Electrochemical Engineering of Nanoporous Materials for Photocatalysis: Fundamentals, Advances, and Perspectives

: Photocatalysis comprises a variety of light-driven processes in which solar energy is converted into green chemical energy to drive reactions such as water splitting for hydrogen energy generation, degradation of environmental pollutants, CO 2 reduction and NH 3 production. Electrochemically engineered nanoporous materials are attractive photocatalyst platforms for a plethora of applications due to their large e ﬀ ective surface area, highly controllable and tuneable light-harvesting capabilities, e ﬃ cient charge carrier separation and enhanced di ﬀ usion of reactive species. Such tailor-made nanoporous substrates with rational chemical and structural designs provide new exciting opportunities to develop advanced optical semiconductor structures capable of performing precise and versatile control over light–matter interactions to harness electromagnetic waves with unprecedented high e ﬃ ciency and selectivity for photocatalysis. This review introduces fundamentaldevelopmentsand recent advancesofelectrochemicallyengineered nanoporous materials and their application as platforms for photocatalysis, with a ﬁnal prospective outlook about this dynamic ﬁeld.


Introduction
Solar light is one of the most promising green energy resources, which can theoretically provide more than enough energy to address emerging global challenges such as climate change nanoporous structure . Advances in electrochemical oxidation (anodization) of valve metals such as aluminium and titanium enable new opportunities to precisely modulate and engineer the effective medium of semiconductor oxides to harness light-matter interactions for photocatalysis. Nanoporous anodic films can be produced with well-defined straight cylindrical nanopores or nanotubes as well as other advanced PC structures [45,46]. EENMs can also be produced with a unique set of physical and chemical properties, including chemical resistance, thermal stability, mechanical robustness, optoelectronic properties and large specific surface area, all of which are essential prerequisites to achieve high-performance photocatalytic devices [37,45]. In this context, this review provides a comprehensive perspective on recent advances in non-structurally and structurally engineered nanoporous photocatalyst materials produced by anodization ( Figure 1). Electrochemical fabrication processes are first introduced, followed by a detailed description of chemical and structural modifications used to enhance the photocatalytic efficiency of these systems. We also provide an overview of the current state of the photocatalytic capabilities of electrochemically engineered nanoporous materials. Finally, this review concludes with a general overview and a prospective outlook on future trends in this field.
Anodization of valve metals is generally performed in a temperature-controlled electrochemical cell, in which two electrodes (i.e., anode = valve metal (M) and cathode = platinum (Pt)) are submerged in an electrolyte (Figure 2a) [69]. When the cell is externally supplied either by constant potential (i.e., potentiostatic oxidation) or by constant current (i.e., galvanostatic oxidation), an oxidation reaction M→ M z+ + z e − is initiated. Depending on the main anodization parameters (i.e., voltage/current, temperature and electrolyte), three possible reactions exist: (i) the metal is continuously dissolved to produce M z+ ions (soluble anodic oxide); (ii) M z+ ions react with O 2from H 2 O in the electrolyte and form compact non-porous oxide (MO) (non-soluble anodic oxide); and (iii) competition between formation and dissolution of oxide occurs, leading to the formation of nanoporous MO structures (partially soluble anodic oxide). Under specific experimental conditions for certain valve metals, (iv) disorganized rapid growth of bundles of nanopores/nanotubes, and (v) formation of thick self-organized mesoporous structures can occur. (ii) formation of compact anodic oxides; (iii) self-organized oxides (nanopores or nanotubes); (iv) rapid (disorganized) oxide nanotube formation; and (v) ordered nanoporous layers. Reproduced from [69], with copyright permission from Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim, 2011. (b) Structural and geometric features of EENMs produced by anodization with tilted view of an EENM (left) (L p -nanopore length and L b -barrier layer thickness) and top view of EENMs (right) (d int -interpore distance and d p -nanopore diameter). Reproduced from [70], with copyright permission from MDPI, 2018.
Anodization is a highly versatile 3D nanofabrication approach, which can be applied to multiple metal substrates. Fabrication of nanoporous anodic alumina (NAA) structures by electrochemical oxidation of aluminium is the most developed anodization technology as compared to other metal oxides due to its controllability and well-defined nanoporous geometry [45]. However, there is a significant progress in anodization of other valve metals such as Ti, Fe, W, Zr, Nb, Hf, V and Co. Efforts have been made to develop Ti anodization technology due to the high stability against photo-corrosion and excellent optoelectronic properties for photocatalysis of anodic titanium dioxide (TiO 2 ) structures [16,37,69]. Anodization of valve metals under suitable conditions can result in anodic films featuring ordered nanopores or nanotubes with high aspect ratio. Figure 2b shows a graphic definition of the structural features of nanopore/nanotube geometry such as diameter (d p ), pore length (L p ), interpore distance (d int ) and barrier layer thickness (L b ). These characteristics can be precisely engineered by controlled manipulation of the anodization parameters [70]. The most representative examples of the self-organizing anodization conditions (i.e., voltage (V), temperature (T), electrolyte type) and main structural features (i.e., d p ) for multiple valve metals are summarized in Table 1. These studies demonstrate the fabrication of highly controllable nanoporous or nanotubular structures of various metal oxides through simple modification of the anodization parameters. For most of the valve metals, the presence of fluoride (F-) ions in the acid electrolyte is key in the formation of nanoporous or nanotubular structures due to the high chemical stability of their anodic oxides.

Chemical Modification of Electrochemically Engineered Nanoporous Materials
The chemical structure of EENMs is often modified with different entities (e.g., metals, quantum dots, semiconductor oxide films, conducting polymers, chalcogenides) to tune their optical, electrical and chemical properties to achieve desired functionalities and performances for specific photocatalytic and photo-electrocatalytic applications [37,69,71].

Doping
The chemical composition of EENMs can be altered by doping or bandgap engineering to enhance their performance in photocatalysis. Doping involves the introduction of secondary electronically active species into the crystal lattice of EENMs in order to modify the electrical conductivity and narrow the optical bandgap of the anodic oxide [72,73]. The intermediate states introduced from the doping elements (i.e., dopants) are required to be considerably close to the band edges (i.e., conduction or valence bands) of the anodic oxide for successful modification. The relative positions of dopants to the band edges of the intrinsic photocatalyst material can be estimated by the density of states (DOS) calculations [16,37]. Strategies to dope the chemical composition of EENMs include the implantation of non-metal (C, N, B and F) [33,[74][75][76][77] and metal (Cu and lanthanides) [78,79] elements, which are introduced by thermal treatment (annealing) in the gas atmosphere of the dopant element, co-sputtering or sputtering in dopant atmosphere, treatment in solution or melt of dopant, and high-energy ion implantation [69].
The most widespread doping approach is N-doping via ion implantation, which has been demonstrated to be particularly successful to incorporate nitrogen-containing species into as-produced TiO 2 nanotube structures at low-to-medium doping levels (1 × 10 16 ions cm −2 ) [75,80]. However, this doping method is limited by the short ion penetration depth, recrystallization requirement and inhomogeneous dopant distribution across the inner and top surface of EENMs [37,81,82]. Despite these constraints, well-defined buried junctions can be created into the inner surface of EENMs by rational utilization of the implantation profiles [80].
A particular straightforward doping approach to modify the chemical structure of EENMs in-situ the anodization process is using a metal alloy as a substrate. The alloy can either be prepared by arc melting or by co-sputtering of the pure valve metal and dopant metal [37]. Alloys can then be anodized to produce metal-doped EENMs. For instance, TiO 2 nanotubular structures doped with N [83], W [84,85], Mo [4], Nb [7], Ta [86], Ru [87,88] and noble metal [89] fabricated by anodization of different alloy substrates have been demonstrated. These electrochemically synthesized doped anodic oxides have shown improved open-circuit photocatalytic (OCP) and photo-electrochemical (PEC) activities. Enhancement effects have been ascribed to modification of the band-or surface-state distribution, increase in conductivity, functionality as a co-catalyst or localized surface plasmon resonance (LSPR). Although considerable efforts have gone into solution-based doping approaches (i.e., ion incorporation from the anodization electrolyte during anodization) for P [90] and N [91], efforts targeting N-doping are questionable due to XPS peaks located at 400 eV (adsorbed species), which were mostly obtained for nitrogen and unclarified electronic coupling of the doping species [69]. Though doping the structure of EENMs with non-metal and metal elements can be a suitable approach to increase the photocatalytic activity of these anodic oxides, the overall conversion efficiencies are often limited by thermal instability, decrease in charge carrier lifetime and unreliability of bandgap engineering [92]. Recent studies have demonstrated that visible light photo-response of EENMs for photocatalysis can be further enhanced by co-doping with a suitable combination of metals and/or non-metals [3,93,94]. For example, Yan et al. demonstrated enhanced photocatalytic and PEC properties under visible light irradiation by the synergetic effect in TiO 2 nanotubes co-doped with N and S ( Figure 3a) [3]. These TiO 2 nanostructures were prepared by a combination of two-step anodization and treatment with thiourea and calcination under vacuum at 500 • C for 3 h.
Reduction or annealing treatment under atmospheric pressure of TiO 2 -based EENMs can result in self-doping of TiO 2 nanotubes by the formation of Ti 3+ /Ov (oxygen vacancy) lattice defects [37]. Ti 3+ formation is beneficial to improve visible light photoresponse, increasing conductivity for improved charge separation and formation of surface states that facilitate charge transfer. This approach has been demonstrated as an optimal means of improving photocatalytic and PEC performances in reduced TiO 2 nanotubes [8,95,96]. Liu et al. showed that TiO 2 nanotubes exposed to high-pressure hydrogen treatment can produce reduced (black) TiO 2 nanotubes for OCP photocatalytic H 2 evolution without the need of co-catalysts ( Figure 3b) [8]. Ion implantation of H and N into reduced TiO 2 nanotubes can further enhance noble metal-free photocatalytic H 2 generation by co-catalytic effect [97,98]. Despite the advantages of Ti 3+ , the surface states of Ti 3+ are less stable and can readily be oxidized by air. Nevertheless, it has been reported that Ti 3+ states in anatase can stabilize these configurations [99].

Physical Vapour Deposition (PVD)
Physical vapour deposition (PVD) is an atomistic deposition process in which atoms or molecules of material are physically vaporized and transported through a vacuum or low pressure gaseous (or plasma) environment to form a thin film on the surface of EENMs by condensation [100]. Although PVD produces a non-conformal coating of deposited materials on EENMs, the deposition process is relatively inexpensive. PVD is typically used to deposit films elements, alloys and compounds with thicknesses in the range of a few nano-s to micrometres. PVD is categorized into five main categories, including electron beam evaporation, molecular beam epitaxy, pulsed laser deposition, sputtering and thermal deposition [101]. Ag, Au, Pt, W, Mo, Ni, Cu, semiconductor oxides and quantum dots are often deposited into nanoporous materials by thermal evaporation [102], sputter deposition [21,28,29,89,[103][104][105] or pulsed laser deposition [106] to improve photocatalytic and PEC performances by extending the lifetime of photogenerated charge carriers, narrowing the bandgap or by integrating LSPR effects. Nanoporous materials, such as NAA, can also be used as templates to produce various nanostructured materials (i.e., nanorods, nanowires and nanotubes) in combination with PVD processes. The resulting nanostructures have broad applicability in photocatalysis and the NAA templates can either be selectively removed or kept after the deposition process [101].

Electrochemical Deposition (ECD)
Electrochemical deposition (ECD)-commonly known as electrodeposition-is compared to the other deposition techniques a relatively simple and inexpensive deposition method where a solid coating on the surface of conductive EENMs is produced from ionic electrolytes under current/voltage-driven electrochemical reduction reactions [107]. Although ECD is limited by low deposition rate and single use of host template, this process is cost-effective, environmentally friendly and can be performed in any wet chemistry laboratory [72,101,107]. Various metals [35,108], oxides [30,109,110] and sulphides [31] can be electrodeposited to enhance the properties of EENMs for photocatalysis applications, including conductivity, chemical stability, PEC and photocatalytic properties [71].

Atomic Layer Deposition (ALD)
Atomic layer deposition (ALD) is a chemical gas phase technique used to deposit thin layers of materials such as oxides, nitrides, phosphates, sulphides and metals onto the entire surface of EENMs [111]. By this approach, the inner surface of EENMs is coated with atomic layers of an arbitrary material by repeating cyclic exposures to precursors and reactants, and purging. This method enables the precise control over the thickness and composition of the deposited film at the atomic scale, allowing the formation of homogeneous monolayers of the deposited material over large surfaces and high aspect ratios [112]. Nevertheless, the slow deposition rate, challenges in theoretical modelling of reactions and film growth, and potential cross-contamination of thin films by residual precursors can limit the application of ALD [113,114]. Despite these limitations, the ALD-assisted modification of EENMs has demonstrated excellent photocatalytic activities in degrading organic pollutants and OCP and PEC H 2 evolution [26,[115][116][117].

Sol-Gel Chemistry
The sol-gel method is a wet chemical deposition technique involving three main steps: (i) hydrolysis and partial condensation of a precursor by dip or spin coating; (ii) gel formation by polycondensation; and (iii) solvent evaporation and gel drying [118,119]. The sol-gel process can be used to fabricate films with high specific surface area and rich surface chemistry that allows for functionalization under low synthesis temperature, controllable reaction conditions and simple equipment [120]. Nevertheless, thin films produced by this method have some drawbacks, including expensive production cost, limited thickness controllability and poor mechanical properties [121]. Despite these drawbacks, the sol-gel method has been extensively used to tune the surface chemistry of the nanoporous materials.
For instance, sol-gel derived synthesized coatings based on TiO 2 and WO 3 have been used to modify the inner surface of NAA-based structures to attain photocatalytic activity to effectively photodegrade organic pollutants under UV or visible light irradiation [22][23][24][25]122].

Decoration with Nanoparticles
The inner and/or outer surfaces of nanoporous materials are often decorated with metal-, semiconductor-and polymer-based nanoparticles to improve their photocatalytic properties by: (i) heterojunction formation and sensitization to change surface band bending (metal clusters or other semiconductors), (ii) suitable surface mediators for enhanced charge carrier transfer (co-catalytic effects), and (iii) LSPR effects for electromagnetic field enhancements around metal nanoparticles to achieve efficient light-harvesting properties [37].
The surface of EENMs can be decorated with noble metal nanoparticles (i.e., Au, Ag, Pt, Pd and mixtures) by PVD [21,89,123,124], photoreduction method [36] or chemical reduction techniques [125]. This approach makes it possible to achieve co-catalytic effects to enhance photocatalytic reactions such as OCP or PEC H 2 evolution and degradation of pollutants. Furthermore, EENMs can also be decorated with noble metal particles by anodization of low concentrations of valve metal-noble metal alloys [20,29,124]. This method provides a controllable distribution of particles over the surfaces of the EENMs with very uniform and defined particle diameters for efficient photocatalysis. Lee et al. demonstrated the in-situ formation of Au nanoclusters on TiO 2 nanotubes grown by anodizing Ti-Au alloys ( Figure 3c) [20]. The homogeneous distribution of Au clusters with a typical particle size of 5-7 nm was achieved. Cluster spacing can be controlled by the Au concentration within the alloy and the anodization time. The remarkable enhancement in H 2 evolution from ethanol solution was achieved by these noble metal-decorated TiO 2 nanotubular structures, with an H 2 production ratẽ 30 times faster than that of bare TiO 2 nanotubes and~50 times more efficient than compact (flat) TiO 2 films decorated with the same number of Au clusters. Although Pt-decorated TiO 2 nanotubes can be produced by anodization of Ti-Pt alloys, poor photocatalytic H 2 generation was obtained due to the poisoning effect [20]. However, this study demonstrated that this decoration technique can be readily transferrable to other noble metals and alloys.
High photocatalytic rates have also been reported for EENMs decorated with semiconductor nanoparticles (i.e., CuO, Fe 2 O 3 , ZnO, Bi 2 O 3 , ZnTe or NiO). These hybrid structures were produced by slow hydrolysis of precursors [126], dip coating [127], chemical vapor deposition [128] and electrodeposition [30,109,110,129]. Enhanced photocatalytic performances were attributed to heterojunction formation, increase in surface area and/or possible charge injection from the electronic states of decorated nanoparticles. Nevertheless, careful considerations on the long-term stability of these composite nanoporous materials are required since some of these materials are susceptible to electrolytic corrosion, photo corrosion and instability under an applied voltage. Other common narrow bandgap semiconductors such as Bi 2 S 3 , PbS, ZnS, CdS and CdSe in the form of various nanostructures (i.e., nanoparticles, nanoclusters or quantum dots) can also be deposited on the inner surface of EENMs by successive ionic layer adsorption and reaction (SILAR) [130][131][132] as well as electrodeposition [31,131,133] methods. These composite photocatalyst materials provide enhanced photocatalytic and photo-electrocatalytic performances. More recent works have decorated EENMs with C60 [134], graphene [135], Ag/AgCl [136], AgBr [137] and BiOI [138] for enhanced photocatalysis. These studies demonstrate that photocatalytic enhancements in these composite photocatalyst materials are associated with different effects such as redox capability of Ag halogenides, effective charge transportation of C60 or graphene, or strong visible light absorption of BiOI. Note, not only inorganic compounds but also organic monolayers such as silanes and phosphonates deposited on the surfaces of EENMs have been investigated for photocatalytic applications [37].

Structural Engineering of Nanoporous Materials
In 1987, Yablonovitch and John demonstrated the concept of photonic crystals (PCs), which are multi-dimensional periodically structured materials that can control the propagation of electromagnetic waves across their structure [139,140]. Since these pioneering studies, many theoretical concepts and technological applications of PCs in the field of sensing [141][142][143][144][145][146][147][148][149][150], photocatalysis [22][23][24][25][34][35][36][38][39][40][41][42][43] and dye-sensitized solar cells [151][152][153] have been demonstrated. PCs feature regularly distributed variations of dielectric constant, and thus the refractive index, in a 1D, 2D or 3D fashion [45]. PCs feature a characteristic photonic stopband (PSB) which corresponds to a wavelength range within which incoming photons are not allowed to propagate through the PC's structure. This light-matter interaction can be tailor-made engineered by the geometric features and chemical composition of the PC structures, enabling different forms of light control such as Bragg diffraction, multiple scattering, light confinement and slow photon (SP) effect. This approach provides new opportunities to enhance light-dielectric interactions for photocatalysis by confining, controlling and manipulating the propagation of incident photons of specific energies or wavelengths [139,154,155].
Bragg diffraction forbids the propagation of photons within the PSB wavelength range, while multiple scattering increases trapping of photons propagating at wavelengths that are away from the PC's PSB. The SP effect slows down the group velocity of photons at the frequency edges of the PSBs. SPs generated on the red and blue edges of the PSB are primarily localized in the high and low dielectric parts of the PCs, respectively, increasing the lifetime of photons and the overall light absorption of the PC structure at specific spectral regions. In general, PC structures can be mechanistically described as an optical effective medium, in which the macroscopic optical properties of the composite PCs can be represented by model approximations such as Bruggeman, Drude, Looyenga-Landau-Lifshiftz, Lorentz-Lorenz and Maxwell Garnett [45]. For instance, the macroscopic optical properties of the composite NAA-PCs (i.e., air and alumina) can be estimated by averaging the properties of individual components (i.e., effective refractive index, effective dielectric constant) through effective medium approximation. In recent years, advances in nanofabrication technology have made it possible to develop conceptual PC structures. The most widespread techniques used to synthesize PCs are lithography and dry etching, vertical selective oxidation, wet chemical etching, fibre-pulling, embossing, self-organization and anodization [156]. Of all nanofabrication approaches, anodization has been demonstrated as a suitable method to fabricate EENM-based multi-dimensional PC structures with finely engineered geometric and optical properties. These PCs provide large surface area, versatility, scalable production, high throughput and resolution, making them attractive platform materials for a broad range of light-based technologies. However, the fabrication of EENMs by anodization still faces challenges. For instance, anodization of valve metals such as Fe, Zr, W, Nb, Hf, V and Co are limited to straight nanotubular structures produced under constant potential and Fcontaining electrolytes [16,157].
NAA-BPFs can be produced by pseudo-stepwise pulse anodization (PSPA). Such PC structures allow for the transmission of light at certain bands of wavelengths in a selective manner while forbidding the propagation of photons of certain energies [164]. NAA-LVBPFs can be produced by the combination of SPA and selective chemical etching, where the effective medium of these EENM-based PC structures is engineered perpendicularly to the nanopores' growth direction [165]. Some detailed review articles describing the fundamental concepts, representative examples and realization of NAA-PCs are provided in [45,70].
3D periodically structured TiO 2 nanotubes with broadband omnidirectional photonic stopbands can be prepared by periodic current pulse anodization, where the currents are alternated between high and low current density values [46,150,152,164,165]. A bilayered TiO 2 PC structure featuring dense arrays of smooth-walled nanotubes and a periodic structure along axial direction can be obtained by periodic current pulse anodization followed by heat treatment and a single constant current anodization step [153]. Aperiodic TiO 2 nanotube PCs with a gradually decreasing lattice constant (in an arithmetic sequence) can be synthesized by time-decreasing current density pulse anodization [166]. These TiO 2 PCs have demonstrated their applicability in a variety of applications, including sensing [150], photonics [46,164,165], dye-sensitized solar cells [151][152][153], photovoltaics [166] and PEC H 2 evolution [34][35][36], to name a few. Some excellent review articles introducing fundamental concepts and realization of TiO 2 nanotube structures are provided in [16,37,69].  Figure 4b presents the geometric and optical features as well as the anodization profile employed to fabricate (i) NAA-µCVs and (ii) periodic TiO 2 -PCs. NAA-µCVs composed of a physical cavity layer sandwiched between two highly reflective mirrors (i.e., NAA-GIFs) were fabricated by sinusoidal pulse anodization and feature a resonance band within the characteristic PSB (Figure 4b(i)). Periodic TiO 2 -PCs featuring small voids around each stack interface and broad PSBs over the entire visible spectrum were obtained under current pulse anodization (Figure 4b(ii)).
Although anodization of valve metals such as Fe, Zr, W, Nb, Hf, V and Co has been demonstrated successfully, the production of PC structures based on the anodic oxides of these valve metals remains challenging [16,57,[60][61][62][63][64][65][66][67][68]157]. The development of advanced PC structures with precisely controlled optical properties is still limited. Therefore, future developments in anodization of these valve metals to fabricate PCs will be essential to spread the applicability of EENMs across many technological disciplines, including photocatalysis and photovoltaics.

Electrochemically Engineered Nanoporous Materials for Photocatalysis
EENMs have been devised as platform materials for photocatalysis due to their unique physical, chemical, electrical and optical properties. EENMs can be used under open-circuit conditions or as a photoanode together with an inert or catalytic cathode such as Pt, C and others. The effective medium, surface chemistry and crystallinity of EENMs can be precisely engineered for enhanced photocatalytic applications, including open-circuit dye degradation, and OCP and PEC H 2 generation. In many cases, photocatalytic performances of EENMs are superior to those of benchmark photocatalyst nanoparticles due to different phenomena, including an increase in effective surface area, enhanced orthogonal carrier separation, the formation of electronic junctions and optimized geometry and crystallinity [37]. This section summarizes the most important factors affecting the photocatalytic activity of EENMs and gives a brief overview of current efforts towards EENM-based photocatalytic and photo-electrocatalytic systems.

Key Factors Influencing Photocatalysis
A common factor influencing the photocatalytic performance of EENMs is the crystal structure (i.e., polymorph and facets). EENMs fabricated by anodization are amorphous in nature but their crystalline structure can be modified by post-annealing treatment. Crystallinity affects the conductivity of the anodic oxide and lifetime of charge carriers in the structure of the photocatalyst material. For instance, annealing TiO 2 nanotube structures in air with increasing annealing temperature >300 • C results in better photocatalytic activities due to the formation of anatase crystals and the higher crystallinity degree of TiO 2 [18]. When TiO 2 nanotubes are annealed at 650 • C, anatase-rutile crystal junctions are formed, enhancing photocatalytic degradation of pollutants and OCP H 2 evolution under solar light or UV laser illumination due to band offsets and increment in light absorption. However, careful control over annealing temperature and oxidizing conditions (i.e., in O 2 ) is required since these conditions may deteriorate the nanostructures or lead to the formation of other crystal structures that might detrimentally affect photocatalytic performances [37]. Exposure of different crystal planes of crystalline structures also influences the photocatalytic activity of EENMs in photocatalysis due to the intrinsically different energetic nature between distinct planes and the formation of microjunctions. Lee et al. observed photocatalytic enhancements in anatase TiO 2 nanotubes featuring predominantly highly energetic {001} facets (i.e., TNT-11 fabricated at an anodization time of 11 h) using the photocatalytic degradation of Rhodamine B as an indicator [19]. Enhancements were attributed to these high energetic facets, which provide sites to boost the production of active oxygen species such as ·OH, ·O 2 − and H 2 O 2 upon UV irradiation (Figure 5a). The structural features of EENMs play a key role in enhancing photocatalysis. In general, a strong increase in photocatalytic degradation kinetics of organic pollutants can be observed with increasing nanopore length, which is attributed to an enhanced light absorption efficiency and a specific surface area increase (Figure 5b) [18,22,37]. Other geometric features influencing the photocatalytic performance of EENMs are the nanopore/nanotube diameter, surface features and sidewall roughness [37]. EENMs in the form of PC structures provide other approaches to further enhance the utilization of light for further boosting photocatalytic or PEC reactions. [22][23][24][25][39][40][41][42][43]. The chemical structure of EENMs can be intrinsically or extrinsically modified to improve their photocatalytic performance by formation of heterojunctions to change surface band bending, creation of suitable surface states for efficient charge exchange, integration of co-catalytic effects for photocatalytic and PEC reactions, incorporation of LSPR effects to concentrate electromagnetic fields in the vicinity of metal particles for more efficient light-harvesting, and increasing specific surface area [37]. These phenomena are strongly dependent on the elements or materials incorporated into the structure of EENMs and the type of treatment applied. For instance, OCP H 2 evolution by TiO 2 nanotubes can only be significantly enhanced with the presence of co-catalysts such as Pt nanoparticles and by exposure of the EENM to activating treatments such as H 2 -annealing and ion-implantation. On the other hand, the conductivity of TiO 2 nanotubes can be adjusted to attain optimal PEC activities by doping their structure with visible-light transition metals (i.e., Nb, Ta, Ru, etc.) or exposure to reductive treatments (Figure 5c).
The photocatalytic performance of ALD TiO 2 coated TiO 2 nanotubes (i.e., TNT with 200 cycles of ALD) was also found to be superior to their uncoated or TiO 2 nanoparticle-decorated counterparts by 72% and 88%, respectively, in degrading methylene blue under 365 nm UV light illumination. TiO 2 nanotubes modified with P, W, Mn, Pt, TiO 2 , Bi 2 S 3 and lanthanides by doping, anodization of alloy substrates, decoration with quantum dots and ALD have also shown successful photocatalytic degradation of gaseous pollutants (i.e., toluene, hexane) [27,79,90,130,175] and the photoreduction of Cr 6+ [176]. In many of these cases, doped TiO 2 nanotubes achieved better photocatalytic performance than their pristine analogues. Despite the advantages of chemically modified TiO 2 nanotubes for photocatalysis, the addition of coating onto TiO 2 nanotubes may not necessarily have the same beneficial effect for the photocatalytic degradation of organic pollutants in air and water. For instance, Sopha et al. have reported that ALD TiO 2 coatings in TiO 2 nanotube layers do not enhance the photocatalytic degradation of hexane in air due to a reduction of active surface area (Figure 6b) [27]. Several studies have explored the application of TiO 2 nanotubes for photocatalytic conversion of CO 2 into CH 4 (in a wet gas phase) or CH 3 OH (in liquid H 2 O) in the presence of H 2 O [177,178], including modified TiO 2 nanotube crystal structures with dopants [9] and decoration with co-catalyst nanoparticles [9,30,179] under UV, UV-visible and solar sunlight irradiation conditions. For example, Varghese et al. developed N-doped TiO 2 nanotubes decorated with Pt and/or Cu co-catalyst nanoparticles to photo-reduce CO 2 with H 2 O vapour under outdoor sunlight irradiation (Figure 6c) [9]. Strategies to enhance photocatalytic CO 2 conversion rates in TiO 2 nanotubular structures include (i) employing high specific surface area of TiO 2 nanotubes with a wall thickness thin enough to facilitate efficient transportation of photogenerated charge carriers to surface species; (ii) introducing N dopants to modify TiO 2 bandgap for absorbing and utilizing visible light; and (iii) distributing co-catalyst nanoparticles on nanotubes to efficiently adsorb reactants and aid redox reactions. Upon natural sunlight irradiation, a very high hydrocarbon production rate of~111 ppm cm -2 h -1 was achieved, which is at least 20 times higher than those reported for Pt/TiO 2 under UV illumination. Nevertheless, conversion of CO 2 to useful fuels is still kinetically hindered by the energetically disadvantageous nature of the two electron transfer steps involved in the conversion process.
Chemically modified TiO 2 nanotubes have also been widely investigated for their potential in photocatalytic [20,28,29,[97][98][99]103,117,123] and photo-electrocatalytic [7,31,74,[86][87][88]109] water splitting to generate clean H 2 . TiO 2 nanotubes split water efficiently as photoanodes in a photo-electrochemical setting with an ideal electrode (such as Pt) for "slow" cathodic H 2 evolution reaction under voltage bias [179]. Photoanodes based on TiO 2 nanotubes have been reported to be more promising than nanoparticulate layers due to their well-defined geometry and incorporation with co-catalysts (i.e., CuO and NiO) [31,109] or dopants (i.e., Nb, Ru, C and Ta) [7,74,[86][87][88]. Particularly promising results regarding the modification of TiO 2 nanotubes for water splitting have been reported by Roy et al. [87] and Yoo et al [88]. TiO 2 nanotubes were synthesized by anodizing low concentration Ti-Ru alloys, where Ru in TiO 2 can either act as a dopant for PEC water splitting or be present as RuO 2 to act as a co-catalyst for O 2 evolution. In PEC experiments, Roy et al. successfully demonstrated very high light-to-H 2 conversion efficiencies of an almost six-fold increase for intrinsically Ru-doped TiO 2 nanotubes in comparison with undoped TiO 2 nanotubes in terms of photocurrent density (Figure 6d) [89]. While TiO 2 can be used as a platform material to perform the photocatalytic synthesis of organics, the high selectivity required for organic synthesis is often not reached due to the non-selective character of generated free radicals (e.g., ·OH radicals) upon light irradiation [37]. However, modification of the electronic properties and nanoporous geometry of TiO 2 can steer carrier energies and lifetimes to overcome this inherent constraint [13,14]. Recently, Tripathy et al. demonstrated modification of crystal structures (i.e., anatase or rutile) and/or incorporation of Ru dopant into TiO 2 nanotubes for photocatalytic oxidation of toluene under 325 nm UV light illumination (Figure 6e) [13]. A significant change in the main reaction product (i.e., benzoic acid versus benzaldehyde) can be achieved, where certain undesired reaction pathways can be completely shut down due to the change in the electronic properties by Ru doping. Ru states prevent the formation of intermediate superoxide radicals, switching the reaction product from toluene to benzoic acid. The geometric features of TiO 2 nanotubes also show higher selectivity than that of nanoparticle-based layers, indicating that the combination of nanoarchitecture and tailor-engineered electronic bands of photocatalyst materials are a promising approach for achieving high selectivity in organic synthesis. Nanotubular TiO 2 structures also provide new opportunities to induce chain scission in functional organic monolayers (i.e., silanes and phosphonates) for drug delivery applications [180][181][182]. When exposed to UV light, scission of organic chains occurs on anchoring groups, leaving the inorganic section of these molecules attached to the surface of TiO 2 nanotubes. As-produced and Fe 3 O 4 nanoparticle-decorated TiO 2 nanotubes have also shown potential for photocatalytic killing of cancer cells under UV light irradiation [182,183]. Faraji et al. have shown that TiO 2 nanotubes modified with Ag/Benzene have effective sterilization effects by degrading a resilient bacterium such as Escherichia coli (E. coli) under UV and visible light irradiation due to LSPR effect and efficient charge separation (Figure 6f) [184].
Other pristine and chemically modified oxide nanotubes such as ZrO 2 [185], WO 3 [186,187] and Co 3 O 4 [68] have also been investigated for various photocatalytic and photo-electrocatalytic applications, including degradation of organic pollutants, water splitting and CO 2 reduction, due to their favorable band-edge positions, superior chemical stability, and low cost. Although NAA is an electronic insulator with a wide energy bandgap (i.e., 8.0-9.5 eV), the inner surface of this EENM can be modified with semiconductor materials for photocatalysis. Several studies have demonstrated successful photocatalytic applications of semiconductor-modified NAA-based platforms, including CO 2 reduction [10] and photodegradation of organic pollutants [122,188].  [27], with copyright permission from Elsevier B.V., 2018. (c) (i) Schematic illustration of sunlight-driven photocatalytic CO 2 reduction using N-doped TiO 2 nanotubes surface-loaded with Cu and/or Pt co-catalyst nanoparticles; (ii) Product evolution rates from N-doped TiO 2 nanotubes annealed at 600 • C and surface loaded with 52% Cu and 48% Pt nanoparticles. Reproduced from [9], with copyright permission from American Chemical Society, 2009.

Structurally Engineered Nanoporous Materials Produced by Anodization
Although various structurally engineered forms of TiO 2 -PCs have been synthesized by anodization [34][35][36]46,[150][151][152][153][164][165][166], only pristine and noble metal (i.e., Ag or Au) decorated hierarchical nanotubes have been investigated for clean H 2 energy generation via PEC water splitting [34][35][36]. Zhang et al. developed hierarchical TiO 2 nanotubes by varying the voltage during a two-step constant anodization for PEC water splitting (Figure 7a) [34]. Hierarchical TiO 2 nanotube arrays were composed of a periodically organized nanoporous layer on top of a uniform nanotube array, in which the top nanoporous structure served as a PC to enhance optical absorption of the overall photocatalyst platform. Of all the hierarchical TiO 2 structures assessed in this study (i.e., nanoring (top)/nanotube (bottom), nanopore (top)/nanotube (bottom) and nanohole-nanocave (top)/nanotube (bottom)), the nanopore (top)/nanotube (bottom) configuration achieved the highest photocurrent density value (i.e., 1.59 mA cm −2 at 1.23 V vs RHE) under AM 1.5G illumination. Lian et al. prepared Ag quantum dots-modified hierarchical TiO 2 nanostructures (i.e., nanoring (top)/nanotube (bottom)) utilizing a similar fabrication method for PEC H 2 evolution under visible light irradiation [35]. A high photocatalytic H 2 evolution rate (i.e., 124.5 µmol cm −2 h −1 ) was achieved, which was attributed to synergetic effects of Ag and the hierarchical PC structure. This smart design makes it possible to harness strong LSPR effect and anti-shielding effect of ultrafine Ag quantum dots within the same photocatalyst platform (Figure 7b).
Zhang et al. developed TiO 2 nanotube-based PCs by voltage variation during the second anodization step flowing a stepwise profile between high and low voltage to compensate for the loss of electric field intensity [36]. This fabrication approach improves periodicity and uniformity of the nanoporous top layer of hierarchical TiO 2 nanostructures, which are critical factors for the formation of an effective PC structure with well-resolved photonic stopband (PSB). Nanoring (top)/nanotube (bottom) hierarchical TiO 2 structures feature well-defined PSBs, which can be finely tuned across high-irradiance spectral regions for enhanced photocatalytic and PEC performances [34]. Furthermore, Zhang et al. assembled 20 nm Au nanocrystals onto these PC structures and achieved a photocurrent density of~150 µA cm −2 at 1.23 V vs RHE, which is the highest value reported to date in hybrid plasmonic Au/TiO 2 system for PEC water splitting under visible light irradiation (Figure 7c) [36]. This significant enhancement was ascribed to match the SPR absorption to the PSB of the PC structure, where the PC structure can localize, trap and provide multiple passes for plasmonically active photons generated around the Au nanocrystals. Thus, the average photon path length is increased and results in an enhanced SPR intensity of the Au nanocrystals for efficient PEC water splitting.
Despite  NAA-GIFs were fabricated by sinusoidal pulse anodization and functionalized with photoactive layers of TiO 2 through the sol-gel method. The photocatalytic performance of TiO 2 -NAA-GIFs was assessed by studying the photodegradation of three model organic dyes (i.e., methyl orange-MO, methylene blue-MB and rhodamine B-RhoB) with well-defined absorption bands across different spectral regions under visible-NIR irradiation conditions. TiO 2 -NAA-GIFs achieved high photocatalytic performances (k) in the degradation of MO, RhoB and MB (i.e., k MO = 0.25 h −1 , k RhoB = 0.39 h −1 and k MB = 2.10 h −1 ). These NAA-based PCs outperformed other forms of semiconductor photocatalysts such as pristine TiO 2 nanotubes, 3D TiO 2 PCs and P25 nanoparticles under simulated solar light irradiation. Performance enhancements were ascribed to the "slow photon" effect, where a group of incoming photons propagate with strongly reduced group velocity and are localized in high (i.e., photocatalysts) and low (i.e., dye and pores) dielectric parts of the red and blue edges of the PSB of these PC structures, respectively. Slow photons collected by the underlying NAA-GIF structure can be efficiently utilized by the photoactive TiO 2 layer to generate extra e -/h + pairs and increase photon-to-electron conversion rates. Spectral alignment of the characteristic PSBs of TiO 2 -NAA-GIFs to the specific absorption bands of the organic dyes was demonstrated to significantly enhance the photodegradation rate of these organic dyes.
Lim et al. have also prepared NAA-DBRs and functionalized with photoactive TiO 2 layers for assessing their photocatalytic performance associated with "slow photon" effect in the degradation of model organics under visible-NIR irradiation conditions ( Figure 9) [23]. These TiO 2 -NAA-DBRs achieved better photocatalytic performances (i.e., k MO = 0.32 h −1 , k RhoB = 0.35 h −1 and k MB = 3.04 h −1 ) than those of TiO 2 -NAA-GIFs. The measured enhanced performances were attributed to a better light collection associated with the broader and more intense characteristic PSB of NAA-DBR structures fabricated by stepwise pulse anodization (STPA). Real-life application of TiO 2 -NAA-DBRs to degrade resilient organochlorine compounds such as 4-chlorophenol in various environmental matrices and the reusability of these composite PC structures over five cycles were also demonstrated [23]. have recently fabricated TiO 2 -NAA-µCVs by a combination of rationally designed sinusoidal pulse anodization and the sol-gel method and explored their applicability in photodegradation of the model organic dyes (i.e., MO, RhoB and MB) under visible-NIR irradiation conditions ( Figure 10) [24]. Unlike the optical phenomena (i.e., "slow photon" effect) of other NAA-PC-based photocatalyst platforms (i.e., TiO 2 -NAA-GIFs and TiO 2 -NAA-DBRs) used to enhance the photocatalytic reactions, photocatalytic enhancements in TiO 2 -NAA-µCVs are attributed to a highly efficient recirculation and confinement of incoming photons with energies within the resonance band of these PC structures. This optical effect extends the lifetime of incident electromagnetic waves within the microcavity structure, thus increasing the probability of photon-to-semiconductor interactions and speeding up photocatalytic reactions due to an increased generation of charge carriers. Photocatalytic enhancements in TiO 2 -NAA-µCVs were found to strongly rely on the relative position between the resonance band of these NAA-PCs and the absorbance band of organic dyes, and on the light confinement quality. These composite PC structures have shown outstanding photocatalytic performances (i.e., k MO = 0.77 h −1 , k RhoB = 1.34 h −1 and k MB = 3.55 h −1 ) and demonstrated superior photocatalytic activities to other NAA-PC-based and TiO 2 -PC-based photocatalyst platforms.
The effect of the crystallographic phase of TiO 2 (i.e., amorphous, anatase and rutile) modified by annealing treatment (i.e., 50, 300 and 600 • C) on the photocatalytic degradation of MB by TiO 2 -NAA-µCVs was also investigated. TiO 2 -NAA-µCVs featuring the rutile phase outperformed their counterparts containing amorphous and anatase phases of TiO 2 due to the higher tendency to trap holes and electrons for photocatalytic reactions. Lim et al. developed hybrid Au-TiO 2 -NAA-DBRs photocatalyst platforms integrating slow photons and LSPR effects in the same photocatalyst platforms ( Figure 11) [25]. Au films were deposited on the top surfaces of TiO 2 -NAA-DBRs, and photocatalytic degradation of the MB molecules under visible-NIR illumination was used as reference reaction to assess enhancements associated with slow photon-LSPR coupling effects. However, this study demonstrated that photocatalytic enhancements are more strongly determined by "slow photon" effect than by LSPR effect due to the localized nature of generated surface plasmons on the top surface of the composite PC structures. Note, the obtained results revealed that these optical phenomena must be finely coupled in order to improve the design of noble metal-NAA-based photoactive PCs.

Conclusions
This review provides a comprehensive overview of fundamental aspects and recent progress in EENMs as photocatalyst platforms for photocatalytic and photo-electrocatalytic applications. Over the past ten years, the potential of these nanoporous materials in pollution degradation and hydrogen evolution has been demonstrated. EENMs provide many advantages to drive photocatalytic reactions efficiently, including highly defined tailor-made self-organized nanoporous structures, outstanding control over light to harness electromagnetic waves and efficient management of charge carrier separation.
The structure of EENMs can be precisely tailored in various types of photonic crystal by anodization-an inexpensive, simple and fully scalable nanofabrication approach. These PC structures provide new opportunities to harvest incident photons by different optical phenomena such as strong localization, trapping, slowing and recirculation. The optimal structural and chemical design of EENMs makes it possible to significantly enhance photon-to-electron conversion rates for several photocatalytic applications. However, more systematic experimental investigations will be needed to fully develop this technology for high performance, real-life photocatalytic applications. Furthermore, the ability to modify EENMs by a broad range of chemical and physical methods can endow these materials with desirable properties to achieve unprecedented performances by rational design and integration of optical and electronic structures. There remain fundamental questions regarding the optimal design of photonic and electronic band structures that can maximize the performance of hybrid semiconductor PC structures. The use of new crystalline structures, geometric features and chemical composition of EENMs, including modification of the inner surface with advanced carbonaceous materials, will be crucial aspects to overcome existing constraints in photocatalysis technology. The existing studies shown throughout this review demonstrate that EENMs have promising potential for a broad range of photocatalytic and photo-electrocatalytic applications. The photocatalytic performance and selectivity of these unique materials are found in many cases to be superior to existing photocatalysts such as nanoparticulate systems and inverted opals. Therefore, it is expected that future developments in EENMs might pave the way to new and exciting opportunities to expand the applicability of this technology across classic photocatalysis fields such as pollution remediation and hydrogen generation and cutting edge applications such as CO 2 reduction, highly selective organic synthesis, biomedical devices and self-cleaning surfaces.

Conflicts of Interest:
The authors declare no conflict of interest.