Self-supported one-dimensional materials for enhanced electrochromism

Abstract

A reversible, persistent electrochromic change in color or optical parameter controlled by a temporarily applied electrical voltage is attractive because of its enormous display and energy-related applications. Due to the electrochemical and structural advantages, electrodes based on self-supported one-dimensional (1D) nanostructured materials have become increasingly important, and their impacts are particularly significant when considering the ease of assembly of electrochromic devices. This review describes recent advances in the development of self-supported 1D nanostructured materials as electrodes for enhanced electrochromism. Current strategies for the design and morphology control of self-supported electrodes fabricated using templates, anodization, vapor deposition, and solution techniques are outlined along with demonstrating the influences of nanostructures and components on the electrochemical redox kinetics and electrochromic performance. The applications of self-supported 1D nanomaterials in the emerging bifunctional devices are further illustrated.

---

Zhongqiu Tong

Zhongqiu Tong received MS and PhD degrees under the supervision of Prof. Yao Li in the Center for Composite Materials and Structure, Harbin Institute of Technology (HIT). In 2016, he joined the School of Materials Science and Engineering as a faculty member at Southwest Petroleum University in Chengdu, China. His research interests focus on fabrication of metal oxide nanomaterials and their applications in electrochemical energy storage and electrochromism.

Shikun Liu

Shikun Liu received his Master degree from QiLu University of Technology. He is currently pursuing his PhD degree under the co-supervision of Prof. Jiupeng Zhao and Prof. Yao Li at School of Chemistry and Chemical Engineering, Harbin Institute of Technology. His research interests mainly focus on the design and synthesis of nanomaterials for electrochemical energy storage and electrochromism.

Yao Li

Yao Li has been a Professor of Materials Science in the Center for Composite Materials and Structure, Harbin Institute of Technology (HIT) and a council member of Chinese Materials Research Society. He was named as “New Century Excellent Talents Program Scholar” in 2006, “Youth leader in Science and Technology Innovation” in 2015, “Yangtze River Scholar” in 2016, and “Ten Thousand Talent Program Scholar” in 2017. He was awarded “Science and Technology Award for the Youth of China” in 2013 and “National Award of the outstanding Scientific and Technological Workers” in 2014. His research interests include fabrication of inorganic materials and conductive polymers with well-defined nano/micro structures for energy storage and electrochromism. He is the author or co-author of over 130 papers in peer-reviewed journals, 62 patents, and 3 books in the field of materials science. His international standing can be substantiated by his delivery of over 30 keynotes or invited talks in international conferences including the International Meeting on Electrochromism and IUMRS-ICA, and organization of >10 international conferences/symposia, etc.

---

1. Introduction

Over the past decades, chromism-related phenomena have received immense research attention due to their broad display and energy-related applications.^1,2 Chromism is regarded as reversible color and optical changes of a material or composite materials derived from an external stimulus. Based on the types of stimulus, chromogenic technologies involving electrochromic, photochromic, thermochromic and gasochromic technologies can be used in various different fields.³ Electrochromism can be defined as color and optical parameter changes in the visible spectrum controlled by a temporarily applied electrical voltage.^4,5 In some cases, the optical parameter changes in the near-infrared (NIR) and infrared (IR) spectrum regions are also used. Compared to other chromogenic types, electrochromism demonstrates some unique advantages, such as low energy consumption and operating voltage, multiple and high chromogenic states, high and reversible cycling stability, and reasonable memory effect. Thus, electrochromism has been demonstrated in various commercial applications, such as in smart windows, display devices, anti-glazing mirrors and spacecraft thermal control.

Because the electrochromism in materials is from reversible electric-field-induced redox processes, nanostructuring is an effective method to improve the performance.^4,5 Among various nanostructures, one-dimensional (1D) morphologies are very applicable for electrochromism.⁶ For basic 1D nanostructures such as nanorods, nanofibers and nanoribbons, their width and thickness (or diameter when the 1D nanostructures exhibit a cylindrical morphology) are confined to the nanoscale range between 1 and 100 nm, while their lengths can be several micrometers, even up to hundreds of micrometers or a few millimeters. The small diameter scale of 1D nanostructures is rather suitable for accelerated electrochromic redox kinetics, while the large scale of length in 1D nanostructures reasonably matches the macroscopic world for many electrochemical and physical measurements, including electrochromic tests.⁷ In addition, the long length but short diameter characters of basic 1D nanostructures indicate the ease and high efficiency of fabricating nanorod, nanofiber and/or nanoribbon-knitted porous complex nanoarchitectures, such as nest- and urchin-like morphologies.

On the other hand, to investigate the nanomaterials’ electrochromic performance or further assemble electrochromic devices, uniformly dispersed electroactive materials on the surfaces of transparent conductive oxide (TCO) electrodes, such as indium tin oxide (ITO), fluorine-doped tin oxide (FTO) and aluminum-doped zinc oxide (AZO) substrates are required. When the electrochromic materials are in a powder form, an additional procedure is needed to disperse the materials on TCO substrates, which could give rise to two possible adverse effects. (1) The weak physical adhesion strength among electrochromic nanomaterials may cause the release of electroactive materials, resulting in optical contrast loss in long-term testing. (2) The poor physical and chemical contacts between the electroactive materials and TCO substrate may hinder efficient electron transport for electrochromism. Thus, preparation of 1D electroactive nanomaterials directly grown on TCO substrates is quite desirable for electrochromism, due to not only the decease of complexity in the electrode and device preparation process but also the ease of improving the electrochromic performance derived from the strong chemical and physical contact between the electroactive materials and substrate.^7,8

Various self-supported 1D nanostructured morphologies of electrochromic materials have been developed and investigated. This review will focus on the recent advancements on the self-supported 1D nanostructures for electrochromic devices. The first part of the review is centered on electrochromic advantages of self-supported 1D nanostructures. Then these nanostructures are separately discussed in two main categories including template-derived and template-free morphologies, followed by discussion about 1D core/shell nanostructures, which are a special type of complex 1D nanostructure with two components or phases. The emphasis is to correlate the morphologies, components and interfacial interactions of the electroactive materials to their electrochromic properties and illustrate how these nanostructures influence the electrochromic redox kinetics and offer advantages. A brief discussion about the application of self-supported 1D nanostructures in electrochromism-involving multifunctional devices is further presented. A future outlook for the self-supported 1D electrochromic nanomaterials will also be presented.

2. Materials for self-supported 1D nanostructures

The origin of studies on electrochromism is usually traced to the pioneering work by Deb in 1969.⁹ He found that tungsten oxide (WO_x) films can be blued in acid solution once a negative electric stimulus was applied, while the blue color was bleached under positive electric stimulus. This facile optical parameter modulation aroused worldwide research enthusiasm. Electrochromism was found in many transition metal oxides, such as titanium dioxide (TiO₂), vanadium pentoxide (V₂O₅), molybdenum trioxide (MoO₃), nickel oxide (NiO), cobalt oxide (Co₃O₄), niobium pentoxide (Nb₂O₅), and tantalum pentoxide (Ta₂O₅).^10,11 For example, in 1989 Cogan et al. found that the color of a vanadium oxide film can be changed from yellow to deep blue in a lithium salt organic electrolyte.¹⁰ The authors attributed the color changes to the double injection of lithium ions and electrons into the vanadium oxide crystalline lattice.

The merits of transition metal oxides for electrochromism typically include: high electrochromic memory effects, long-term cycling stability, good mechanical strength, desirable environmental durability, and especially high durability under ultraviolet exposure outdoors when used as smart window electrode materials. However, their electrochromic kinetics is rather unsatisfactory when the transition metal oxides are not in the nano-region. In addition, the relatively low coloration efficiency of bulk transition metal oxides is a non-negligible obstacle for electrochromic applications, because a low coloration efficiency means that a high energy consumption is needed to fulfill coloration state switching.

Conductive polymers are another large family of materials used for electrochromism.^12,13 The typical structure of these polymers includes a conjugated π bond on the main chain. The optical modulation arises when electrochemical doping/de-doping occurs on the π bonds. For example, in 1984 Kobayashi et al. found that a polyaniline film exhibited four color changes (transparent yellow at −0.2 V, green at 0.5 V, dark blue at 0.8 V, and black at 1.0 V) in 1 M HCl solution.¹² These polymeric electrochromic materials demonstrate the advantages of fast redox kinetics, vivid color versatility, high optical modulation, rapid response times, low power consumption, and ease of manipulation of properties through structural modifications. However, the cycling stability, mechanical strength, and environmental durability are not as desirable as for transition metal oxides. Commonly used conductive polymers for electrochromism include polyaniline (PANI), polypyrrole (PPy), polythiophene (PT), poly(3,4-ethylenedioxythiophene) (PEDOT), polycarbazole, and their derivatives.

Hybrid materials, including inorganic–inorganic and organic–inorganic categories, have been a hotspot of the research community.^4,14 The prime reasons behind this popularity are the synergetic properties offered by the resulting hybrids. They often combine the elasticity and functionality of each component to overcome the drawbacks of components. Hybrid materials are not simple physical mixtures. The interactions holding the two components together include weak interactions such as van der Waals forces and hydrogen bonds, and/or strong chemical interactions, which are beneficial for improving the structural integrity for long-term measurements. For inorganic–inorganic hybrids, their redox kinetics could be enhanced due to the efficient electron and ion transport derived from the interactions between electron bands of the two components. While for organic–inorganic hybrids, they can exhibit high optical modulation and rapid switching response derived from the conductive polymers as well as high thermal and chemical stability derived from the mechanical strength of the inorganic component. In this review, the main hybrids considered are self-supported 1D core/shell nanostructures.

3. Electrochromic advantages of self-supported 1D nanostructures

The electrochromic properties of transition metal oxides come from the electric-field-induced double injection of electrons and small ions, such as H⁺, Li⁺, and OH⁻ (eqn (1) and (2)),^3–5 while alternating electrochemical doping/de-doping processes of the conjugated main chains are the origin of conductive polymers’ electrochromism (eqn (3)).^12,13

MO_y + xA⁺ + xe ↔ A_xMO_y (A: H, Li) (1)

NiO + OH⁻ ↔ NiOOH + e (2)

PANI + nClO⁻ ↔ (PANIⁿ⁺)(ClO⁻) + ne (3)

Such redox-based optical response makes nanostructuring an effective approach to boost the electrochromic performance. The preparation of the electrochromic materials directly grown on TCO substrates (“self-supported” electrodes) is rather favorable for electrochromism due to the elimination of the incorporation of conductive additives or electrode binders and the added step of slurry casting during the electrode fabrication process.^15,16 Furthermore, the formation of 1D nanoarchitectures, such as nanoneedle, nanorod, nanofiber, nanowire, and nanotube arrays, could significantly improve the electrochromic performance due to the electrochemical superiorities of these morphologies.^6–8,17 Thus, the synthesis of 1D nanomaterials employing a self-supported strategy has become fascinating for electrochromism. Self-supported 1D nanostructures possess the general nanostructure-derived electrochemical and electrochromic advantages such as large surface area providing plenty of accessible electroactive sites for high coloration contrast, short ion diffusion distance for fast switching response, and enough voids for efficient strain relaxation, etc., while self-supported 1D nanoarchitectures also demonstrate several unique electrochemical and electrochromic advantages. Fig. 1 shows a schematic representation of an idealized self-supported 1D electrode.

Fig. 1 Schematic diagram illustrating the advantages of an ideal self-supported 1D nanostructure for electrochromism.

(i) Controllable areal density of the electroactive materials: the loose stacking of electroactive materials is particularly important for electrochromic electrode films because of the favorable electrolyte penetration and enough voids for strain relaxation. On the other hand, the ideal areal density of the electrode materials is a key factor to realize high electrochromic performance. Low areal density of electroactive materials could give rise to unsatisfactory coloration contrast while too high areal density of electroactive materials might result in slow redox kinetics. For the majority of methods used to synthesize self-supported 1D nanomaterial films, the areal density of the electroactive materials can be facilely controlled.^18–20

(ii) Strong attachment to the substrate and efficient 1D electron transport: for the majority of methods used to fabricate self-supported 1D nanostructured films, the electrochromic materials are directly grown on substrates, resulting in strong physical and chemical attachment between the electroactive materials and substrate, as well as continuous conductive pathways to the substrate, guaranteeing efficient electron transport for electrochromism. In addition, strong bonds ensure the structural integrity of the whole electrode films, efficiently preventing the detachment of electrochromic materials from the substrate, beneficial for the long-term cycling performance.

(iii) Underlying enhanced redox kinetics: due to their structural advantages, the majority of delicately designed self-supported 1D nanostructures can exhibit enhanced redox kinetics, compared to bulk and many other nanostructured materials. Because of efficient 1D electron transport, large surface area, and other characteristics such as a large number of surface defects which are formed during the preparation process or post-treatment, the kinetics of redox reactions occurring on the surface and near surface layer are significantly improved. Such fast redox kinetics leads to an enhanced pseudocapacitive effect which is beneficial for fast switching response. An enhanced surface contribution has been found in electrode films of self-supported nanofiber,²¹ nanorod²² and nanotube arrays.²³ In addition, for some delicately designed self-supported 1D nanostructures, their ion diffusion coefficient and electronic conductivity also can be significantly improved due to the surface defects, guest ion doping and single-crystal nature of the nanofibers and nanorods.²⁴ Furthermore, some studies demonstrate that 1D nanostructures also exhibit a short characteristic relaxation process, indicating the high electrochemical stability of redox states in 1D nanostructures.²⁵

(iv) Ease of realizing the surface modification or fabrication of hybrid nanostructures: reactive liquids and gases (such as hydrogen, ammonia, and plasma) can easily penetrate and flow in the interconnected voids among the nanofibers or nanorods. Then reactions occur and may lead to mixed valences, new phases, defects, changes in the band structures, and/or production of new functional groups on the surfaces of the materials.

Using pre-prepared self-supported nanofibers and/or nanorods as templates to deposit another electrochromic material gives rise to self-supported core/shell hybrid nanostructures, which could exhibit enhanced coloration contrast and cycling stability, shortened switching response time, and increased coloration states. The coating layer can be the same material as that of the core but in a different phase, generating a “crystalline/amorphous core/shell” nanostructure.²⁶ However, in most cases, the coating material is different from that of the core.²⁷ As for the anodic preparation method, mixed metal oxides or doped nanotube arrays can be facilely prepared from anodization treatment on alloys.

(v) Ease of characterization and assembly of electrochromic devices: the strong structural integrity of self-supported 1D nanostructured electrode films indicates that these films can be directly used for electrochemical-optical characterization or device assembly.

(vi) Ease of realizing multifunctional electrodes or devices. Both electrochemical energy storage and electrochromism are from redox reactions, causing bifunctional integration into one film or device to be possible.^28,29 However, the existing conflicts between electrochromism and electrochemical energy storage (such as fast switching response for electrochromism and high energy density for electrochemical energy storage) result in the necessity of delicately designed nanostructured morphologies. Fabrication of self-supported 1D nanostructures with high ion insertion/extraction kinetics is an ideal approach to solve these problems, as described in later sections. Furthermore, given the significance of self-supported 1D nanostructures for the emerging miniaturization of power sources aimed at integration into micro- and nano-electronic devices,⁸ fabrication of such self-supported 1D morphologies is becoming more and more vital and important for miniaturized electrochromic energy storage devices.

4. Template-derived self-supported 1D nanostructures for electrochromism

Employing porous templates is a direct and effective method to prepare nanostructures for high-performance electrochemical-optical materials. The general fabrication process of nanostructures relies on the replication of the well-confined voids (such as holes, pores, channels or other hollow spaces) of the templates.^18,30 Various methods, for instance electrodeposition, sol–gel chemistry, hydrothermal deposition, and physical/chemical vapor deposition have been developed to effectively fill the voids with precursors. Removing the templates and transforming the precursors into the targeted materials with post treatments produces the replicated nanostructures. The large variety of templates with diverse porous morphologies can produce various nanomaterials from 1D nanostructures (nanofibers, nanorods, nanotubes, etc.) to 3D ordered or disordered meso-/macro-porous materials. Also, some biological species (such as leaves, butterfly wings, wood, viruses, proteins, and DNA) are used as templates to form nanowires and other nanostructures.³¹

The templates widely used to fabricate self-supported 1D nanostructures are membranes with 1D parallel nanochannels. There are two types of typically used templates: polymetric (such as PC and nitrocellulose) and oxide-based (AAO) membranes. Their high pore density (up to 10¹¹ cm⁻²) and wide pore diameter range (from tens to hundreds of nanometers) could ensure membrane-derived self-supported 1D nanostructures with a desirable areal density of electroactive materials, a controlled diameter scale and enough interspacing voids for efficient electrolyte penetration to achieve satisfactory coloration saturation. Meanwhile, the large thickness range of the membranes (from several to hundreds micrometers) gives feasible choice to prepare 1D nanostructures with controlled length.

4.1 Template-derived self-supported 1D metal oxide nanostructures

Typical membrane-assisted preparation of metal oxide self-supported nanofiber and nanorod arrays is shown in Fig. 2a. Attaching the template to the TCO substrate, filling the nanochannels with precursors and then transforming them into the targeted metal oxides with subsequent annealing treatment whilst removing the membrane leads to the successful synthesis of self-supported metal oxide nanorod, nanowire, or nanofiber arrays (Fig. 2b).³⁰ The tight attachment of the membrane to the substrate could effectively ensure strong physical and chemical contacts between the electroactive materials and the TCO substrate, resulting in efficient electron transport and long-term cycling stability. Meanwhile using different experimental details can control the morphologies of the 1D nanostructures, simply from nanorod arrays to nanofiber arrays. Notably, when the mechanical strength of the targeted 1D electroactive materials is not high enough or the length-to-diameter ratio of the targeted 1D electroactive materials is rather high, the tops of the nanorods or nanofibers could aggregate together, forming nanorod or nanofiber bunches instead of free standing and isolated nanorod or nanofiber arrays.

Fig. 2 (a) Schematic diagram illustrating the template-assisted synthesis of self-supported 1D metal oxide nanostructures. Step 1 is attachment of membranes on the TCO substrates, followed by a precursor filling process. Step 2 is the removal of the membranes and transformation of the precursor into the targeted metal oxide. (b) SEM image of TiO₂ nanorod arrays prepared using an AAO membrane. Reprinted from ref. 30. Copyright 2015 Springer. (c and d) SEM image of open-ended (nanorod bunches) and close-ended Ni(OH)₂ nanorod arrays prepared using an AAO membrane. (e) Comparison of the transmittance modulation for the annealed Ni(OH)₂ thin film, and close-ended and open-ended nanorod structures. Reprinted from ref. 35. Copyright 2014 Wiley.

Zheng et al. prepared arrays of WO₃ nanorods with lengths of ca. 1.8 μm by direct current (DC) magnetron sputtering of WO₃ on an aluminum lattice membrane (an AAO membrane with short thickness).³² The confined growth of WO₃ in the 1D nanochannels led to the diameter of the self-aligned parallel WO₃ nanorods being almost the same as the pores of the aluminum lattice membrane (about 200 nm). In 1 M NaCl solution, sodium-ion insertion under −0.8 V (vs. SCE) led to blue coloration, while bleaching under 1.2 V polarization resulted in a transparent state. The WO₃ nanorod arrays demonstrated obvious transmittance modulation in the visible spectrum range with a maximum value of about 50% at λ = 600 nm. By using a two-step-oxidation prepared AAO membrane with a pore size of 80 nm as a template and a colloidal suspension as a precursor, arrays of WO₃ nanorods with nanosized pores distributing regularly along the length of the nanowires were prepared.³³ A two-step electrochemical anodization on Al–W overlapped metal layers to prepare Al-doped WO₃ nanorod arrays was proposed by Park et al.³⁴ The outer Al layer was anodized into AAO, followed by anodization of the W layer creating Al-doped WO_x in the pre-formed nanochannels of AAO. The removal of the AAO template and annealing treatment produced self-supported nanorod arrays on a W substrate. With the assistance of AAO membranes on ITO substrates, Guo et al. prepared two types of self-supported Ni(OH)₂ nanorod arrays by electrodeposition filling and removal of the membranes with 10% NaOH solution. Over deposition gave rise to a dense film on the top of the nanorod arrays, producing a close-ended morphology (Fig. 2c and d).³⁵ Open-ended nanowire arrays demonstrated good electrochemical activity with superior transmittance modulation of ∼35% at λ = 635 nm, better than the close-ended nanowire arrays and dense film (∼20% and 17%, respectively) (Fig. 2e). Also, by comparison with other literature reports, the authors also found that the open-ended nanowire arrays exhibited higher coloration efficiency (50.5 cm² C⁻¹) than many other Ni(OH)₂ nanostructures, further demonstrating the efficient redox kinetics for low energy consumption in electrochromic nanowire arrays. Yamada et al. used a two-step process to prepare branched Au/NiO nanorod arrays for accelerated electrochromic color changes.³⁶ After Au nanorod arrays were prepared by AAO membrane-assisted electrodeposition, a branched NiO layer was deposited on the Au surface. Due to the high reflectivity, mechanical strength and electrical conductivity of Au, as well as the thin thickness of the NiO layer (15 nm), the branched Au/NiO electrode exhibited high and stable reflectance contrast over 0.4 at λ = 600 nm.

Polymetric membranes (PC membrane as representative) are another category of widely used templates to prepare self-supported 1D nanostructures. The preparation process also can be reflected in Fig. 2a. These polymetric membranes can be easily wiped out by annealing treatment in air or dissolution by organic solvents. Limmer et al. demonstrated a general technique to synthesize metal oxide nanorod arrays using sol electrophoretic deposition with appropriate track-etched hydrophilic PC membranes attached on the ITO substrates.^37,38 Electrochromic TiO₂, V₂O₅, and Nb₂O₅ polycrystal nanorods with diameters of ∼100 nm and a length of ∼10 μm were prepared. These oxide nanofiber arrays exhibited absorbance modulation in the visible spectrum range with obvious color contrast during the Li-ion insertion/extraction processes. Furthermore, Limmer et al. also found that the diameter of the V₂O₅ nanorods can decrease to 50 nm; meanwhile these nanorods were changed to be single crystals by using delicately designed experimental parameters.³⁹ These single-crystal V₂O₅ nanorod arrays exhibited higher transmittance modulation and faster switching response speed compared to polycrystal V₂O₅ nanorod arrays and a sol–gel-derived V₂O₅ dense film. Li-ion storage properties detected by charge/discharge tests showed that single-crystal V₂O₅ nanorod arrays exhibited superior lithium storage.^39,40 Higher lithium storage of the single-crystal V₂O₅ nanorod arrays confirmed their electrochromic advantages because the larger amount of inserted Li-ions indicated that a larger amount of V⁵⁺ ions took part in the redox reactions for electrochromism. Furthermore, compared to the polycrystalline V₂O₅ (10⁻³–10⁻² S cm⁻¹), the high electrical conductivity of single-crystal V₂O₅ nanofibers (0.5 S cm⁻¹)⁴¹ positively influenced the Li-ion insertion/extraction kinetics, which was beneficial for improving the electrochromic performance. By using PC membrane-assisted sol electrophoretic deposition, single-crystal TiO₂ nanofiber arrays were also prepared.⁴² Furthermore, self-supported mixed transition metal oxide nanorod or nanofiber arrays, such as V₂O₅–TiO₂,⁴³ can also be prepared using a mixed metal oxide sol solution. Self-supported 1D nanostructures of mixed transition metal oxides were believed to possess better electrochromic performance compared to the corresponding 1D nanostructures of a single metal oxide component, because of the multiple types of color centers and enhanced redox kinetics derived from doping effects.

Nanotube arrays are another type of self-supported 1D nanostructure that can be prepared using PC and AAO membranes. Compared to nanofiber and/or nanorod arrays, nanotube arrays exhibit several electrochemical advantages. The internal cylindrical pores not only significantly increase the amount of surface electroactive sites to enhance optical modulation, but also obviously decrease the ion diffusion distance to improve the switching response speed, as well as provide more voids for efficient strain relaxation.^18,44,45 In addition, for metal oxide nanotubes, the crystal size of the electroactive materials is usually smaller than that of nanofibers or nanorods, which is believed to be feasible for ultrafast redox kinetics.¹⁸

The uniformly preferential growth of metal oxide precursors on the surface of nanochannels is a prerequisite for synthesizing nanotubes employing PC or AAO membranes. Atomic layer deposition (ALD) combining AAO membranes is a well-developed process to prepare self-supported metal oxide nanotube arrays.^46,47 In an ALD process, the gaseous precursors travel into the voids of the membrane and are preferentially absorbed on the surface of the nanochannels to decrease the system energy, producing nanotubes attached on the surface of the nanochannels (Fig. 3a). Furthermore, a surface layer covering on the membrane is usually produced in the ALD process. The removal of the membrane and the surface layer produces metal oxide nanotube arrays. Fig. 3b and c demonstrate typical SEM and TEM images of TiO₂ nanotube arrays prepared by the approach combining an AAO template and ALD.⁴⁶ The capability to prepare perfectly ordered nanotube arrays with a uniform pore size, length and wall thickness through the entire film is an outstanding advantage of this method. On the other hand, further excessive deposition of precursors in the nanochannels could make the voids fully filled, producing nanorod or nanofiber arrays. Preparation of nanotube arrays employing low concentration of metal oxide sol solution and AAO membranes is also due to the preferential absorption of colloid particles on the surface of the nanochannels.^44,48 Physical and chemical interactions between the metal oxide sol solution (precursors) and the templates, such as capillary forces, hydrophobic interactions and chemical bonds could also enable the successful preparation of nanotubes. These chemical and physical interactions were also the main forces for the production of nanotubes during the filling process of membranes using hydrolysis⁴⁹ and layer-by-layer (LBL) coating⁵⁰ methods.

Fig. 3 (a) Schematic diagram illustrating metal oxide nanotube arrays prepared using a membrane-assistant ALD process. (b and c) SEM and TEM images of TiO₂ nanotube arrays prepared by the approach combining an AAO template and ALD method. Reprinted from ref. 46. Copyright 2015 American Chemical Society. (d) Schematic diagram illustrating metal oxide nanotube arrays by using a membrane-assistant electrodeposition process. A sputtered Au layer is used as a conductive binder between the membrane and substrate. (e) SEM image of TiO₂ nanotube arrays prepared from an AAO template-assisted electrodeposition method. Reprinted from ref. 52. Copyright 2014 American Chemical Society.

Electrodeposition with the assistance of PC and AAO membranes under constant current density is another well-developed approach to fabricate self-supported metal oxide nanotube arrays. Because of the insulated characteristic of the membranes, a conductive layer needs to be coated on one surface of the membrane as a substrate to realize electrodeposition. Usually, a thin gold layer is sputtered or evaporated on the surface of the membrane to produce a conductive substrate. Researchers found that the deposition speed around the joint sites between the membrane and substrate was faster than other positions, which finally created nanotube arrays after the removal of the membrane (Fig. 3d).^51–53 A rational theory explaining the fast growth speed was the high electric field around the joint sites. Additionally, by-products produced during the electrodeposition, such as H⁺, OH⁻, and hydrogen bubbles, could also be beneficial factors for producing nanotubes. Fig. 3e shows the SEM image of TiO₂ nanotube arrays prepared by an AAO membrane-assisted galvanostatic deposition process.⁵² The bottom-up growth of TiO₂ led to nanotube arrays tightly bonded to the substrate; while uniform current density on the whole substrate ensured that the nanotubes showed relatively uniform diameter and length. In addition, when the deposition time is enough long or the deposition current density is rather high, excessive deposited metal oxide could be filled in the nanochannels, resulting in nanorod or nanofiber arrays.

Other 1D nanostructures, such as carbon nanotube arrays⁵⁴ and ZnO nanorod arrays⁵⁵ could also be used as templates to prepare metal oxide nanotube arrays. Notably, the mechanism using ZnO nanorod arrays as templates to prepare metal oxide nanotube arrays was sacrificial template-accelerated hydrolysis, a process of gradual dissolution of ZnO and slow deposition of amorphous metal oxide particles on the surface of the nanorods. The nanotubes prepared by this hydrolysis process were mesoporous and composed of several nanometer-sized crystals, leading the as-prepared nanotube arrays to demonstrate high optical modulation and fast switching response, as well as high-rate Li-ion storage, as shown in the following part about electrochromic energy storage devices.

Until now, scattered studies about the use of template-derived self-supported metal oxide nanotube arrays for electrochromism have been reported. However, it is believed that these nanostructured electrodes are desirable for high electrochromism because they have exhibited highly reversible and stable redox kinetics as reflected by the high-rate and long-term electrochemical energy storage performance.^56,57

4.2 Template-derived self-supported 1D conductive polymer nanostructures

Self-supported 1D conductive polymer nanostructures can also be fabricated employing PC and AAO membranes. The preparation process of these conductive polymer 1D nanostructures is similar to the synthesis process used to fabricate transition metal oxide 1D nanostructures. Electropolymerization is widely used to fill the nanochannels due to the advantages of bottom-up growth and ease of producing uniform conductive polymer 1D nanostructures. Similar to the AAO membrane-assisted electrodeposition of metal oxides, conductive polymer nanorod (or nanofiber) and nanotube arrays can be selectively synthesized by controlling the experimental parameters.⁵⁸Fig. 4a and b demonstrate SEM images of self-supported PPy nanofiber⁵⁹ and poly(3-methylthiophene) nanotube⁶⁰ arrays prepared by AAO membrane-assisted electropolymerization, respectively.

Fig. 4 (a) SEM image of PPy nanofiber arrays prepared by AAO-membrane assisted electropolymerization. Reprinted from ref. 59. Copyright 2016 by Wiley. (b) SEM image of poly(3-methylthiophene) nanotube arrays prepared by AAO-membrane assisted electropolymerization. Inset top-view SEM image indicating the hollow character. Reprinted from ref. 60. Copyright 2016 by Elsevier. (c) Schematic representation of an ultrafast electrochromic device based on PEDOT nanotube arrays. (d) Plots of reflectivity of the PEDOT electrochromic window monitored at λ = 530 nm for coloration and bleaching. Patterned letter “N” on the device to demonstrate the color contrast between coloration and bleaching states. Reprinted from ref. 65. Copyright 2005 by Wiley. (e) The nanotube array architecture, double-walled structure, and high conductivity in the electrode provide ion and electron “highways” and a high utilization rate of the electrode. Reprinted from ref. 80. Copyright 2013 American Chemical Society.

Besides the above mentioned electrochemical advantages of the nanostructures, 1D nanostructuring can give the conductive polymers two interesting merits for electrochromism. (1) It was found that the nanofibers and nanotubes fabricated with AAO membranes demonstrated higher electric conductivity than dense films, and the value increased with the decrease of the diameter for nanofibers (or nanorod, and nanowire) or the wall thickness for nanotubes.⁶¹ For example, detected by using four-probe technology on platinum micro-electrodes, when the diameter of PEDOT nanofibers decreased from 190 to 25 nm, the electric conductivity increased from ca. 11 to 550 S cm⁻¹.⁶²p-Toluene sulfonic acid doped PPy microtubes with 560–400 nm outer diameters exhibited poor conductivity of only 0.13–0.29 S cm⁻¹ (the inner pore diameter was 80 nm), while a high conductivity of 73 S cm⁻¹ was achieved when the outer thickness was about 130 nm.⁶³ Research implied that confined growth in the narrow nanochannels of the membranes altered the extent of disorder and polarons on the main chains of the conductive polymers, resulting in high electrical conductivity.^61–63 (2) The AAO membrane-derived conductive polymer 1D nanostructures demonstrated enhanced mechanical strength due to the extent of arrangement of the conductive polymer main chains.⁶¹ For instance, research indicated that PPy nanotubes with thicker walls demonstrated higher elastic modulus.⁶⁴ When the PPy nanotube thickness was between 20 and 16 nm (outer diameter of nanotube: between 100 and 70 nm), the elastic modulus of a single nanotube was around 5 GPa. When the nanotube wall thickness decreased under 16 nm (outer diameter of nanotube: <70 nm), the elastic modulus strongly increased with decreasing wall thickness. It reached a value close to 60 GPa for a thickness of 6.5 nm (outer diameter of nanotube: 35 nm). The high electric conductivity and mechanic strength are beneficial for fast electrochromic redox kinetics and long-term cycling performance.

Because of the intrinsic fast redox kinetics of conductive polymers and short ion diffusion distance, conductive polymer nanotube arrays could demonstrate fast color and optical switching response under a voltage stimulus. Fast color switching is a prerequisite, especially for display applications. The desired video rate of movies for human eyes is about 24 frames per second, i.e. about 42 ms to accomplish color switching is the minimum requirement. Fig. 4c illustrates arrays of electropolymerized PEDOT nanotubes attached on the inner wall surface of an AAO membrane for reflectance-mode electrochromic cells.⁶⁵ The AAO membrane was not removed to maintain the well-ordered arrangement of the nanotube arrays. Because of the short ion diffusion distance of the nanotube walls, these PEDOT nanotube arrays demonstrated fast redox doping/de-doping (bleaching/coloration) speeds with rapid switching of ca. several milliseconds to fulfill reflectivity modulation of ca. 23% at λ = 530 nm, as well as strong coloration contrast (between colorless and blue, contrast of 6) (Fig. 4d). The fast switching response and strong coloration contrast meant that the as-prepared PEDOT nanotube arrays could be used for display applications. In contrast, the PEDOT dense films exhibited a rather long switching response time of 1–2 s because of long ion diffusion distance.⁶⁶ By using a PC membrane attached on an Au/Cu bilayer metal foil and dissolving it after electropolymerization, Cho et al. prepared flexible reflectance-mode PEDOT electrochromic electrodes.⁶⁷ Although the removal of the membrane made some nanotubes become curved and fall down, resulting in a somewhat unexpected adverse influence on the electrochromic performance, this flexible electrode still demonstrated fast switching responses (20 and 30 ms for bleaching and coloration processes, respectively) with high reflectivity modulation of ca. 50% at λ = 600 nm. Attaching an AAO membrane on an ITO substrate and dissolving the template after electropolymerization produced PEDOT nanotube array electrodes for window-mode electrochromic devices.⁶⁸ The as-prepared PEDOT nanotube arrays showed obvious transparent/blue color contrast and high transmittance modulation of 45% at λ = 600 nm during the electrochemical doping/de-doping switching cycles. However, compared to the Au/Cu bilayer metal foil substrate, the relatively low electric conductivity of the FTO substrate resulted in slightly increased response times (50 and 70 ms for bleaching and coloration, respectively). Nanofiber, nanorod, and nanotube arrays of other conductive polymers (such as PANI, PPy and PT) for electrochromism were also successfully prepared by electropolymerization with AAO and PC membranes.^69–71

Liquid crystals,⁷² porous block copolymer films,⁷³ Au nanorod arrays,⁷⁴ silicon nanowire arrays,⁷⁵ nanopatterned polydimethylsiloxane (PDMS) molds,⁷⁶ anodic metal oxide nanotube arrays,⁷⁷ ZnO nanorod arrays,⁷⁸ and MnO₂ nanostructures⁷⁹ have also been used to fabricate self-supported conductive polymer 1D nanostructures with high electrochemical and electrochromic performance. In particular, the ZnO nanorod arrays were also used to fabricate mixed conductive polymer double-walled nanotube arrays (PPy@PANI hybrid NTAs) (Fig. 4e).⁸⁰ The PPy@PANI hybrid NTAs exhibited extraordinarily higher electrochemical activity and stability compared to the nanotube arrays with a single component due to the underlying synergistic effect.^81,82 Such mixed double-walled nanotube arrays bring interesting insights into the design of nanostructures for electrochromism.

5. Template-free self-supported 1D nanostructures for electrochromism

5.1 Solution process-deposited self-supported 1D metal oxide nanostructures

Solution deposition is one of the most widely used methods to fabricate electrochromic self-supported 1D metal oxide nanostructured electrodes. In a typical synthesis, a TCO substrate is immersed into a solution containing precursors. With thermodynamics and kinetics activations, metal oxide nuclei are produced at the low energy sites of TCO substrates. After enough time to fulfil crystal growth, self-supported 1D metal oxide nanostructures on TCO substrates are obtained. A subsequent annealing treatment is used to make the as-prepared metal oxides crystallized if fully crystalline samples are required. By varying the experimental details, the thickness of the nanostructured films and the areal density of the electrode materials could be optimized for high optical modulation and fast switching response. In particular, when used for transmittance-mode electrochromic devices, as-fabricated electrode films with high transparency are greatly expected because of the enhancement of the bleaching/coloration contrast.^4,5

The solution process-based preparation process of self-supported 1D metal oxide nanostructures typically includes two types: hydrothermal and ambient pressure methods. Hydrothermal (or solvothermal when the solvent is not water) growth of metal oxides on substrates usually operates in a sealed autoclave above ambient temperature (typically from 120 to 200 °C) in a solution with metal-element-containing precursors (such as metal salts and organometallic compounds). Using a suitable synthesis temperature and pressure, fully crystallized even single-crystal metal oxide nanostructures could be produced, so post heat treatment is often not necessary. The hydrothermal method possesses many advantages including: (1) relatively lower temperature than that of solid reactions; (2) ease of preparing single phase nanomaterials; (3) ease of preparing various morphologies by simply using different tunable experimental parameters, such as temperature, type and concentration of the precursors, and solvents; (4) ease of preparing uniformly doped materials. However, the high sensitivity of the morphology and phase of the products on the experimental details as well as the still not fully understood nanostructure growth mechanism under hydrothermal processes could bring some difficulties for the reproduction of the morphology and/or performance. Furthermore, the relatively not large volume of autoclaves may limit the areal size of the as-synthesized electrodes, which could bring some unpleasant difficulties for practical applications.

An ambient pressure solution processes to prepare nanostructures are conducted in open vessels. Similar to the hydrothermal process, metal oxide nanostructures are also deposited on the TCO substrates from the metal-element-containing precursor solution, but the synthesis temperature under ambient pressure is lower than that of the hydrothermal method due to the limitation of the boiling point of the solvents. The growth of metal oxide nanostructures in an ambient pressure solution process is derived from the hydrolysis of metal ions in the solution or crystallization process of the sol–gel-based solution. In addition, the as-deposited nanostructures usually either contain crystallographic and/or absorbed water, or are amorphous and poorly crystallized, or are metal hydroxides, indicating that a subsequent annealing is needed. However, nanostructured electrodes with a large size can be facilely prepared using an ambient pressure solution process because of no requirement on the high experimental temperature and pressure as well as the size of experimental vessels. In addition, flexible electrochromic film electrodes can be prepared using an ambient pressure solution method, because typically used flexible transparent conductive substrates are polymer films (such as polyethylene terephthalate (PET)) coated by ITO or FTO layers. These flexible substrates are not stable under hydrothermal and solid state reaction conditions.

5.1.1 Hydrothermal preparation of self-supported 1D nanostructures.
Tungsten oxide. Tungsten oxide (WO_x) is one of the most common transition metal oxides used to hydrothermally fabricate self-supported 1D nanostructures for a high cathodic electrochromic electrode. Crystalline WO_x materials are a category of oxide with perovskite-like structures mainly formed with WO₆ octahedra arranged in various corner-sharing or edge-sharing configurations.^3,10 Theoretical calculations on the electronic structures of WO₃ and M_xWO₃ (M = H, Li, Na) demonstrated that the electrochromism of WO_x was from small-ion insertion-derived band changes which had a strong relationship with the changes of oxygen ions of W–O–W bridges.^3,83,84 Depending on the different hydrothermal conditions for crystal growth, WO₆ octahedra can be organized in various sharing (edges, corners and planes) configurations in different tilting angles and distortions, generating a variety of WO_x materials with different phases and/or W : O stoichiometric ratios, such as tetragonal WO₃, orthorhombic WO₃, monoclinic WO₃, triclinic WO₃, cubic WO₃, W₃O₈, and W₁₈O₄₉.⁸³ Although the electrochromic color changes of these WO_x materials are the same (between blue and transparent), WO_x materials with different phases exhibit different electron band structures and lattice void types, giving rise to different ion insertion/extraction kinetics which leads to different optical modulation and switching response speeds. In addition, using different hydrothermal parameters can produce various self-supported WO_x nanostructured morphologies with high research significance. These morphologies include basic 1D nanostructures (such as nanorod,^85–88 nanofiber^89–93 and nanoribbon^94–96 arrays), complex 1D nanostructures (such as nano-/micro-flowers,^97,98 nanoforests,^99,100 honeycomb-like nanostructures,^101,102 and nest-like nanostructures¹⁰³) and other nanostructures (such as nanosheet,^104–106 nanobricke,¹⁰⁷ nanoplate,^108,109 and nanocuboid^110,111 arrays). Fabricating WO_x nanostructures with different phases and/or morphologies then investigating their electrochemical-optical performance is meaningful for not only the theoretical study of the hydrothermal crystallization mechanism and redox kinetics of WO_x, but also the promotion of practical application of nanostructured WO_x. Electroanalysis exhibited that self-supported 1D WO₃ nanostructures showed superior redox kinetics, including improved lithium-ion diffusion coefficient and enhanced pseudocapacitive effect which were beneficial for electrochromism.^84,109

Fig. 5a demonstrates typical SEM images of hydrothermally prepared WO₃ arrays of nanorods with a typical diameter of ∼80 nm and length of ∼1.1 μm on an FTO substrate.⁸⁶ Investigation of the morphology changes in the samples sequentially prepared under different reaction times indicated the WO₃ nanorod array growth mechanism. It was found that the substrate provided low energy sites for nucleation, and the shape of the initially grown WO₃ nanocrystals was irregular. With further growth, these irregular nanocrystals transformed into a rod shape to expose the low energy crystal faces to decrease the system energy. It was also noteworthy that the initial generated small WO₃ nanorods randomly stacked on the substrate, but the nanorods which were not perpendicular to the substrate were dissolved with the increase of experimental time also due to a decrease of system energy, finally producing WO₃ nanorod arrays grown on the FTO substrate, as shown in the inset cross-sectional view SEM image. Arrays of WO₃ nanorods with a diameter of ∼22 nm and length of ∼240 nm prepared using similar hydrothermal parameters exhibited a high transmittance modulation of 41.2% at λ = 632.8 nm and fast switching responses (3.6 s for coloration and 4.2 s for bleaching).⁸⁷ Using substrates with pre-deposited WO_x seeds is another efficient hydrothermal method to grow WO₃ nanorod arrays.^85,88 It exhibits several advantages, including controllable areal density of nanomaterials and more uniformly produced morphology. A seed-based cylindrical WO₃ nanorod array film exhibited a large transmittance modulation of 66% at λ = 632.8 nm and a high coloration efficiency of 106.8 cm² C⁻¹.⁸⁸

Fig. 5 (a) SEM images of self-supported WO₃ nanorod arrays. Reprinted from ref. 86. Copyright 2013 by Royal Society of Chemistry. (b) SEM images of self-supported WO₃ nanofiber arrays. Reprinted from ref. 92. Copyright 2014 Elsevier. (c) SEM images of self-supported W₁₈O₄₉ nanofiber arrays. Reprinted from ref. 93. Copyright 2016 Royal Society of Chemistry. (d and e) SEM image and electrochromic performance of a nanofiber stacked honeycomb WO₃ architecture. Reprinted from ref. 101. Copyright 2015 Royal Society of Chemistry.

Self-supported WO₃ nanofiber arrays were also prepared as electrochromic electrodes. Fig. 5b demonstrates typical SEM images of WO₃ nanofiber arrays on an FTO substrate prepared in hydrothermal solution containing similar reagents to that used to prepare nanorod arrays.⁹² The production of nanofibers can be regarded as the further growth stage of nanorods. As discussed above, the preferred orientation leads to the growth of nanorods along the length direction being faster than that along the diameter direction, finally giving rise to a nanofiber array morphology. The as-prepared WO₃ nanofiber arrays showed remarkable enhancement of the transmittance modulation in the visible spectrum (66.5% at λ = 633 nm) and IR region (73.8% and 57.7% at λ = 0.2 and 8 μm, respectively), as well as high cycling stability. To further improve the length-to-diameter ratio, especially decrease the scale of the diameter, chemicals restricting the growth along the diameter direction were added to the hydrothermal solution. For instance, Zhang et al. developed a sulfate-assisted hydrothermal method to prepare arrays of WO₃ nanofibers with a long length of ∼1.5 μm with a small diameter of 20–40 nm.⁹⁰ The as-prepared nanofiber films exhibited a high coloration efficiency of 102.8 cm² C⁻¹ and fast switching speeds (7.6 and 4.2 s for coloration and bleaching, respectively). Employing mixed solvents or pure organic solvent was another effective method to prepare nanofibers with high length/diameter ratio. Hung et al. prepared nanofiber network films on an FTO substrate in a water/isopropyl alcohol mixture solution.⁹¹ The fast Li-ion insertion/extraction kinetics with a Li-ion diffusion coefficient of 2.14 × 10⁻⁹ cm² s⁻¹ led to a desirable transmittance modulation of 57% at λ = 632 nm and a high coloration efficiency of 120.3 cm² C⁻¹. A solvothermal process seems to be more favorable to synthesize tungsten oxide nanofibers with a smaller diameter. Lu et al. prepared W₁₈O₄₉ nanofiber arrays with diameter <25 nm in polyethylene glycol as a solvent.⁸⁹ The as-prepared W₁₈O₄₉ nanostructures exhibited a high transmittance modulation of 49.64% at λ = 632.8 nm with high cycling stability (>3000 cycles). Lu et al. found that the solvothermal preparation in methanol solvent could further downsize the diameter of W₁₈O₄₉ nanofibers to ∼6 nm (Fig. 5c).⁹³ A complementary electrochromic device based on self-supported W₁₈O₄₉ nanofiber arrays and a Prussian blue thin film showed a high transmittance contrast (59.05% at λ = 632.8 nm) and fast switching response (coloration time of 6.9 s and a bleaching time of 1.2 s). A nanofiber stacked honeycomb WO₃ architecture with high Li⁺ diffusion coefficient of 3.091 × 10⁻⁹ cm² s⁻¹ was prepared by using a seed and sulfate co-assisted hydrothermal process (Fig. 5d).¹⁰¹ A assembled half-cell electrochromic device (bare TCO substrate as counter electrode) exhibited continuous color changes with high optical modulation of 60.74% at λ = 630 nm and fast switching response time (4.29 s for coloration and 3.38 s for bleaching) (Fig. 5e).

Titanium oxide. Titanium oxide (TiO_x) is a type of cathodic electrochromic material with color changes between blue (coloration) and transparent (bleached). TiO₂, the most common oxide type of hydrothermal or heat-treated product, has been regarded as a promising electrochromic material due to its stable crystal structure, low cost, high mechanical stability, environmental friendliness, safety, and fast charge transport and collection abilities.^112,113 TiO₂ naturally exhibits three common types of polymorphs, i.e., rutile, anatase, and brookite. All these three TiO₂ phases use Ti–O octahedrons as the fundamental building block and share hollow lattice channels for facile ion (H⁺ and Li⁺) insertion/extraction. In addition, the significantly improved electron diffusion coefficient^114,115 and enhanced surface redox contribution¹¹⁶ in TiO₂ nanorods and nanofibers make the hydrothermally fabricated TiO₂ self-supported 1D nanostructures expected to be high-performance electrochromic electrode films.

Using delicately designed experimental parameters, Liu et al. reported a three-step hydrothermal process to prepare arrays of TiO₂ single crystalline anatase nanofibers oriented in the [100] direction with a diameter of 105 ± 10 nm and length of 12.16 ± 0.56 μm on FTO substrates.¹¹⁷ Such [100] direction oriented nanofibers were believed to possess fast electron transport.^112,113 Anatase TiO₂ nanofiber arrays with a high-porosity cross-linked geometry directly grown onto FTO substrates were prepared through hydrothermal processes under mild alkali conditions.¹¹⁸ The TiO₂ nanofiber array-based half-cell electrochromic device demonstrated a desirable transmittance change of 28.20% at λ = 600 nm and an acceptable coloration efficiency of 13.87 cm² C⁻¹ with fast switching response (11.3 s for coloration and 14.3 s for bleaching). Electroanalysis indicated that the enhanced Li⁺ diffusion coefficient is an important feature. In addition, a low refractive index of 1.37 made the as-prepared TiO₂ nanofiber arrays become ideal electrode films for transmittance-mode electrochromic devices. Qiang et al. hydrothermally fabricated densely packed rutile TiO₂ nanorod arrays on the FTO substrate.¹¹⁹ The high areal density of nanorods led the electrode film to exhibit large transmittance modulation in the visible spectrum range with a maximum value of approximately 64% at λ = 600 nm. Furthermore, a self-powered system integrating an electrochromic device and a dye-sensitized solar cell was also assembled and tested. To further improve the electrochromic performance, Liu et al. fabricated a micro-/nanostructured film of self-supported rutile TiO₂ nanorod arrays decorated by anatase TiO₂ nanoparticles (TiO₂ NR-NP).¹²⁰ This micro-/nano-structured film simultaneously possessed the high electron transport and mechanical stability properties from the nanorod arrays, and fast and reversible redox kinetics from the small decorated nanoparticles, giving rise to enhanced transmittance modulation as well as accelerated switching responses, compared to pure TiO₂ nanorod arrays (TiO₂ NR) (Fig. 6a). Using a two-step hydrothermal process and pre-fabricated self-supported TiO₂ nanofiber arrays as skeletons, hierarchical TiO₂ nanofiber arrays with densely-packed and omnidirectional branches were synthesized (Fig. 6b).¹²¹ A typical synthesis process was via high-concentration TiCl₄ treatment of upright backbone TiO₂ nanofibers to produce seeds on the surface followed by hydrothermal growth, giving rise to the growth of secondary TiO₂ nanobranches (such as nanoneedles and nanosheets) on the TiO₂ nanofiber backbone and in all directions (Fig. 6c). Such a preparation method brings new insights to prepare novel hierarchical nanofiber arrays where the nanofiber and nanobranches could be the same oxides or different types like WO₃ nanostructures on a TiO₂ nanofiber backbone. Hierarchical nanofiber arrays are desirable for electrochromic applications.

Fig. 6 (a) Comparison of the transmittance modulation in the visible spectra of the TiO₂ NW and NW-NP films. Inset figures demonstrate the coloration and bleaching colors of the TiO₂ NW-NP film. Reprinted from ref. 120. Copyright 2014 Royal Society of Chemistry. (b and c) Schematic diagram illustrating preparation and SEM images of densely-branched TiO₂ NWs. Reprinted from ref. 121. Copyright 2013 Royal Society of Chemistry.

Nickel oxide and cobalt oxide. Nickel oxide (NiO_x) and cobalt oxide (CoO_x) are two representative anodic electrochromic metal oxides. Crystalline NiO_x and CoO_x are formed by NiO₆ or CoO₆ octahedra connected by sharing common corners and/or by sharing common edges.¹⁰ Because of the high electrical conductivity of NiO_x and CoO_x,^122,123 as well as their stable lattice structure, crystalline NiO_x and CoO_x nanomaterials exhibit faster coloration/bleaching switching and higher cycling stability compared to many other electrochromic metal oxides.

NiO, the common electrochromic nickel oxide, has been widely investigated in alkali solutions. In electrochromic devices, NiO has been successfully explored as a competent anodic counter electrode (also called an ion storage electrode) in conjunction with WO₃ as a cathodic working electrode.¹²⁴ However, practical applications of NiO as a promising electrochromic material are still difficult because of the low color contrast, poor cycling durability and unsatisfactory ion storage capacity when the NiO is in bulk particles or a dense film. Fabrication of self-supported 1D NiO nanostructures on the TCO substrate is an ideal method to solve these problems. Typically, the initial hydrothermal crystalline products are hydrous nickel oxides or nickel hydroxides; an annealing treatment is needed to transform the hydrothermal products to NiO. However, because the initial generated hydrothermally synthesized crystalline products on TCO substrates were usually in morphologies of nanosheet arrays or nanosheet-stacked complex nanoarchitectures (such as rose-like), the annealing-produced NiO nanostructures were inevitably in nanosheet arrays or nanosheet-stacked complex nanoarchitectures.^125–129 These NiO nanosheet-based architectures exhibited desirable electrochromic performance. For instance, mesoporous nickel oxide nanosheet arrays demonstrated a high transmittance modulation of 77% at λ = 550 nm, fast switching response (2 and 2.5 s for coloration and bleaching, respectively), and highly stable cycling performance of negligible degradation after 3000 cycles.¹²⁷ But the hydrothermal fabrication of self-supported NiO 1D nanostructures is still a challenge.

On the other hand, hydrothermal fabrication of self-supported Co₃O₄ (the common electrochromic cobalt oxide) 1D nanostructures is much easier. Additionally, as demonstrated in the literature reports about electrochemical energy storage applications using Co₃O₄ nanostructures, the better rate and cycling performance of self-supported Co₃O₄ 1D nanostructures than many other Co₃O₄ nanoarchitectures indicated their superiority of highly reversible and fast redox kinetics,^130,131 implying the possible achievement of satisfactory color saturation for fast switching response. Self-supported Co₃O₄ nanorod and nanofiber arrays can be hydrothermally grown on TCO substrates.^132–135 Furthermore, by pre-preparing seeds on the substrates, uniform morphology and controlled areal density of Co₃O₄ nanofiber or nanorod arrays can be prepared.^136,137 Xia et al. fabricated self-supported Co₃O₄ nanorod arrays on an ITO substrate using a seed-assisted hydrothermal process.¹³⁶ The authors delicately controlled the diameter and length of the Co₃O₄ nanorods to optimize the electrochromic properties. The arrays of Co₃O₄ nanorods with diameters varying from 70 to 100 nm and a length of about 550 nm showed obvious color changes from light brown (bleached) to black (colored) with fast switching response (1.8 and 1.4 s for colored and bleached states, respectively) (Fig. 7a and b). Furthermore, this Co₃O₄ nanorod arrays also demonstrated quite good cycling stability with a transmittance modulation maintenance proportion of 88% after 4500 cycles. Such high satisfactory color saturation achieved in short switching time and high cycling stability was consistent with the high-rate electrochemical energy storage performance of self-supported Co₃O₄ 1D nanostructures.^130,131

Fig. 7 (a) SEM image of Co₃O₄ nanorod arrays on an ITO substrate prepared by a seed-assisted hydrothermal process. Inset digital photos demonstrate the color contrast between coloration and bleaching. (b) Transmittance modulation response of Co₃O₄ nanorod arrays under alternative voltages. Reprinted from ref. 136. Copyright 2010 Elsevier.

Vanadium oxide and molybdenum oxide. Crystalline vanadium oxide and molybdenum oxide have been widely studied for electrochemical redox-based applications because of their attractive layered structure.¹³⁸ The two-dimensional crystalline sheet structures are formed by MO₆ (VO₆ or MoO₆) octahedra sharing corners and/or edges. Due to the layered structure, vanadium oxide and molybdenum oxide can demonstrate fast ion insertion/extraction kinetics in the interlamination of (001) planes. In addition, as typical extrinsic pseudocapacitor materials, 1D vanadium oxide and molybdenum oxide nanomaterials could exhibit enhanced pseudocapacitive effect, high Li-ion diffusion coefficient, and short characteristic relaxation process,¹³⁹ which are beneficial for high optical modulation and fast switching response.

V₂O₅, as the common hydrothermally prepared vanadia type showing both cathodic and anodic coloration, has been widely investigated as a counter electrode material in electrochromic devices,^10,139,140 because of not only its unique electrochromic performance but also its high Li-ion storage capacity. 1D V₂O₅ nanostructures can be facilely prepared using a hydrothermal method due to the strong preferential growth derived from the obvious lattice anisotropy of the layer crystalline structure.^141–143 Chu et al. hydrothermally prepared V₂O₅ nanorod array films on FTO substrates and investigated their electrochromic performance (Fig. 8a).¹⁴⁴ It was found that the V₂O₅ nanorod array films demonstrated obvious absorbance modulation in the visible spectrum range and vivid color changes between pale blue (cathodic coloration) and yellow-green (anodic coloration) (Fig. 8b).

Fig. 8 (a) SEM image of V₂O₅ nanorod arrays directly grown on an FTO substrate by using a hydrothermal method. (b) Optical modulation of V₂O₅ nanorod arrays derived from a Li-ion insertion/extraction process. Inset digital photos and plot of absorbance vs. time under alternative voltages indicate the electrochromic color contrast and switching response performance. Reprinted from ref. 144. Copyright 2016 Elsevier.

Molybdenum oxide can be used as a cathodic electrochromic material with color changes between transparent and blue.^10,138 MoO₃ is the most common form of hydrothermally prepared molybdenum oxide. To date, self-supported MoO₃ nanorod arrays have been successfully prepared using hydrothermal methods.^145,146 MoO₃ nanorods, nanofibers, and nanobelts powders were also hydrothermally prepared for high-performance electrochemical energy storage and electrochromism.^147–149 However, research about the fabrication and investigation of self-supported MoO₃ 1D nanostructures on TCO substrates for electrochromism is still rare.

Doped metal oxides. Doping the metal oxides with guest ions or molecules is an effective approach to affect the coordination of host metal-oxide octahedra, generating a new electronic structure or band structure which could bring unusual electrochromic performance differing from pure host metal oxides.^3,150–152 In addition, the guest ions and molecules could bring lattice distortion and defects, affecting the diffusion behaviors of inserted ions in the host metal oxides.^{151,153–156} Research has indicated that the widened interlamellar spacing of V₂O₅, MoO₃ and TiO₂ not only facilely accelerated the Li-ion insertion speed but also made the insertion of metal ions with larger radius (such as Na⁺ and K⁺) possible.^{151,154–156}

A hydrothermal process is an effective and facile method to synthesize 1D uniformly doped metal oxide nanostructures, broadening the research horizon of electrochromism. For instance, the Ni-doping significantly increased the optical modulation of WO₃ nanorods.¹⁵⁷ When the Ni atomic concentration was 1.5%, the Ni-doped WO₃ nanorods exhibited maximum transmittance modulation and charge density of 86.0% (at λ = 600 nm) and 24.6 mC cm⁻², compared to the corresponding relatively low values of pure WO₃ nanorods (50.9% and 18.2 mC cm⁻², respectively). In addition, the switching response time was shortened to around 2 s. The detected increase of electrical conductivity was believed to be an important feature. Doping also can significantly enhance the optical modulation of metal oxides in the NIR and IR spectrum range. Zhou et al. reported that 2% Mo-doped WO₃ nanofibers showed a high transmittance modulation in the visible and NIR spectrum range (56.7%, 83% and 48.5% at λ = 750 nm, 1600 nm and 10 μm, respectively), while the pure WO₃ nanofibers only exhibited 44.4%, 52.6% and 25.1% at the three above mentioned wavelengths.¹⁵⁸ Additionally, the enhanced electrical conductivity accelerated the switching speed to around 4 s. Nevertheless until now, the majority of these hydrothermally prepared 1D doped metal oxide nanostructures were in the powder form, and the fabrication of electrochromic 1D metal oxide nanostructures directly grown on TCO substrates for electrochromism deserves to be investigated.

5.1.2 Ambient pressure solution preparation of self-supported 1D nanostructures. Chemical bath deposition (CBD) is a typical and effective ambient pressure solution method to prepare metal oxide nanostructures on solid substrates. The CBD process refers to the thermohydrolysis (or “forced hydrolysis”) of metal salts involving two steps, nucleation and particle growth at low temperatures.¹⁵⁹ Typical electrochromic metal oxide films or nanostructures can be easily prepared on the substrates using CBD,^160–163 while the synthesis of well-defined self-supported 1D nanostructures seems to be difficult. For instance, electrochromic NiO nanostructures prepared by CBD are usually in the morphology of nanosheet arrays on TCO substrates.^164,165 However, self-supported 1D nanostructures of some metal oxides can be obtained by using delicately designed experimental reagents and procedures. Hosono et al. found that employing a mixed solution of water–methanol as solvents could produce cobalt oxide nanorod arrays on glass slides, compared to nanosheet arrays prepared in a solvent of water.¹⁶⁶ The reason that nanorods were produced in the water–methanol solvent was due to the increased amount of dissolved oxygen molecules, which converted some of the divalent cobalt ions into trivalent cobalt ions through oxidation. Li et al. proposed a modified CBD process to synthesize vertically aligned H₂Ti₅O₁₁·3H₂O nanowire arrays on arbitrary substrates of stainless steel, glass, silicon wafer and carbon cloth at a low temperature of 80 °C and in an open atmosphere (Fig. 9a and b).¹⁶⁷ Anatase TiO₂ nanowire arrays can be obtained by a further annealing treatment. The authors attributed the successful preparation of H₂Ti₅O₁₁·3H₂O nanowire arrays in the modified CBD process to the addition of an appropriate concentration of NaOH in the solution. In the presence of NaOH, Ti(IV) derived from the Ti–H₂O₂ interaction was transformed to protonated titanate nanowires after an alkali-hydrothermal-like process followed by a proton exchange, giving rise to H₂Ti₅O₁₁·3H₂O nanowires at a low temperature. Dhara et al. employed a seed-based chemical bath deposition to successfully synthesize single-crystal MoO₃ nanorod arrays vertically grown on glass (Fig. 9c).¹⁶⁸ The pre-coated seed layer was a key factor to obtain MoO₃ nanorods vertical to the substrate because they defined the vertical direction as the only preferred growth direction for MoO₃ nanorods (Fig. 9d). It is meaningful to note that these CBD methods were substrate-independent, providing the possibility of using flexible ITO/PET films as substrates.

Fig. 9 (a and b) Cross-sectional and top-view SEM images of self-supported H₂Ti₅O₁₁·3H₂O nanowire arrays prepared by a modified CBD method. Reprinted from ref. 167. Copyright 2014 Royal Society of Chemistry. (c and d) Top-view SEM image of MoO₃ nanorod arrays and the corresponding schematic diagram illustrating the preparation process of a seed-based CBD. Reprinted from ref. 168. Copyright 2014 Royal Society of Chemistry.

5.2 Vapor process-deposited self-supported 1D metal oxide nanostructures

Vapor process deposition is a direct and bottom up method to prepare self-supported 1D metal oxide nanostructures on substrates.¹⁶⁹ Because the metal oxide nanostructures are grown from gaseous precursors, solid substrates are needed to provide low energy sites for nucleation and subsequent crystal growth. In a vapor deposition process, thermal activation under high temperature is usually employed to facilitate crystal nucleation and growth on the TCO substrate. The high experimental temperature makes the as-prepared self-supported 1D metal oxide nanostructures fully crystallized, even in single-crystal nature. Similar to a crystal growth mechanism under hydrothermal conditions, the production of 1D crystalline metal oxide nanostructures on TCO substrates is naturally formed due to the preferential direction of the anisotropic crystal lattice with or without external physical forces. In addition, employing catalysts or surface modification on substrates can affect the nucleation and crystal growth processes, controlling the morphology or phase of the produced self-supported 1D metal oxide nanostructures.^170,171 Typically, vapor process deposition can be categorized into physical and chemical vapor deposition (PVD and CVD) processes.

Physical vapor deposition. PVD is a process in which the solid metal oxide and substrate are placed in a chamber; meanwhile, the solid metal oxide is volatilized into a vapor phase precursor and then transferred onto the surface of the solid substrate with lower temperature. These precursors condense on the surface of the substrate followed by nucleation and crystal growth processes, generating self-supported 1D metal oxide nanostructures (Fig. 10a). Based on the methods employed to volatilize solid metal oxides or to assist the metal oxide nucleation and crystal growth, PVD technologies include thermal evaporation, sputtering, laser ablation, plasma assistant, confinement growth, lithography, glancing angle deposition, etc.¹⁶⁹ All these PVD technologies can be employed to fabricate self-supported 1D metal oxide nanostructures.^172–178 To prepare electrochromic film electrodes, FTO or ITO substrates are usually used. Cheng et al. successfully prepared V₂O₅ arrays of nanowires on ITO substrates by thermal evaporation PVD.¹⁷⁶ The diameter of the V₂O₅ nanowires was from 10 to 100 nm and the length was up to several hundred nanometers. Between −1 and +1 V (vs. Ag/AgCl) in 1 M LiClO₄/propylene carbonate (PC), V₂O₅ nanowire arrays exhibited vivid color changes between pale blue (−1 V) and yellow (+1 V) with the maximum transmittance modulation of 37.4% at λ = 415 nm and fast switching response (6 and 5 s for cathodic and anodic coloration, respectively). The as-deposited V₂O₅ nanowires demonstrated high structural stability. The morphology and redox current response of the nanowire arrays did not alter after 1000 Li-ion insertion/extraction cycles observed by SEM and CV examinations. Liao et al. reported the successful preparation of W₁₈O₄₉ nanofiber arrays on the ITO substrate using a similar PVD process.¹⁷⁷ The average diameter and length of the nanofibers were about 52 ± 13 nm and up to 5 μm, respectively. The high areal density of the nanofibers (about 1.68 × 10¹⁰ per square centimeter) led to a high transmittance modulation in the visible spectrum range with a maximum value of 34.5% at λ = 700 nm. The W₁₈O₄₉ nanofiber arrays also demonstrated fast switching response of less than 2 s both for coloration and bleaching with long-term cycling stability. Xiao et al. prepared three-dimensional, high-porous, and oriented WO₃ nanocolumn layers on ITO/PET and FTO substrates employing a glancing angle PVD process when the WO₃ layer was prepared with a deposition angle of 75° (Fig. 10b).¹⁷⁸ By dynamically varying the substrate incline and revolution regarding the incoming tungsten oxide vapor in the glancing angle PVD process, the WO₃ layer morphological and physical properties were precisely adjusted. The WO₃ nanocolumn film anchoring ITO/PET substrate exhibited high flexibility and high transmittance modulation of 52% at λ = 700 nm with fast switching response (4.2 and 7.8 s for coloration and bleaching, respectively). It is interesting to note that the as-deposited WO₃ nanocolumns can improve the broadband antireflective performance of the substrates due to the high porosity of WO₃ nanocolumns, which can increase coloration/bleaching contrast.

Fig. 10 (a) Schematic diagram of tube furnace for synthesis of self-supported 1D metal oxide nanostructures by a PVD process. (b) Cross-sectional SEM image of three-dimensional, high-porous, and oriented WO₃ nanocolumn layer prepared by a glancing angle PVD method. Left inset figure demonstrates the optical transmittance spectrum of bare ITO and ITO/WO₃ films at normal incidence. Right inset figure demonstrates the transmittance modulation switching response curves of dense and three-dimensionally porous PET/ITO/WO₃ films. Reprinted from ref. 178. Copyright 2016 American Chemical Society. (c) Electrochemical and electrochromic performance of self-supported NiO nanorods prepared by the HFMOVD method. Inset figure is the SEM image of the NiO sample. Reprinted from ref. 184. Copyright 2013 Elsevier. (d) Transmittance modulation of self-supported brookite TiO₂ nanoneedles prepared by the HFMOVD method. Inset figure is the SEM image of the TiO₂ sample. Reprinted from ref. 185. Copyright 2015 Elsevier.

Chemical vapor deposition. CVD is a process in which the condensed precursors for metal oxide crystal growth are derived from gaseous chemical reactions. Thus, the ability to produce compounds with high purity is a unique advantage of CVD over PVD. Based on differences in the used reactants and assisting energy sources, CVD technologies can be divided into various categories such as atmospheric-pressure, hot-filament, metal–organic, microwave-plasma, plasma-enhanced, and photo-assisted types, etc.¹⁶⁹ To prepare 1D metal oxide nanostructures for electrochromism, an oxygen atmosphere is a vital factor. Until now, various electrochromic self-supported 1D nanostructures of metal oxides (such as WO₃,^179,180 Ta₂O₅,¹⁸¹ MoO₃,¹⁸² V₂O₅,¹⁸³ NiO,¹⁸⁴ and TiO₂¹⁸⁵) on TCO substrates have been fabricated by a hot-filament metal-oxide vapor deposition (HFMOVD) technique. In the HFMOVD process, the metal source is placed on a heater in a vacuum chamber. The substrate is placed apart from the heater at an appropriate distance. When the metal source is heated to be volatilized, the volatilized metal atoms react with the residual oxygen molecules in the chamber or low oxygen flow supplied by external equipment, creating highly active metal–oxygen-containing gaseous precursors. These precursors travel to the TCO substrates, followed by metal oxide 1D crystal growth. The crystal growth mechanism in the HFMOVD process is quite different from other methods, resulting in the preparation of some unique self-supported 1D metal oxide nanostructures for high electrochromism. For instance, self-supported single-crystal NiO nanorod arrays which were hard to synthesize using a hydrothermal method, were simply prepared on an ITO substrate using the HFMOVD method (inset SEM image in Fig. 10c).¹⁸⁴ The NiO nanorods exhibited a length of ∼500 nm and a width of ∼100 nm, with an areal density of ∼100 nanorods per square micrometer. In 0.5 M KOH electrolyte, the NiO nanorod arrays exhibited outstanding electrochromic properties, including great optical transmittance modulation (60% at λ = 630 nm) and fast coloration and bleaching times (1.55 and 1.22 s, respectively). Electrochemical analysis indicated that the NiO nanorod arrays showed fast redox kinetics with a large diffusion coefficient of ∼6.33 × 10⁻⁸ cm² s⁻¹. 1D TiO₂ nanostructures (nanoneedles with length of ∼650 nm and width of ∼30 nm) in a pure brookite phase which is really hard to prepare by common synthesis processes such as hydrothermal methods and electrodeposition, were successfully prepared by the HFMOVD process (inset SEM image in Fig. 10d).¹⁸⁵ Besides the high transmittance modulation and fast switching response, the TiO₂ nanoneedles exhibited large optical density of 0.85, great coloration efficiency of 226 cm² C⁻¹, and high cycling retention of ∼99% after 250 cycles in a Li-ion organic electrolyte. Also, a fair diffusion coefficient of 1.56 × 10⁻¹¹ cm² s⁻¹ was believed to be an advantageous electrochemical feature.

Aerosol-assisted chemical vapor deposition (AACVD) is another widely used CVD process employing metal-containing compounds as sources to prepare self-supported 1D metal oxide nanostructures for high electrochromism. In AACVD, metal-containing reactants are transported in solution as a mist, and therefore, compound volatility is less critical than in conventional CVD, which has an important influence on the morphology of the products.¹⁸⁶ Various electrochromic self-supported 1D metal oxide nanostructures, such as NiO nanorod arrays,¹⁸⁶ WO₃ nanoneedle arrays¹⁸⁷ and WO₃ nanorod arrays,¹⁸⁸ were prepared using AACVD on TCO substrates.

5.3 Anodic oxidation prepared self-supported metal oxide nanotube arrays

Fabrication of metal oxide nanotube arrays by anodic oxidation on metal foils has received enormous interest due to the wide applications of metal oxide nanotube arrays in the fields of energy storage and conversion, sensors, and biomedical applications.¹⁸⁹ The anodic formation mechanism of metal oxide nanotubes on the clean surface of metal foil in an electrolyte has been previously detailed.¹⁸⁹ Typically, once a sufficient potential is applied, the metal atoms (M) on the surface are oxidized into metal ions (Mⁿ⁺) while the O²⁻ are generated by deprotonation of water in the electrolyte, producing a thin barrier metal oxide (MO_x) layer on the foil surface. Field-assisted dissolution of the metal oxide occurs at the oxide/electrolyte interface as the M–O bonds undergo polarization and are weakened. In the presence of some active ions (such as F⁻) which can react with Mⁿ⁺ forming water-soluble complexes, small pits in the oxide layer form. In the pits, O²⁻ ions continue to diffuse through the oxide layer and then react with the metal to produce fresh metal oxide at the bottom. When the rate of oxide growth at the metal/metal oxide interface and the rate of oxide dissolution at the metal oxide/electrolyte interface ultimately become equal, amorphous nanowalls between the pits are formed. The metal oxide growth/dissolution process moves further into the metal making the pore deeper, finally generating nanotube arrays on the foil. An annealing treatment is needed to make the amorphous products fully crystallized (Fig. 11a). Furthermore, anodic oxidation of alloy foils is a facile way to prepare doped (or mixed) metal oxide nanotube arrays, which can exhibit unusual electrochemical kinetics and/or new electrochromism.^189,190 Mott–Schottky analysis indicated that the doped crystalline metal oxide nanotube arrays demonstrated higher electrical conductivity than the pure crystalline metal oxide nanotube arrays.¹⁹⁰

Fig. 11 (a) Typical SEM image of anodic TiO₂ nanotube arrays. (b) Reflectance modulation curves of mixed Ti–V–O nanotube arrays with different vanadium atom concentrations. Reprinted from ref. 196. Copyright 2011 Elsevier. (c) Schematic diagram illustrating preparation of anodic metal oxide nanotube arrays directly grown on TCO substrates. (d) The influence of coated amount of MoO₃ on the transmittance modulation of anodized Nb₂O₅/electrodeposited α-MoO₃ binary films. Reprinted from ref. 208. Copyright 2014 American Chemical Society. (e and f) Digital photos of free anodic TiO₂ nanotube arrays transferred on the FTO substrate and their transmittance modulation performance. Reprinted from ref. 213. Copyright 2016 Elsevier.

The initial electrochromic applications of anodic metal oxide nanotube arrays were in reflectance-mode because of the difficulty in peeling off the nanotube arrays from the metal substrate.^191–198 For instance, H⁺ insertion/extraction led to a reversible reflectance modulation of over 50% and a fast switching response of 2 s at λ = 480 nm in the TiO₂ nanotube arrays (film thickness of 1 ± 0.1 μm with an individual tube diameter of 100 ± 10 nm and a tube wall thickness of 10 ± 2 nm),¹⁹¹ while Li⁺ insertion/extraction could result in a reflectance modulation of approximately 35% at the same wavelength with a switching time of 5 s.¹⁹² CV measurements indicated that the TiO₂ nanotube arrays showed improved H⁺ and Li⁺ chemical diffusion coefficients. Yang et al. reported that the V₂O₅ nanotube arrays with a length of 20 μm and pore diameter of 12 nm could demonstrate a reflectance modulation of over 30% with vivid color change from yellow to green and then dark accompanied by an increased amount of inserted Li-ions.¹⁹³ In addition, the as-prepared V₂O₅ nanotube arrays also exhibited high cycling stability with negligible degradation of reflectance contrast after 250 cycles.

Anodic oxidation of alloy foils produced doped metal oxide (or mixed metal oxides) nanotube arrays. Ghicov et al. synthesized tungsten-doped titanium oxide nanotube arrays using Ti–W alloys.¹⁹⁴ It was found that the doped products exhibited higher reflectance modulation and charge density than the pure TiO₂ nanotube arrays. Ghicov et al. also found that the Ti–Nb–O nanotube arrays anodically prepared from β-Ti45Nb alloy exhibited a pure phase of anatase TiO₂, while the lattice interlamellar spacing was widened due to the Nb-doping.¹⁹⁵ Such widened lattice interlamellar spacing facilitated the Li⁺ insertion/extraction process, leading to increased reflectance modulation of approximately 80% at λ = 600 nm and high charge density of 126.1 mC cm⁻¹, compared to the corresponding values of approximately 50% and 76.4 mC cm⁻¹ for pure TiO₂ nanotube arrays. Yang et al. anodically synthesized mixed TiO₂–V₂O₅ nanotube arrays from Ti–V alloys.¹⁹⁶ Maximum values of reflectance modulation and charge density were achieved for the nanotube arrays prepared from a Ti–V alloy with 3 at% vanadium during the Li-ion insertion/extraction process (Fig. 11b). In addition, the doping of vanadium also significantly enhanced the color contrast by increasing color saturation under cathodic polarization. Anodically prepared mixed Ti–Mo–O and W–Ta–O nanotube arrays also demonstrated enhanced reflectance modulation performance.^197,198 In addition, the nanotube arrays provided large surface area for decoration of a second type of electrochromic material, such as silver phosphate crystals¹⁹⁹ and WO₃ nanocrystals,²⁰⁰ resulting in enhanced reflectance modulation and multicolor changes. Furthermore, mesoporous NiO²⁰¹ and WO₃²⁰² films were also prepared by anodic oxidation processes for reflectance-mode electrochromic electrodes.

Anodic oxidation of metal layers deposited on the TCO substrates is an easy approach to directly prepare nanotube array electrodes for transmittance-mode electrochromic devices (Fig. 11c). The typical process to deposit metal layers was magnetron sputtering, and meanwhile heat-stable TCOs, such as ITO and FTO, were used as substrates.^203,204 Barredo-Damas et al. prepared anodic Nb₂O₅ nanotube arrays with a thickness of 1 μm from a thin Nb-layer on FTO glass.²⁰⁵ The as-prepared nanotube arrays exhibited a high transmittance modulation of ca. 90% at λ = 600 nm with obvious color changes in 0.1 M HClO₄. Yao et al. synthesized self-supported TiO₂ nanotube arrays for electrochromic applications by anodic oxidation of an RF sputtered titanium layer on an FTO substrate.²⁰⁶ To further increase the optical modulation, α-MoO₃ of ∼5 to 15 nm thickness was electrodeposited on the wall surfaces of the TiO₂ nanotube arrays. The MoO₃/TiO₂ system showed a 4-fold increase in optical density over bare TiO₂ when the thickness of the MoO₃ coating was optimized. The enhancement was ascribed to (I) the desirable electron band coupling effect between α-MoO₃ and TiO₂ for feasible charge carrier transfer, (II) the increased amount of inserted Li-ions derived from the layered structure of α-MoO₃ and (III) enhanced electron transport derived from the high electrical conductivity of α-MoO₃ layers acting as efficient pathways for charge carrier transfer.^206,207 Yao et al. also synthesized anodized Nb₂O₅/electrodeposited α-MoO₃ binary films for high electrochromism.^208,209 The thickness of the MoO₃ layer was controlled by different electrodeposition cycles (a cycle was a chronoamperometric deposition of 60 s). It was found that the increased thickness of the MoO₃ layer led to an increase of transmittance modulation, reaching a favorable value of approximately 45% at λ = 550 nm between ±1 V in a 0.1 M LiClO₄/PC electrolyte (Fig. 11d). WO₃ mesoporous film electrodes were prepared by anodic oxidation treatment on the corresponding metal layers deposited on FTO substrates.²¹⁰ These films also exhibited high transmittance modulation with fast switching response.

Free TiO₂ nanotube array membranes can be peeled off from the Ti substrates using a modified anodic oxidation process.^211,212 Transfer of these membranes on TCO substrates also achieves transmittance-mode electrochromic electrodes. Lv et al. developed a two-step anodization process to prepare free TiO₂ nanotube array membranes with various thicknesses of 539, 1200 and 1930 nm, and attached them on FTO substrates for electrochromic tests (Fig. 11e).²¹³ It was found that the thickness had a strong effect on the transmittance modulation in the whole visible spectrum range (Fig. 11f). Typically, the 539 nm-thick TiO₂ nanotube array membrane exhibited a transmittance modulation of 65% at λ = 750 nm. In addition, the TiO₂ membrane also exhibited high cycling stability with a high maintained transmittance modulation of 57% and a negligible degradation of coloration efficiency after 1000 cycles. Furthermore, decoration of WO₃ nanocrystals on the surface of TiO₂ nanotube arrays further increased the transmittance modulation and color contrast between coloration and bleaching states.²¹⁴

5.4 Other processes for self-supported metal oxide 1D nanostructures

Phase transformation-induced self-supported 1D nanostructures. For metal oxides demonstrating a layered crystalline structure, annealing treatment on their amorphous or poorly crystallized metal oxide films could produce self-supported 1D nanostructures. By annealing treatment on the 3D macroporous amorphous vanadium oxide films prepared by electrodeposition with the assistance of polystyrene colloidal crystals as templates (Fig. 12a),^215–217 V₂O₅ nanorod architectures^218,219 and a V₂O₅ nanofiber grassland²²⁰ can be prepared by using different annealing temperatures. Colloidal sphere-assisted heterogeneous nucleation during electrodeposition and the anisotropic bonding of the V₂O₅-layered structure were two important factors for the generation of the V₂O₅ nanorod architecture at 350 °C (Fig. 12b), while for the growth of the V₂O₅ nanofiber grassland morphology obtained under a higher annealing temperature of 450 °C (Fig. 12c), coalescence of adjacent nanorods with a similar crystallographic orientation, atom rearrangement, and Ostwald ripening were responsible for the production of nanofibers under thermal activation. The V₂O₅ nanorod architecture exhibited a five-color change in the voltage range of ±1 V (yellow, yellow-green, dark slate gray, steel blue, and Prussian blue) with transmittance modulation of 27.43% at λ = 450 nm and acceptable switching response time (8.8 s for coloration and 9.3 s for bleaching).²¹⁵ The V₂O₅ nanofiber grassland displayed promising transmittance modulation (34% and 25% at λ = 460, 1000 nm, respectively) and high color contrast (yellow and dark green) upon the Li-ion insertion/extraction processes (Fig. 12d).²²⁰ Alsawafta et al. also reported a similar process to prepare high-performance electrochromic V₂O₅ nanorod arrays using stacked polystyrene spheres as a template.²²¹ Self-supported nanorod arrays of MoO₃^222–224 and Nb₂O₅²²⁵ which showed an intrinsic layered structure, also have been prepared by annealing treatment on the corresponding metal oxide amorphous films. These films could exhibit high electrochromic performance.

Fig. 12 (a) SEM image of 3D macroporous amorphous vanadium oxide films prepared by electrodeposition with the assistance of polystyrene colloidal crystals as templates. (b and c) SEM images of V₂O₅ nanorods and nanofiber grassland prepared by annealing treatment on 3D macroporous amorphous vanadium oxide films. Reprinted from ref. 218. Copyright 2015 Wiley. Reprinted from ref. 220. Copyright 2015 Macmillan Publishers Limited. (d) Electrochromic performance of the V₂O₅ nanofiber grassland. (e) Cross-sectional SEM image of a flexible WO_x–NbO_x composite film electrode. Inset digital photos demonstrate the three-stage transmittance modulation in the visible and NIR regions. Reprinted from ref. 232. Copyright 2017 American Chemical Society. (f and g) SEM image and electrochromic performance of coassembled Ag/W₁₈O₄₉ nanofiber networks on the PET substrate. Reprinted from ref. 239. Copyright 2017 American Chemical Society.

Self-supported nanorod arrays of other electrochromic metal oxides, including Co₃O₄,²²⁶ NiO²²⁷ and WO₃²²⁸ have also been prepared by annealing treatment on the corresponding amorphous films, while the morphology and phase transformation mechanisms are still not clear.

Drop-assembly of self-supported 1D nanostructures. When the 1D metal oxide nanomaterials are in a powder form, they can be assembled into porous films on TCO substrates by a drop-assembly process. Typically, the nanomaterials are uniformly dispersed into solvents, creating a suspension with high concentration. Then the suspension is dip-coated onto a clean TCO substrate and dried, producing a porous 1D metal oxide nanomaterials-stacked film.^229–232 When efficient electron transport between the nanomaterials and substrate was achieved, these drop-assembled films could exhibit high electrochromic performance. For example, Xiong et al. prepared silver vanadium oxide and V₂O₅ nanofibers with a length of over 30 μm and a diameter of 10–20 nm by a hydrothermal process, and then fabricated the corresponding nanofiber-stacked porous films with a thickness of ∼500 nm on ITO substrates by a drop-assembly method.²³⁰ The as-prepared silver vanadium oxide film exhibited a higher transmittance modulation (60% at λ = 1020 nm) and faster switching response (0.2 and 0.1 s for coloration and bleaching, respectively) than the V₂O₅ film (corresponding values of 33.7%, 6 s and 5 s, respectively). The high electrical conductivity (0.5 S cm⁻¹) and chemical Li⁺ diffusion coefficient (5.3 × 10⁻¹⁰ cm² s⁻¹) of silver vanadium oxide nanofibers were the important features, compared to the corresponding values of V₂O₅ nanofibers (0.08 S cm⁻¹ and 7.5 × 10⁻¹¹ cm² s⁻¹, respectively). Heo et al. drop-assembled WO_x nanorods on ITO/PET substrates then coated NbO_x nanoparticles on the WO_x nanorods, forming flexible electrochromic WO_x–NbO_x composite films (Fig. 12e).²³² The as-prepared film showed high and multi-stage transmittance modulation in the visible and NIR regions in the voltage range from 1.5 to 4 V vs. Li⁺/Li (inset figures in Fig. 11d). At 4 V (“Bright mode”), transmission of the film was high in both the visible and NIR regions as WO_x and NbO_x are in their oxidized states. When 2.4 V (“Cool mode”) was applied, transmittance in the NIR region was lowered due to charging of WO_x, blocking the solar heat gained while largely maintaining the visible transmittance which was useful for daylighting. At 1.5 V (“Dark mode”), NbO_x was also intercalated with Li⁺ so the visible transmittance also decreased.

Langmuir–Blodgett assembly of self-supported 1D nanostructures. Metal oxide 1D nanostructured films formed by Langmuir–Blodgett (LB) assembly of nanorods and nanofibers are attractive for electrochromism because not only can the assembled films almost completely possess the electrochemical advantages of nanorods or nanofibers, but also new features (such as coupling effect and synergistic effect) can arise when the 1D nanostructures are in a ordered arrangement. The LB assembly of metal oxide 1D nanorods or nanofibers on the substrate typically includes two steps: (1) ordered arrangement of metal oxide 1D nanorods or nanofibers into a continuous film at the phase interface between the solution and air because of surface tension; (2) transfer of the film onto a TCO substrate by slowly withdrawing the immersed TCO substrate out of the solution.^233,234 The thickness of the assembled films can be controlled by repeating the withdrawing cycles. Metal oxide 1D nanostructures with various length-to-diameter ratios can be assembled by LB technology and all assembled films exhibited high electrochromic performance.^235–239

Yoo et al. solvothermally prepared 1D W₁₈O₄₉ nanobundles formed by parallelly stacking tens of W₁₈O₄₉ nanowires in the diameter direction.²³⁵ Then a series of nanobundle-assembled porous films on ITO substrates were prepared by LB technology under different maximum surface pressures. A superior transmittance of ca. 33% at λ = 633 nm with a coloration efficiency of 47.5 cm² C⁻¹ was achieved when the proper maximum surface pressure was used. Wang et al. assembled the hydrothermally prepared WO₃ nanorods (diameters of ∼100 nm and lengths of ∼2 μm) into a film on ITO substrates using an LB process.²³⁷ The as-assembled WO₃ film exhibited multicolor changes (green, green-blue, and blue), high transmittance modulation of ∼ 66% at λ = 632.8 nm, and high intercalated charge of ∼133 mC cm⁻² mg⁻¹ compared to many other WO₃ nanostructured films. The authors also found that the assembly was not affected by the surface properties of the substrates, but was dependent on the drying rate of the film, the concentration of the suspension, and the aspect ratio of the nanorods. Liu et al. found that the use of a surfactant was the key factor to effectively assemble metal oxide nanofibers with high length-to-diameter ratios.²³⁸ By using LB assembly with W₁₈O₄₉ nanofibers (sub 5 nm in diameter and tens of micrometers in length) and poly(vinyl pyrrolidone) (PVP), the authors achieved high electrochromic films on ITO substrates. Interestingly, W₁₈O₄₉ nanowires and Ag nanowires can be co-assembled as hybrid networks on PET substrates by an LB process, creating ITO-free flexible electrochromic electrodes (Fig. 12f).²³⁹ The m_Ag : m_W18O49 mass ratio was 14 : 4. The hybrid film with optimized thickness exhibited high conductivity (7 Ω per square), high transmittance modulation of 68% at λ = 632.5 nm, fast switching response (3.92 and 7.78 s for coloration and bleaching, respectively) and good bending strength stability (Fig. 12g).

5.5 Solution process-deposited self-supported 1D conductive polymer nanostructures

Chemical oxidative and electrochemical polymerizations are the two most general methods used for synthesizing self-supported 1D conductive polymer nanostructures from solutions containing monomers. The generation of conductive polymer 1D nanostructures during the polymerization process seems to have a strong relationship with the intrinsic polymerization mechanism.²⁴⁰ For the typical conductive polymer of PANI, it was found that the polymerization of aniline monomers on substrates covered two different stages.^241,242 In the first stage, a compact PANI layer was formed on the substrate because of the 2D nucleation manner. While in the second stage, the nucleation changed to a 1D manner, producing branched PANI nanostructures (i.e. 1D nanostructures) due to autocatalytic oxidation of aniline on PANI formed in the first stage. In some cases, surfactants were added into the monomer-containing solutions to facilitate the production of PANI 1D nanostructures when the surfactant concentration was higher than the value of the critical micelle concentration.²⁴³

1D conductive polymer nanostructures were much easier to be synthesized when chemical oxidative polymerization occurred under low temperature such as from 0 to 5 °C.^244,245 Zhang et al. prepared randomly stacked PANI nanotubes about 80–180 nm in diameter in the presence of D-10-camphorsulfonic acid as the dopant under ice bath conditions.²⁴⁴ Vertically aligned PANI nanowires with ∼50 nm average diameter and ∼600 nm length were synthesized by Wang et al. using a similar polymerization process.²⁴⁵ Chiou et al. developed a novel dilute polymerization method to produce a large quantity of interconnected and networked PANI nanofibers in bulk solutions on various substrates, including conducting and non-conducting substrates.^246–248 In a typical dilute polymerization process, substantially lower concentrations of aniline and oxidant than the usually employed values were used. The diameters of the tips of the nanofibers can be controlled within the range of 10–40 nm, and the average length can be controlled within the range of 70–360 nm (Fig. 13a).²⁴⁸ These PANI nanofiber arrays grown on TCO substrates exhibited high redox kinetics, leading them to be desirable for bi-functional electrochromic energy storage applications, as shown in the following part about electrochromism-containing multifunctional devices.

Fig. 13 (a) SEM image of PANI nanofiber arrays prepared by using a dilute polymerization method. Reprinted from ref. 248. Copyright 2018 Elsevier. (b) SEM image of PPy nanofiber arrays electropolymerized by a galvanostatic method. Reprinted from ref. 253. Copyright 2016 Elsevier. (c and d) SEM image and electrochromic color changes of a PANI/inorganic salt binary nanofiber network. Reprinted from ref. 255. Copyright 2016 Springer.

Galvanostatic, potentiostatic, and cyclic voltammetry methods are three common technologies for electropolymerization of conductive polymers.^249–251 The morphologies of electropolymerized nanostructures varied a lot from the used techniques. Due to the ease of 1D nucleation, a galvanostatic method was widely used to prepare vertically aligned conductive polymer 1D nanostructures, such as PANI and PPy nanorod and/or nanowire arrays on various conductive substrates.^251–253 The morphology characters, such as areal density, length and diameter of 1D active materials can be easily controlled by varying the experimental parameters such as electropolymerization current densities and times. Fig. 13b showed a typical SEM image of PPy nanofiber arrays prepared using galvanostatic electropolymerization.²⁵³ By simply changing the electropolymerization time, the length of the PPy nanofibers varied from 1.5 to 4 μm accompanying the increase of diameter from 80 to 100 nm.^252,253 When the conductive polymers were polymerized using potentiostatic, and cyclic voltammetry methods, the products usually exhibited a porous nanofiber network morphology built by random twisting of nanofibers.^{250,254–256}Fig. 13c demonstrates a typical SEM image of a PANI/inorganic salt (LiClO₄) binary nanofiber (diameter of 80–110 nm) network electropolymerized at a constant potential of 0.9 V versus Ag/AgCl.²⁵⁵ This nanofiber network exhibited three-color changes and high transmittance modulation in the visible spectrum range with a maximum value of ca. 50% at λ = 560 nm (Fig. 13d). Osuna et al. prepared a PEDOT nanofiber network on a flexible transparent single-wall carbon nanotube/PET electrode.²⁵⁶ This flexible electrochromic electrode showed a high transmittance modulation of ca. 38% at λ = 550 nm, fast switching response time (2.4 and 1.1 s for coloration and bleaching, respectively), and high cycling stability with negligible performance degradation after 1000 cycles.

6. Core/shell 1D nanostructures for electrochromism

6.1 Inorganic/inorganic core/shell 1D nanostructures

Crystalline/amorphous metal oxide (MO_x) core/shell 1D nanostructures. Crystalline/amorphous metal oxide (MO_x) core/shell 1D nanostructures are attractive for high electrochromism due to the simultaneous possession of the electrochemical advantages of crystalline and amorphous MO_x (here MO_x means that the materials of the core and shell are the same oxide). Amorphous MO_x electrodes exhibit fast coloration/bleaching response, high coloration efficiency, high ion diffusion coefficient, and fast redox kinetics on the surface or near surface layer derived from the surface defects and disordered atom arrangements, while crystalline MO_x electrodes demonstrate fast electron transport, high mechanical strength and slow dissolution rate in electrolytes.^257–260 Furthermore, there is a significant difference in the electrochromic kinetics between crystalline and amorphous MO_x. Taking WO₃ as an example, redox-reduced W-ions (i.e. W⁵⁺) and interstitial oxygen ions in amorphous WO₃ can be separated and frequently hop from one site to another between two adjacent sites of W⁵⁺ and W⁶⁺ accompanying transfer of electrons between these sites, while no spatial separation occurs in crystalline WO₃. This difference in the diffusion of ions and electrons gives rise to different electrochromic kinetics and performance in amorphous and crystalline WO₃.^11,84 Therefore, developing an electrochromic MO_x with balanced electrochromic properties has become a major target pursued by researchers. Antonaia et al. proposed an amorphous/crystalline WO₃ double layer and reported that this double layer showed a faster coloration response and a higher transmittance asymptotic value for the bleaching phase than the amorphous or crystalline single layer.²⁵⁷ Qu et al. demonstrated that the amorphous/crystalline WO₃ double layer exhibited higher coloration efficiency than the amorphous or crystalline single layer.²⁶⁰ Analysis of the color centers of electrochromism indicated that the double layer demonstrated a higher number of surface and body color centers than the amorphous or crystalline single layer. Given the structural advantages, fabrication of self-supported crystalline/amorphous core/shell 1D nanostructures is a rational strategy to further enhance the electrochromic performance of MO_x.

Her et al. synthesized crystalline/amorphous WO₃ core/shell nanorod powders using a two-step hydrothermal process and then fabricated a porous film by the drop assembly method.²⁶ Compared to pure nanorod and amorphous films, the crystalline/amorphous WO₃ core/shell nanorods assembled film exhibited increased transmittance modulation of 44% at λ = 550 nm with acceptable switching response (41 and 6 s for coloration and bleaching respectively). Crystalline/amorphous WO₃ core/shell nanorod arrays directly grown on FTO substrates exhibited better electrochromic performance than the film fabricated by drop assembly.²⁶¹ Because of efficient electron transport as well as the enhanced synergistic effect between the amorphous shells and crystalline cores, the nanorod array electrode showed improved optical modulation both in the visible and IR regions (70.3% at λ = 750 nm, 42.0% at λ = 2000 nm and 51.4% at λ = 10 μm), fast switching speed (3.5 and 4.8 s for coloration and bleaching, respectively), and excellent cycling performance (maintained transmittance modulation of 68.5% after 3000 cycles). Self-supported core/shell nanorod arrays of crystalline/amorphous TiO₂ on FTO substrates fabricated by hydrothermal then layer-by-layer methods were demonstrated by Chen et al. (Fig. 14a).²⁶² The layer-by-layer method was beneficial to produce uniform and thin amorphous shells as well as high-quality chemical and physical contacts at the phase interfaces (Fig. 14b). The arrays of crystalline/amorphous TiO₂ core/shell nanorods with an optimal amorphous layer thickness of 13 nm possessed notably larger optical contrast (43%) and higher coloration efficiency (16.2 cm² C⁻¹) than crystalline TiO₂ nanorod arrays (18%, 8.8 cm² C⁻¹) at λ = 800 nm. In addition, it was interesting and meaningful to note that the coating of an amorphous shell significantly enhanced the redox stability of crystalline TiO₂ nanorods in a larger potential window.

Fig. 14 (a) Schematic diagram illustrating preparation of self-supported crystalline/amorphous TiO₂ core/shell nanorod arrays. (b) HRTEM image of a crystalline/amorphous TiO₂ core/shell nanorod. Reprinted from ref. 262. Copyright 2017 Elsevier. (c and d) Cross-sectional SEM images of TiO₂ nanorod arrays and TiO₂/Prussian blue core/shell (TiO₂@PB) nanorod arrays. (e) In situ transmittance response at λ = 700 nm of the PB-based and TiO₂@PB-based electrochromic devices (ECDs). Reprinted from ref. 280. Copyright 2017 Elsevier.

Mixed metal oxide (M1O_x/M2O_y) core/shell 1D nanostructures. Mixed metal oxide (M1O_x/M2O_y) core/shell 1D nanostructures are also attractive for high electrochromism because of the simultaneous possession of the physical and electrochemical advantages of M1O_x and M2O_y (here M1O_x/M2O_y means that the materials of the core and shell are the different oxides). These advantages can include: (1) realization of high surface area and short ion diffusion distance both for M1O_x and M2O_y, compared to many nanostructures such as overlapped double layer M1O_x/M2O_y films; (2) formation of continuous electron bands or consecutive redox reactions in a large potential range to obtain large optical modulation;^206,207,232 (3) facile charge transfer between M1O_x and M2O_y to accelerate the redox kinetics.^208,263

Based on whether the oxides of the cores are electrochromic or not, M1O_x/M2O_y core/shell 1D nanostructures can be separated into two categories. In the first category, the 1D M1O_x oxide cores act as strong mechanical and highly conductive skeletons to deposit M2O_x shells with large surface area and efficient redox reactions. To obtain high color contrast in the M2O_x shells, the 1D M1O_x skeletons are expected to demonstrate high transparency and negligible redox reactions when the M2O_x shells show electrochromism. Thus, 1D ZnO and SnO₂ nanostructures are ideal skeletons. Wang et al. presented a flexible electrochromic display electrode based on amorphous WO₃ nanoparticle-modified ZnO nanorod arrays on ITO/PET substrates prepared by using a facile hydrothermal process and pulsed laser deposition (PLD) method.²⁶⁴ The as-prepared ZnO/WO₃ core/shell nanorod arrays exhibited faster switching response time of ca. 5 s for both coloration and bleaching with higher coloration efficiency than many WO₃ nanorod arrays. Bi et al. investigated the influence of WO₃ shell thickness and morphology on the electrochromic performance of ZnO/WO₃ core/shell nanorod arrays prepared on AZO/PET substrates employing a similar synthesis method.²⁶⁵ By using an optimized thickness and morphology of the WO₃ shells, the ZnO/WO₃ core/shell nanorod arrays exhibited a high transmittance modulation of 68.2% at λ = 633 nm and a large coloration efficiency of 80.6 cm² C⁻¹ with unusual color changes between transparent and black, compared to transparent/blue changes for pure WO₃ nanostructures. Zhang et al. prepared hierarchical SnO₂/NiO core/shell nanostructures on FTO substrates by a two-step hydrothermal process.²⁶⁶ Compared to the pure NiO nanostructures, the SnO₂/NiO core/shell nanostructures exhibited higher transmittance modulation in the whole visible spectrum range, better cycling stability and a notably better memory effect under open-circuit conditions due to the enhancements of charge storage in NiO derived from the electron band structure and high electrical conductivity of SnO₂ cores.

As for the second category where M1O_x and M2O_y are two electrochromic oxides, the M1O_x/M2O_y core/shell 1D nanostructures could demonstrate enhanced and/or new electrochromic performance compared to the pure M1O_x and M2O_y nanoelectrodes. Until now, NiO/V₂O₅,²⁶⁷ WO₃/TiO₂,^53,268,269 TiO₂/NiO,^270,271 TiO₂/WO₃,^272,273 lithium-titanate/WO₃,²⁷⁴ TiO₂/V₂O₅,²⁷⁵ and TiO₂/MoO₃²⁷⁶ core/shell 1D nanostructures have been fabricated for electrochromic applications. The electrochromic performance of these M1Ox/M2Oy core/shell 1D nanostructures was highly dependent on the mass ratio of the two oxides. For example, as shown in Table 1, TiO₂/WO₃ core/shell nanorod arrays prepared by a two-step approach including hydrothermal and electrodeposition processes exhibited better performance than pure WO₃ films and TiO₂ nanorod arrays. Their transmittance modulation and switching response time varied a lot when the thickness of the WO₃ shell changed.²⁷² The M1O_x/M2O_y core/shell 1D nanostructures could also exhibit better cycling stability and higher coloration efficiency than the pure M1O_x and M2O_y nanostructures as demonstrated in the electrochromic performance of WO₃/TiO₂,^53,269 TiO₂/NiO,²⁷⁰ TiO₂/WO₃,²⁷² and TiO₂/V₂O₅²⁷⁵ core/shell nanorod arrays. Electrochemical analysis indicated that the enhanced redox kinetics including the improved Li⁺ diffusion coefficient and efficient interfacial electron transport were important features.

Table 1 Comparison between transmittance modulation of the WO₃ film, TiO₂ nanorod and TiO₂/WO₃ core/shell nanorod arrays. The thicknesses of the WO₃ shells were controlled by different electrodeposition times at −0.7 V (vs. Ag/AgCl). TiO₂@200WO₃ meant a time of 200 s. Reprinted from ref. 272. Copyright 2013 Elsevier

Sample	Transmittance modulation (%)			Switching time (s)
	λ = 700 nm	λ = 1800 nm	λ = 10 μm	Coloration	Bleaching
Pure WO3	31.6	35.5	6.9	5.0	5.2
Pure TiO2	4.3	3.1	0.6	0.6	3.6
TiO2@200WO3	40.3	70.3	38.4	2.4	4.4
TiO2@400WO3	52.2	46.3	37.2	3.2	4.6
TiO2@600WO3	57.2	38.2	17.8	6.4	5.4

---

Metal oxide/polynuclear compound core/shell 1D nanostructures. Hexacyanoferrates, Prussian blue as a representative, are redox-derived electrochromic compounds.^277–279 Because the electrochromism of hexacyanoferrates is from the ion insertion/extraction process, nanostructuring is an effective process to improve the performance. Chen et al. prepared TiO₂/Prussian blue core/shell nanorod arrays through two-step soft chemical methods (Fig. 14c and d).²⁸⁰ The remarkably enlarged surface area, notably shortened ion diffusion length and porous character of Prussian blue shells resulted in highly increased storage capacity of K⁺ ions in electrochromic processes. Compared with the half-cell electrochromic device based on dense Prussian blue film, higher coloration efficiency (131.5 cm² C⁻¹), higher optical contrast (48% at λ = 700 nm) and faster response speed (coloration and bleaching time of 6.2 and 2.2 s, respectively) were achieved in the device fabricated by TiO₂@ Prussian blue core/shell structures (Fig. 14e).

Polyoxometalates are a newly emerging class of transition metal-containing nanoclusters with intriguing structures and diverse redox properties for electrochromism.^281,282 Coating polyoxometalates on metal oxide 1D nanostructures can significantly enlarge their surface area and shorten the ion diffusion distance, resulting in enhanced electrochromic performance. Liu et al. synthesized Dawson-type polyoxometalate K₆P₂W₁₈O₆₂@TiO₂ hybrid nanowire arrays on FTO substrates by combining hydrothermal and layer-by-layer assembly methods.²⁸³ The as-prepared hybrid nanowire arrays exhibited higher transmittance modulation (45.1% at λ = 650 nm) and faster switching times (1.9 and 6.7 s for coloration and bleaching, respectively) than the K₆P₂W₁₈O₆₂ dense film (corresponding values of 31.2%, 5.3 s and 10.5 s, respectively).

6.2 Inorganic/organic core/shell 1D nanostructures

Metal oxide/conductive polymer core/shell 1D nanostructures. Self-supported metal oxide/conductive polymer core/shell 1D nanostructures are a meaningful category of hybrid materials for enhanced electrochromism,⁴ because of the structural and componential advantages of the hybrids as described in Section 2. The electrochemical and electrochromic superiorities of these hybrid 1D nanostructures has made related research hot in recent years. Typically, these hybrid nanostructures are fabricated by a two-step process including the initial synthesis of metal oxide 1D nanostructures then polymerization of conductive polymers on the surface of metal oxides. Both chemical oxidative and electrochemical polymerizations can be used to produce the conductive polymer shells,^284–287 while the latter is more popular due to its unique advantages. These advantages include: (I) the uniform and morphology-independent coating process even in small voids such as mesopores, (II) the nucleation process only occurring on the metal oxide surfaces resulting in highly tight interaction between the conductive polymer shell and metal oxide core and meanwhile avoiding the generation of free conductive polymer nanoparticles derived from the nucleation in the solution like in chemical oxidative polymerization proccess,^284,285 (III) facile control of shell thickness and generation of different conductive polymer shell morphologies such as coaxial and branched PANI shells simply by varying the electropolymerization time (Fig. 15a and b).^286,287 “Solvophobic” and “electrostatic” interactions are proposed to account for the preferential growth of conductive polymers along metal oxides to form core/shell nanostructures.^286,287

Fig. 15 (a and b) SEM images of coaxial and branched V₂O₅/PANI core shell nanorod arrays, respectively. The color of the samples indicates the different loaded mass of PANI shells. Reprinted from ref. 287. Copyright 2017 Elsevier. (c) Transmittance modulation of TiO₂/PANI core shell nanorod arrays under different voltages in 0.1 M HCl solution. Inset digital photos indicate the color changes under different voltages. Reprinted from ref. 286. Copyright 2013 American Chemical Society. (d and e) Schematic and electrochromic color contrast of electrochromic devices with viologen-modified ZnO NT (or NR) arrays as the working electrode. Reprinted from ref. 299. Copyright 2016 Royal Society of Chemistry.

Various electrochromic metal oxide/conductive polymer core/shell 1D nanostructures including NiO/PANI (or PEDOT),^286,288 Co₃O₄/PANI (or PEDOT),²⁸⁶ TiO₂/PANI (or PEDOT),^286,289,290 WO₃/PANI (or PEDOT),^291–293 ZnO/PEDOT (or poly(3-methylthiophene) (PMeT)),^294,295 and MoO₃/PANI²⁹⁶ have been fabricated for electrochromism. When the metal oxide 1D cores only acted as skeletons, the conductive polymer coating layers (i.e. shells) exhibited enlarged surface area and shortened ion diffusion distance, leading to enhanced electrochromism. For instance, the TiO₂/PANI core/shell nanorod arrays showed similar but enhanced electrochemical/elecrochromic performance to pure PANI films in a 0.1 M HCl aqueous electrolyte.^12,286 The as-prepared hybrid nanorod arrays exhibited high transmittance modulation in the whole visible spectrum range, four distinct color changes, fast switching speeds (1.3 and 1.2 s for coloration and bleaching, respectively), and ultra-high cycling stability (Fig. 15c). Using metal oxides with higher electrical conductivity such as ZnO and SnO₂ nanorod arrays, the metal oxide/conductive polymer core/shell 1D nanostructures can exhibit much enhanced electrochromic performance. Kateb et al. prepared a ZnO/PEDOT core/shell nanorod array film as an electrochromic electrode.²⁹⁴ This electrode showed an almost homogeneous transmittance modulation of ca. 48% in the whole visible spectrum range, ultrafast switching response time of <2.2 ms to achieve 100% transmittance saturation modulation (faster than PEDOT nanotube arrays of 50 and 70 ms for bleaching and coloration, respectively⁶⁸), and ultrahigh cycling stability with no transmittance modulation degradation after 100 000 cycles. Electrochemical analysis indicated that the ZnO/PEDOT core/shell nanorods exhibited a significantly improved Li⁺ diffusion coefficient of 2.01 × 10⁻⁴ cm² s⁻¹ which was 4 orders of the value of the PEDOT dense film.

When both metal oxide cores and conductive polymer shells demonstrate electrochromism, the hybrid core/shell 1D structures can exhibit unusual optical modulation characteristics. For instance, the WO₃/PANI nanorod arrays exhibited color changes differing from the pure WO₃ and PANI electrodes.²⁹¹ The PANI showed three colored redox states under positive voltages and a light yellow or transparent state under negative voltages,²⁵⁰ while the WO₃ presented a transparent state under positive voltages and deep blue state under negative voltages.⁸⁶ This meant that the WO₃/PANI nanorod arrays had a complementary dual-electrochromic effect to each other, giving rise to a new coloration phenomenon, i.e. deep blue color under both positive and negative voltages while bleaching under voltages around 0 V accompanying unique transmittance modulation characters.²⁹¹ Similar electrochromic phenomena were also found in the NiO/PANI²⁸⁸ and MoO₃/PANI²⁹⁶ core/shell nanostructures.

Viologen-modified metal oxide 1D nanostructures. Viologen and its derivatives are a category of bipyridilium-containing molecules. Electrochromism is derived from the electrochemical doping/de-doping process of π bonds in their chains.²⁹⁷ To improve the electrochromic performance, adsorption of viologen molecules on metal oxide 1D nanostructured skeletons is an effective method. Technically speaking, these viologen-modified metal oxide 1D nanostructures do not demonstrate core/shell morphology because the absorbed viologen layer is very thin, around molecular scale in most cases. However, we regard them as core/shell nanostructures due to their theoretical and operating similarities to that of the above mentioned MO_x/conductive polymer core/shell 1D nanostructures.

ZnO nanorod or nanofiber arrays are desirable skeletons due to their large surface area, high transparency, high electrical conductivity, electrochemical stability in the electrochromic voltage window of viologens, and easy fabrication. Sun et al. assembled an electrochromic device (ECD) using viologen-modified ZnO nanowire arrays on FTO glass as the electrochromic electrode and TiO₂ nanoparticles on FTO glass as the counter electrode.²⁹⁷ The as-assembled ECD demonstrated obvious transparent/blue color changes, fast switching time (170 and 142 ms for coloration and bleaching respectively) and desirable cycling stability. Hu et al. fabricated a flexible viologen-based electrochromic electrode using ZnO nanowire arrays on an ITO/PET substrate as a skeleton.²⁹⁸ Li et al. made a comparative study of the electrochromic performance of methyl-viologen modified ZnO nanotube (NT) and nanorod (NR) arrays (Fig. 15d).²⁹⁹ Due to the larger surface area of the NTs than that of the NRs, the electrochromic device based on ZnO NTs exhibited higher transmittance modulation of 70% at λ = 608 nm, higher coloration efficiency of 93.5 cm² C⁻¹, faster switching response (6 and 3 s for coloration and bleaching, respectively), and better coloration/bleaching color contrast than the electrochromic device based on ZnO NRs (Fig. 15e).

7. 1D nanostructures for bifunctional electrochromism-involving applications

Electrochromic energy storage devices

Given that the electrochemical energy storage of metal oxides, conductive polymers, and hexacyanoferrates is also derived from the redox process, research on the integration and applications of electrochromic energy storage are becoming more and more popular.^28,29,300 Many electrochromic materials, including PANI, PEDOT, TiO₂, WO₃, NiO, V₂O₅, MnO₂, Prussian blue, and some hybrids have been used to fabricate electrochromic energy storage electrodes and devices.^28,29,300 In addition, fabrication of self-supported 1D nanostructures could significantly enhance redox kinetics due to the electrochemical and structural advantages.^8,17,252

Wang et al. fabricated a symmetric electrochromic supercapacitor using PANI nanowire arrays (prepared by dilute polymerization method) on flexible PEDOT:PSS/PET substrates with gel electrolyte (Fig. 16a and b).³⁰¹ In the voltage range from 0 to 1 V, the as-assembled device exhibited light yellow green (0 V)/dark blue (1 V) color changes with high areal capacitance rate performance (0.017 and 0.004 F cm⁻² at scan rate of 5 and 100 mV s⁻¹ under CV examination) (Fig. 16c).

Fig. 16 (a) SEM images of PANI nanowire arrays on a flexible PEDOT:PSS/PET substrate. (b and c) Electrochromic color contrast and energy storage performance of a symmetrical electrochromic supercapacitor using PANI nanowire arrays as electroactive materials and polymer gel as an electrolyte. Reprinted from ref. 301. Copyright 2012 Royal Society of Chemistry. (d) Schematic diagram illustrating preparation of mesoporous TiO₂ nanotube arrays using STAH process. Step I was electrodeposition of ZnO nanorod arrays. Step II was immersion of ZnO into STAH solution. Step III was preparation of TiO₂ mesoporous nanotube arrays by STAH and annealing treatment. (e and f) SEM and TEM images of mesoporous TiO₂ nanotube arrays. (g) Transmittance modulation during the charge/discharge process of the TiO₂ nanotube array electrode at a current density of 15C. Reprinted from ref. 302. Copyright 2018 Royal Society of Chemistry.

As for the application of metal oxides for electrochromic energy storage, the relatively slow redox kinetics derived from low electrical conductivity and extrinsic pseudocapacitive character could seriously restrict the electrode to simultaneously deliver high desirable energy density accompanying high and fast optical modulation.¹³⁹ Fabrication of 1D metal oxides in a scale of several nanometers with plenty of pores is a promising strategy. Sacrificial template-accelerated hydrolysis (STAH) using ZnO nanorod arrays as templates to prepare mesoporous metal oxide nanotube arrays stacked by several nanometer-sized oxide crystals is an ideal strategy to synthesize electrochromic energy storage electrodes with high energy and power densities.^54–56 For instance, an electrodeposited ZnO nanorod arrays on an FTO substrate was used to fabricate TiO₂ nanotube arrays in a simple aqueous solution containing (NH₄)₂TiF₆ and H₃BO₃ (Fig. 16d and e).³⁰² As expected, the as-prepared mesoporous TiO₂ nanotubes were formed by several nanometer-sized TiO₂ crystals (Fig. 16f). It was found that the TiO₂ mesoporous nanotube arrays showed bi-functionality of high transmittance modulation during the galvanostatic charge/discharge process delivering high-rate lithium storage. Fig. 16g demonstrated the in situ transmittance response during the charge/discharge at a high current density of 15C (1C = 168 mA g⁻¹) at λ = 700 nm, and the TiO₂ nanostructured electrode showed a high charge specific capacity of ca. 98 mA h g⁻¹ achieved in the short time range of 144 s with high Coulombic efficiency of 97% and a desirable transmittance modulation of ca. 30%. Electrochemical analysis indicated that the significantly enhanced pseudocapacitive effect (51.4% of capacitive contribution at scan rate of 0.4 mV s⁻¹, e.g.) and high Li⁺ diffusion coefficient (8.8 × 10⁻¹⁵ cm² s⁻¹) were two important merits of such morphology. In addition, the ZnO/WO₃ core/shell nanorod arrays also delivered high areal capacitance rate performance due to the ZnO cores for fast electron transport and thin thickness of WO₃ shells for fast redox kinetics.²⁶⁵

Photoelectrochromic devices

Taking into account the fact that both ECDs and dye-sensitized solar cells (DSSCs) share sandwiched structures, the integration of ECD and DSSC functionalities into one device (i.e. photoelectrochromic device (PECD) or photovoltachromic device (PVCD)) becomes possible by simple coating EC materials on the cathode of the DSSCs and adding lithium salts into the electrolytes.²⁹ Given that the short-circuit current (J_sc) and open-circuit photovoltage (V_oc) were two dominating electrochemical parameters causing electrochromism on the cathode and these two parameters were relatively low in typical DSSCs, a 1D nanostructuring strategy could be an effective approach to improve the optical modulation performance of PECDs. (I) 1D TiO₂ nanostructures with high surface area (such as TiO₂ nanowire arrays and mesoporous TiO₂ nanotube arrays) in the photoanode can increase the amount of absorbed dyes to improve J_sc and V_oc.¹¹⁹ (2) 1D electrochromic nanomaterials with enhanced electrochemical kinetics in the cathode can decrease the energy barriers of redox reactions, significantly decreasing the required J_sc and V_oc to achieve electrochromism.^303–305Fig. 17a illustrates the schematic design and working principle of a PECD using electropolymerized poly(3,4-(2,2-dimethylpropylenedioxy)thiophene) (PProDOT-Me₂) nanofiber networks (Fig. 17b) as an electrochromic layer.^304,305 This PECD operated in the following manner. Under sunshine, when in the open-circuit state, the electrons accumulate in the TiO₂ photoanode, while the PProDOT-Me₂ film was in an oxidized state and therefore transparent. When in the short-circuit state, electrons flowed into the PProDOT-Me₂ film along the external circuit and the PProDOT-Me₂ film existed in a reduced state and was therefore colored. The transfer of electrons to the EC film depended kinetically on the redox couple (redox couple of Br⁻/Br₃⁻ was used in the literatures) in the electrolyte employed. Finally, after reaching the static equilibrium state, the dye molecules continued to generate electrons into the TiO₂ film, leading the device to behave as a pure DSSC. When the outer electron transport circuit was cut off (i.e. in the open-circuit state), the PProDOT-Me₂ nanofibers were oxidized, resulting in a transparent state again. From the comparative study results, the authors found that due to the intrinsic fast redox kinetics of the porous PProDOT-Me₂ nanofiber network and desirable DSSC characters derived from the Br⁻/Br₃⁻ redox couple, the PECD exhibited a high V_oc of 0.702 V and an acceptable J_sc of 0.633 mA cm⁻², leading the device to exhibit desirable and stable optical modulation performance. As shown in Fig. 17c, the maximum optical contrast of 44% was achieved at λ = 582 nm between the coloration and bleaching states.³⁰⁴ Furthermore, the as-assembled PECD also exhibited a fast switching response of 2.5 and 2.6 s for coloration and bleaching, respectively. Yu et al. fabricated a photoelectrochromic cell by using a polyaniline nanofiber film as the chromic electrode.³⁰⁶ The as-assembled cell demonstrated a fast coloration time of 7 s due to the high values of V_oc (0.58 V) and J_sc (0.47 mA). Other structural types of PECDs can be found in a related review.²⁹

Fig. 17 (a) Schematic design and working principle of a PECD employing a dye-adsorbed TiO₂ nanoparticle film as the photoanode and an electropolymerized PProDOT-Me₂ film as a cathode. (b) SEM image of the electropolymerized PProDOT-Me₂ film. Reprinted from ref. 305. Copyright 2017 Springer. (c) Transmittance modulation and long-term cycling performance of the as-assembled PECD. Inset digital photos demonstrate the color contrast of the as-assembled PECD. Reprinted from ref. 304. Copyright 2016 Elsevier.

8. Conclusions and outlook

Electrochromism has a wide range of applications in the various fields such as antireflection coatings, smart thermal control coatings, displays, and smart windows. Due to the fact that the electrochromism of metal oxides, hexacyanoferrates and conductive polymers comes from redox reactions, fabrication of self-supported 1D nanostructures becomes an effective category of morphologies to improve their performance due to their structural, electrochemical and processible advantages. The past few years have witnessed fast development and remarkable achievements of various morphologies and materials of self-supported 1D nanostructures for high electrochromism. Furthermore, apart from the basic function of electrochromism, additional functions have been successfully incorporated into electrochromic devices to broaden the devices’ functionality.

To further accelerate the application of self-supported 1D nanostructures for electrochromism, some aspects may be improved by future study. (1) Investigation about the influence of morphology on the redox kinetics is desirable to be added. Better and deeper understanding of the relationship between the morphology and redox kinetics could give some useful guidance to design and prepare 1D nanostructures with desirable morphologies. (2) Fabrication of self-supported 1D nanostructures with high and reversible electrochromism derived from redox reactions of large-radius ions such as Na⁺, K⁺, Al³⁺, and Mg²⁺ is urgent, given the following facts: (I) economical favorability for large-scale applications due to the low cost and abundant metal resources; (II) different redox mechanisms and/or kinetics of large-radius ions compared to Li⁺ and H⁺; (III) the increasing popularity and economic requirement to develop materials and nanostructures for high Na⁺, K⁺, Al³⁺, and Mg²⁺ electrochemical energy storage devices such as batteries and supercapacitors.^307–312 (3) Exploiting electrochromic materials with selective and independent optical modulation in the visible, NIR, and IR spectrum ranges is practically attractive. For smart window applications, the photons in the visible, NIR, and IR spectrum ranges carry different energies. Selective and independent optical modulation of a smart window in these different spectrum ranges could give multiple choices about selective control of indoor brightness and temperatures, making living more comfortable.^151,232,313

Conflicts of interest

There are no conflicts to declare.

Acknowledgements

We thank the National Natural Science Foundation of China (No. 51572058 and 51502057), the National Key Research & Development Program (2016YFB0303903), the Foundation of Science and Technology on Advanced Composites in Special Environment Laboratory, the Science and Technology Foundation of Guizhou Province of China (No. qian ke he ji chu [2017] 1065), the Scientific Research Starting Project of SWPU (2017QHZ019), and the Young Scholars Development Fund of SWPU (201799010003).

References

1. C. G. Granqvist, P. C. Lansåker, N. R. Mlyuka, G. A. Niklasson and E. Avendaño, Sol. Energy Mater. Sol. Cells, 2009, 93, 2032–2039.
2. D. R. Rosseinsky and R. J. Mortimer, Adv. Mater., 2001, 13, 783–793.
3. J. N. Yao, Y. A. Yang and B. H. Loo, J. Phys. Chem. B, 1998, 102, 1856–1860.
4. V. K. Thakur, G. Ding, J. Ma, P. S. Lee and X. Lu, Adv. Mater., 2012, 24, 4071–4096.
5. C. G. Granqvist, Thin Solid Films, 2016, 614, 90–96.
6. J. N. Tiwari, R. N. Tiwari and K. S. Kim, Prog. Mater. Sci., 2012, 57, 724–803.
7. R. S. Devan, R. A. Patil, J.-H. Lin and Y.-R. Ma, Adv. Funct. Mater., 2012, 22, 3326–3370.
8. B. L. Ellis, P. Knauth and T. Djenizian, Adv. Mater., 2014, 26, 3368–3397.
9. S. K. Deb, Appl. Opt., Suppl., 1969, 3, 192–195.
10. S. F. Cogan, N. M. Nguyen, S. J. Perrotti and R. D. Rauh, J. Appl. Phys., 1989, 66, 1333.
11. C. G. Granqvist, Sol. Energy Mater. Sol. Cells, 2012, 99, 1–13.
12. T. Kobayashi, H. Yoneyama and H. Tamura, J. Electroanal. Chem., 1984, 161, 419–423.
13. C. M. Amb, A. L. Dyer and J. R. Reynolds, Chem. Mater., 2011, 23, 397–415.
14. S. Xiong, S. Yin, Y. Wang, Z. Kong, J. Lan, R. Zhang, M. Gong, B. Wu, J. Chu and X. Wang, Mater. Sci. Eng., B, 2017, 221, 41–53.
15. H. Zhang, G. Duan, G. Liu, Y. Li, X. Xu, Z. Dai, J. Wang and W. Cai, Nanoscale, 2013, 5, 2460–2468.
16. R. Leones, R. C. Sabadini, F. C. Sentanin, J. M. S. S. Esperanç, A. Pawlicka and M. M. Silva, Sol. Energy Mater. Sol. Cells, 2017, 169, 98–106.
17. K. Wang, H. Wu, Y. Meng and Z. Wei, Small, 2014, 10, 14–31.
18. G. Cao and D. Liu, Adv. Colloid Interface Sci., 2008, 136, 45–64.
19. G. K. Mor, O. K. Varghese, M. Paulose, K. Shankar and C. A. Grimes, Sol. Energy Mater. Sol. Cells, 2006, 90, 2011–2075.
20. J. Martín, M. Martín-González, J. F. Fernández and O. Caballero-Calero, Nat. Commun., 2014, 5, 5130.
21. H. Zhang, G. R. Li, L. P. An, T. Y. Yan, X. P. Gao and H. Y. Zhu, J. Phys. Chem. C, 2007, 111, 6143–6148.
22. Y. Tang, Y. Zhang, J. Deng, J. Wei, H. L. Tam, B. K. Chandran, Z. Dong, Z. Chen and X. Chen, Adv. Mater., 2014, 26, 6111–6118.
23. K. Zhu, Q. Wang, J.-H. Kim, A. A. Pesaran and A. J. Frank, J. Phys. Chem. C, 2012, 116, 11895–11899.
24. X. Lu, G. Wang, T. Zhai, M. Yu, J. Gan, Y. Tong and Y. Li, Nano Lett., 2012, 12, 1690–1696.
25. A. Dey, S. De, A. De and S. K. De, Nanotechnology, 2004, 15, 1277–1283.
26. Y.-C. Her and C.-C. Chang, CrystEngComm, 2014, 16, 5379–5386.
27. X. H. Xia, J. P. Tu, J. Zhang, X. H. Huang, X. L. Wang, W. K. Zhang and H. Huang, Electrochem. Commun., 2009, 11, 702–705.
28. P. Yang, P. Sun and W. Mai, Mater. Today, 2016, 19, 394–402.
29. Z. Tong, Y. Tian, H. Zhang, X. Li, J. Ji, H. Qu, N. Li, J. Zhao and Y. Li, Sci. China: Chem., 2017, 60, 13–37.
30. Z. Yao, C. Wang, Y. Li and N.-Y. Kim, Nanoscale Res. Lett., 2015, 10, 166.
31. H.-W. Shim, Y.-H. Jin, S.-D. Seo, S.-H. Lee and D.-W. Kim, ACS Nano, 2011, 5, 443–449.
32. H.-J. Zheng, X.-D. Wang and Z.-H. Gu, Acta Phys.-Chim. Sin., 2009, 25, 1650–1654.
33. Z. Xiao, L. Zhang, X. Tian and X. Fang, Nanotechnology, 2005, 16, 2647–2650.
34. S. H. Park, J. Y. Song, H. M. Park and H. Yu, Mater. Res. Soc. Symp. Proc., 2010, 1258, 1258.
35. L. Guo, Y. Ren, J. Liu, S. Y. Chiam and W. K. Chim, Small, 2014, 10, 2611–2617.
36. K. Yamada, Y. Tanaka and S. Akimoto, Bull. Soc. Photogr. Imag. Jpn., 2015, 25, 38–40.
37. S. J. Limmer, K. Takahashi and G. Cao, Proc. SPIE., 2003, 25, 5224.
38. S. J. Limmer and G. Cao, Adv. Funct. Mater., 2003, 15, 427–431.
39. K. Takahashi, Y. Wang and G. Cao, Appl. Phys. Lett., 2005, 86, 053102.
40. K. Takahashi, S. J. Limmer, Y. Wang and G. Cao, J. Phys. Chem. B, 2004, 108, 9795–9800.
41. J. Muster, G. T. Kim, V. Krstić, J. G. Park, Y. W. Park, S. Roth and M. Burghard, Adv. Mater., 2000, 12, 420–424.
42. Z. Miao, D. Xu, J. Ouyang, G. Guo, X. Zhao and Y. Tang, Nano Lett., 2002, 2, 717–720.
43. K. Takahashi, Y. Wang, K. Lee and G. Cao, Appl. Phys. A: Mater. Sci. Process., 2006, 82, 27–31.
44. Y. Wang and G. Cao, J. Mater. Chem., 2007, 17, 894–899.
45. C. Yan, G. Chen, X. Zhou, J. Sun and C. Lv, Adv. Funct. Mater., 2016, 26, 1428–1436.
46. A. Al-Haddad, Z. Wang, R. Xu, H. Qi, R. Vellacheri, U. Kaiser and Y. Lei, J. Phys. Chem. C, 2015, 119, 16331–16337.
47. Y. Zhang, M. Liu, W. Ren and Z.-G. Ye, Appl. Surf. Sci., 2015, 340, 120–125.
48. C. C. Chen, C.-H. Cheng and C.-K. Lin, Ceram. Int., 2013, 39, 6631–6636.
49. L. Yuan, S. Meng, Y. Zhou and Z. Yue, J. Mater. Chem. A, 2013, 1, 2552–2557.
50. Z. Liang, A. S. Susha, A. Yu and F. Caruso, Adv. Mater., 2003, 15, 1849–1853.
51. L. Li, S. Pan, X. Dou, Y. Zhu, X. Huang, Y. Yang, G. Li and L. Zhang, J. Phys. Chem. C, 2007, 111, 7288–7291.
52. G. Y. Teo, M. P. Ryan and D. J. Riley, Electrochem. Commun., 2014, 47, 13–16.
53. H. Wang, Y. Song, W. Liu, S. Yao and W. Zhang, Mater. Lett., 2013, 93, 319–321.
54. N. M. Vuong, D. Kim and H. Kim, J. Mater. Chem. C, 2013, 1, 3399–3407.
55. J. Liu, K. Song, P. A. van Aken, J. Maier and Y. Yu, Nano Lett., 2014, 14, 2597–2603.
56. J. Liu, Y. Li, H. Fan, Z. Zhu, J. Jiang, R. Ding, Y. Hu and X. Huang, Chem. Mater., 2010, 22, 212–217.
57. K. Wang, M. Wei, M. A. Morris, H. Zhou and J. D. Holmes, Adv. Mater., 2007, 17, 3016–3020.
58. S. I. Cho and S. B. Lee, Acc. Chem. Res., 2008, 41, 699–707.
59. M. Xue, F. Li, D. Chen, Z. Yang, X. Wang and J. Ji, Adv. Mater., 2016, 28, 8265–8270.
60. Y. K. Hong, S. Kim, H. T. Kim, S. Kim, B.-G. Kim, S. Lee, D. H. Park and B.-H. Kim, Org. Electron., 2016, 32, 59–64.
61. Y.-Z. Long, M.-M. Li, C. Gu, M. Wan, J.-L. Duvail, Z. Liue and Z. Fan, Prog. Polym. Sci., 2011, 36, 1415–1442.
62. J. Duvail, Y. Long, S. Cuenot, Z. Chen and C. Gu, Appl. Phys. Lett., 2014, 90, 102114.
63. Y. Long, L. Zhang, Z. Chen, K. Huang, Y. Yang and H. Xiao, Phys. Rev. B: Condens. Matter Mater. Phys., 2005, 71, 165412.
64. S. Cuenot, S. Demoustier-Champagne and B. Nysten, Phys. Rev. Lett., 2000, 85, 1690.
65. S. I. Cho, W. J. Kwon, S.-J. Choi, P. Kim, S.-A. Park, J. Kim, S. J. Son, R. Xiao, S.-H. Kim and S. B. Lee, Adv. Mater., 2005, 17, 171–175.
66. H. Zhang, H. Qu, H. Lv, S. Hou, K. Zhang, J. Zhao, X. Li, E. Frank and Y. Li, Chem. – Asian J., 2016, 11, 2882–2888.
67. S. I. Cho, D. H. Choi, S.-H. Kim and S. B. Lee, Chem. Mater., 2005, 17, 4564–4566.
68. S. I. Cho, R. Xiao and S. B. Lee, Nanotechnology, 2007, 18, 405705.
69. R. Xiao, S. I. Cho, R. Liu and S. B. Lee, J. Am. Chem. Soc., 2007, 129, 4483–4489.
70. S. Xiong, Q. Wang and H. Xia, Mater. Res. Bull., 2004, 39, 1569–1580.
71. C. Mijangos, R. Hernández and J. Martín, Prog. Polym. Sci., 2016, 54-55, 148–182.
72. L. Huang, Z. Wang, H. Wang, X. Cheng, A. Mitra and Y. Yan, J. Mater. Chem., 2002, 12, 388–391.
73. J. I. Lee, S. H. Cho, S.-M. Park, J. K. Kim, J. K. Kim, J.-W. Yu, Y. C. Kim and T. P. Russell, Nano Lett., 2008, 8, 2315–2320.
74. G. Kimura and K. Yamada, Synth. Met., 2009, 159, 914–918.
75. Y. Kim, J. Baek, M.-H. Kim, H.-J. Choi and E. Kim, Ultramicroscopy, 2008, 108, 1224–1227.
76. Y. Kim, Y. Kim, S. Kim and E. Kim, ACS Nano, 2010, 4, 5277–5284.
77. R. B. Ambade, S. B. Ambade, N. K. Shrestha, R. R. Salunkhe, W. Lee, S. S. Bagde, J. H. Kim, F. J. Stadler, Y. Yamauchi and S.-H. Lee, J. Mater. Chem. A, 2017, 5, 172–180.
78. M. Döbbelin, R. Tena-Zaera, P. M. Carrasco, J.-R. Sarasua, G. Cabañero and D. Mecerreyes, J. Polym. Sci., Part A: Polym. Chem., 2010, 48, 4648–4653.
79. L. J. Pan, L. Pu, Y. Shi, S. Y. Song, Z. Xu, R. Zhang and Y. D. Zheng, Adv. Mater., 2007, 19, 461–464.
80. Z.-L. Wang, X.-J. He, S.-H. Ye, Y.-X. Tong and G.-R. Li, ACS Appl. Mater. Interfaces, 2014, 6, 642–647.
81. U. Abacia, H. Y. Guney and U. Kadiroglu, Electrochim. Acta, 2013, 96, 214–224.
82. W. Ap. Christinelli, A. B. Trench and E. C. Pereira, Sol. Energy Mater. Sol. Cells, 2016, 157, 703–708.
83. D. S. Dalavi, R. S. Devan, R. A. Patil, R. S. Patil, Y.-R. Ma, S. B. Sadale, I. Kim, J.-H. Kim and P. S. Patil, J. Mater. Chem. C, 2013, 1, 3722–3728.
84. K. Ghosh, A. Roy, S. Tripathi, S. Ghule, A. K. Singh and N. Ravishankar, J. Mater. Chem. C, 2017, 5, 7307–7316.
85. C. Y. Ng, K. A. Razak and Z. Lockman, Thin Solid Films, 2015, 595, 73–78.
86. F. Zheng, M. Guo and M. Zhang, CrystEngComm, 2013, 15, 277–284.
87. C.-H. Lu, M. H. Hon and I.-C. Leu, J. Electron. Mater., 2017, 46, 2080–2084.
88. D. Ma, G. Shi, H. Wang, Q. Zhang and Y. Li, J. Mater. Chem. A, 2013, 1, 684–691.
89. C.-H. Lu, M. H. Hon, C.-Y. Kuan and I.-C. Leu, J. Mater. Sci., 2015, 50, 5739–5745.
90. J. Zhang, J.-P. Tu, X.-H. Xia, X.-L. Wang and C.-D. Gu, J. Mater. Chem., 2011, 21, 5492–5498.
91. C. J. Hung, Y. H. Huang, C. H. Chen, P. Lin and T. Y. Tseng, IEEE Trans. Compon., Packag., Manuf. Technol., 2014, 4, 831–838.
92. G. F. Cai, J. P. Tu, D. Zhou, X. L. Wang and C. D. Gu, Sol. Energy Mater. Sol. Cells, 2014, 124, 103–110.
93. C.-H. Lu, M.-H. Hon, C.-Y. Kuan and I.-C. Leu, RSC Adv., 2016, 6, 1913–1918.
94. K. Hong, W. Yiu, H. Wu, J. Gao and M. Xie, Nanotechnology, 2005, 16, 1608–1611.
95. H. Wang, X. Quan, Y. Zhang and S. Chen, Nanotechnology, 2008, 19, 065704.
96. J. Su, X. Feng, J. D. Sloppy, L. Guo and C. A. Grimes, Nano Lett., 2011, 11, 203–208.
97. R. Yu, Z. Meng, M. Ye, Y. Lin, N. Lin, X. Liu, W. Yu and X. Liu, CrystEngComm, 2015, 17, 6583–6590.
98. N. Y. Bhosale, S. S. Mali, C. K. Hong and A. V. Kadam, Electrochim. Acta, 2017, 246, 1112–1120.
99. M. Shibuya and M. Miyauchi, Chem. Phys. Lett., 2009, 473, 126–130.
100. J. Zhang, X. L. Wang, X. H. Xia, C. D. Gu and J. P. Tu, Sol. Energy Mater. Sol. Cells, 2011, 95, 2107–2112.
101. V. V. Kondalkar, S. S. Mali, R. R. Kharade, K. V. Khot, P. B. Patil, R. M. Mane, S. Choudhury, P. S. Patil, C. K. Hong, J. H. Kim and P. N. Bhosale, Dalton Trans., 2015, 44, 2788–2800.
102. D. Ma, T. Li, Z. Xu, L. Wang and J. Wang, Sol. Energy Mater. Sol. Cells, 2018, 177, 51–56.
103. H. Li, G. Shi, H. Wang, Q. Zhang and Y. Li, J. Mater. Chem. A, 2014, 2, 11305–11310.
104. G. Cai, J. Tu, D. Zhou, L. Li, J. Zhang, X. Wang and C. Gua, CrystEngComm, 2014, 16, 6866–6872.
105. S. Poongodi, P. S. Kumar, Y. Masuda, D. Mangalaraj, N. Ponpandian, C. Viswanathana and S. Ramakrishnad, RSC Adv., 2015, 5, 96416–96427.
106. H. Li, J. Wang, G. Shi, H. Wang, Q. Zhang and Y. Li, RSC Adv., 2015, 5, 196–201.
107. V. V. Kondalkar, R. R. Kharade, S. S. Mali, R. M. Mane, P. B. Patil, P. S. Patil, S. Choudhury and P. N. Bhosale, Superlattices Microstruct., 2014, 73, 290–295.
108. J. Chu, J. Lan, D. Lu, J. Ma, Xi. Wang, B. Wu, M. Gong, R. Zhang and S. Xiong, Micro Nano Lett., 2016, 11, 749–752.
109. Z. Jiao, X. W. Sun, J. Wang, L. Ke and H. V. Demir, J. Phys. D: Appl. Phys., 2010, 43, 285501.
110. S. Adhikari and D. Sarkar, RSC Adv., 2014, 4, 20145–20153.
111. Z. Jiao, J. Wang, L. Ke, X. Liu, H. V. Demir, M. F. Yang and X. W. Sun, Electrochim. Acta, 2012, 63, 153–160.
112. K. Lee, A. Mazare and P. Schmuki, Chem. Rev., 2014, 114, 9385–9454.
113. Z. Zhao, J. Tian, Y. Sang, A. Cabot and H. Liu, Adv. Mater., 2015, 27, 2557–2582.
114. S. H. Kang, S.-H. Choi, M.-S. Kang, J.-Y. Kim, H.-S. Kim, T. Hyeon and Y.-E. Sung, Adv. Mater., 2008, 20, 54–58.
115. B. H. Lee, M. Y. Song, S.-Y. Jang, S. M. Jo, S.-Y. Kwak and D. Y. Kim, J. Phys. Chem. C, 2009, 113, 21453–21457.
116. B. Hao, Y. Yan, X. Wang and G. Chen, ACS Appl. Mater. Interfaces, 2013, 5, 6285–6291.
117. B. Liu, J. E. Boercker and E. S. Aydil, Nanotechnology, 2008, 19, 505604.
118. J.-Z. Chen, W.-Y. Ko, Y.-C. Yen, P.-H. Chen and K.-J. Lin, ACS Nano, 2012, 8, 6633–6639.
119. P. Qiang, Z. Chen, P. Yang, X. Cai, S. Tan, P. Liu and W. Mai, Nanotechnology, 2013, 24, 435403.
120. S. Liu, X. Zhang, P. Sun, C. Wang, Y. Wei and Y. Liu, J. Mater. Chem. C, 2014, 2, 7891–7896.
121. D. Lee, Y. Rho, F. I. Allen, A. M. Minor, S. H. Ko and C. P. Grigoropoulos, Nanoscale, 2013, 5, 11147–11152.
122. H.-L. Chen and Y.-S. Yang, Thin Solid Films, 2008, 516, 5590–5596.
123. C.-S. Cheng, M. Serizawa, H. Sakata and T. Hirayama, Mater. Chem. Phys., 1998, 53, 225–230.
124. D. Zhou, D. Xie, X. Xia, X. Wang, C. Gu and J. Tu, Sci. China: Chem., 2017, 60, 3–12.
125. S. H. Park, J. W. Lim, S. J. Yoo, I. Y. Cha and Y.-E. Sung, Sol. Energy Mater. Sol. Cells, 2012, 99, 31–37.
126. Z. Chen, A. Xiao, Y. Chen, C. Zuo, S. Zhou and L. Li, J. Phys. Chem. Solids, 2013, 74, 1522–1526.
127. F. Cao, G. X. Pan, X. H. Xia, P. S. Tang and H. F. Chen, Electrochim. Acta, 2013, 111, 86–91.
128. C. Zhao, F. Du and J. Wang, RSC Adv., 2015, 5, 38706–38711.
129. Y. Chen, Y. Wang, P. Sun, P. Yang, L. Du and W. Mai, J. Mater. Chem. A, 2015, 3, 20614–20618.
130. Y. Li, B. Tan and Y. Wu, Nano Lett., 2008, 8, 265–270.
131. W. Mei, J. Huang, L. Zhu, Z. Ye, Y. Mai and J. Tu, J. Mater. Chem., 2012, 22, 9315–9321.
132. X. Xia, J. Tu, Y. Mai, X. Wang, C. Gu and X. Zhao, J. Mater. Chem., 2011, 21, 9319–9325.
133. Z. Wen, L. Zhu, W. Mei, L. Hu, Y. Li, L. Sun, H. Cai and Z. Ye, Sens. Actuators, B, 2013, 186, 172–179.
134. Y. Li, B. Tan and Y. Wu, Nano Lett., 2008, 8, 265–270.
135. S. Mishra, P. Yogi, S. Saxena, S. Roy, P. R. Sagdeo and R. Kumar, Adv. Mater. Processes Technol., 2017, 3, 627–631.
136. X. H. Xia, J. P. Tu, J. Zhang, J. Y. Xiang, X. L. Wang and X. B. Zhao, Sol. Energy Mater. Sol. Cells, 2010, 94, 386–389.
137. S. Elhag, Z. H. Ibupoto, X. Liu, O. Nur and M. Willander, Sens. Actuators, B, 2014, 203, 543–549.
138. N. A. Chernova, M. Roppolo, A. C. Dillon and M. S. Whittingham, J. Mater. Chem., 2009, 19, 2526–2552.
139. V. Augustyn, P. Simon and B. Dunn, Energy Environ. Sci., 2014, 7, 1597–1614.
140. Y. Lu, L. Liu, D. Mandler and P. S. Lee, J. Mater. Chem. C, 2013, 1, 7380–7386.
141. S. D. Perera, B. Patel, J. Bonso, M. Grunewald, J. P. Ferraris and K. J. Balkus Jr., ACS Appl. Mater. Interfaces, 2011, 3, 4512–4517.
142. G. Wang, X. Lu, Y. Ling, T. Zhai, H. Wang, Y. Tong and Y. Li, ACS Nano, 2012, 6, 10296–10302.
143. H. Wang, X. Gao, J. Feng and S. Xiong, Electrochim. Acta, 2015, 182, 769–774.
144. J. Chu, Z. Kong, D. Lu, W. Zhang, X. Wang, Y. Yu, S. Li, X. Wang, S. Xiong and J. Ma, Mater. Lett., 2016, 166, 179–182.
145. L. Zheng, Y. Xu, D. Jin and Y. Xie, J. Mater. Chem., 2010, 20, 7135–7143.
146. B. Hui, G. Li, X. Zhao, L. Wang, D. Wu, J. Li and B. K. Via, J. Mater. Sci.: Mater. Electron., 2017, 28, 3264–3271.
147. X. W. Lou and H. C. Zeng, Chem. Mater., 2002, 14, 4781–4789.
148. L. Zheng, Y. Xu, D. Jin and Y. Xie, Chem. Mater., 2009, 21, 5681–5690.
149. L. Wang, X. Zhang, Y. Ma, M. Yang and Y. Qi, Mater. Lett., 2016, 164, 623–626.
150. D. Ma and J. Wang, Sci. China: Chem., 2017, 60, 54–62.
151. C. J. Dahlman, Y. Tan, M. A. Marcus and D. J. Milliron, J. Am. Chem. Soc., 2015, 137, 9160–9166.
152. Y. Zhan, M. R. J. Tan, X. Cheng, W. M. A. Tan, G. F. Cai, J. W. Chen, V. Kumar, S. Magdassi and P. S. Lee, J. Mater. Chem. C, 2017, 5, 9995–10000.
153. S. Lou, Y. Ma, X. Cheng, J. Gao, Y. Gao, P. Zuo, C. Du and G. Yin, Chem. Commun., 2015, 51, 17293–17296.
154. M. Najdoski, V. Koleva and A. Samet, J. Phys. Chem. C, 2014, 118, 9636–9646.
155. X. Xiao, C. Zhang, S. Lin, L. Huang, Z. Hu, Y. Cheng, T. Li, W. Qiao, D. Long, Y. Huang, L. Mai, Y. Gogotsi and J. Zhou, Energy Storage Mater., 2015, 1, 1–8.
156. K. Kalantar-zadeh, J. Z. Ou, T. Daeneke, A. Mitchell, T. Sasaki and M. S. Fuhrer, Appl. Mater. Today, 2016, 5, 73–89.
157. J. Zhou, Y. Wei, G. Luo, J. Zheng and C. Xu, J. Mater. Chem. C, 2016, 4, 1613–1622.
158. D. Zhou, F. Shi, D. Xie, D. H. Wang, X. H. Xia, X. L. Wang, C. D. Gu and J. P. Tu, J. Colloid Interface Sci., 2016, 465, 112–120.
159. G. Hodes, Phys. Chem. Chem. Phys., 2007, 9, 2181–2196.
160. X. H. Xia, J. P. Tu, J. Zhang, X. H. Huang, X. L. Wang, W. K. Zhang and H. Huang, Electrochem. Commun., 2008, 10, 1815–1818.
161. M. Z. Najdoski and T. Todorovski, Mater. Chem. Phys., 2007, 104, 483–487.
162. T. Berger, T. Lana-Villarreal, D. Monllor-Satoca and R. Gómez, Chem. Phys. Lett., 2007, 447, 91–95.
163. M. Najdoski, V. Koleva and S. Demiri, Mater. Res. Bull., 2012, 47, 737–743.
164. X. H. Xia, J. P. Tu, J. Zhang, X. L. Wang, W. K. Zhang and H. Huang, Sol. Energy Mater. Sol. Cells, 2008, 92, 628–633.
165. S. Akinkuade, B. Mwankemwa, J. Nel and W. Meyer, Physica B, 2017 DOI:10.1016/j.physb.2017.06.021.
166. E. Hosono, S. Fujihara, I. Honmaa and H. Zhou, J. Mater. Chem., 2005, 15, 1938–1945.
167. B. Li, J.-M. Wu, T.-T. Guo, M.-Z. Tang and W. Wen, Nanoscale, 2014, 6, 3046–3050.
168. A. Dhara, G. Hodes and S. K. Sarkar, RSC Adv., 2014, 4, 53694–53700.
169. R. S. Devan, R. A. Patil, J.-H. Lin and Y.-R. Ma, Adv. Funct. Mater., 2012, 22, 3326–3370.
170. J. M. Velazquez and S. Banerjee, Small, 2009, 5, 1025–1029.
171. L.-C. Tien and Y.-J. Chen, Appl. Surf. Sci., 2012, 258, 3584–3588.
172. C.-C. Liao, F.-R. Chen and J.-J. Kai, Sol. Energy Mater. Sol. Cells, 2007, 91, 1258–1266.
173. R. Nechache, M. Nicklaus, N. Diffalah, A. Ruediger and F. Rosei, Appl. Surf. Sci., 2014, 313, 48–52.
174. S. M. Nicaise, J. J. Cheng, A. Kiani, S. Gradečak and K. K. Berggren, Nanotechnology, 2015, 26, 075303.
175. R. K. Sharma, P. Kumar and G. B. Reddy, J. Alloys Compd., 2015, 638, 289–297.
176. K.-C. Cheng, F.-R. Chen and J.-J. Kai, Sol. Energy Mater. Sol. Cells, 2006, 90, 1156–1165.
177. C.-C. Liao, F.-R. Chen and J.-J. Kai, Sol. Energy Mater. Sol. Cells, 2006, 90, 1147–1155.
178. L. Xiao, Y. Lv, W. Dong, N. Zhang and X. Liu, ACS Appl. Mater. Interfaces, 2016, 8, 27107–27114.
179. S. Gubbala, J. Thangala and M. K. Sunkara, Sol. Energy Mater. Sol. Cells, 2007, 91, 813–820.
180. J. Thangala, S. Vaddiraju, R. Bogale, R. Thurman, T. Powers, B. Deb and M. K. Sunkara, Small, 2007, 3, 890–896.
181. R. S. Devan, S.-Y. Gao, W.-D. Ho, J.-H. Lin, Y.-R. Ma, P. S. Patil and Y. Liou, Appl. Phys. Lett., 2011, 98, 133117.
182. P. Meduri, E. Clark, J. H. Kim, E. Dayalan, G. U. Sumanasekera and Ma. K. Sunkara, Nano Lett., 2012, 12, 1784–1788.
183. M.-C. Wu and C.-S. Lee, J. Solid State Chem., 2009, 182, 2285–2289.
184. R. A. Patil, R. S. Devan, J.-H. Lin, Y.-R. Ma, P. S. Patil and Y. Liou, Sol. Energy Mater. Sol. Cells, 2013, 112, 91–96.
185. R. A. Patil, R. S. Devan, Y. Liou and Y.-R. Ma, Sol. Energy Mater. Sol. Cells, 2016, 147, 240–245.
186. M. Z. Sialvi, R. J. Mortimer, G. D. Wilcox, A. M. Teridi, T. S. Varley, K. G. U. Wijayantha and C. A. Kirk, ACS Appl. Mater. Interfaces, 2013, 5, 5675–5682.
187. T. Stoycheva, F. E. Annanouch, I. Gràcia, E. Llobet, C. Blackman, X. Correig and S. Vallejos, Sens. Actuators, B, 2014, 198, 210–218.
188. H. Kim, R. O. Bonsu, C. O’Donohue, R. Y. Korotkov, L. McElwee-White and T. J. Anderson, ACS Appl. Mater. Interfaces, 2015, 7, 2660–2667.
189. D. Kowalskia, D. Kim and P. Schmuki, Nano Today, 2013, 8, 235–264.
190. N. A. Kyeremateng, F. Vacandio, M.-T. Sougrati, H. Martinez, J.-C. Jumas, P. Knauth and T. Djenizian, J. Power Sources, 2013, 224, 269–277.
191. A. Ghicov, H. Tsuchiya, R. Hahn, J. M. Macak, A. G. Muñoz and P. Schmuki, Electrochem. Commun., 2006, 8, 528–532.
192. R. Hahn, A. Ghicov, H. Tsuchiya, J. M. Macak, A. G. Muñoz and P. Schmuki, Phys. Status Solidi A, 2007, 204, 1281–1285.
193. Y. Yang, D. Kim and P. Schmuki, Electrochem. Commun., 2011, 13, 1198–1201.
194. Y.-C. Nah, A. Ghicov, D. Kim, S. Berger and P. Schmuki, J. Am. Chem. Soc., 2008, 130, 16154–16155.
195. A. Ghicov, M. Yamamoto and P. Schmuki, Angew. Chem., Int. Ed., 2008, 47, 7934–7937.
196. Y. Yang, D. Kim and P. Schmuki, Electrochem. Commun., 2011, 13, 1021–1025.
197. N. K. Shrestha, Y.-C. Nah, H. Tsuchiy and P. Schmuki, Chem. Commun., 2009, 2008–2010.
198. R. Kirchgeorg, S. Berger and P. Schmuki, Chem. Commun., 2011, 47, 1000–1002.
199. S. Adhikari, Y.-Y. Song, Y.-M. Wang, M. Niraula, K. Cho, S. K. Dhungel, Z.-D. Gao, N. K. Shrestha and S.-H. Han, Opt. Mater., 2015, 40, 112–117.
200. A. Benoit, I. Paramasivam, Y.-C. Nah, P. Roy and P. Schmuki, Electrochem. Commun., 2009, 11, 728–732.
201. N. K. Shrestha, M. Yang and P. Schmukiz, Electrochem. Solid-State Lett., 2010, 13, C21–C24.
202. J. Z. Ou, S. Balendhran, M. R. Field, D. G. McCulloch, A. S. Zoolfakar, R. A. Rani, S. Zhuiykov, A. P. O'Mullane and K. Kalantar-zadeh, Nanoscale, 2012, 4, 5980–5988.
203. S. Berger, A. Ghicov, Y.-C. Nah and P. Schmuki, Langmuir, 2009, 25, 4841–4844.
204. X. Yang, X. Wei, S. Wu, T. Wu, Z. Chen and S. Li, J. Mater. Sci.: Mater. Electron., 2015, 26, 7081–7085.
205. S. Barredo-Damas, K. Lee, R. Kirchgeorg, R. Sánchez-Tovar and P. Schmuki, ECS Electrochem. Lett., 2013, 2, C4–C6.
206. D. D. Yao, M. R. Field, A. P. O'Mullane, K. Kalantar-zadeh and J. Z. Ou, Nanoscale, 2013, 5, 10353–10359.
207. D. Guan, J. Li, X. Gao and C. Yuan, RSC Adv., 2014, 4, 4055–4062.
208. D. D. Yao, R. A. Rani, A. P. O’Mullane, K. Kalantar-zadeh and J. Z. Ou, J. Phys. Chem. C, 2014, 118, 10867–10873.
209. D. D. Yao, R. A. Rani, A. P. O’Mullane, K. Kalantar-zadeh and J. Z. Ou, J. Phys. Chem. C, 2014, 118, 476–481.
210. J. Zhang, X. L. Wang, X. H. Xia, C. D. Gu, Z. J. Zhao and J. P. Tu, Electrochim. Acta, 2010, 55, 6953–6958.
211. J. Lin, X. Liu, M. Guo, W. Lu, G. Zhang, L. Zhou, X. Chen and H. Huang, Nanoscale, 2012, 4, 5148–5153.
212. I. S. Cho, J. Choi, K. Zhang, S. J. Kim, M. J. Jeong, L. Cai, T. Park, X. Zheng and J. H. Park, Nano Lett., 2015, 15, 5709–5715.
213. H. Lv, N. Li, H. Zhang, Y. Tian, H. Zhang, X. Zhang, H. Qu, C. Liu, C. Jia, J. Zhao and Y. Li, Sol. Energy Mater. Sol. Cells, 2016, 150, 57–64.
214. K. R. Reyes-Gil, Z. D. Stephens, V. Stavila and D. B. Robinso, ACS Appl. Mater. Interfaces, 2015, 7, 2202–2213.
215. Z. Tong, J. Hao, K. Zhang, J. Zhao, B.-L. Su and Y. Li, J. Mater. Chem. C, 2014, 2, 3651–3658.
216. Z. Tong, H. Yang, L. Na, H. Qu, X. Zhang, J. Zhao and Y. Li, J. Mater. Chem. C, 2015, 3, 3159–3166.
217. Z. Tong, H. Xu, G. Liu, J. Zhao and Y. Li, Electrochem. Commun., 2016, 69, 46–49.
218. Z. Tong, X. Zhang, H. Lv, N. Li, H. Qu, J. Zhao, Y. Li and X.-Y. Liu, Adv. Mater. Interfaces, 2015, 2, 1500230.
219. Z. Tong, N. Li, H. Lv, Y. Tian, H. Qu, X. Zhang, J. Zhao and Y. Li, Sol. Energy Mater. Sol. Cells, 2016, 146, 135–143.
220. Z. Tong, H. Lv, X. Zhang, H. Yang, Y. Tian, N. Li, J. Zhao and Y. Li, Sci. Rep., 2015, 5, 16864.
221. M. Alsawafta, A. Almoabadi, S. Badilescu and V.-V. Truong, J. Electrochem. Soc., 2015, 162, H466–H472.
222. C.-S. Hsu, C.-C. Chan, H.-T. Huang, C.-H. Peng and W.-C. Hsu, Thin Solid Films, 2008, 516, 4839–4844.
223. M. Dhanasankar, K. K. Purushothaman and G. Muralidharan, Solid State Sci., 2010, 12, 246–251.
224. S. Cong, T. Sugahara, T. Wei, J. Jiu, Y. Hirose, S. Nagao and K. Suganuma, Cryst. Growth Des., 2015, 15, 4536–4542.
225. C. Yan and D. Xue, Adv. Mater., 2008, 20, 1055–1058.
226. L. He, Z. Li and Z. Zhang, Nanotechnology, 2008, 19, 155606.
227. K. K. Purushothaman and G. Muralidharan, Sol. Energy Mater. Sol. Cells, 2009, 93, 1195–1201.
228. R. Seelaboyina, Nanotechnology, 2016, 27, 112502.
229. H. Tokudome and M. Miyauchi, Angew. Chem., Int. Ed., 2005, 44, 1974–1977.
230. C. Xiong, A. E. Aliev, B. Gnade and K. J. Balkus, Jr., ACS Nano, 2008, 2, 293–301.
231. W. Kang, C. Yan, X. Wang, C. Y. Foo, A. W. M. Tan, K. J. Z. Chee and P. S. Lee, J. Mater. Chem. C, 2014, 2, 4727–4732.
232. S. Heo, J. Kim, G. K. Ong and D. J. Milliron, Nano Lett., 2017, 17, 5756–5761.
233. P. Yang and F. Kim, ChemPhysChem, 2002, 3, 503–506.
234. F. Pan, J. Zhang, C. Cai and T. Wang, Langmuir, 2006, 22, 7101–7104.
235. S. J. Yoo, Y. H. Jung, J. W. Lim, H. G. Choi, D. K. Kim and Y.-E. Sung, Sol. Energy Mater. Sol. Cells, 2008, 92, 179–183.
236. X. Changa, S. Suna, Z. Li, X. Xu and Y. Qiu, Appl. Surf. Sci., 2011, 257, 5726–5730.
237. J. Wang, E. Khoo, P. S. Lee and J. Ma, J. Phys. Chem. C, 2008, 112, 14306–14312.
238. J.-W. Liu, J. Zheng, J.-L. Wang, J. Xu, H.-H. Li and S.-H. Yu, Nano Lett., 2013, 13, 3589–3593.
239. J.-L. Wang, Y.-R. Lu, H.-H. Li, J.-W. Liu and S.-H. Yu, J. Am. Chem. Soc., 2017, 139, 9921–9926.
240. D. N. Nguyen and H. Yoon, Polymers, 2016, 8, 118.
241. L. Zhao, L. Zhao, Y. Xu, T. Qiu, L. Zhi and G. Shi, Electrochim. Acta, 2009, 55, 491–497.
242. S. Yunus, A. Attout and P. Bertrand, Langmuir, 2009, 25, 1851–1854.
243. S. Zhang, R. Fu, S. Wang, Y. Gu and S. Chen, Mater. Lett., 2017, 202, 127–130.
244. L. Zhang and M. Wan, Nanotechnology, 2002, 13, 750–755.
245. K. Wang, P. Zhao, X. Zhou, H. Wu and Z. Wei, J. Mater. Chem., 2011, 21, 16373–16378.
246. N.-R. Chiou, C. Lu, J. Guan, L. J. Lee and A. J. Epstein, Nat. Nanotechnol., 2007, 2, 354–357.
247. K. Wang, W. Zou, B. Quan, A. Yu, H. Wu, P. Jiang and Z. Wei, Adv. Energy Mater., 2011, 1, 1068–1072.
248. Y. Wang, S. Xu, H. Cheng, W. Liu, F. Chen, X. Liu, J. Liu, S. Chen and C. Hu, Appl. Surf. Sci., 2018, 428, 315–321.
249. J. Desilvestro and W. Scheifele, J. Mater. Chem., 1993, 3, 263–272.
250. Z. Tong, H. Lv, J. Zhao and Y. Li, Chin. J. Polym. Sci., 2014, 32, 1040–1051.
251. K. Wang, J. Huang and Z. Wei, J. Phys. Chem. C, 2010, 114, 8062–8067.
252. J. Huang, K. Wang and Z. Wei, J. Mater. Chem., 2010, 20, 1117–1121.
253. S. Jiang, B. Yi, L. Cao, W. Song, Q. Zhao, H. Yu and Z. Shao, J. Power Sources, 2016, 329, 347–354.
254. E. M. Erro, A. M. Baruzzi and R. A. Iglesias, Polymer, 2014, 55, 2440–2444.
255. Y. E. Firat and A. Peksoz, J. Mater. Sci.: Mater. Electron., 2017, 28, 3515–3522.
256. R. M. Osuna, V. Hernández, J. T. L. Navarrete, E. I. Kauppinen and V. Ruiz, J. Phys. Chem. Lett., 2010, 1, 1367–1371.
257. A. Antonaia, M. L. Addonizio, C. Minarini, T. Polichetti and M. Vittori-Antisari, Electrochim. Acta, 2001, 46, 2221–2227.
258. M. Deepa, A. G. Joshi, A. K. Srivastava, S. M. Shivaprasad and S. A. Agnihotry, J. Electrochem. Soc., 2006, 153, C365–C376.
259. H.-T. Fang, M. Liu, D.-W. Wang, T. Sun, D.-S. Guan, F. Li, J. Zhou, T.-K. Sham and H.-M. Cheng, Nanotechnology, 2009, 20, 225701.
260. H. Qu, X. Zhang, L. Pan, Z. Gao, L. Ma, J. Zhao and Y. Li, Electrochim. Acta, 2014, 148, 46–52.
261. D. Zhou, D. Xie, F. Shi, D. H. Wang, X. Ge, X. H. Xia, X. L. Wang, C. D. Gu and J. P. Tu, J. Colloid Interface Sci., 2015, 460, 200–208.
262. Y. Chen, X. Li, Z. Bi, X. He, X. Xu and X. Gao, Electrochim. Acta, 2017, 251, 546–553.
263. W. Dong, Y. Lv, L. Xiao, Y. Fan, N. Zhang and X. Liu, ACS Appl. Mater. Interfaces, 2016, 8, 33842–33847.
264. M. Wang, G. Fang, L. Yuan, H. Huang, Z. Sun, N. Liu, S. Xia and X. Zhao, Nanotechnology, 2009, 20, 185304.
265. Z. Bi, X. Li, Y. Chen, X. Xu, S. Zhang and Q. Zhu, Electrochim. Acta, 2017, 227, 6–68.
266. J. Zhang, J. Tu, D. Zhou, H. Tang, L. Li, X. Wang and C. Gu, J. Mater. Chem. C, 2014, 2, 10409–10417.
267. Y. Liu, C. Jia, Z. Wan, X. Weng, J. Xie and L. Deng, Sol. Energy Mater. Sol. Cells, 2015, 132, 467–475.
268. B.-R. Huang, T.-C. Lin and Y.-M. Liu, Sol. Energy Mater. Sol. Cells, 2015, 133, 32–38.
269. L. Cheng, X. Zhang, B. Liu, H. Wang, Y. Li, Y. Huang and Z. Du, Nanotechnology, 2005, 16, 1341–1345.
270. G. Cai, J. Tu, D. Zhou, L. Li, J. Zhang, X. Wang and C. Gu, J. Phys. Chem. C, 2014, 118, 6690–6696.
271. G. Bo, X. Wang, K. Wang, R. Gao, B. Dong, L. Cao and G. Su, J. Alloys Compd., 2017, 728, 878–886.
272. G. F. Cai, D. Zhou, Q. Q. Xiong, J. H. Zhang, X. L. Wang, C. D. Gu and J. P. Tu, Sol. Energy Mater. Sol. Cells, 2013, 117, 231–238.
273. Y. Li, Z. Liu, L. Zhao, T. Cui, B. Wang, K. Guo and J. Han, Electrochim. Acta, 2015, 173, 117–123.
274. Y. Dong, C. Xiong, Y. Zhang, S. Xing and H. Jiang, Nanotechnology, 2016, 27, 105704.
275. W. He, Y. Liu, Z. Wan and C. Jia, RSC Adv., 2016, 6, 68997–69006.
276. H. Yu, Y. Li, L. Zhao, G. Li, J. Li, H. Rong and Z. Liu, Mater. Lett., 2016, 169, 65–68.
277. A. P. Baioni, M. Vidotti, P. A. Fiorito, E. A. Ponzio and S. I. Córdoba de Torresi, Langmuir, 2007, 23, 6796–6800.
278. K.-M. Lee, H. Tanaka, A. Takahashi, K. H. Kim, M. Kawamura, Y. Abe and T. Kawamoto, Electrochim. Acta, 2015, 163, 288–295.
279. H.-Y. Liao, T.-C. Liao, W.-H. Chen, C.-H. Chang and L.-C. Chen, Sol. Energy Mater. Sol. Cells, 2016, 145, 8–15.
280. Y. Chen, Z. Bi, X. Li, X. Xu, S. Zhang and X. Hu, Electrochim. Acta, 2017, 224, 534–540.
281. S.-M. Wang, L. Liu, W.-L. Chen, Z.-M. Zhang, Z.-M. Su and E.-B. Wang, J. Mater. Chem. A, 2013, 1, 216–220.
282. B. Zhang, W. Guan, S. Zhang, B. Li and L. Wu, Chem. Commun., 2016, 52, 5308–5311.
283. S. Liu and X. Qu, Appl. Surf. Sci., 2017, 412, 189–195.
284. C. Zhou, Y. Zhang, Y. Li and J. Liu, Nano Lett., 2013, 13, 2078–2085.
285. Z. Tong, Y. Yang, J. Wang, J. Zhao, B.-L. Su and Y. Li, J. Mater. Chem. A, 2014, 2, 4642–4651.
286. X. Xia, D. Chao, X. Qi, Q. Xiong, Y. Zhang, J. Tu, H. Zhang and H. J. Fan, Nano Lett., 2013, 13, 4562–4568.
287. Z. Tong, S. Liu, X. Li, Y. Ding, J. Zhao and Y. Li, Electrochim. Acta, 2016, 222, 194–202.
288. X. H. Xia, J. P. Tu, J. Zhang, X. L. Wang, W. K. Zhang and H. Huang, Nanotechnology, 2008, 19, 465701.
289. G. Cai, J. Tu, D. Zhou, J. Zhang, Q. Xiong, X. Zhao, X. Wang and C. Gu, J. Phys. Chem. C, 2013, 117, 15967–15975.
290. X. Fu, C. Jia, Z. Wan, X. Weng, J. Xie and L. Deng, Org. Electron., 2014, 15, 2702–2709.
291. G. F. Cai, J. P. Tu, D. Zhou, J. H. Zhang, X. L. Wang and C. D. Gu, Sol. Energy Mater. Sol. Cells, 2014, 122, 51–58.
292. D. Ma, G. Shi, H. Wang, Q. Zhang and Y. Li, J. Mater. Chem. A, 2014, 2, 13541–13549.
293. C. Dulgerbaki, N. N. Maslakci, A. I. Komur and A. U. Oksuz, Electroanalysis, 2016, 28, 1873–1879.
294. M. Kateb, S. Safarian, M. Kolahdouz, M. Fathipour and V. Ahamdi, Sol. Energy Mater. Sol. Cells, 2016, 145, 200–205.
295. J.-C. Chou, C.-J. Yang, Y.-H. Liao, P.-A. Ho, H.-T. Chou and C.-H. Huang, J. Disp. Technol., 2015, 11, 430–437.
296. K. Zhang, Y. Wang, X. Ma, H. Zhang, S. Hou, J. Zhao, X. Li, L. Qiang and Y. Li, New J. Chem., 2017, 41, 10872–10879.
297. X. W. Sun and J. X. Wang, Nano Lett., 2008, 8, 1884–1889.
298. A. Hua, F. Wu, J. Liu, J. Jiang, R. Ding, X. Li, C. Cheng, Z. Zhu and X. Huang, J. Alloys Compd., 2010, 507, 261–266.
299. S. Li, Y. Wang, J.-G. Wu, L. Guo, M. Ye, Y.-H. Shao, R. Wang, C. Zhao and A. Wei, RSC Adv., 2016, 6, 72037–72043.
300. G. Cai, J. Wang and P. S. Lee, Acc. Chem. Res., 2016, 49, 1469–1476.
301. K. Wang, H. Wu, Y. Meng, Y. Zhang and Z. Wei, Energy Environ. Sci., 2012, 5, 8384–8389.
302. Z. Tong, S. Liu, X. Li, L. Mai, J. Zhao and Y. Li, Nanoscale, 2018, 10, 3254–3261.
303. J.-J. Wu, M.-D. Hsieh, W.-P. Liao, W.-T. Wu and J.-S. Chen, ACS Nano, 2009, 3, 2297–2303.
304. X. Wu, J. Zheng and C. Xu, Electrochim. Acta, 2016, 191, 902–907.
305. X. Wu, J. Zheng and C. Xu, Sci. China: Chem., 2017, 60, 84–89.
306. X. F. Yu, Y. X. Li, N. F. Zhu, Q. B. Yang and K. Kalantar-zadeh, Nanotechnology, 2007, 18, 015201.
307. Y. Tian, W. Zhang, S. Cong, Y. Zheng, F. Geng and Z. Zhao, Adv. Funct. Mater., 2015, 25, 5833–5839.
308. Z. Tong, S. Liu, Y. Zhou, J. Zhao, Y. Wu, Y. Wang and Y. Li, Energy Storage Mater., 2018, 13, 223–232.
309. S. Lou, X. Cheng, J. Gao, Q. Li, L. Wang, Y. Cao, Y. Ma, P. Zuo, Y. Gao, C. Du, H. Huo and G. Yin, Energy Storage Mater., 2018, 11, 57–66.
310. S. Lou, X. Cheng, L. Wang, J. Gao, Q. Li, Y. Ma, Y. Gao, P. Zuo, C. Du and G. Yin, J. Power Sources, 2017, 361, 80–86.
311. S. Lou, X. Cheng, Y. Zhao, A. Lushington, J. Gao, Q. Li, P. Zuo, B. Wang, Y. Gao, Y. Ma, C. Du, G. Yin and X. Sun, Nano Energy, 2017, 34, 15–25.
312. Y. Huang, X. Li, J. Luo, K. Wang, Q. Zhang, Y. Qiu, S. Sun, S. Liu, J. Han and Y. Huang, ACS Appl. Mater. Interfaces, 2017, 9, 8696–8703.
313. A. Llordés, G. Garcia, J. Gazquez and D. J. Milliron, Nature, 2013, 500, 323–326.

---

Footnote

† These two authors contributed equally to this work.

---