array heterojunction photoanode by a wet transfer method and achieved a photon to current conversion efficiency of 4.18% which is far higher than that of pristine TiO 2 nanotube photoanodes. 229 The formation of a heterojunction with graphene not only reduced the bandgap but also enhanced visible light absorption. Mohammadi and coworkers modied TiO 2 nanotube arrays with g-C 3 N 4 and ZnO nanoparticles through a hydrothermal route. Even though there is no signicant improvement in PCE by this composite formation, the I sc and V oc of the hybrid photoanode showed considerable improvement. In addition to the above-mentioned investigations, some distinct approaches were also attempted by researchers.  204,219,234 and (e) secondary semiconductor particle decoration. 205 An overview of DSSC applications of the tubular morphology of TiO 2 nanomaterials is given in Table 3.

Effect of one-dimensional nanomaterial morphology on electronic and charge transport properties
Compared to bulk TiO 2 having a wide bandgap, one dimensional TiO 2 nanostructures exhibited enhanced surface area, more active sites and quantum connement effects. Size, specic crystal facets, morphology and alignment are the main factors determining the electronic properties of these onedimensional TiO 2 structures. In the case of nanowires, as the size reduces to the nanoscale regime, the bandgap starts increasing as per quantum connement. 66 Doping of foreign elements is the main strategy employed for the tuning of the bandgap and such dopants can align the valence band and conduction band positions in favour of the sensitizer energy levels. Also, the best DSSC performances were obtained for vertically aligned nanowire based photoanodes. In the case of nanorods, signicant reduction in grain boundaries which act as electron traps caused an increment in the electron diffusion length and thereby enhanced charge transfer. 174,178 The bandgap adjustment can be made in TiO 2 nanorods by varying the aspect ratio of nanorods. With an increase in the aspect ratio of nanorods, there will be a downshi of the conduction band edge. 66 Similarly, surface area enhancement can be achieved by introducing surface roughness and mesoporosity. 235 In the case of TiO 2 nanotubes, those having a wall thickness less than 5 nm are capable of manifesting quantum connement effects. 236 Only hydrothermal tubes fall under this category. The presence of Ti 3+ and oxygen vacancies are another reason for enhanced charge transport. Annealing temperature is a crucial factor determining the charge transport ability of nanotubes. 236 Bandgap engineering in nanotubes can be made by doping, surface adsorption, heterojunction formation and surface decoration. 76,205,222,223 3.5. Signicance of one-dimensional TiO 2 nanomorphology for exploitation of the photoanode performance TiO 2 one-dimensional (1D) morphologies were introduced as photoanode materials to overcome the limitations encountered by TiO 2 nanoparticulate electrodes. These one-dimensional morphologies are found to have profound photoelectrode characteristics with certain constraints which limit their use as photoanodes. We address this problem to highlight a particular one-dimensional morphology as the best choice by analysing various TiO 2 nanomaterials with 1D morphologies. A clear distinction of the best among nanowires, nanorods and nanotubes as photoanode materials requires a detailed investigation of several aspects. The main aspects to be considered are the band gap, surface area, aspect ratio, charge recombination time, transport properties and stability. In terms of surface area, nanowires, nanorods and nanotubes are inferior to zerodimensional TiO 2 nanoparticles. 89 However, they possess high aspect ratios and better charge transport pathways compared to conventional multidimensional TiO 2 NPs. As a result of limited surface area, the amount of dye (sensitizer) intake into these one-dimensional morphologies would be lesser. Even then, DSSCs based on TiO 2 one-dimensional photoanodes exhibit better IPCEs. It was found that the TiO 2 NP layer, having thrice the surface area of a TiO 2 nanotube layer, exhibited a much lower IPCE compared to the nanotube layer. 237 This enhancement in IPCE of nanotubes can be attributed to the much larger electron transport time prevailing in them. 237 In the case of nanotube photoanode materials, the fraction of dye intake as well as electron diffusion lengths increases with an increase in the tube length. The electron diffusion lengths in nanotubes are almost thirty times that of NPs. 236,237 Thus, an increased tube length can have the synergic effect of both surface area enhancement and smoother electron transport, irrespective of the high density of electron trap states in TiO 2 nanotubes. Similar results are obtained in the case of TiO 2 nanowires also. The best efficiencies are obtained upon arranging the nanowires vertically rather than randomly, which ensures effective contact between individual nanowires and the electrode. 66 With an increase in the length of nanowires, the number of electron traps and dye loading increases. So, nding the optimal nanowire length ensures the best IPCE. The TiO 2 nanorod-based photoanode exhibited an electron lifetime 8 times that of a TiO 2 NP-based one. 174 TiO 2 nanorods showed increased electron diffusion lengths and reduced grain boundaries compared to NP-based electrodes. 191 The above information leads us to the conclusion that the selection of the best one-dimensional TiO 2 photoanode material among the three i.e., nanorods, nanowires and nanotubes is rather difficult. This dilemma arises due to the vast number of modications possible over one dimensional TiO 2 nanomaterials using various design strategies under different experimental conditions. Several of these investigations have led to signicant and valuable ndings regarding the photoanode architecture, performance parameters and stability. Upon a detailed survey of the literature, it can be seen that the combination of one dimensional TiO 2 nanomaterials with other nanomaterials exhibited the best conversion efficiencies. These nanomaterials are chosen in such a way that the limitations of one-dimensional TiO 2 morphologies can be made up by them. Particle decoration, heterojunction formations and multi-layered architectures are superior design strategies employed for the fabrication of modied onedimensional TiO 2 photoanodes. The limited surface area of the one-dimensional structures is improved by the addition of TiO 2 NPs. These added layers cause enhanced light scattering and better dye loading capability. Decoration with metal nanoparticles makes use of surface plasmon resonance and yields better light absorption and electron transport. Heterostructures of one-dimensional TiO 2 with materials having suitable band structures and electron transport properties exhibit improved efficiency. From the above literature survey (Tables 1, 2 and 3), it can be seen that top IPCE values were obtained for multi-layered photoanodes involving one-dimensional TiO 2 morphologies. Each layer was incorporated with a specic function to be carried out, and such a division of labour led to enhanced outputs. Single layered one-dimensional TiO 2 morphologies prepared by advanced techniques like electrospinning, anodization, etc. or a combination of these techniques with sol-gel and hydrothermal methods yield higher efficiencies.

Summary and future prospects
This review discusses the design and working principle of various TiO 2 nanomorphologies and surveys various aspects of TiO 2 based photoanode materials in dye-sensitized solar cells. The total reliance of mankind on non-renewable energy sources has resulted in their depletion as well as serious environmental pollution. In this scenario, sunlight is considered a universal source of energy which is abundant in nature and free of cost. Thus, solar energy and its conversion into electricity and fuel (hydrogen) can be exploited. Solar cells which convert sunlight into electricity can be considered a future remedy for the energy crisis. In recent years a wide variety of TiO 2 nanostructures have been synthesized as photoanodes for DSSC applications such as nanoparticles, nanowires, nanorods, nanotubes, etc. Due to their excellent scattering ability and high electron transport rate, these one-dimensional nanostructures are found to be efficient candidates for DSSC photoanode fabrication. By overcoming their limitations with surface area, even higher efficiencies can be achieved. The suitable band alignment of TiO 2 with the sensitizer and the subsequent charge transport properties make it more attractive among other semiconductors in addition to its non-toxicity and viable functional architecture. TiO 2 nanomaterials have a band gap in the UV region of solar spectra which restricts their use in the eld of visible active applications. In DSSCs, the electrons in the LUMO of the dye are transferred to the conduction band of TiO 2 nanomaterials. Anion and cation doping in TiO 2 nanomaterials are remarkable routes to make suitable alignments of the LUMO of the dye and conduction band position of TiO 2 for fast electron transfer. Recent reports show that the IR absorption capability of TiO 2 can be enhanced by extending its absorption edge into the infrared region which provides a high charge collection. 116,238

Conflicts of interest
There are no conicts to declare.