Flexible electronics based on inorganic nanowires

Abstract

Flexible electronics have gained considerable research interest in the recent years because of their special features and potential applications in flexible displays, artificial skins, sensors, sustainable energy, etc. With unique geometry, outstanding electronic/optoelectronic properties, excellent mechanical flexibility and good transparency, inorganic nanowires (NWs) offer numerous insights and opportunities for flexible electronics. This article provides a comprehensive review of the inorganic NW based flexible electronics studied in the past decade, ranging from NWs synthesis and assembly to several important flexible device and energy applications, including transistors, sensors, display devices, memories and logic gates, as well as lithium ion batteries, supercapacitors, solar cells and generators. The integration of various flexible nanodevices into a self-powered system was also briefly discussed. Finally, several future research directions and opportunities of inorganic NW flexible and portable electronics are proposed.

---

Zhe Liu

Zhe Liu is currently a visiting student at the Institute of Semiconductors, Chinese Academy of Sciences. He is a PhD candidate at the Huazhong University of Science and Technology (HUST). His current scientific interests are focused on the synthesis of metal oxide nanowires for transparent and flexible transistors and photodetector applications.

Jing Xu

Jing Xu is currently a PhD candidate at the Huazhong University of Science and Technology (HUST) and a visiting student at the Institute of Semiconductors, Chinese Academy of Sciences. Her current scientific interests focus on electrochemical supercapacitors and lithium-ion batteries.

Di Chen

Di Chen received her BS degree from Anhui Normal University (1999) and a PhD degree from the University of Science and Technology of China (2005). Currently, she is a professor in the University of Science and Technology, Beijing. Her research interests focus on designing nanostructures for sustainable energy applications, including energy storage, solar cells and photocatalysts.

Guozhen Shen

Guozhen Shen received his BS degree from Anhui Normal University (1999) and a PhD degree from the University of Science and Technology of China (2003). Currently, he is a Principal Investigator at the Institute of Semiconductors, Chinese Academy of Sciences. He has edited two themed issues on flexible electronics, flexible energy storage and conversion for the Journal of Materials Chemistry A & C and is on the editorial boards for several journals, including Nanoscale Research Letters. His research interests include flexible/printable nanowire electronics and sensing systems.

---

1 Introduction

Over the last few decades, advancements in flexible electronics have spread across an expansive area ranging from the development of fundamental transistors, different kinds of sensing devices, and flexible organic light-emitting diode displays in various kinds of flexible substrates.^1–3 This academic interest in flexible electronics will continue for some years, which is driven by the growing demand for electronics permitting lightweight design, portability, and low manufacturing cost as compared to their rigid substrate counterparts, and supported by techniques for the ceaseless miniaturization of individual elements in microelectronics, for example, reductions in the critical dimensions (i.e. channel lengths and dielectric thicknesses) of transistors in integrated circuits.^4,5 Commercially, during approximately the past 15 years, the most arresting examples of macroelectronics are flat-panel displays (FPDs) that use thin film transistors (TFTs) to drive and address active matrix pixels.⁶ However, the novel commercial success of FPDs opens an era of flexible electronics, the desire for extra characteristics including lightweight, low cost and portability lead to the emergence of new applications in rollable displays.⁷ Moreover, thin film solar cells,^8,9 artificial skins^10,11 and other new applications have created intensive interest in building electronic devices directly on flexible substrates like thin plastics in scalable ways. However, the requirements for flexibility and compatibility with plastic substrates that limit the process temperature to be lower than 300 °C are a formidable challenge to the fabrication techniques, especially in the choice of materials.^12,13

Organic semiconductors appear to be quite suitable for these applications because of their good flexibility, low-temperature synthesis process and inherent compatibility with plastic substrates.^14–18 Organic electronics based on these have materials have been widely investigated and employed in light-emitting diodes,^19–21 radio frequency identification tags,^22,23 sensors,^24–26etc. However, an existing problem that limits the application of organic devices still persists, namely, the restrictions of mobility and stability. As compared with conventional silicon based materials and organic semiconductors,^27–32 low-dimension materials with excellent electrical and optical properties have attracted intensive attention in the recent years.^33–38 Because of their relatively higher carrier mobility and size-related intriguing physical properties, these low-dimensional materials present widely potential applications in high performance electronic devices. For example, significant interest has been drawn to one-dimensional (1-D) single-walled carbon nanotube (SWCNT) based devices. With intrinsic mobility over 100 000 cm⁻² V⁻¹ s⁻¹, good mechanical flexibility and optical transparency, SWCNTs are promising candidates for high-speed electronics.^39–48 However, there are still several big challenges of the SWCNTs integration process such as removing metallic carbon nanotubes from semiconducting ones as well as controlling the conducting type and doping level reproducibly.^49,50 In contrast, inorganic semiconducting NWs are ideal materials with controllable size and electrical/optoelectronic properties, and have presented one of the most interesting research directions in nanoscience and nanotechnology.^51–53 High quality single-crystalline inorganic semiconducting NWs have been synthesized via various methods.^54–56 Superior electrical properties such as high carrier mobility over amorphous silicon, amorphous oxide thin films and organic semiconductors have been found.^57–60 In addition to the advantages of mobility, the problems caused by metallic carbon nanotubes have also been tackled.

Besides their unique geometric structures, excellent mechanical flexibility and superior electrical properties (as inorganic NWs are capable to be transferred to all kinds of substrates), it is highly desirable to use them as the building blocks for flexible electronics in a broad and highly interdisciplinary area. First, by providing 1-D monocrystalline pathways for carrier flows, NWs have been demonstrated as reliable and efficient conductive channels for flexible electronics. As carriers do not have to pass through many boundaries between adjacent nanocrystals, effective mobility in flexible electronics is largely improved. Second, for the large specific surface, inorganic NWs have also been demonstrated to act as highly sensitive materials for flexible electronics, especially for sensing devices. Third, inorganic NWs exhibit excellent mechanical flexibility that mainly comes from three aspects: (1) with the materials' dimensions reducing to the nanometer scale, the mechanical strength of inorganic NWs is usually far superior to their bulk counterparts because of a reduction of defect incorporation. This characteristic is very common and indeed expected in many nanoscale materials. For example, the average Young's modulus of ZnO NWs is very close to that reported for its bulk and thin-film values, while the ultimate strengths are up to 40 times that of their bulk counterparts.⁶¹ (2) It is well known that the propensity for crack formation in semiconductor materials is linearly reduced with the dimensions.⁶² In other words, inorganic NWs can be more reliable and robust under bending conditions when compared with their bulk counterparts. (3) Inorganic NWs themselves only cross over a span of the micrometer scale and often serve as vital but tiny parts of an electronic device or a system. This feature allows inorganic NWs to avoid macroscopic deformations caused by structural damage in bulk materials.

To fulfil fully flexible electronics, flexible, lightweight, and miniaturized energy conversion/storage devices are recognized as one of the key components. However, the conventional energy conversion/storage devices are too bulky and inflexible to integrate with the above flexible electrical/optoelectronic devices even on the common forms of flexible substrates. The lack of flexible energy conversion/storage units gives a natural limit on the further development of next-generation flexible devices; as a result, several enabling technologies must be developed to realize flexible energy conversion/storage devices with optimized performance. Very recently, several simple but inspirational models have been demonstrated including supercapacitors and lithium ion batteries. Flexible photovoltaic systems have also been considered for harvesting energy from sunlight and subsequent conversion into electrical power.^63,64

This article reviews the state-of-art research progress in the field of flexible electronics based on inorganic NW materials. It begins with the preparation and assembly techniques of basic materials in Section 2 and their various applications are discussed from Sections 3 to 5 including basic flexible electronic device elements like transistors, sensors, displays, memories, logic gates, etc. Since flexible energy conversion/storage devices are vital to realize fully flexible electronic–optoelectronic systems, to this effect, a number of recently developed flexible energy conversion/storage devices were discussed and summarized in the following sections. Finally, we will discuss the future challenges and perspectives of inorganic NWs based flexible electronics.

2 Inorganic NW synthesis and assembly

2.1 NW synthesis

Inorganic NWs can be prepared through a variety of methods, among which the bottom-up approach is a summary of various routes using basic elements to synthesize NW materials.^65,66 Numerous kinds of materials such as metals, inorganic compounds and polymers have been successfully prepared by the bottom-up approach.^67–69 In general, the preparation of 1-D nanomaterials relies on the ability to condense and grow an assembly of atoms by breaking the symmetry of their crystal lattices that can be realized through controlling the surface energies of different crystal faces, templates assisted growing or catalyzed synthesis. The synthesis of NWs has been well-documented in the last several years; hence, they are not the focus of this review and we only provide a very brief introduction about NWs synthesis in this section.

The initial strategies that have been widely applied are solution-based methods including hydrothermal method, solvothermal method, reflux method, microwave synthesis method and so on.^70–72 At the beginning of the synthesis, nucleation results in the formation of massive nanocrystals with different crystal faces exposed to the solution. The diversity of surface energies will lead to anisotropic condensation or the growth of these nanocrystals. If one of the crystal faces has a distinct surface energy, a preferred growth on this face will result in the formation of one dimensional morphology.⁷³ It is worth noting is that the surface energies are not only the intrinsic properties of materials but also affected by the solution environment or the choice of source reagents. The pH values of the solution are often investigated to control the growth of NWs, while some specific cations or anions are also observed to play important roles in the formation of 1-D structures. For example, Xu et al. studied the effect of capping agent, metal cations as well as pH values on the hydrothermal growth of 1-D ZnO nanostructures.⁷⁴ It was found that the long chains of oleate ions led to the formation of ZnO NWs (Fig. 1a) and nanorods as compared to CH₃COO⁻ or Cl⁻. Similar phenomena were also found in Birkel's result.⁷⁵

Fig. 1 SEM and TEM images of various types of 1-D nanostructures grown from both solution and vapor phase processes. (a) ZnO NWs obtained via hydrothermal method. Reprinted with permission from ref. 74 © 2011, The Royal Society of Chemistry. (b, c) Ag nanocables via hydrothermal method. Reprinted with permission from ref. 76 © 2005, Wiley-VCH. (d, e) In₂O₃ NW thin films via VLS method. Reprinted with permission from ref. 79 © 2010, Wiley-VCH. (f) ITO NW arrays via an epitaxial growth method. Reprinted with permission from ref. 92 © 2006, Wiley-VCH. (g) SnO₂ nanobelts via VS method. Reprinted with permission from ref. 99 © 2003, American Institute of Physics. (h) Well-aligned ZnO NWs via the VS method. Reprinted with permission from ref. 102 © 2005, American Institute of Physics.

Solution phase methods have been extensively explored to synthesize various NWs. Metallic NWs like Ag (Fig. 1b and c), Cu, Au, and Pt^76,77 and semiconducting oxide NWs such as ZnO, SnO₂, In₂O₃, Cu₂O, Fe₂O₃ and V₂O₅ have been successfully synthesized.⁷⁸ In spite of the decreased crystallinity and weakened electrical properties, it can be predicted that solution phase methods are promising ways for mass production and large scale application of NWs with the advantages of low cost and easy fabrication.

Large aspect ratio often helps to enhance the contact of the electrodes with the active materials in electronic devices like transistors.⁷⁹ It also facilitates the adsorption–desorption of O₂ in optoelectronic devices such as photodetectors.⁸⁰ That is why NWs find such a wide application in nano-devices. However, it is usually a challenge to synthesize uniform NWs, especially highly ordered NW arrays, by using general hydro/solvothermal methods. In this case, templates are adopted to control the material's growth in a confined domain or along a specific direction. Among all the adopted templates, anodic aluminium oxide (AAO) membranes are widely utilized. As a typical example, highly ordered CdS NWs arrays with lengths up to 1 μm and diameters as small as 9 nm have been obtained by using AAO templates. During the process, CdS was electrochemically deposited into the pores of AAO films from an electrolyte containing Cd²⁺ and S in dimethyl sulfoxide and the deposited material was found to be hexagonal CdS with the crystallographic c-axis preferentially oriented along the length of the pore.⁸¹ A similar method was also employed to synthesize several other 1-D nanostructures, such as In₂O₃, SnO₂, Fe₂O₃ and Fe₃O₄ NWs.^79,80–84

For the vapor phase processes, an important one is using catalysts to guide the growth of NWs that is generally described as the vapor–liquid–solid (VLS) method.⁸⁵ In the VLS growth process, metal droplets, usually formed by annealing a gold thin film prepared by sputtering (1–10 nm) or gold colloids,⁸⁶ are indispensable to serve as catalysts. These metal droplets capture the source vapor and form liquid alloy droplets; then, supersaturation and nucleation occur at the liquid/solid interface and lead to axial crystal growth. Owing to the catalysis mechanism, the NWs often have diameters equal to the droplets, and only grow at sites activated by metal droplets. The VLS method was first developed by Wagner and Ellis to synthesize single crystal Si NWs and has become one of the most successful approaches for the preparation of various kinds of NWs,⁸⁷ including element semiconductors like Si and Ge, metal oxides like ZnO and In₂O₃ (Fig. 1d and e), II–VI group semiconductors like ZnS and CdSe, III–V group materials like GaAs and InP, and others.^88–91

Using an appropriate substrate, well-aligned NW arrays can be fabricated with the orientation normally perpendicular to the substrate surface through catalyzed epitaxial growth. For example, vertically aligned tin-doped indium oxide (ITO) NW array has been synthesized using an ITO thin film as the substrate, as shown in Fig. 1f.^92,93 Besides the ability to tune the diameters of NWs by changing the catalysts' sizes, one of the advantages of the VLS method is the possibility to control the orientation, position and density of NW arrays that are very important to the assembly of NWs for device applications, through patterning the catalyst nanoparticles. Combining the epitaxial growth technique and nano-scale patterned metal catalysts fabricated by electron-beam lithography, photolithography, mask lithography by porous alumina or nanosphere lithography, highly ordered vertical aligned ZnO NW arrays with different densities have been successfully synthesized in hexagonal, orthogonal or other kinds of arrangements.^94,95

In addition, different kinds of heterostructures along axial or radial directions are also available through clever synthesis strategies. Branched NWs can be fabricated using a two-step VLS growth method, which has been utilized for SnO₂ hierarchical NWs.⁹⁶ Besides, NWs can be formed with different materials in the core–shell layout by controlling the growth conditions to induce homogeneous reactant decomposition on the core NW surface.⁹⁷ Through the repeated modulation of reactants, even multishells can be obtained. Axial heterostructures have also been formed by the time-dependent modulation of NW composition and doping during the growing process. These heterostructured NWs are essential for many potential applications in nanoscale optoelectronic devices.

Another process related to the VLS method that has also been employed in the growth of NWs is named vapor–solid (VS) technique. The VS technique does not require metal droplets serving as catalysts. Meanwhile, a higher temperature is usually needed to overcome the much higher activation energy when compared with the VLS method.⁶³ Metal oxide NWs such as ZnO,⁹⁸ SnO₂ (Fig. 1g),⁹⁹ In₂O₃,¹⁰⁰ CdO¹⁰¹ and many other NWs have been successively synthesized via the VS process in recent years, demonstrating its universality towards various NWs. In catalysts-assistant growth, the remains of catalysts or additives will unavoidably influence the purity of the final products; therefore, the fundamental properties of the materials will inevitably be affected. NWs synthesized via the VS process may also have high crystallinity and outstanding microstructure. For example, Zhang et al. have reported a simple vapor solid deposition approach to synthesize well-aligned ZnO NWs on c-oriented ZnO thin films using Zn as the vapor source.¹⁰² As shown in Fig. 1h, well-aligned NWs are in high density and are uniformly distributed over the substrate.

2.2 NW assembly and device fabrication

Between NW synthesis and device fabrication, a key challenge is the controlled assembly of NWs. In the initial phase of NW device research, studies concentrated mainly on individual devices. NWs are directly suspended in solutions and dropped onto the substrates. Randomly distributed and arranged NWs are obtained after the solutions are volatilized. These individual devices are fabricated to investigate the electrical transport properties of NWs. However, large scale device arrays and integrated circuits are impossible to create using this method due to the disordered arrangement of NWs. In order to overcome this challenge and tap the potential of NWs in flexible electronics, several post-growth techniques are developed. NWs assembly is not the focus of this review and Javey in UC Berkeley and Fan in HKUST have already provided very good reviews focusing on NWs assembly; hence, we only provide a brief introduction on NWs assembly here.

One of the earliest approaches to assemble NWs is the fluidic alignment in microchannels. Huang et al. constructed fluidic channels on flat substrates using poly(dimethylsiloxane) (PDMS) mold.¹⁰³ NW arrays are then assembled by passing suspensions of NWs through these channels. Fig. 2a shows the SEM image of InP NWs assembled on a flat substrate, showing that NWs are aligned along the flow direction. By controlling the flow rate, more than 80% of the NWs can be aligned within ±5° of the flow direction. Combined with surface-patterning methods to modify the surface chemical functionality, the fluid flow can generate well-ordered NW parallel arrays. With the use of a layer-by-layer scheme, this approach can also be used to organize NWs into more complex crossed structures, which are critical for building dense electronic device arrays.

Fig. 2 (a) SEM images of parallel arrays of InP NWs aligned by channel flow. Reprinted with permission from ref. 103 © 2001, American Association for the Advancement of Science. (b) Illustration of the blown bubble film NW alignment. Inset shows the dark-field optical image showing Si NWs in the film. Reprinted with permission from ref. 104 © 2007, Macmillan Publisher Ltd. (c) Optical image of a novel diode structure fabricated on parallel arrays of p-Si NWs assembled by the contact printing method. The upper right corner shows the schematic of the Pd-Al contacts and the lower right corner shows the regular array assembly of single Si NWs. Reprinted with permission from ref. 107 © 2008, American Chemical Society. (d) Schematic of the differential roll printing setup. Reprinted with permission from ref. 109 © 2007 American Institute of Physics.

NWs can also be organized within blown-bubble films. The shear stress created by expanding the NW suspension bubble is utilized by Yu et al. to align the high aspect ratio NWs along the principle direction of strain, as shown in Fig. 2b.¹⁰⁴ In this approach, a homogeneous polymer suspension of NWs is first prepared and then expanded using a circular die to form a bubble at the controlled pressure and expansion rate. The shear force formed during the expansion leads to the alignment of NWs. The aligned NW density can also be controlled by the concentration of suspensions. The organized NW blown-bubble film can be transferred to flexible plastic sheets as well as highly curved surfaces, showing great potential for large area flexible electronic applications.

Other methods to assemble NWs include the Langmuir–Blodgett (LB) technique and field-assisted orientation. The LB technique is widely used to deposit organic monolayers from a liquid surface to a solid substrate. In order to form NW monolayers, NWs are first functionalized with the surfactant and dispersed on the water surface. When the NWs are compressed, they will organize as parallel arrays with their longitudinal axes perpendicular to the compression direction. Magnetic field is also utilized by Hangarter et al. to assemble NWs.¹⁰⁵ However, their methods are limited to metallic or magnetic NWs. In another way, Freer et al. used a dielectrophoretic approach to achieve 98.5% yield of single NWs assembled over 16 000 patterned electrode sites with submicrometer alignment precision. By carefully controlling the hydrodynamic and dielectrophoretic forces, single NWs can be assembled on each pair of electrodes with high probability.¹⁰⁶

Recently, a new NW assembly approach called contact printing has drawn considerable attention, having the advantages of high efficiency, large scale applicability and compatibility with printing electronics. As shown in the lower right corner in Fig. 2c, the alignment of NWs is realized by directionally sliding the grown substrate with a dense “lawn” of NWs on top of the receiver substrate coated with a patterned photoresist. During the process, NWs are anchored on the receiver substrate by contacting and aligning by the sliding step. Multilayer NWs were assembled by Javey et al. using this approach and used as building blocks for three-dimensional, multifunctional electronic devices.¹⁰⁷ Fan et al. improved this dry transfer method by introducing a lubricant during the transfer process, which serves as a spacing layer between two substrates and minimizes the friction between NWs.¹⁰⁸ As a result, the uncontrolled breakage and detachment of NWs are prevented and highly ordered NW arrays are obtained. The printed NW density can also be modulated by appropriate surface chemical modification. Both dense and single NW arrays can be aligned through pre-patterning the receiver substrates. To further extend this approach for scalable and large area applications, a differential roll printing (DRP) process is developed. As shown in Fig. 2d, an important aspect of this strategy, as compared to previous methods, is the use of a cylindrical grown substrate.¹⁰⁹ NWs are deposited using the VLS process on glass or quartz tubes. The grown tube is then used as the roller and brought into contact with the pre-patterned receiver substrate. By rolling the tube with a controlled velocity, NWs are detached and transferred to the receiver substrate. This process provides a new approach for NW transfer and assembly in a cost-effective and large area manner.

Electrospinning is an efficient technology to fabricate randomly oriented and coiled NWs on certain substrates. A modified electrospinning method was introduced to print large-area organic semiconductor NW arrays uniformly with the desired orientation and position.^110–112 This kind of method has many outstanding advantages: (1) it fits for different types of NWs, such as organic NWs arrays with poly(3-hexylthiophene) (P3HT) and poly(ethylene oxide) (PEO) as the precursors, etc. (2) Semiconductor NW arrays can be successfully printed on almost any kind of substrate, such as SiO₂/Si, flexible polyarylate (PAR) or even common paper. (3) High-speed and large-area printing is another attractive advantage of this method. (4) The diameter of the NWs, the distance between NWs, and the NW density can be precisely controlled. All these advantages guarantee highly significant device fabrication work on the following steps, and a piece of typical work based on this method will be fully discussed in the following transistor section.

The successful synthesis of NWs and printing of NWs arrays make it possible to fabricate flexible electronic devices. The fabrication process for flexible electronic devices is similar to that of rigid electronic ones, that is, photolithography process followed by micro/nano-electrode metallization and lift-off. First, traditional photolithography is used to pattern contact areas onto both the ends of the NW or NW arrays. Specifically, after NWs growth and assembly on the flexible substrate, the photoresist (positive or negative) is spin coated onto the substrate followed by exposure using UV lithography machine with a patterned photolithography mask. The photoresist over unwanted areas is then removed using the developing solution. Then, the metal film is evaporated over the surface and the remaining photoresist is removed using acetone. At last, the electrodes over the contact areas are annealed at a proper temperature in an inert atmosphere to enhance the contact between the NWs and the electrodes. This photolithography, metallization and lift-off processes have shown themselves to be very resourceful for the fabrication of various flexible electronic devices before. For example, by adjusting the electrode materials, symmetric or asymmetric contacts between the NWs and the electrodes can be obtained. Fig. 2c shows a Schottky diode with asymmetric PA-Al contacts at different ends of the NWs, that is, the Pd forming a near Ohmic contact to the valence band and Al resulting in a Schottky interface. Besides, by adopting different modes of electrodes (T-shaped gate, top-gate, Ω-gate, etc.), by changing different dielectric films, by improving the surface passivation quality, or even via e-beam lithography techniques, a more sophisticated flexible device with smaller sizes can be fabricated.

3 Inorganic NWs based flexible electronic devices

Because organic materials can be formed into NWs with excellent flexibility, it is necessary to give a brief introduction on flexible electronics based on organic NWs, with a focus on their advantages as well as disadvantages. Recent reports have demonstrated that organic materials in a 1-D structure ranging from nanotubes to NWs can be really used for various applications in flexible electronics.^113–117 These 1-D structures can be grown or fabricated by both bottom-up or top down strategies. For example, Fig. 3a gives a detailed description of the home-built printer used to print organic NWs.¹¹⁰ In this electrospinning process, organic precursors were injected through a metal nozzle at a predefined rate. An x–y stage driven by a high-speed motor served as the collector. When applying a high voltage between the nozzle tip and the collector and moving the target substrate along a certain direction at the same time, organic NWs can be well aligned on the receiver substrates with the desired orientation, as shown in the inset. Fig. 3b shows the large-area organic NW based transistors array on a flexible PAR substrate, potentially offering the performance that is required for a new type of electronics along with desired transparency and flexible characteristics. Fig. 3c shows the large-area uniform copper phthalocyanine (CuPc) NWs with proper density and a high degree of orientation on the polyurethane acrylate (PUA) grating substrate, which was an ideal candidate for flexible optoelectronic devices.¹¹⁶

Fig. 3 (a) Schematic diagram of organic semiconductor NW printing and inset is the optical image of the well aligned organic NWs. (b) Large-area organic NWs FET array on a flexible PAR substrate. Reprinted with permission from ref. 110 © 2013, Macmillan Publisher Ltd. (c) SEM image of organic CuPc NW arrays. Reprinted with permission from ref. 116 © 2012, The Royal Society of Chemistry.

Organic NWs offer the possibility of direct NWs printing and NW-based flexible devices' integration. Organic NWs have the following advantages. (1) Organic NWs can be synthesized from simple, cost-effective, and scalable processes, which have shown promise for their assembly into functional systems. (2) The physical parameters of organic NWs such as size, orientation, surface states, defects and polarization can be easily controlled. (3) The relatively high degree of transparency has been proven to be very important for designing next-generation flexible electronic devices. For example, “see-through” electronic displays, as a completely new conceptual product, have attracted a lot of attention in recent years. On the other hand, organic NWs still have some drawbacks. (1) The relatively low mobility of organic NWs often leads to poor performance of the final devices, thereby restricting the range of application possibilities. (2) High contact resistance as well as mismatched interfaces between organic NWs and the metal electrodes also lead to the further worsening of the final device's performance. (3) Organic NWs do not enjoy full compatibility with existing IC manufacturing processes because of their poor heat tolerance. (4) If organic NWs are introduced to the large-scale commercial production, organic pollution is an obvious result, and sometimes unavoidable.

As we mentioned in the introduction part that inorganic NWs have several superior features when compared with organic NWs, in the following parts, we will, therefore, give details on flexible electronics based on inorganic NWs, ranging from basic electronic devices such as transistors to sensors, photodetectors and other complicated devices.

3.1 Electrodes

The development of flexible electronics requires high quality electrodes with a combination of high electrical conductivity and high flexibility, as well as optical transparency. Indium tin oxide (ITO) is the commonest choice for most current flexible nanodevices because of its high transmittance and conductivity. However, considering the scarcity of indium and the brittlement of ITO film, a new class of conducting material that can replace the ITO film should be designed. Recently, uniform conducting films such as carbon nanotube films, graphene films and especially metal NW films have been employed as stretchable and foldable conductors. For example, Madaria et al.¹¹⁸ fabricated Ag NWs electrode using a dry transfer printing technique on flexible polyethylene terephthalate (PET) substrates. The obtained conducting films show strong adhesion to a PET substrate over a large area with a low sheet resistance of 10 Ω sq⁻¹ and transparency of 85%, comparable to the conventional ITO films. Such a process also allows the preparation of high quality patterned films of silver NWs with different line widths and shapes, allowing the possible application of flexible electronic circuits. Details about the preparation of Ag conducting film were illustrated in Fig. 4a–c. Another example is the use of a Meyer rod coating method for the fabrication of Ag NWs based transparent electrodes.¹¹⁹ Ag NWs ink in a methanol solution with a concentration of 2.7 mg mL⁻¹ was prepared as the coating agent. As shown in Fig. 4d–f, pre-sonicated Ag NWs ink is dropped onto a PET substrate. Then, a Meyer rod is either pulled or rolled over the solution, leaving a uniform, thin layer of NWs network. After drying the film at a proper temperature, Ag NWs based electrodes with specular transmittance of ∼80% and with sheet resistance of 20 Ω sq⁻¹ can be achieved. Such a method is also applicable for other transparent metal NWs electrodes, such as copper,¹²⁰ a metal with the electrical resistivity almost as good as that of silver.

Fig. 4 (a) Schematic illustration of the transfer process of Ag NW networks from growing AAO substrate to the PET or glass substrate. (b, c) Photograph of a NW film on PET substrate and SEM image of the network, respectively. Reprinted with permission from ref. 118 © 2010, © Springer. (d–f) Schematic illustration of Meyer rod assisted coating method for Ag electrodes. Reprinted with permission from ref. 119 © 2010, American Chemical Society.

3.2 Transistors

A field-effect transistor (FET) is the most important and basic component in electronics. NWs have been widely investigated recently as the building blocks for flexible FETs, such as silicon, metal oxides and other compound semiconductors. To meet the demand for high speed devices, inorganic semiconductor NWs with high mobility are desired. Flexible InAs NW array FETs working on radio frequency were demonstrated by Takahashi et al. (Fig. 5a and b).¹²¹ The high-frequency performance of the InAs NW array FETs were analyzed by measuring two-port scattering parameters (S parameters) in the common-source configuration over a frequency range from 40 MHz to 10 GHz. Derived from these S parameters, the current gain (h₂₁), maximum stable gain (MSG), and unilateral power gain (U) extracted from the measured S parameters as a function of frequency are shown in Fig. 5c. The InAs NW FETs exhibit an impressive maximum frequency of f_max = 1.8 GHz and a cutoff frequency f_t = 1.08 GHz. At 1 GHz, the MSG reaches ∼6 dB, indicating that designing an RF amplifier in the gigahertz region is plausible based on InAs NW FETs. The high-frequency response of the devices is due to the high saturation velocity of electrons in high-mobility InAs NWs. More importantly, the InAs NW FETs do not exhibit significant electrical degradation even when bent to a radius of ∼18 mm, showing excellent mechanical flexibility. The unique characteristics in terms of the large on-current (I_on), large transconductance (g_m), high mobility and impressive maximum frequency from the InAs FETs pave the way to develop next-generation high-performance semiconducting elementary units for multifunctional optoelectronic device applications in the future.

Fig. 5 Optical image of (a) the printed InAs NW array FET fabricated on a flexible PI substrate and (b) the InAs NW region. (c) High-frequency behaviour of the InAs NW array FET. Reprinted with permission from ref. 121 © 2010, American Chemical Society. (d) Schematic illustration of the p-type Si NWs array based TFT and (e) the SEM image of the aligned NWs. (f) I_DSversus V_GS (V_DS = 1 V) for a p-type Si NWs array based TFT. Inset is the I_DSversus V_DS curves correspond to V_GS = −5, −4, −3, −2, −1 and 0 V. Reprinted with permission from ref. 122 © 2007, Macmillan Publisher Ltd.

In order to build power-efficient flexible logic gates and integrated devices, both n-type and p-type semiconductor NWs are required. However, most of the currently investigated NWs are n-type semiconductors. Flexible p-type NWs FETs are rarely reported. Doping is the first-thought routine to modify the conducting type of NWs that are used in flexible FETs, from n-type to p-type. Mcalpine et al. present a scalable and parallel process for transferring hundreds of pre-aligned p-type silicon NWs onto plastic to yield highly ordered films for flexible FETs.¹²²Fig. 5d shows the schematic illustration of flexible FETs with the electrodes and various labeled layers. Plastic served as a bendable substrate and 2 μm-thick SU-8 epoxy was used as a gate dielectric in this device. Fig. 5f shows the electronic performance of the top-gate FETs containing about 200 p-type Si NWs. The inset of Fig. 5f shows the I_ds–V_ds curves at different V_gs from 0 V to −5 V, with negatively increasing gate voltage: the conductance of the device was gradually suppressed, demonstrating the p-type nature of the B-doped Si NWs.

Another interesting work on flexible p-type FETs are PbSe NWs FETs encapsulated in polymethyl methacrylate (PMMA).¹²³ PbSe NWs were aligned under an electric field on varying dielectric stacks and fabricated into flexible FETs operated on kapton films, and aluminum oxide (30 nm) was used as the back gate dielectric and octadecylphosphonic acid (ODPA) was used to passivate the aligned NWs. The obtained device exhibited ambipolar characteristics, as evidenced by the electron transport in the I_ds–V_gs curves. A predominantly p-type nature under a negative bias voltage was observed; moreover, a predominantly n-type nature for positive gating was also confirmed by the ambipolar characteristics of this PbSe device. Contrary to the colloidal PbSe NW (NW) and nanocrystal (NC) FETs, this PMMA encapsulated PbSe NW arrays showed a fairly low hysteresis value (<1 V).¹²⁴

3.3 Sensors

With controllable geometry and high surface areas, NWs exhibited excellent response to outer environments, such as optical radiation, temperature variation, pressure differential, chemical molecules, pH value, proteins and DNA.¹²⁵ In this section, we highlight several recent significant results that have made impacts on the field of flexible and stretchable sensors, including photodetectors, stress sensors and artificial skins.

3.3.1 Photodetectors. A photodetector is one of the most important applications of semiconductor NWs and has attracted extensive attention recently. The photoresponse of NWs to light irradiation with different wavelengths mainly depends on their band gap. For instance, with large band-gaps, metal oxide NWs such as ZnO,¹²⁶ SnO₂¹²⁷ and In₂O₃¹²⁸ NWs are widely studied for the fabrication of high performance ultraviolet (UV) photodetectors. Metal chalcogenides NWs with moderate band gaps such as ZnSe,¹²⁹ Sb₂Se₃¹³⁰ and In₂Se₃¹³¹ are chosen as active materials for visible light photodetectors. Narrow band-gap semiconductors such as In₂Te₃¹³² and InAsSb¹³³ NWs have been demonstrated as sensitive materials for infrared photodetectors. In this section, the latest developments of flexible NW photodetectors are summarized, focusing on three types of photodetectors: NW networks based photodetectors, individual NW based photodetectors and NW array based photodetectors.

NW networks, or NW thin films, are emerging as promising structures for photodetectors as their advantages include easy device fabrication process. High performance photodetectors can be fabricated on various NWs thin films such as ZnO, SnO₂, In₂O₃, and multicomponent metal oxides. For example, Yan et al. synthesized Zn₂GeO₄ NW networks and fabricated efficient ultraviolet photodetectors on a SiO₂/Si substrate by using the photolithography technique.^134,135 However, expensive and tedious photolithography techniques were required in Yan's work, which would certainly act as a hurdle in the practical applications of NW networks based photodetectors. Recently, flexible photodetectors were fabricated on Zn₂GeO₄ and In₂Ge₂O₇ NW networks, as shown in Fig. 6. High quality single crystalline Zn₂GeO₄ and In₂Ge₂O₇ NW networks were synthesized via a conventional CVD method. By simply transferring the NW mats to a transparent adhesive PET tape and printing the silver paste as parallel lines (which served as electrodes), flexible photodetectors were successfully fabricated. Both Zn₂GeO₄ and In₂Ge₂O₇ NW networks based flexible photodetectors exhibited excellent photoconductive performance in terms of high sensitivity, excellent stability, and fast response and recovery time to the UV light.

Fig. 6 SEM images of (a) Zn₂GeO₄ and (b) In₂Ge₂O₇ NW networks prepared for flexible photodetectors. (c, d) TEM images of these two as-synthesized NWs, respectively. (e) NW networks transfer and flexible device fabrication process for a representative photodetector based on PET substrate. Reprinted with permission from ref. 38 © 2012 Optical Society of America.

Single NW based photodetectors are the most common NW based optoelectronic devices. They can be easily fabricated via UV photolithography or e-beam lithography techniques followed by metal film deposition techniques. A single NW photodetector is thought to be the most effective, which could directly and quantitatively study light acquisition as well as the optical-to-electrical conversion process. As an example, single ZnTe NW based flexible photodetectors were recently reported.⁵⁶ ZnTe NWs (Fig. 7a) exhibited p-type conductivity with an effect mobility of 11.3 cm² V⁻¹ s⁻¹ and the corresponding flexible photodetectors were proven to have the features of excellent flexibility, stability and sensitivity to visible incident light. As shown in Fig. 7b, the flexible device has very stable photocurrent-switching behavior under different light irradiance conditions, indicating its sensitivity to very slight variations in the incident light. In Fig. 7c, the device was bent to different curvature radii to precisely assess the optoelectronic performance of the device. The photocurrents were maintained at about 30 ± 1.2 nA and the on/off ratio stayed at the level of 98 ± 4 under certain test conditions. Besides, Fig. 7d shows the I–V curves of the flexible ZnTe photodetectors after 50, 100, 150, 200 and 258 cycles of bending; no obvious current degradation was recorded, indicating the mechanical robustness and electronic stability of the flexible device.

Fig. 7 (a) Photoresponse of the flexible ZnTe NWs photodetector under an improved incident light level and (b) at different bias values under the same illumination (532 nm, 1.2 mW cm⁻²). (c) The magnitude of photocurrents (red) and the corresponding on/off ratio (blue) under different bending radii. Inset shows the optical micrograph of the experimental setup. (d) I–V curves measured before and after different bending cycles. The operation voltage of Fig. 5(a) and 6(c) are 12.5 V and 9 V, respectively. The light intensity of (c, d) is 1.4 mW cm⁻². Reprinted with permission from ref. 56 © 2013 Optical Society of America.

An ordered NW array is a simple and efficient remedy for low-level photocurrent faced in single NW based photodetectors. Many technical methods have been summarized above to design arranged NWs, and the obtained NW arrays could be transferred to the flexible substrate directly for the fabrication of flexible nanodevices. In the report of Bai et al., integrated ZnO NW UV photodetectors were fabricated on rigid glass and flexible PET substrates at the macroscopic scale.¹³⁶ As shown in Fig. 8a and b, cross-finger shaped Ag electrodes were pressed onto the aligned ZnO NWs and provided an Ohmic contact between the Ag electrodes and the ZnO NWs on the PET substrate, which is confirmed by the I–V measurements in Fig. 8c. We can find that the photocurrent is about five orders of magnitude larger than the dark current, giving a very large ON/OFF ratio. Time-resolved UV photoresponse measurements under UV irradiance of 4.5 mW cm⁻² and 3 V bias are shown in Fig. 8d, confirming the ultrahigh ON/OFF ratio. This ultrahigh ON/OFF ratio implies the ability to sense weak UV incident light even in a high noise level environment. Using contact printed aligned NW arrays (Zn₃As₂, Zn₃P₂, etc.) as the active materials, different types of high performance flexible photodetectors were also successfully fabricated, with typical response to light irradiation with different wavelengths.¹³⁷ In fact, flexible NWs array based photodetectors are promising building blocks for next generation stretchable optoelectronic devices. Undoubtedly, more research results and exciting discoveries lie ahead.

Fig. 8 (a) Schematic design of an integrated UV sensor. (b) Optical image of ZnO NWs connected in parallel between Ag electrodes. (c, d) I–V curves and time-dependent photoresponse of the flexible UV sensor with/without UV illumination, respectively. Inset of (c) is the optical image of a flexible integrated NW UV sensor. Reprinted with permission from ref. 136 © 2011, Wiley-VCH.

3.3.2 Stress sensors. Stress sensing or pressure sensing is another important sensing property of NWs. Experiments have demonstrated the potential of NWs as stress sensors. For example, in Yang's work, the piezoresistance effect of silicon NWs, instead of Si bulk that can be applied to stress sensors, was systematically studied.¹³⁸ To be specific, individual p-type Si NW with 〈111〉 or 〈110〉 growth directions were grown in trenches on silicon-on-insulator (SOI) wafers to form bridge structures. Then, a quantitative relation between conductivity and stress/strain was measured by a four-point bending method and the corresponding piezoresistance coefficient along the growth direction was calculated. It was found that the conductance increases under compression and decreases under tension for a p-type 〈111〉 oriented Si NWs. Enhanced longitudinal piezoresistance coefficients as high as −3550 × 10⁻¹¹ Pa⁻¹ were obtained, in comparison with a bulk value of −94 × 10⁻¹¹ Pa⁻¹. This giant piezoresistance effect in Si NWs may lead to good application prospects of NW-based flexible stress sensors.

Technically, the Si-NW bridges on a SOI substrate adhered on the outer surface of a steel plate is not a precise flexible stress sensor, but a means of estimation in strain/stress and pressure. Another report on a fully packaged strain sensor device based on a single ZnO piezoelectric fine-wire (NWs, microwires) gives us a vivid description of real flexible stress sensors.¹³⁹ As shown in Fig. 9a, the flexible and optically transparent strain sensor device was fabricated by bonding a ZnO piezoelectric fine-wires laterally on a polystyrene (PS) substrate. Silver paste was applied at both ends of the ZnO wires to fix its two ends tightly on the PS substrate while serving as the source and drain electrodes. Further, a thin layer of polydimethylsiloxane (PDMS) was introduced to package the device. The I–V behaviors of the sensor device were measured under various strains, as shown in Fig. 9b. We can see that the I–V curves shift upward with tension strain and downward with compressive strain under various strain values. Besides, the current response of the sensor over many cycles of repeated compressing and stretching were also tested, and the results are recorded in Fig. 9c. Fig. 9d shows the partially enlarged drawing. The highly regular response curves indicate the high reproducibility and good stability of this kind of flexible sensor device. Gauge factor, which is the key parameter for a practical strain sensor, is defined as [ΔI(ε)/I(0)]/Δε, where Δε and ΔI(ε) are the amount of variation change of strain and the corresponding current. The highest gauge factor demonstrated for this kind of flexible stress sensor device is 1250, far exceeding the conventional doped-Si strain sensor (∼200). The outstanding performance of this flexible strain sensor based on ZnO nano/microwires shows its good application future in strain and stress measurements in the fields of science and technology.

Fig. 9 (a) Schematic (up) and optical image (bottom) of the flexible single ZnO based strain sensor device. (b) Typical I–V curves of the sensor at different strain conditions. (c, d) Current response of the flexible stress device that was repeatedly stretched at a frequency of 2 Hz under fixed bias of 2 V. Reprinted with permission from ref. 139 © 2008, American Chemical Society.

3.3.3 Artificial skins. Touch is an efficient way in which we interact with our surrounding environment and skin serves as a natural interface between humans and other things. The emulation of the stress senses by artificial means is a great challenge as well as an intriguing task. By applying a certain pressure over a tress sensor, a corresponding current or voltage signal could be obtained. Therefore, it is easy to imagine a kind of primitive artificial skin that is composed of neatly arranged stress sensor arrays, in which every single pressure sensing pixel is independently driven. However, the flexibility is still the most basic demand of artificial skins. This requires that we fabricate fully flexible electronic stress sensor arrays or at least sensor arrays based on some flexible substrate. As recently reported by Bao et al., flexible pixel-type pressure-sensor arrays with high sensitivity and very short response times have been designed.¹⁴⁰ In their work, organic field-effect transistors (OFETs) including cleverly designed thin, regularly structured rubber (PDMS film with 3-D microstructures) served as the active elements in press sensor devices. Fig. 10a and b shows the layout of the organic FETs employing a PDMS film as the dielectric layer. Under different applied pressures, the organic single-crystal transistors have different output currents (Fig. 10c). The change in current is proportional to the measured relative change in capacitance, which depends directly on the applied pressure (Fig. 10d). A primitive capacitive matrix-type pressure sensor was also designed to prove the possibility for use in multi-touch devices of this flexible electronic stress sensor. Details of the structure and its performance are shown in Fig. 10e and f.

Fig. 10 (a) Schematic of the pressure-sensitive organic field-effect transistors device that uses a microstructured PDMS film as the dielectric layer. (b) Optical image of the aluminium-coated microstructured PDMS film, indicating the high flexibility of the dielectric layer film. (c, d) Pressure response of the pressure-sensing organic single-crystal transistors based on the microstructured PDMS dielectric layer. (e) Flexible 8 × 8 pixel pressure sensor arrays fabricated by sandwiching the microstructured PDMS film between two PET substrates. (f) The pressure response to an above-placed tripod. Reprinted with permission from ref. 140 © 2010, Macmillan Publisher Ltd.

Parallel NW arrays were proven to be promising candidates for flexible and robust artificial skin, as reported by Takei et al.¹⁴¹ In their work, macroscale (7 × 7 cm²) integration of parallel NW arrays as the active-matrix backplane of a flexible pressure-sensor array were reported, as shown in Fig. 10b and 11a. Laminated pressure-sensitive rubber (PSR) is used as the sensing element and the source electrodes of printed Si/Ge NW-array based FETs (inset of Fig. 11a) are connected to drive the PSR. With variations of external pressure, the conductance of PSR rises or falls, leading to the different modulation index of the NW FETs and eventually results in the specific output signal of the pixel. By setting a certain frequency of the applied pressure, the corresponding time dependent output electrical signal of a pixel could be obtained. Fig. 11c shows the conductance-pressure profile with a frequency of 5 Hz, indicating the fast and stable response of this stress sensing unit. Finally, a 19 × 18 pixels array based artificial skin was successfully fabricated, as shown in Fig. 11d. A normal pressure of ∼15 kPa was applied on the surface of a ‘C’ shaped PDMS film that covered the artificial skin. By measuring the conductance of each pixel and introducing normalizations, the authors formulated the two-dimensional intensity profile (Fig. 11e). The contrast represents the normalized output signal, with the blue pixels representing the defective points. The character ‘C’, corresponding to the applied pressure area, illustrate the stable response of this flexible artificial skin. The uniform assembly of ordered Ge/Si NW arrays ensured the outstanding mobility of (∼20 cm² V⁻¹ s⁻¹) NW-array FETs and provided superb mechanical flexibility of the flexible artificial skin. NW also exhibits the single-crystalline nature, which could help in the low-voltage operation mode for the sensor circuitry. All these attributes make the parallel NW arrays an ideal choice for more sophisticated artificial skin.

Fig. 11 (a) Schematic of the passive and active layers of NWs based artificial skin. (b) Optical image of the individual stress sensing pixel. (c) Time dependent output signal for an applied pressure frequency of 5 Hz. (d) As-fabricated 19 × 18 pixel array artificial skin with a ‘C’ shaped PDMS mould placed on the top for applying pressure. (e) Corresponding two-dimensional output signal profile. Reprinted with permission from ref. 141 © 2010, Macmillan Publisher Ltd.

Although the above work demonstrated the fulfillment of artificial skin with NW arrays, it only involved the integration of pressure sensors with an NW based electronic readout circuit. It was very inconvenient to obtain an intuitive grasp on the performance of the flexible artificial skins, as the output of each pixel was individually measured and then statistically analyzed. Considerable progress was recently obtained by Wang et al., who designed a user-interactive artificial skin that not only spatially maps the applied pressure but also provides an instantaneous visual response through a tricolour OLED pixels array.¹⁴² This artificial skin basically consists of three parts, namely pressure-sensitive rubber, active-matrix backplane circuitry and OLED arrays. As shown in Fig. 12a, each sensing and display pixel consists of a TFT, an OLED and a PSR integrated vertically on a polyimide substrate. The enlarged optical micrographs of one pixel is shown in Fig. 12c, where we can see that the drain of the TFT is connected to an ITO pad that serves as the anode electrode for the corresponding OLED, while the cathode of each OLED is connected to the ground through the PSR. Fig. 12b illustrates an entire 16 × 16 pixels array of pixels, in which the scan lines (5/35 nm Ti/Au back-gate electrodes) and data lines (0.5/40 nm Ti/Pd source–drain contacts) are connected to 5 and 10 V, respectively, as shown in Fig. 12d. In Fig. 12e, letter shaped PDMS slabs were placed onto the PSR and uniform external force was applied on the PDMS slabs. The conductivity of the PSR increased accordingly and subsequently resulted in the underlying OLED turning on. A uniform performance of each TFT and single-pixel OLED allowed the pressure signal to be both spatially mapped and visually seen, as corresponding C-, A- and L-shaped bright light emitting areas were clearly recognized on the OLED panel (Fig. 12e, right). Besides, the color of the OLED can be effectively tuned by applying different emissive layer materials and the brightness of each OLED can also be sensitively controlled by the magnitude of the local pressure. Although the spatial resolution is far from usable in practical applications, this flexible artificial skin still demonstrated the practicability of user-interactive sensing system. By further improving the pixel integration level and using multiple sensor components, a superior kind of artificial skin that enables more sophisticated bionic operations can be successfully planned.

Fig. 12 (a) Schematic illustration of the user-interactive artificial skin. LiF/Al is the cathode that connects the PSR to each pixel of AMOLED. (b) Photograph of a fabricated device (16 × 16 pixels) artificial skin. (c) Enlarged view of the active area. (d) Circuit schematic of the artificial skin matrix. (e) As-fabricated 16 × 16 pixel array artificial skin with ‘C/A/L’ shaped PDMS slabs placed on the top for applying pressure. Reprinted with permission from ref. 142 © 2013, Macmillan Publisher Ltd.

3.4 Display devices

As an important display component, light emitting diodes have developed rapidly in the recent years. NWs based on various semiconductors offer the added advantage of high material quality leading to improved device performance. Basically, research in the area of NW based display components was focused on two aspects.

On one hand, semiconductor NWs were used as the optically active component in display devices. Individual gallium nitride (GaN) NWs have been configured as UV-light-emitting diodes on flexible plastic substrates by Lieber's group.¹⁴³ In this report, the authors assembled crossed n-type GaN NWs and p-type Si NWs onto a plastic substrate to form flexible crossed-NW LEDs, as shown in Fig. 13a and b. This kind of n-GaN/p-Si crossed diodes not only exhibited stable emission (Fig. 13c) but also could maintain their emissive properties upon repeated cycles of bending. In another work, vertically oriented ZnO NWs were fabricated into a highly flexible light-emitting device.¹⁴⁴ From the design scheme in Fig. 13d, we could see that the flexible LED was composed of four layers, namely, supporting (PET/ITO) substrate, vertically oriented ZnO NWs layer, insulating polystyrene (PS) film and the hole-injection anode layer. In particular, the hole-injection anode layer consisted of a thin poly(3,4-ethylene-dioxythiophene) poly(styrenesulfonate) (PEDOT:PSS) layer and an evaporated Au film. Fig. 13e and f show the overall image of the active ZnO NWs layer. All the NWs showed a uniform size with typically ∼2 μm length and 70–120 nm diameter. With or without Al doping, this diode structure showed excellent optical emission performance in the range of 500–1100 nm (Fig. 13g). These results suggest that semiconductor NWs could serve as high-performance building blocks for future lightweight display devices.

Fig. 13 (a) Schematic and (b) SEM image of a crossed-NW LED array on a flexible plastic substrate, in which p-SiNWs are vertically aligned and GaN NW is horizontally aligned. (c) Electroluminescence (EL) images of localized emission from forward-biased Si–GaN junctions. Reprinted with permission from ref. 143 © 2003, American Chemical Society. (d) Scheme for a flexible LED structure consisting of vertically oriented ZnO NWs grown on a polymeric ITO-coated substrate. (e, f) SEM images of the ZnO NWs grown on a planar transparent substrate and on a bent Au film, respectively. (g) Normalized electroluminescence spectra versus wavelength based on various LED structures. Reprinted with permission from ref. 144 © 2008, American Chemical Society.

On the other hand, semiconductor NWs were used as the transistor driving circuitry. NW transistor driving circuitry is of particular interest for future display devices because it offers high carrier mobility, optical transparency, as well as mechanical flexibility. Zhou et al. reported high-performance arsenic (As)-doped indium oxide (In₂O₃) NWs for transparent active-matrix organic light-emitting diode (AMOLED) displays.⁶ Thin-film transistors (TFTs) based on As-In₂O₃ acted as the key components of the driving circuitry. The synthesized As-doped In₂O₃ NWs was shown in Fig. 14a, which have high electron mobility of ∼1490 cm² V⁻¹ s⁻¹. The as-designed TFTs can be used to generate the driving circuit to control a variable-intensity OLED, as shown in the inset of Fig. 14b. Regulating the input voltage could efficiently control the light intensity of the OLED. Fig. 14c shows the circuit diagram that indicates the physical layout of the connections for the seven-segment AMOLED display. The most important components of this circuitry were the switching transistors (T₁), driving transistors (T₂), and one storage capacitor (C_st). They were introduced here to select a specified pixel, control the current and store data during one period for time-varying operations, respectively. This fully integrated seven-segment AMOLED display was proven to have good transparency of 40%, as shown in Fig. 14d. The background image could be partly seen through the AMOLED display arrays. The optical images of the seven-segment pixels with lighted numerical digits (Fig. 14e–g) indicated the outstanding and stable performance of the NWs transistor driving circuitry. By changing the ITO glasses with the PET substrate, the authors further proved the successful fabrication of flexible SWNT transistors for LED driving circuitry.⁴⁹

Fig. 14 (a) SEM image of As-In₂O₃ NWs. (b) Output current through the loaded phosphorus OLED (I_OLED) versus V_in with V_dd at 5.0 V in linear scale (red line) and log scale (blue line), respectively. (c) Circuit diagram of seven-segment AMOLED display unit. (d) Optical photograph of fully transparent AMOLED display driving circuitry. (e–g) The seven-segment AMOLED display unit displays numbers 1, 3, and 6, respectively. Reprinted with permission from ref. 6 © 2009, American Chemical Society.

3.5 Memories and logic gates

Over the past several years, semiconductor NWs have also represented attractive building blocks for memories and logic gates, which are essential components for the next generation circuits. Traditional memories and logic gates based on silicon have greatly promoted the development of the very large scale integrated circuit (VLSI) technology. However, in terms of flexible devices, for example, that are designed to mount on human skin or planted into internal organs, the constraint of traditional electronic units would certainly limit their potential applications. Therefore, fully flexible memories and logic gates have been greatly welcomed by scientific researchers for NW based electronic devices.

Lieber's group first reported the monolithic integration of individual and parallel arrays of germanium–silicon (Ge–Si) core–shell NWs as multilayer flexible circuits.¹⁴⁵ As shown in Fig. 15a and b, the final circuit consists of a lower layer of PMOS inverters and an upper layer of floating gate memory elements. From the input–output characteristics of the lower layer of PMOS inverter (Fig. 15c), it was found that these inverters exhibited p-type transfer characteristics with a maximum voltage gain of 3.5. Fig. 15d shows the AC inverter characteristics of the lower layer of PMOS inverters when driven by a 50 MHz sine wave supply of 4 V. At a frequency of 50 kHz, the input signals are ideally inverted to the output signals without any loss. Fig. 15e and f show the current vs. voltage sweeps recorded on the upper layer of the floating gate memory elements. A large hysteresis (i.e. memory window) and stable writing and erasing operation can be observed, respectively, indicating the good memory retention of the floating gate memory elements.

Fig. 15 (a) Schematics and circuit diagrams of inverter (top) and floating gate memory (bottom) elements. (b) Optical image of inverters (layer 1) and floating gate memory (layer 2) on Kapton. (c) DC inverter characteristics. (d) AC inverter characteristics. (e) Hysteresis in current–voltage characteristics of a memory layer element. (f) Switching characteristics of memory at V_CG = 5 V with ±15 V pulses (pulsing time: 1 ms). Reprinted with permission from ref. 145 © 2007, American Chemical Society.

Flexible logic gates were also fabricated on the base of well aligned NWs that were generated by top-down fabrication steps.¹⁴⁶ By using anisotropic chemical etching steps, Rogers et al. have once obtained GaAs nano/microwires with triangular cross sections from GaAs wafers. A slightly oxidized PDMS stamp was used to transfer the GaAs wire arrays from the original base onto a PET/PU (polyurethane) substrate (Fig. 16a). Simple circuits consist of various logic gates and individual MESFETs on a PET substrate are shown in Fig. 16b and c. Logic functions, such as NOR and NAND, can be performed by these flexible logic circuits. For example, Fig. 16d and e show the scheme of combining two GaAs wire switching transistors with another GaAs wire load transistor to yield NOR logic functions, as the logic status of V_out can only be logic 1 (high positive output voltage) when both inputs are logic 0 (at high negative voltages). Fig. 16f and g show the optical image of circuit diagrams and the voltage transfer characteristics of the NAND gate, respectively. It can be seen that the flexible NAND gate also shows very stable logic operation performance. Furthermore, Rogers and his colleagues have reported a medium-scale integrated digital circuit composed of up to about 100 SWNT transistors fabricated on a flexible plastic substrate.¹⁴⁷ This flexible digital circuit could be used to decode a binary-encoded input of four data bits into sixteen individual data output lines. This outstanding ability suggests that well-designed flexible integrated circuits based on NWs could be applied widely in many areas where conventional, rigid circuits cannot meet the requirements, for example, implantable electronics and prosthetics in the near future.

Fig. 16 (a) Transfer print process of GaAs wires and logic device fabrication process. (b) Optical images of various logic gates and individual MESFETs fabricated on PET substrates. (c) Magnified optical micrograph of an individual Ti/n-GaAs Schottky diode on the PET sheet. (d, e) Optical images and input–output characteristics of a NOR gate. (f, g) Optical images and input–output characteristics of a NAND gate. Reprinted with permission from ref. 146 © 2006, Wiley-VCH.

4 NW flexible energy conversion and storage devices

Accompanied with the development of flexible electronics, great interest has been aroused in integral flexible/bendable electronic equipment such as wearable devices, rollup displays and bendable mobile phones. However, the current energy storage devices containing lithium ion batteries and supercapacitors are usually too heavy, rigid and bulky to match the particular requirements of flexible electronics. Therefore, the trend in the next generation of energy storage devices development is to realize light, flexible and small units with shape-conformability, aesthetic diversity and excellent mechanical properties. NWs with high aspect ratio, fast charge transfer pathways and stable frameworks have attracted increasing research interest in the area of energy conversion and storage. In this section, we review some recent advances that NWs have made in the field of lithium ion batteries (LIBs), supercapacitors (SCs), solar cells, and generators.

4.1 Lithium ion batteries

Among the newly emerging energy storage devices, lithium ion batteries (LIBs) are one of the most popular candidates due to their relatively high energy density, long cycling life, good power performance and low cost.^148–178 For LIBs, the charges stored are mainly contributed by the intercalation of Li⁺ on the surface and in the bulk of the active materials, delivering higher energy density (120–180 W h kg⁻¹) than other energy storage devices.¹⁴⁸ NWs with a highly elastic structure demonstrated appropriate characteristics for novel flexible LIBs, which are in urgent demand to meet the development and promotion of flexible electronics. Recently, several excellent comprehensive reviews were given to introduce the fast development of flexible LIBs; therefore, we will not provide the details here, but only provide a brief introduction in this area.^179–182

Planar type flexible LIBs are one of the most important and earliest developed batteries. This kind of flexible LIBs has the features of very small thickness, light weight, good energy density and power density, and excellent cycling stability. One good example is the flexible planar LIBs with 3D hierarchical ZnCo₂O₄ NW arrays grown on a carbon cloth as the flexible electrode.¹⁶⁷ As shown in Fig. 17, the 3D nanostructure exhibits a distinct plateau between 0.8 and 1.2 V in the voltage window of 0.01–3.0 V and only 16.5% decay of initial capacity (1530 mA h g⁻¹) after 100 cycles (Fig. 17c). The specific capacity is almost constant in the range of about 1200–1340 mA h g⁻¹ with 99% capacity retention from 3 to 160 cycles. Fig. 17d illustrates the typical voltage profiles of the flexible battery device composed of ZnCo₂O₄/carbon cloth as anode and LiCoO₂/Al foil as the cathode by bending it from different directions for hundreds of cycles. The constant capacity, even after 120 cycles of bending, reveals that the electrical stability of the fabricated flexible full battery is hardly affected by external bending stress. The good adhesion and electrical contact between the NWs and carbon cloth, the facile diffusion of the electrolyte and accommodation of the strain induced by the volume change and the 3D hierarchical nanostructures resulted in better performance of the flexible battery. An enhanced power density was achieved once urchin-like ZnCo₂O₄ NWs were used.¹⁷²

Fig. 17 (a, b) Typical FESEM images of the ZnCo₂O₄ NW arrays grown on a carbon cloth at different magnifications. (c) Typical voltage versus specific capacity profiles for the first, second, 50th, and 100th discharge and charge cycles. (d) Typical voltage versus specific capacity profiles for the first, second, 50th, and 100th discharge and charge cycles. (e) Schematic representation and operating principles of rechargeable lithium-ion battery based on ZnCo₂O₄ NW arrays/carbon cloth. Reprinted with permission from ref. 167 © 2012, American Chemical Society.

Besides ZnCo₂O₄ NWs, other metal oxide nanowires were also used to fabricate flexible LIBs. Although these types of planar flexible LIBs exhibit many advantages when compared with conventional rigid LIBs, they still suffer from several drawbacks. For example, the size of planar flexible LIBs is still quite large and the device is too thick, which makes them unfit for integration with other flexible electronic or optoelectronic devices, especially in the case of on-chip integration for microelectronics applications. Besides, this kind of planar flexible LIBs cannot stand a large degree of curving or bending, which also makes it impossible for real flexible electronics applications.

One solution to overcome these problems is to develop ultra-small and stretchable LIBs.¹⁸³ Stretchability represents a rigorous challenging of mechanical stability. Stretchable devices must bear large strain deformation and large shape deformation, including not only bending but also twisting, stretching, compressing and other forms of deformation. Recently, Rogers et al. developed a very fancy type of highly stretchable LIBs with the active materials segmented design, and unusual ‘self-similar’ interconnect structures between them. The stretchable LIB was fabricated on thin silicone elastomers as substrates, as illustrated in Fig. 18a. As a result, the as-fabricated LIB obtains stretchability of up to 300%, and it still maintains a capacity as high as 1.1 mA h cm⁻² at the rate of C/2. Although the output power of the battery decreased slightly with strain resulting from increased internal resistances with strains at these large levels, surprisingly, the stretchable battery could turn on a commercial LED even when it was stretched by up to 300%, as shown in Fig. 18b and c. The high stretchability of the fabricated battery satisfies the requirements of many applications that are being contemplated for stretchable electronics.

Fig. 18 (a) Schematic illustration and exploded view layout of completed stretchable LIBs. (b, c) Lighting a red LED without and with strain. Reprinted with permission from ref. 183 © 2013, Macmillan Publisher Ltd.

Although significant progresses have been made in flexible LIBs, the following items should be considered before their real applications in flexible electronics. First, both the energy density and power density should be further improved to make them comparable with conventional rigid LIBs. Second, it is quite important and sometimes essential to fabricate ultra-small batteries to realize chip-level integration with flexible electronic or optoelectronic devices. Third, more delicate devices should be designed in case of special applications.

4.2 Supercapacitors

Supercapacitors (SCs), also known as electrochemical capacitors or ultracapacitors, are another kind of vital energy storage devices because of their long cycle life (>10⁵ cycles), high power density (>10 kW kg⁻¹) and high rate capability with fast charge–discharge capability within seconds. Similarly, with the configuration of LIBs, a SC is also composed of an anode, cathode, separator and electrolyte. The working mechanism of SCs can be classified into electrical double layer capacitors (EDLCs) and pseudocapacitors.^184,185 Since both the EDLCs and pseudocapacitors are surface electrochemical related devices, high surface area electrode materials are necessary in SCs, fitting well with the features of NWs. Similar with the construction of flexible LIBs, several kinds of flexible SCs have been demonstrated based on NWs (carbon materials, transition metal oxides and conducting polymers), including paper-like SCs, transparent SCs, woven SCs, micro-SCs and fiber SCs, which will be discussed in sequence in the following sections.

4.2.1 Paper-like supercapacitors. Besides high flexibility, paper-like SCs also exhibit ultrathin and lightweight properties. Free-standing CNT papers possess high conductivity and excellent flexibility, which can serve as both electrode materials and current collectors. However, the free-standing CNT paper with connecting bundles usually suffer from a low specific surface area, and therefore, a poor capacitive performance (less than 100 F g⁻¹), which limits the energy density of the SCs.^187–189 There are two ways to improve the performance of free-standing CNT papers. The first one is to incorporate active carbons and etched carbon materials.^{186,190–192} For example, Zheng et al. increased the specific surface area of a hierarchical CNT film from 670 to 871 m² g⁻¹ by using the CO₂ etching method, reaching the maximum value of 267.6 F g⁻¹ after loading 5 wt% of etched CNT into the ultralong CNT networks. The second way is to form hybrid transition metal oxides–CNTs or conducting polymers–CNTs to obtain higher energy densities.^193–200 Liu et al. designed flexible and ultrathin all-solid-state supercapacitors with two slightly separated PANI–CNT nanocomposite electrodes effectively solidified in the H₂SO₄–polyvinyl alcohol (PVA) gel electrolyte.¹⁹⁸Fig. 19a displays the configuration of the as-fabricated SC. The free-standing CNT networks, whose electrical conductivity is as high as 15 000 S m⁻¹, served as the scaffold for the incorporation of polyaniline and current collectors. The volume density of the CNTs was only 30%, leaving 70% of the total networks in the form of a porous structure. Even though the thickness of the device was estimated to be only 113 μm, the entire device shows superior mechanical flexibility. At 1.0 A g⁻¹, the discharge specific capacitance of the flexible device is 332 F g⁻¹ based on the total mass of the device, which is only 7.8% smaller than the one in 0.5 M H₂SO₄ solution (360 F g⁻¹) (Fig. 19b). A high energy density of 7.1 W h kg⁻¹ and a high power density of 2189 W kg⁻¹ were achieved, which is far superior to those of current conventional supercapacitors (Fig. 19d). After 1000 charge–discharge cycles, the device shows good cycling stability with only 8.1% decay and stable Coulombic efficiency ranging from 95.2% to 102.8%. During the time stability test displayed in Fig. 19e, the capacitance shape and redox peaks were effectively retained after 1 week and even 2 months, leading to a much stable device.

Fig. 19 (a) Schematic illustration of the PANI/CNT nanocomposite electrodes effectively solidified in the polymer gel electrolyte and digital pictures that show the all-solid-state device (size ∼ 0.5 cm × 2.0 cm) under normal conditions (top) and its highly flexible (twisting) state under electrochemical measurements (bottom). (b) Comparison of discharge abilities of the flexible PANI/CNT nanocomposite thin film electrodes in the H₂SO₄–PVA gel electrolyte and in the 0.5 M H₂SO₄ aqueous solution. The inset in (b) shows one cycle of galvanostatic charge–discharge curves at 1 A g⁻¹. (c) Comparison of CV curves at 5 mV s⁻¹ for the all-solid-state device tested in the normal, twisting and even folded condition, respectively. (d) Ragone plots for the electrode materials and for the entire all-solid-state device. (e) Time life stability of the device. Reprinted with permission from ref. 198 © 2010, American Chemical Society. (f) Schematic of Xerox brand printer paper treatment with PVDF. (g) Photograph of a Meyer rod-coated supercapacitor on Xerox paper. (h) Photograph of ink-jet printed supercapacitor on Xerox paper. (i) Specific capacitance at different current densities. Reprinted with permission from ref. 201 © 2010, AIP Publishing LLC.

Another kind of paper-like supercapacitors can be fabricated by coating conductive and electroactive CNTs or semiconductor NWs onto paper substrates.^{161,163,201–204} Among the reported achievements, Cui's group has dedicated a lot of efforts on paper-like supercapacitors by using the low-cost Xerox paper as the substrate. For instance, Cui et al. reported carbon nanotube thin film-based supercapacitors fabricated with printing methods, where electrodes and separators are integrated into single sheets of commercial paper.²⁰¹ Before coating the SWCNT ink on to the paper, the paper substrates were first treated with polyvinylidene fluoride (PVDF) to prevent the penetration of ink into the paper. The schematic illustration is presented in Fig. 19f. Fig. 19g shows a printed supercapacitor using the Meyer rod coating method. For large scale fabrication, an ink-jet printer can be utilized to print any patterns (Fig. 19h). Based on the voltage profile and mass density of SWCNT, the specific capacitance is calculated to be 33 F g⁻¹ (Fig. 19i) at a specific power of 250 000 W kg⁻¹. Due to the lightweight SWNT films and the absence of heavy metals, the capacitance of the assembled device could be greatly improved when compared with traditional supercapacitors.

4.2.2 Transparent supercapacitors. Flexible and transparent supercapacitors have recently attained great research interest for applications in flexible and transparent electronics such as transparent and flexible active matrix organic light-emitting diode displays that may find applications in heads-up displays, automobile wind-shield displays, and conformable products.²⁰⁵ NW based flexible and transparent electrodes were fabricated by depositing NWs (CNTs, carbon nanocup arrays, polyaniline NWs and In₂O₃ NWs) onto flexible and transparent substrates.^206–210 Chen et al. demonstrated a supercapacitor with the features of optical transparency and mechanical flexibility using In₂O₃ NWs–CNT heterogeneous film, as shown in Fig. 20a–d.²⁰⁶ The device delivered high specific capacitance of 64 F g⁻¹ and good stability.

Fig. 20 (a) Photograph of a flexible and transparent supercapacitor fabricated using CNT films. (b) AFM image of entangled CNT networks sitting on a transparent PET substrate. (c) Schematic of a flexible and transparent supercapacitor. The gray color represents a Nafion film as the separator between two In₂O₃ NWs/CNT heterogeneous film electrodes. (d) Cycle-life data of In₂O₃ NWs/CNT heterogeneous film electrochemical capacitor measured at 0.5 A g⁻¹. Reprinted with permission from ref. 206 © 2009, AIP Publishing LLC. (e) SEM images of branched nanocup film, where short carbon nanotubes (25 nm in diameter and 330 610 nm in length) are branched from the bottom of the nanocup. (f, g) Transparent and flexible natures of the supercapacitor devices. (h) CNC devices are compared with various energy storing devices by the Ragone plot that shows the values of energy density for power density. (SG: single layer graphene, RMGO: reduced multilayer graphene oxide, HGO: hydrated graphitic oxide, LSG-EC: laser-scribed graphene electrochemical capacitor). Reprinted with permission from ref. 208 © 2012, Macmillan Publisher Ltd.

More recently, Jung et al. constructed mechanically flexible and optically transparent thin film solid state supercapacitors by assembling nano-engineered carbon electrodes.²⁰⁸ Through the control of the nanopore dimension in anodic aluminum oxide (AAO) templates and CVD conditions, precisely tailored carbon nanocups can be fabricated, as shown in Fig. 20e. The as-fabricated electrode is highly conductive (117 S m⁻¹) and transparent. After being sandwiched by polyvinyl alcohol–phosphoric acid gel electrolyte, transparent and highly flexible supercapacitors are demonstrated (Fig. 20f and g). With the increasing measured temperature, a high areal capacitance of 1220 μF cm⁻² was obtained at 80 °C, which is almost three times than that (409 μF cm⁻²) at room temperature. The Ragone plot is shown in Fig. 20h. The volumetric peak power and energy densities of the branched carbon nanocup based supercapacitors are 19 mW cm⁻³ and 47 mW h cm⁻³, respectively, which is similar to some high performance graphene-based supercapacitors.²¹¹

4.2.3 Wearable supercapacitors. Smart textiles have unique functions such as thermal energy conversion, energy storage, sensors and communication.²¹² Among them, wearable supercapacitors are considered as an important element to drive other textile electronics. So far, much progress has been achieved on the design of wearable supercapacitors and NWs have contributed a lot towards their applications as conductive and electroactive materials. The substrates used in wearable supercapacitors include carbon cloth,^213–216 cotton fabrics,^217–219 non-woven cloths and so on.^220,221

Carbon cloth, as flexible, conductive and stable substrates, has attracted much interest to be used as current collectors in flexible supercapacitors. Our group has devoted some efforts on the development of flexible SCs by using carbon cloth current collectors to support NWs.^213,216 For example, we developed asymmetric supercapacitors (ASCs) based on acicular Co₉S₈ nanorod arrays as positive materials and Co₃O₄@RuO₂ nanosheet arrays as negative materials.²¹³ Acicular Co₃O₄ nanorod arrays were first grown on carbon cloth, acting as a template for further Co₉S₈ nanorod arrays growth or RuO₂ nanosheets deposition (Fig. 21a and b). The schematic illustration of the as-fabricated ASC is shown in Fig. 21c. Remarkably, the as-fabricated ASCs could be cycled reversibly in the range of 0–1.6 V in an aqueous electrolyte (LASC) and solid-state (SASC) electrolyte. The performance of SASC was preserved after being bended and twisted, demonstrating the desirability of carbon cloth-based current collectors (Fig. 21d). Both capacitors reveal excellent rate performance through five steps of charge–discharge rates, changing successively from 2.5 to 50 mA cm⁻². Moreover, high volumetric capacitance of 4.28 F cm⁻³ for SASC remained at 90.2% even after 2000 charge–discharge cycles. Electrochemical performance with an energy density of 1.44 mW h cm⁻³ at the power density of 0.89 W cm⁻³ for SASC was demonstrated.

Fig. 21 (a, b) SEM images of Co₃O₄ and Co₉S₈ acicular nanorod arrays grown on carbon cloth. (c) Schematic illustration of the asymmetric supercapacitors utilizing Co₉S₈–carbon cloth as positive electrode and RuO₂–carbon cloth as negative electrode. (d) CV curves of the ASC tested under normal, bent and twisted conditions. Reprinted with permission from ref. 213 © 2013, American Chemical Society. (e–g) Photograph and SEM images of CNT coated cotton cloth. (h) Ragone plot of commercial SCs and SWNT SC on porous conductors including all the weights. (i) The specific capacity for a stretchable SC before and after stretching to 120% strain for 100 cycles. The current density is 1 mA cm⁻². Reprinted with permission from ref. 217 © 2010, American Chemical Society. (j) Schematic illustration of the laminated structure of fabric SC. (k) Ragone plot of the laminated capacitors. (l) Schematic illustration of the tandem structure of fabric SC. (m) Galvanostatic charge–discharge curves measured at 0.566 mA cm⁻² for tandem MnO₂/CNT fabric SC with 1, 2, 3, 4, and 10 unit(s), respectively. Reprinted with permission from ref. 220 © 2013, Wiley-VCH.

Low-cost, flexible, stretchable and lightweight cotton cloth and non-woven cloth were also proven to be ideal substrates for wearable supercapacitors. Cui's group conformally coated single-walled carbon nanotubes (SWNTs) on cotton cloth to fabricate porous conductors.²¹⁷ As shown in Fig. 21e–g, the uniformly coated SWNTs make these textiles highly conductive with sheet resistance less than 4 Ω sq⁻¹. Supercapacitors made from these conductive textiles with large CNT loading mass (up to 8 mg cm⁻²) showed high areal capacitance, up to 0.48 F cm⁻² and good cycling stability (<2% decrease after 130 000 cycles). In the Ragone plot (Fig. 21h), the masses of the electrode materials (∼16 mg cm⁻²), cotton cloth (∼24 mg cm⁻²), electrolyte (∼6 mg cm⁻²), and separator (∼2 mg cm⁻²) were included in the complete SC device, which achieves high energy density of 20 W h kg⁻¹ at specific power of 10 kW kg⁻¹. When substituting the cotton cloth with stretchable fabrics, a flexible and stretchable SC is also feasible, showing excellent stability even after thousands of cycles (Fig. 21i).

Currently available energy storage systems often occupy more space than the electronic devices that they power. Therefore, the areal capacity and output voltage are key factors and can be tuned by textile electrodes. Recently, building on non-woven cloth electrodes, we presented two novel configurations (i.e., laminated and tandem) for SCs, which were expected to increase the energy density and output voltage, respectively.²²⁰ The fabric electrodes were prepared by dipping the non-woven cloth into a dispersion of carbon nanotubes and subsequent MnO₂ electrodeposition. In the lamination configuration (Fig. 21j), several pieces of MnO₂/CNT cloth were laminated in a KOH aqueous electrolyte to construct individual electrodes that enable fold-increased areal capacitances and excellent cycling stability. Expectedly, the areal energy density increased with the laminations while the power density decreased, accounting for the increased resistance, as shown in Fig. 21k. In the tandem configuration, each unit supercapacitor was made of two pieces of MnO₂/CNT electrodes sandwiched with H₃PO₄/PVA solid-state electrolyte. A high-output-voltage device could then be readily obtained by using a layer-by-layer assembly, as displayed in Fig. 21l. The charge–discharge curves of the MnO₂–CNT-based SCs with different numbers of unit cells (Fig. 21m) suggested that these tandem stack structures precisely follow the principle of tandem capacitors.

Another kind of wearable supercapacitor is a planar-shaped supercapacitor built on woven individual supercapacitor fibers or parallel arranged supercapacitor fibers. By using CNT coating on the common fibers (cellulose, metals and plastics) or CNT yarns, robust and flexible electrodes were obtained. To enhance the capacity, pseudocapacitive materials such as ZnO–MnO₂ nanorods arrays, ZnCo₂O₄ NWs and polyaniline NWs have been elaborated on these fibers.^222–226 Fiber-shaped supercapacitors can work independently or be woven into any desired shape. For example, Wang and co-workers presented a simple design and fabrication method for a high-performance, thread-like supercapacitor, taking the form of a two-ply composite yarn consisting of two carbon nanotube (CNT) single yarns that are infiltrated with polyaniline NW arrays.²²³ The CNT yarn possesses good mechanical properties, with tensile strength normally in the range of 500–800 MPa. The optical microphotograph of the fiber supercapacitor by twisting two CNT@PANI single yarns with the PVA gel electrolyte is shown in Fig. 22a. Fig. 22b shows an optical microphotograph of a model fabric composed of four conventional two-ply cotton yarns and four two-ply CNT@PANI@PVA yarn supercapacitors, giving a prototype of woven electronics. In Fig. 22c, the areal capacitance of CNT@PANI yarn-based supercapacitor and CNT yarn-based supercapacitor were compared, and a capacitance of 38 mF cm⁻² at a current density of 0.01 mA cm⁻² was achieved by the former one. When evaluated under different bending conditions, the capacitance of the yarn supercapacitor changed marginally (Fig. 22d). On the other hand, our group recently designed a flexible all-solid-state planar integrated fiber supercapacitor based on hierarchical ZnCo₂O₄ nanowire arrays/carbon fiber electrodes (Fig. 22e).²²⁶Fig. 22f depicts the measured specific capacitances of the flexible devices made of 2, 6, 10, 14, 20, and 30 composite fiber electrodes, presenting a significant rising trend with increasing fiber numbers. As shown by the equivalent circuit of the planar-integrated fiber supercapacitors (inset of Fig. 22f), this new structure exhibited an enhanced distributed-capacitance effect caused by the interaction among the non-adjacent opposite fiber electrodes, which can largely reduce the size of the device and achieve superior utilization efficiency and maximum functionality.

Fig. 22 (a) Two CNT@PANI@PVA single yarns twisted together to form a thread-like, two-ply yarn supercapacitor. Inset shows the SEM images of CNT@PANI composite yarn. Ordered PANI NW arrays are on the surface of the CNT yarn. (b) Yarn supercapacitors are co-woven with conventional cotton yarns to form a flexible electronic fabric with a self-sufficient power source. (c) Areal capacitance plots of two-yarn capacitors. (d) Capacitance retention under different bending states. Reprinted with permission from ref. 223 © 2013, Wiley-VCH. (e) Schematic illustration of the configuration of the device. (f) The analysis of the equivalent circuit and enhanced distributed-capacitance effect for the flexible planar-integrated fiber supercapacitors. Reprinted with permission from ref. 226 © 2013, Wiley-VCH.

4.3 Dye-sensitized solar cells

In 1991, the concept of using dye-sensitized colloidal TiO₂ nanoparticle films for photo-generated electrons was first proposed by Grätzel's group.²²⁷ It is a highly successful nanostructured solar cell that delivers energy conversion efficiency up to 15% till now.^228–231 Focusing on dye-sensitized solar cells (DSSC), we will review the most recent progress on NWs-based DSSCs. When compared with traditional mesoporous TiO₂ nanoparticle networks, nanowires offer much shorter and faster electron transport paths. Besides, NWs also scatter light and enhance light harvesting.²³² Moreover, in flexible DSSCs, due to the robust structure for bendable stress release and unique functions through space adjustment, NWs have drawn enormous attention in the investigation of traditional flat flexible solar cells and fibrous wearable electronics.²³³

4.3.1 Flat flexible DSSCs.
4.3.1.1 Photoanodes in flat flexible DSSCs. TiO₂ has been widely used as the photoanodes of DSSCs due to its suitable physical and chemical properties. Recently, some other semiconductor materials with a similar bandgap and properties to TiO₂ such as ZnO, SnO₂, and Zn₂SnO₄ have also been reported as photoanodes.^233–249 Transparent polymer substrates are efficient substrates for flexible DSSCs. However, the photovoltaic performance is hampered by slow electron transport and large interface resistance caused by the low temperature sintering that ITO polymer substrates undergo. Recently, Jiang et al. demonstrated that NW based bendable electrodes can release the bending stress effectively through space adjustment, while the mesoporous network suffers the cracking or peeling, as shown in the schematic in Fig. 23a and b.²³³ ZnO-NW arrays grown on PET/ITO substrates showed no visible cracks or signs of peeling off after bending (Fig. 23c and d). To overcome the low efficiency of NW-based DSSC caused by the lower surface area and lower dye-loading, ZnO nanoparticles were filled in the gaps between the nanowires. As expected, the modified DSSC combining both advantages exhibits higher efficiency and good flexibility (Fig. 23e). Later, Li et al. also developed a PVDF-nanofiber-reinforced TiO₂ electrode with high bendability.²³⁶ PVDF nanofibers could release the stress concentration in the electrode and minimize the occurrence of mechanical failure. After 500 or even 1000 cycles of bending tests for composite film based cells fabricated on ITO/PET, only a few cracks were visible, which is much more robust than the PVDF-free TiO₂ cells. Moreover, PVDF polymers containing fluorine atoms, which have the smallest ionic radius and largest electronegativity, are expected to facilitate the ionic transport of electrolytes and reduce the recombination rate at the semiconductor/electrolyte interface. A remarkable efficiency of 4.78% was achieved. Therefore, NWs open a new way for the fabrication of flexible DSSCs.

Fig. 23 Schematic representation of the bending stress release in nanocrystalline films of a mesoporous network (a) and NW array (b). SEM images of the as prepared ZnO NWs grown on flexible PET/ITO substrates by hydrothermal synthesis (c) and those after 2000 bending cycles with a bending radius of 5 mm (d). (e) Current–voltage (I–V) characteristics of DSSCs constructed using ZnO NWs (triangle) and NP-modified ZnO NWs (circle), and the I–V curve (dashed) of the NP-modified ZnO-NW DSSC after bending to 5 mm radius for 5 cycles under illumination: 100 mW cm⁻², AM 1.5G. Inset: absorptions of dye solutions containing dyes from ZnO-NW film solid and NP-modified ZnO-NW film (dashed), respectively. Reprinted with permission from ref. 233 © 2008, AIP Publishing LLC. (f) Photograph of a CNT/TiO₂ film attached to a PET film. (g) High-resolution SEM image of the CNT/TiO₂ film, showing uniform oxide coatings around CNTs. (h) IPCE spectra of the CNT/TiO₂ film and the P25/CNT electrode. (i) Resistance change of a CNT/TiO₂ film on a PET film after the film was bent to a radius of 0.4 cm for different cycles. Reprinted with permission from ref. 239 © 2013, Wiley-VCH.

A transferred NW film on flexible polymer substrates provides another way to fabricate flexible DSSCs.^237–239 The inverted process enables the thermal treatment of materials and prevents the thermal decomposition of the polymer substrates. An all-flexible DSSC was fabricated by assembling free-standing TiO₂ NW membrane with a Pt/Ti counter electrode and delivered an efficiency of 2.5%, which is comparable with the efficiency (3.8%) of a rigid device constructed using similar materials.²³⁷ In another effort, Di et al. reported a CNT/TiO₂ hybrid architecture for flexible photoelectrodes.²³⁹ As shown in Fig. 23f, the CNT/TiO₂ film can be transferred onto the PET film, exhibiting good flexibility. In the photoanode, CNTs not only act as a scaffold but also provide fast transport paths for photogenerated charges in the oxide. Without using additional transparent conducting oxide (TCO) substrates, this unique feature of the film boosts the incident photon-to-electron conversion efficiency to 32% in the UV light region without dye, outperforming TiO₂ nanoparticle electrodes fabricated on TCO substrates (Fig. 23h). The structural durability of the film was also tested by bending the film to a radius of 0.4 cm hundreds of times. Only 2–3% increase in the resistance is observed after 600 bending cycles (Fig. 23i). The inverting transfer method allows for the high temperature annealing to enhance the crystallinity.

Ti foil has been developed as both the precursor for TiO₂ NW arrays growth and substrates for subsequent DSSC fabrication. By facile anodization or hydrothermal processes, highly ordered meso-structured TiO₂ nanotubes or NWs can be fabricated, in which superior photovoltaic performances can be achieved through the realization of photoanode materials with an ordered structure.^{232,240–245} An excellent photoelectric conversion efficiency up to 8% has been achieved by the TiO₂ nanotube arrays coupled with the Pt-FTO glass counter electrode.²³¹ Recently, Kuang and co-workers reported a flexible DSSC using TiO₂ nanotube arrays on a Ti foil as the working electrode and Pt coated polyethylene naphthalate (ITO/PEN) as the counter electrode in combination with solvent-free ionic liquid electrolyte.²⁴⁰ TiO₂ nanotubes with a pore size of ∼70 nm was fabricated (Fig. 24a). By varying the anodization time, the length of the NWs/nanotubes could be controlled to fit for the best photoelectric conversion efficiency. The photocurrent–voltage curves are shown in Fig. 24b. Under AM 1.5 light irradiation from the back side, the flexible DSSC achieved an efficiency of 3.19%, which is comparable with that of the rigid DSSC assembled with Pt-FTO glass (3.3%). In order to obtain a flexible DSSC on the polymer substrates, Galstyan et al. sputtered Ti on a Kapton HN tape for electrochemical anodization, which enabled the post-treatment at 350 °C.²⁴⁴ The whole process of nanotube formation and cell fabrication is sketched in Fig. 24c. An almost-linear dependence of dye loading on the thickness of the nanotubes was experimentally found (Fig. 24d). A test cell based on a DSSC fabricated using the longest nanotube array (6 μm) was investigated under simulated sunlight irradiation in back-irradiation at AM 1.5G (100 mW cm⁻²) and a mechanical filter to obtain light attenuation (Fig. 24e). Expectedly, a high efficiency of 3.5% was obtained at the light intensity of 63 mW cm⁻². Mesoporous TiO₂ nanotubes have promising prospects for the development of innovative technologies in the field of low-cost flexible photovoltaics.

Fig. 24 (a) Schematic illustration of the flexible DSSC assembled with the TiO₂ nanotube arrays on Ti foil and Pt-coated PEN. (b) Current–voltage characteristics of solar cells based on different TiO₂ nanotube lengths using rigid (Pt/FTO-glass) and flexible (Pt/ITO-PEN) substrates. Solid lines were measured under AM 1.5 full sunlight (100 mW cm⁻²) illumination. The dotted lines were measured in the dark. Reprinted with permission from ref. 240 © 2008, American Chemical Society. (c) Schematic of fabrication of DSCs over flexible Kapton substrates. (d) UV-vis spectrum of N719 dye extracted from a series of TiO₂-nanotube arrays with different thicknesses. (e) J–V curves of a DSC under different simulated sunlight intensities. Back-irradiation geometry has been applied, as illustrated in the inset. Reprinted with permission from ref. 244 © 2011, Wiley-VCH.

4.3.1.2 Counter electrodes of flat flexible DSSCs. Pt is usually used as the catalytic materials for tri-iodide reduction and high conductivity. However, as a noble metal, Pt is expensive and exhibits instability in the methoxy propionitrile electrolyte solvents.²⁵⁰ Recently, several kinds of non-Pt materials have been exploited as catalytic materials, such as carbon materials (activated carbon, graphene, carbon nanotubes etc.),^251,252 conductive polymers (poly(3,4-ethylenedioxythiophene), PANI etc.),^253,254 inorganic materials (CoS, Co₉S₈, etc.²⁵⁵) and their compounds.²⁵⁶ 1D nanostructure based counter electrodes are one of the vital choices to prevent the fragmentation and detachment of materials under tough bending conditions. For example, DSSC with CNT-based coatings on conductive glass exhibited photoelectric conversion efficiency of 10%, which can be attributed to the high electrochemical catalytic activity, excellent conductivity and chemical stability.²⁵⁷ As a promising conductive material for transparent electrodes, CNT can be used as the conductive cover to replace ITO or FTO underlying coating and Pt catalyst at the same time.^258–260 The fabrication methods of CNT based flexible counter electrodes include the ink-coating process, growth-detachment-transfer process and the direct usage of free-standing CNT films.^259–262 For instance, Chang et al. established a Pt nanoparticle (NP)/CNT hybrid counter electrode on Ti foil by using a stable dispersant consisting of poly(oxyethylene)diamine (POEM) segment and imide linkage functionalities.²⁶¹ The solution was coated on the foil by the doctor blade technique, followed by annealing at 390 °C to obtain a flexible electrode (Fig. 25a). A conversion efficiency of about 9.04% was obtained (Fig. 25b). The CNTs not only enhanced the surface area and Pt loading but also served as the electrocatalyst. Recently, Malara et al. proposed a free-standing and ultrathin plate by compositing CNTs with polypropylene for the counter electrode,²⁵⁹ as shown in Fig. 25c. The extended etching treatment conditions dramatically benefit the electrochemical properties of the NC/electrolyte as well as the value of the series resistance of the plate (Fig. 25d). Flexible DSSC based on this plate yielded the highest efficiency of 6.67%. In their later work, a much enhanced efficiency of 7.26% was obtained by transferring a vertically aligned carbon nanotube forest onto the CNT–polymer plates, demonstrating a promising implementation of the CNTs as an engineered flexible counter electrode.²⁶⁰

Fig. 25 (a) Photograph of a flexible CE with Pt NP/MWCNT. (b) Photocurrent density–voltage curves of the DSSCs using Pt NP/MWCNT CEs with various ratios, measured at 100 mW cm⁻² light intensity. s-Pt was also used for comparison. Reprinted with permission from ref. 261 © 2012, The Royal Society of Chemistry. (c) Photograph of a CNT-based monolithic counter electrode and (d) schematic representation of etching treatment effect. Reprinted with permission from ref. 259 © 2011, American Chemical Society. (e) Co₉S₈ acicular nanotube arrays (ANTAs) are prepared from LCCH acicular nanorod arrays (ANRAs) by soaking in Na₂S solution (0.01 M) for 3 h with the help of the Kirkendall effect. (f) CV curves for sputtered Pt and Co₉S₈ ANTAs CEs prepared with different periods of chemical bath deposition. (g) J–V curves of the DSSCs using sputtered Pt and Co₉S₈ ANTAs CEs prepared within different periods of CBD. Reprinted with permission from ref. 264 © 2013, The Royal Society of Chemistry.

Metal oxides and sulphides were also used as counter electrodes of DSSCs.^263–265 A maximum power conversion efficiency of 7.67% was usually achieved for a cell with CoS nanorod arrays grown on FTO-glass, which is nearly the same as that of a cell with a sputtered Pt counter electrode (7.70%).²⁶³ Lately, the authors have fabricated Co₉S₈ acicular nanorod arrays on a conducting plastic substrate by a two-step approach successfully.²⁶⁴ By chemical bath deposition, layered cobalt carbonate hydroxide acicular nanorod arrays (ANRAs) were fabricated on a conducting plastic substrate, followed by a simple ionic-exchange process in Na₂S solution to convert the precursor into Co₉S₈ nanorod arrays under facile conditions (Fig. 25e). In Fig. 25f, all the CV curves for the Co₉S₈ ANTAs-60 min, Co₉S₈ ANTAs-90 min, Co₉S₈ ANTAs-120 min, Co₉S₈ ANTAs-150 min and Co₉S₈ ANTAs-180 min CEs show two pairs of redox peaks, which reveal similar electrochemical behavior to that of the sputtered Pt CE. Finally, a power conversion efficiency of 5.47% was achieved, which was comparable to that of the DSSC using sputtered Pt (5.62%) on the same plastic substrates (Fig. 25g).

4.3.2 Fibrous DSSCs. Recently, there has been a growing interest in making fiber-shaped photovoltaics owing to their promising applications in flexible electronics and wearable power supplies.²⁶⁶ When compared with flat DSSCs, thin fiber-shaped solar cells have the advantages of light weight and wearability. Generally, optical fibers, carbon fibers or metal treads (Ti, stainless steel) are used as fiber substrates. The basic structure of fiber-DSSCs consists of three kinds of assemblies, including the twisted, parallel arrangement by two wires or a single cable-like fiber wrapped by functional layers, as shown in Fig. 26a.^267–278

The twisted architecture is the most employed one among all the fiber-DSSCs, in which one fiber is coated with photoactive materials and the other one is a counter electrode. The counter electrode (Pt wires, carbon fibers, etc.) is twined on the working electrode with intervals to allow light absorption by the dye molecules. Zou's group has devoted considerable efforts on this kind of fiber DSSCs and achieved a high photo-electric conversion efficiency of 7% by using Pt wire as the counter electrode.^132,138 Peng's group reported a further enhanced result by utilizing anodized TiO₂ nanotube arrays on a Ti thread as the photoanode and Pt and Ni particle decorated CNT yarns as the counter electrode.²⁷⁷Fig. 26b displays the schematic illustration and SEM image of the photovoltaic wire. This fiber is fairly strong, exhibiting tensile stress of hundreds of megapascals. Decorated by 18.64 wt% and 9.33 wt% of Pt and Ni nanoparticles, respectively, the highest efficiency of 8.03% was obtained. Given the excellent performance, the twisted structure of fiber DSSCs is a promising candidate for practical applications. Chen et al. constructed a series module by five twisted fiber DSSCs, which can successfully drive three commercial light emitting diodes (LEDs) in parallel connection under illumination.²⁶⁹

Fig. 26 (a) Schematic illustration of twisted, parallel and cable-like solar cells. (b) Photovoltaic wire by twisting a TiO₂ nanotube-modified titanium wire that serves as the working electrode and an Fe₃O₄–CNT composite fiber that functions as the counter electrode together. Reprinted with permission from ref. 277 © 2013, Wiley-VCH. (c) Double-metal-wire/PET, double-sided, transparent DSSC device based on Pt–Fe microwires and ZnO NW–Fe microwires. Reprinted with permission from ref. 267 © 2012, Wiley-VCH. (d) Illustration of a coaxial single-wire structure DSSC, consisting of a core Ti wire, dye-grafted TiO₂ nanotube arrays, and a flexible, transparent CNT film wrapping around the wire. Reprinted with permission from ref. 273 © 2011, American Chemical Society. (e) Design and principle of a three-dimensional DSSC based on an optic fiber. Reprinted with permission from ref. 279 © 2009, Wiley-VCH. (f) Procedure of all-solid, flexible solar textiles fabricated using DSSC with ZnO NR arrays on SS wires. (g) Photograph of a fabricated flexible woven solar cell. Reprinted with permission from ref. 272 © 2013, Elsevier B.V. (h) A fiber cell being woven with the other CNT fibers into a textile. (i) A fiber cell being woven into a textile composed of aramid fibers. Reprinted with permission from ref. 280 © 2012, American Chemical Society.

Since the twisted structures usually lead to a large part of the working electrode being not covered by the Pt counter electrode and slight wastage of the covered area, parallel arrangement was thus proposed.²⁶⁷ Wang et al. developed a novel methodology for the fabrication of an extended-area, flexible, transparent, double-sided, planar DSSC, formed with metal wire/ZnO-NW arrays as the working electrode and a Pt wire as the counter electrode. Highly ordered, single-crystalline, and high-density ZnO-NW arrays were uniformly deposited onto the Fe microwires. The two metal wires were placed parallel to each other and encapsulated in a poly(ethylene terephthalate) (PET) or polydimethylsiloxane (PDMS) chamber containing the electrolyte (Fig. 26c). The double-wire DSSCs remain stable for a long period of time and can be bent at large angles, up to 107°, reversibly, without any loss of performance.

Cable-like solar cells are also proposed by assembling the active materials on a single substrate wire layer by layer. Zhang et al. developed a cable-like DSSC based on a single Ti wire by wrapping a carbon nanotube film around the Ti wire-supported TiO₂ tube arrays as the transparent electrode (Fig. 26d).²⁷³ The CNT film ensures full contact with the underlying active layer, as well as uniform illumination along the circumference through the entire DSSC. The single-wire DSSC shows a power conversion efficiency of 1.6% under standard illumination (AM 1.5, 100 mW cm⁻²). As the cable-like DSSC fiber works on back illumination with the bottom side unilluminated, Wang's group integrated optical fibers and NW arrays as 3D cable-like DSSCs, as shown in Fig. 26e.²⁷⁹ On the lower region of the fiber surface, dye-coated ZnO nanorod arrays are grown on the outer side of the optical fibers, and then assembled with a Pt tube. The key principle is that the light entering from the axial direction inside the fiber experiences multiple internal reflections along the fiber and finally reaches the dye molecules. When the DSSC is illuminated along the fiber axis, an efficiency of 0.44% is achieved, which is much better than the value (0.071%) obtained by vertical illumination.

Based on the successful fabrication of fiber DSSCs, wearable DSSCs can be elaborated via a convenient weaving technology.^272,280 For instance, Chae et al. reported a woven DSSC by using ZnO nanorod vertically grown from fiber-type conductive stainless steel (SS) wires as the photoanode and Pt-coated SS wires as the counter electrode.²⁷² As shown in Fig. 26f, all the solid, flexible solar textiles were fabricated by coating dyes and solid-state electrolyte, followed with PET film package. The satin weaving structure was particularly employed due to a larger illumination area as compared to plain and twill weaving structures. The all-solid, flexible properties of the solar textiles fabricated with ZnO NR-based DSSCs are shown in Fig. 26g. The solar textile with 10 × 10 wires exhibited an energy conversion efficiency of 2.57% with a short circuit current density of 20.2 mA cm⁻² at 100 mW cm⁻² illumination. Chen et al. also reported a woven DSSC by CNT fiber based solar cells.²⁸⁰ A twisted fiber DSSC consists of a CNT fiber and TiO₂ coated CNT fiber can be woven into CNT textiles and aramid textiles successfully (Fig. 26h and i), achieving an efficiency of 2.94%.

4.4 Generators

Since it was first proposed by Wang et al. in 2005, the self-powering nanotechnology has attracted much attention in the energy harvesting area. This concept came into existence to serve the interests for driving portable electronics. With the rapid development and wide use of nanotechnology, it is the batteries rather than the devices that finally determine the total size and weight of the system. Moreover, there is a pressing need to design a kind of rather small generator that could provide power for some micro/nano devices or medicine related devices, for example, implantable health monitors and artificial organs etc. Therefore, to meet these technological challenges, it is highly necessary to develop the science and technology closely related to nanogenerators.

Wang and his co-workers demonstrated an approach to converting mechanical energy into electric power by using vertical aligned piezoelectric (PZ) NWs,²⁸¹ and accordingly, designed a series of nanogenerators with sufficient electric output. Here, we only provide one example, as it illustrates the basic principle and technology of nanogenerators.²⁸² In their experiments, vertically aligned ZnO NWs were transferred to a receiving substrate to form horizontally aligned arrays and then parallel-strips-like electrodes were deposited on the horizontal ZnO NW arrays. By connecting a bridge rectifier and a capacitor, a simple self-powered device that could harvest energy from the mechanical bending of the flexible substrate was successfully fabricated. Fig. 27a shows the SEM image of ZnO NW arrays with deposited Au electrodes and the inset shows the photograph of the as-fabricated nanogenerator. Orderly aligned ZnO NWs with a diameter of ∼200 nm served as the active material of the nanogenerators. By an externally applied physical force, the flexible substrate was bent gradually, and correspondingly, the NWs started to deform. The piezoelectric property of the ZnO NW induced a piezoelectric field along the length, which naturally caused a transient charge flow in the external circuit, as shown in Fig. 27b. Repeated bending and releasing of the substrate induced an alternating flow of charges in the external circuit. Time dependent open circuit voltages and short circuit currents are shown in Fig. 27c and d, respectively. From the figures, we could see that an open-circuit voltage as high as 2.03 V and a peak output power density of ∼11 mW cm⁻³ could be achieved from a single layer NW nanogenerator structure. As shown in the inset of Fig. 27e, an integrated full wave rectifying bridge was connected with a red commercial LED to prove the successful fabrication of the high-output flexible nanogenerators. A total voltage source of 3.7 V and a maximum discharging current of 4.5 mA result in the lighting lasting for 0.1–0.2 s of the LED, as shown in Fig. 27f and g. Through calculations, the effective energy generation efficiency is estimated to be ∼4.6%. Further, by optimizing the density of the NWs on the substrate as well as the use of multilayer integration, many more practical nanogenerators with high open-circuit voltages, high short circuit currents as well as much higher energy generation efficiency are bound to be fabricated.

Fig. 27 (a) SEM image of active area of the nanogenerator. Inset shows the photograph of an as-fabricated nanogenerator. (b) Demonstration of the operating principle of the nanogenerator when mechanical deformation is induced, where the “±” signs indicate the polarity of the local piezoelectric potential created in the NWs. (c, d) Open circuit voltage and short circuit current measurement of the nanogenerator. The measurement is performed at a strain of 0.1% and strain rate of 5% s⁻¹ under the frequency of 0.33 Hz. Insets are the enlarged view of the single cycle of the signal. (e) Output voltage measured when connected with a full wave rectifying bridge. Here, the arrowhead points out that the signals of the negative signs are reversed. Inset shows the schematic of the charging–discharging circuit. (f, g) Images of the red LED before it was lit up and at the moment when it was lit up by the energy generated from the nanogenerator, respectively. Reprinted with permission from ref. 282 © 2010, American Chemical Society.

5 Integrated devices and systems

The successful fabrication of both NW-based flexible devices and energy conversion/storage sources offer numerous opportunities for the fields of energy system, electronics, automobile, and mechanical equipment. The integration of these semiconductor nanodevices, especially passive devices with power sources, would be of great interest in micro/nano systems' applications. The design and fabrication of high-performance integrated nanopower–nanodevice system, especially flexible systems, present several challenges, for example, impedance matching, electrical/mechanical stability and the improvement of operation efficiency etc. The challenges and limitations of these factors complicate the corresponding improvement. However, there are still a few studies that have focused on the assembly of basic nanodevices into an integrated simple system that can work independently. In this section, the methods and the ultimate performances of several integrated systems will be briefly described.

Recently, Wang et al. proposed a flexible self-powered nanosystem, consisting of a multiple lateral-NW-array integrated nanogenerator (VING) and a NW based detector.²⁸³ The eventual aim of this design is to be self-driven, that is, the realization of the stable and persistent operation of passive devices without using a battery or external power source. The key to provide a high output voltage and power of the VING is the rational growth of the NWs during the VING fabrication procedure, as the voltage and power produced by a single NW are insufficient for a real device. Fig. 28a shows the experimental procedures designed to fabricate a flexible VING. Briefly, rows of stripe-shaped ZnO pattern with a top and side layer of chromium was first achieved; then, a wet chemical method was used to obtain horizontally grown ZnO NW arrays. Next, gold electrodes were deposited only on the abovementioned chromium layer side, and finally, a layer of soft polymer was spin-coated onto the device to improve the stability and mechanical robustness of the entire structure. Fig. 28b is the SEM image of a single-row VING structure, indicating the ordered alignment of the NWs and the good contacts between the NWs and the electrodes. The final flexible VING was made of 700 rows of NWs on a 125 μm Kapton film. In response to a low-frequency mechanical strain of 0.19% at a straining rate of 2.13% s⁻¹, the average output voltage and the current pulse from the VING were measured to be ∼1.2 V and ∼26 nA, respectively. To demonstrate the ‘self-powered’ performance, the VING was integrated with a single NW-based nanosensor. First, a ZnO NW-based pH sensor was connected with the VING and the voltage across the nanosensor was monitored while changing the degree of acidity versus alkalinity in the target solution (Fig. 28c). An obvious and stable voltage variation corresponding to a local pH change was observed, indicating the high sensitivity of the ‘self-powered’ pH meter. Then, a ZnO NW-based UV sensor was also integrated with the VING (Fig. 28d). When the UV light is turned off, the resistance of the ZnO NW is measured to be ∼10 MΩ, which is comparable to the inner resistance of the VING. Further, the voltage drop across the ZnO NW-based UV sensor was ∼25 mV. Upon UV light illumination, the resistance of the ZnO NW dropped dramatically, and the voltage drop across the ZnO NW-based UV sensor was very small. This clearly indicates that a VING with an output voltage of 20–40 mV can successfully power up various kinds of single-NW-based nanosensors.

Fig. 28 (a) Schematic of the fabrication process of multiple lateral-NW-array integrated nanogenerators. (b) SEM images of a single row of the NWs bonded by metal electrodes. (c) Voltage drop across a single ZnO NW-based pH sensor powered by a VING with an output voltage of ∼40 mV. (d) Voltage drop across a ZnO NW-based UV sensor powered by a VING with an output voltage of ∼25 mV. Insets show the schematics of the NW-based integrated systems. Reprinted with permission from ref. 283 © 2010, Macmillan Publisher Ltd.

Very recently, our group reported a self-powered photodetector, which was driven by a flexible fiber shaped supercapacitor (Fig. 29a and b).²⁸⁴ In this work, uniform Co₃O₄ NWs (Fig. 29c) were successfully synthesized on different types of metal fibers, namely, nickel and titanium; thereafter, they were fabricated into two kinds of flexible all-solid-state supercapacitors. The as-synthesized Co₃O₄ NWs were proven to exhibit high capacitance of 2.1 F cm⁻³ at a current density of 20 mA cm⁻³ and the final wire shaped all-solid-state supercapacitor was demonstrated to have an impressive capacitive behaviour as well as long cycling performances (as shown in Fig. 29d and e). As discussed above, piezoelectric (PZ) ZnO NWs based nanogenerators could serve as an excellent energy-scavenging unit and could be integrated into a nanosystem to power electrical devices without the need of external batteries. By the same token, Co₃O₄ NWs based all-solid-state supercapacitors here could also act as power sources. The as-prepared visible light photodetector made from carbon fibers/graphene, which also served as the negative electrode of the flexible all-solid-state asymmetric supercapacitor, was proven to have substantial response to white light irradiation (Fig. 29f). Our primary research work here indicates that the wire shaped flexible supercapacitors can efficiently be used to power electronic and optoelectronic devices, and if the material quality and the device structure could be continually improved, more complicated self-powered devices and systems with minimised sizes and promoted performances will be designed. The harvesting of energy from ambient sources, energy storage/conversion and energy utilization will be more efficient.

Fig. 29 (a, b) Photograph and schematic of the self-powered nanosystem powered by the Co₃O₄–graphene based fiber-shaped supercapacitor. 1–3 illustrate the different bending states of our flexible asymmetric supercapacitor. (c) SEM images of the as-synthesized Co₃O₄ NWs grown on nickel fiber used for supercapacitors. (d) CV curves and (e) galvanostatic charge–discharge curves of the fiber-shaped supercapacitor. (f) Current response of the photodetector powered by the supercapacitor. Reprinted with permission from ref. 284 © 2014, Wiley-VCH.

6 Conclusions and outlooks

In this paper, the recent advances in the design and fabrication of NW based flexible devices and their integration have been reviewed. We first present a brief description of different approaches to realize the synthesis of a versatile group of semiconductor NWs and a few methods for the assembly of NWs array. Among the various semiconductors across an expansive area covering from metal oxides to III–V compound, II–VI compound and even multi-metal compounds, many NWs and their combinations such as network, film as well as ordered arrays have shown significant promise for electronic, optoelectronic, photovoltaic, electrochemical and piezoelectric applications. Combining the unique geometric structure of NWs and the superior physical/chemical properties of semiconductors, many kinds of typical nano/micro devices have been designed and fabricated. We show that both individual NW and NW assembled structures offer tempting possibilities in flexible device applications, for example, flexible electrodes and transistors, flexible sensors, flexible display devices, flexible memories and logic gates, as well as flexible energy conversion and storage devices. However, for the NW synthesis and assembly, devices fabrication and system integration processes for economically viable applications require more significant research works.

First, even the research in various NWs has witnessed rapid development and substantial achievements; however, certain issues, such as the following, still persist:

(1) Lack of good control over the size of NWs: the difference in properties caused by size heterogeneity is very obvious. A more efficient synthesis method for more precise control of the NWs' size is needed. Owing to the small dimensions and tendency to aggregate, manipulating NWs grown by bottom-up approaches is still challenging. There are two conceptually different types of approaches that can be used to address this problem. The first, a novel 3D printer can be invented to print quality NWs or uniform NW arrays directly. This scheme overcomes the inherent non-uniformity of NWs produced by traditional bottom-up growth process while retaining the advantages of NWs themselves. The second, improved top-down approaches are still attractive when integration and addressability are concerned. As for the manipulating of high density NW arrays, especially for addressability at the single NW level, a new kind of top-down approach has the advantage of being more amenable for large scale integration.

(2) Difficulties in controlling the chemical or physical properties of NWs: NWs with uniform size do not guarantee uniformity in their properties such as electronic transport characteristics. To minimize the NW-to-NW variability is very important but still remains a challenge.

(3) Combinations of different modifying methods over semiconductor NWs need to be systematically studied. Doping, surface modification, organic–inorganic hybrid, and the formation of multiple compounds or heterojunctions have been extensively studied for more than a decade. Still, a successful modification method would require the consideration of numerous factors such as surface states, defects, traps, catalysts-induced contamination, orientation and polarization of NWs. It seems a bit difficult to select an efficient way to modify the as-prepared NWs based on their ultimate use. Therefore, establishing an effective method for determining the modification process is fairly critical.

(4) The yield of NWs with high quality for scalable device fabrication and integration is significantly insufficient.

(5) In particular, each current NW assembly technique has its advantages and weaknesses, and it creates certain difficulties in printing NWs precisely on flexible substrates. The existing NW assembly techniques are also inefficient and incur high time costs. More appropriate NW assembly techniques need to be improvised and different types of assembly methods should be combined to yield high quality NW arrays.

Second, there is an obvious challenge in optimizing the configuration and fabrication process of flexible devices. By taking the advantages of 1-D geometric features and excellent physical/chemical properties of NWs, a group of nanodevices have been created. However, most of them are just the basic unit that could only be used for NWs' properties measurement. In other words, most existing nanodevices are inappropriate for practical applications. For example, NW based photodetectors are still at an early stage. Most current NW photodetectors have much lower performance in efficiency and bandwidth when compared with their classical counterparts,²⁸⁵ and the device-to-device variability of the performance has largely limited their development and applications. Besides, measurement theory and methods related to device properties should be found to the needs of specific and novel devices, for example, some sensors work in a harsh environment²⁸⁶ or under variable conditions. New measurement and evaluation methods then should be formulated. Although the optimization of the electronic/optoelectronic performance of many flexible devices is still under way, we are of the view that offering a few new, entry-level products provides a new impetus for exploring this largely uncharted territory. Roll-up portable displays, sensory skins and electronically steerable antenna arrays for wireless communication, in our opinion, will be in considerable demand during the coming years. Furthermore, commercializing some of the important applications will undoubtedly stir up more interest in this area.

Third, flexibility is one of the highlights of several devices and much more importance should be continuously attached to it. Flexible display devices have been a part of several promising ventures. If an efficient way can be found to ensure both high performance and outstanding flexibility for nanodevices, much more functional industrial products based on 1-D nanostructures will be designed.

At last, appropriate combinations of two or more as-fabricated NW devices will lead to some particularly exciting achievements. Combining appropriate electronic devices with energy units into a fully flexible system, in particular, is of special significance. Owing to the urgent demand of next-generation consumer electronics and with the rapid development of nanotechnology, integrated 1-D NW devices and systems are the most promising field to be realized in the near future. Besides, miniaturization, high integration and portability are the necessary requirements of future electronics. For example, nanoscale sensors can be equipped with tiny self-powered elements and microwave communication components to serve as body-implantable sensing networks. Considering these, many start-ups want scientists to engage in the applied research of fully flexible systems, rather than fundamental research of individual components. Fundamental breakthroughs, although important, should not dominate our research results list. The future of flexible inorganic NW electronics, in our opinion, will be largely dependent on how effectively we can balance the functionality, stability as well as the cost of NW-based flexible systems. We believe cooperation between entrepreneurs and scientists will certainly accelerate our research in this area. Assembling existing components into functional and smart systems, increasing their availability and reliability and then introducing them to the market is basically the only way to realize the commercial and industrial applications of flexible electronics.

Acknowledgements

This work was supported by the National Natural Science Foundation (61377033 and 91123008).

Notes and references

1. B. K. Kim, S.-K. Lee, M. S. Kang, J.-H. Ahn and J. H. Cho, ACS Nano, 2012, 6, 8646–8651.
2. D. Grimm, C. C. B. Bufon, C. Deneke, P. Atkinson, D. J. Thurmer, F. Schäffel, S. Gorantla, A. Bachmatiuk and O. G. Schmidt, Nano Lett., 2013, 13, 213–218.
3. L. Y. Yuan, J. J. Dai, X. H. Fan, T. Song, Y. T. Tao, K. Wang, Z. Xu, J. Zhang, X. D. Bai, P. X. Lu, J. Chen, J. Zhou and Z. L. Wang, ACS Nano, 2011, 5, 4007–4013.
4. Y. Sun and J. A. Rogers, Adv. Mater., 2007, 15, 1897–1916.
5. R. T. Hamers, Nature, 2001, 412, 489–490.
6. P.-C. Chen, G. Z. Shen, H. Chen, Y. G. Ha, C. Wu, S. Sukcharoenchoke, Y. Fu, J. Liu, A. Facchetti, T. J. Marks, M. E. Thompson and C. W. Zhou, ACS Nano, 2009, 3, 3383–3390.
7. D. R. Cairns and G. P. Crawford, Proc. IEEE, 2005, 93, 1451–1458.
8. J. Yoon, A. J. Baca, S. Park, P. Elvikis, J. B. Geddes, L. F. Li, R. H. Kim, J. L. Xiao, S. D. Wang, T. H. Kim, M. J. Motala, B. Y. Ahn, E. B. Duoss, J. A. Lewis, R. G. Nuzzo, P. M. Ferreira, Y. G. Huang, A. Rockett and J. A. Rogers, Nat. Mater., 2008, 7, 907–915.
9. C. Y. Chang, Y. J. Cheng, S. H. Hung, J. S. Wu, W. S. Kao, C. H. Lee and C. S. Hsu, Adv. Mater., 2012, 24, 549–553.
10. W. T. S. Huck, Nat. Mater., 2005, 4, 271–272.
11. K. Efimenko, M. Rackaitis, E. Manias, A. Vaziri, L. Mahadevan and J. Genzer, Nat. Mater., 2005, 4, 293–297.
12. S. Ju, A. Facchetti, Y. Xuan, J. Liu, F. Ishikawa, P. D. Ye, C. W. Zhou, T. J. Marks and D. B. Janes, Nat. Nanotechnol., 2007, 2, 378–384.
13. K. H. Cherenack, A. Z. Kattamis, B. Hekmatshoar, J. C. Sturm and S. Wagner, IEEE Electron Device Lett., 2007, 28, 1004–1006.
14. R. J. Hamers, Nature, 2001, 412, 489–490.
15. S. R. Forrest, Nature, 2004, 428, 911–918.
16. G. H. Gelinck, H. E. A. Huitema, E. V. Veenendaal, E. Cantatore, L. Schrijnemakers, J. B. P. H. V. D. Putten, T. C. T. Geuns, M. Beenhakkers, J. B. Giesbers, B. H. Huisman, E. J. Meijer, E. M. Benito, F. J. Touwslager, A. W. Marsman, B. J. E. V. Rens and D. M. Leeuw, Nat. Mater., 2004, 3, 106–110.
17. T. Someya, Y. Kato, T. Sekitani, S. Iba, Y. Noguchi, Y. Murase, H. Kawaguchi and T. Sakurai, Proc. Natl. Acad. Sci. U. S. A., 2005, 102, 12321–12325.
18. G. Eda, G. Fanchini and M. Chhowalla, Nat. Nanotechnol., 2008, 3, 270–274.
19. J. H. Burroughes, D. D. C. Bradley, A. R. Brown, R. N. Marks, K. Mackay, R. H. Friend, P. L. Burns and A. B. Holmes, Nature, 1990, 347, 539–541.
20. S. Lamansky, P. Djurovich, D. Murphy, F. A. Razzaq, H. E. Lee, C. Adachi, P. E. Burrows, S. R. Forrest and M. E. Thompson, J. Am. Chem. Soc., 2001, 123, 4304–4312.
21. S. Reineke, F. Lindner, G. Schwartz, N. Seidler, K. Walzer, B. Lüssem and K. Leo, Nature, 2009, 459, 234–238.
22. S. Soeren, D. V. Stijn, M. Kris, L. Martijin, G. Jan and H. Paul, J. Appl. Phys., 2006, 99, 114519.
23. R. Robert, M. Siddharth, O. Viorel, W. Robert, G. Michelle, D. Klaus, S. Oleg and D. Ananth, Appl. Phys. Lett., 2006, 88, 123502.
24. L. E. Kreno, K. Leong, O. K. Farha, M. Allendorf, R. P. V. Duyne and J. T. Hupp, Chem. Rev., 2012, 112, 1105–1125.
25. C. B. Black, B. Andrioletti, A. C. Try, C. Ruiperez and J. L. Sessler, J. Am. Chem. Soc., 1999, 121, 10438–10439.
26. J. Li, Y. J. Lu, Q. Ye, M. Cinke, J. Han and M. Meyyappan, Nano Lett., 2003, 3, 929–933.
27. A. Dodabalapur, L. Torsi and H. E. Katz, Science, 1995, 268, 270.
28. K. Nomura, H. Ohta, K. Ueda, T. Kamiya, M. Hirano and H. Hosono, Science, 2003, 300, 1269.
29. K. Nomura, H. Ohta, A. Takagi, T. Kamiya, M. Hirano and H. Hosono, Nature, 2004, 432, 488.
30. Y. Yamashita, Sci. Technol. Adv. Mater., 2009, 10, 024313.
31. S. Y. Park, B. J. Kim, K. Kim, M. S. Kang, K. H. Lim, T. I. Lee, J. M. Myoung, H. K. Baik, J. H. Cho and Y. S. Kim, Adv. Mater., 2012, 24, 834.
32. N. A. Azarova, J. W. Owen, C. A. McLellan, M. A. Grimminger, E. K. Chapman, J. E. Anthony and O. D. Jurchescu, Org. Electron., 2010, 11, 1960.
33. J. Li, C. Papadopoulos, J. M. Xu and M. Moskovits, Appl. Phys. Lett., 1999, 75, 367.
34. S. Cui, H. B. Liu, L. B. Gan, Y. L. Li and D. B. Zhu, Adv. Mater., 2008, 20, 2918–2925.
35. C. A. Wang, K. M. Ryu, A. Badmaev, J. L. Zhang and C. W. Zhou, ACS Nano, 2011, 5, 1147.
36. C. Wang, A. Badmaev, A. Jooyaie, M. Q. Bao, K. L. Wang, K. Galatsis and C. W. Zhou, ACS Nano, 2011, 5, 4169.
37. G. Z. Shen, B. Liang, X. F. Wang, P. C. Chen and C. W. Zhou, ACS Nano, 2011, 5, 2155.
38. Z. Liu, H. T. Huang, B. Liang, X. F. Wang, Z. R. Wang, D. Chen and G. Z. Shen, Opt. Express, 2012, 20, 2982–2991.
39. D. M. Sun, M. Y. Timmermans, Y. Tian, A. G. Nasibulin, E. I. Kauppinen, S. Kishimoto, T. Mizutani and Y. Ohno, Nat. Nanotechnol., 2011, 6, 156–161.
40. C. Kocabas, H. S. Kim, T. Banks, J. A. Rogers, A. A. Pesetski, J. E. Baumgardner, S. V. Krishnaswamy and H. Zhang, Proc. Natl. Acad. Sci. U. S. A., 2008, 105, 1045–1049.
41. J. Appenzeller, Proc. IEEE, 2008, 96, 201–211.
42. M. Rinkiö, A. Johansson, G. S. Paraoanu and P. Törmä, Nano Lett., 2009, 9, 643–647.
43. A. B. Kaul, E. W. Wong, L. Epp and B. D. Hunt, Nano Lett., 2006, 6, 942–947.
44. T. Durkop, S. A. Getty, E. Cobas and M. S. Fuhrer, Nano Lett., 2004, 4, 35–39.
45. E. T. Thostenson, Z. F. Ren and T. W. Chou, Compos. Sci. Technol., 2001, 61, 1899–1912.
46. T. Takenobu, T. Takahashi, T. Kanbara, K. Tsukagoshi, Y. Aoyagi and Y. Iwasa, Appl. Phys. Lett., 2006, 88, 33511–33513.
47. Q. Cao, S. H. Hur, Z. T. Zhu, Y. Sun, C. Wang, M. A. Meitl, M. Shim and J. A. Rogers, Adv. Mater., 2006, 18, 304–309.
48. E. Artukovic, M. Kaempgen, D. S. Hecht, S. Roth and G. Gruner, Nano Lett., 2005, 5, 757–760.
49. F. N. Ishikawa, H. K. Chang, K. Ryu, P. C. Chen, A. Badmaev, L. G. D. Arco, G. Z. Shen and C. W. Zhou, ACS Nano, 2009, 3, 73–79.
50. Y. Che, C. Wang, J. Liu, B. L. Liu, X. Lin, J. Parker, C. Beasley, H. S. P. Wong and C. W. Zhou, ACS Nano, 2012, 6, 7454–7462.
51. E. Stern, J. F. Klemic, D. A. Routenberg, P. N. Wyrembak, D. B. T. Evans, A. D. Hamilton, D. A. Lavan, T. M. Fahmy and M. A. Reed, Nature, 2007, 445, 519–522.
52. Z. Liu, B. Liang, G. Chen, G. Yu, Z. Xie, L. N. Gao, D. Chen and G. Z. Shen, J. Mater. Chem. C, 2013, 1, 131–137.
53. G. D. Mahan, R. Gupta, Q. Xiong, C. K. Adu and P. C. Eklund, Phys. Rev. B: Condens. Matter Mater. Phys., 2003, 68, 073402.
54. M. S. Gudiksen, L. J. Lauhon, J. F. Wang, D. C. Smith and C. M. Lieber, Nature, 2002, 415, 617–620.
55. R. Q. Zhang, Y. Lifshitz and S. T. Lee, Adv. Mater., 2003, 15, 635–640.
56. Z. Liu, G. Chen, B. Liang, G. Yu, H. T. Huang, D. Chen and G. Z. Shen, Opt. Express, 2013, 21, 7799–7810.
57. A. I. Hochbaum and P. D. Yang, Chem. Rev., 2010, 110, 527–546.
58. P. D. Yang, R. X. Yan and M. Fardy, Nano Lett., 2010, 10, 1529–1536.
59. H. T. Huang, B. Liang, Z. Liu, X. F. Wang, D. Chen and G. Z. Shen, J. Mater. Chem., 2012, 22, 13428–13445.
60. M. Law, J. Goldberger and P. D. Yang, Annu. Rev. Mater. Res., 2004, 34, 83–122.
61. B. M. Wen, J. E. Sader and J. J. Boland, Phys. Rev. Lett., 2008, 101, 175502.
62. D.-H. Kim, N. S. Lu, R. Ghaffari and J. A. Rogers, NPG Asia Mater., 2012, 4, 15.
63. X. F. Wang, B. Liu, Q. F. Wang, W. F. Song, X. J. Hou, D. Chen, Y.-B. Chen and G. Z. Shen, Adv. Mater., 2013, 25, 1479–1486.
64. B. Liu, X. F. Wang, H. T. Chen, Z. R. Wang, D. Chen, Y.-B. Cheng, C. W. Zhou and G. Z. Shen, Sci. Rep., 2013, 3, 1622.
65. X. Wang, J. Zhuang, Q. Peng and Y. D. Li, Nature, 2005, 437, 121–124.
66. Y. Cui, L. J. Lauhon, M. S. Gudiksen, J. F. Wang and C. M. Lieber, Appl. Phys. Lett., 2001, 78, 2214.
67. C. Wang, Y. J. Hu, C. M. Lieber and S. S. Heng, J. Am. Chem. Soc., 2008, 130, 8902–8903.
68. J. M. Cai, P. Ruffieux, R. Jaafar, M. Bieri, T. Braun, S. Blankenburg, M. Muoth, A. P. Seitsonen, M. Saleh, X. L. Feng, K. Müllen and R. Fasel, Nature, 2010, 466, 470–473.
69. G. B. Rossi, G. Beaucage, T. D. Dang and R. A. Vaia, Nano Lett., 2002, 2, 319–323.
70. J. D. Holmes, K. P. Johnston, R. C. Doty and B. A. Korgel, Science, 2000, 287, 1471–1473.
71. Y. G. Sun, B. Gates, B. Mayer and Y. N. Xia, Nano Lett., 2002, 2, 165–168.
72. B. Weintraub, Z. Z. Zhou, Y. H. Li and Y. L. Deng, Nanoscale, 2010, 2, 1573–1587.
73. A. P. Alivisatos, J. Phys. Chem., 1996, 100, 13226–13239.
74. J. Xu, Y. G. Zhu, H. T. Huang, Z. Xie, D. Chen and G. Z. Shen, CrystEngComm, 2011, 13, 2629–2635.
75. A. Birkel, N. Loges, E. Mugnaioli, R. Branscheid, D. Koll, S. Frank, M. Panthöfer and W. Tremel, Langmuir, 2010, 26, 3590–3595.
76. X. M. Sun and Y. D. Li, Adv. Mater., 2005, 17, 2626–2630.
77. Z. P. Zhu, Y. Yang, J. B. Liang, Z. K. Hu, S. Li, S. Peng and Y. T. Qian, J. Phys. Chem. B, 2003, 107, 12658–12661.
78. T. Y. Zhai, X. S. Fang, M. Y. Liao, X. J. Liu, H. B. Zeng, B. Yoshio and D. Golberg, Sensors, 2009, 9, 6504–6529.
79. G. Z. Shen, J. Xu, X. F. Wang, H. T. Huang and D. Chen, Adv. Mater., 2011, 23, 771.
80. G. Chen, Z. Liu, B. Liang, G. Yu, Z. Xie, H. T. Huang, B. Liu, X. F. Wang, D. Chen and G. Z. Shen, Adv. Funct. Mater., 2013, 21, 2681.
81. D. Routkevitch, T. Bibioni, M. Moskovits and J. M. Xu, J. Phys. Chem., 1996, 100, 14037–14047.
82. L. L. Wang, J. N. Deng, Z. Lou and T. Zhang, J. Mater. Chem. A, 2014, 2, 10022–10028.
83. L. Y. Zhang, J. Feng and D. S. Xue, Mater. Lett., 2007, 61, 1363.
84. D. S. Xue, L. Y. Zhang, A. B. Gui and X. F. Xu, Appl. Phys. A: Mater. Sci. Process., 2005, 80, 439.
85. S. M. Roper, H. S. Davis, S. A. Norris, A. A. Golovin, P. W. Voorhees and M. Weiss, J. Appl. Phys., 2007, 102, 034304.
86. K. C. Grabar, R. G. Freeman, M. B. Hommer and M. J. Natan, Anal. Chem., 1995, 67, 735–743.
87. R. S. Wagner, W. C. Ellis, K. A. Jackson and S. M. Arnold, J. Appl. Phys., 1964, 35, 2993.
88. Y. Y. Wu and P. D. Yang, J. Am. Chem. Soc., 2001, 123, 3165–3166.
89. K.-K. Lew, L. Pan, E. C. Dickey and J. M. Redwing, Adv. Mater., 2003, 15, 2073–2076.
90. Y. Ding, P. X. Gao and Z. L. Wang, J. Am. Chem. Soc., 2004, 126, 2066–2072.
91. Y. Li, Y. Bando and D. Golberg, Adv. Mater., 2003, 15, 581–585.
92. Q. Wan, M. Wei, D. Zhi, J. L. MacManus-Driscoll and M. G. Blamire, Adv. Mater., 2006, 18, 234.
93. Q. Wan, P. Feng and T. H. Wang, Appl. Phys. Lett., 2006, 89, 123102.
94. J. H. He, J. H. Hsu, C. W. Wang, H. N. Lin, L. J. Chen and Z. L. Wang, J. Phys. Chem. B, 2006, 110, 50–53.
95. L. E. Greene, M. Law, D. H. Tan, M. Montano, J. Goldberger, G. Somorijai and P. D. Yang, Nano Lett., 2005, 5, 1231–1236.
96. Z. Ying, Q. Wan, Z. T. Song and S. L. Feng, Mater. Lett., 2005, 59, 1670–1672.
97. N. Sköld, L. S. Karlsson, M. W. Larsson, M.-E. Pistol, W. Seifert, J. Trägårdh and L. Samuelson, Nano Lett., 2005, 5, 1943–1947.
98. A. Umar, S. H. Kim, Y.-S. Lee, K. S. Nahm and Y. B. Hahn, J. Cryst. Growth, 2005, 282, 131–136.
99. S. H. Sun, G. W. Meng, Y. W. Wang, T. Gao, M. G. Zhang, Y. T. Tian, X. S. Peng and L. D. Zhang, Appl. Phys. A: Mater. Sci. Process., 2003, 76, 287–289.
100. X. S. Peng, G. W. Meng, J. Zhang, X. F. Wang, Y. W. Wang, C. Z. Wang and L. D. Zhang, J. Mater. Chem., 2002, 12, 1602–1605.
101. D. H. Fan, J. Cryst. Growth, 2009, 311, 2300–2304.
102. L. S. Wang, X. Z. Zhang, S. Q. Zhao, G. Y. Zhou, Y. L. Zhou and J. J. Qi, Appl. Phys. Lett., 2005, 86, 024108.
103. Y. Huang, X. F. Duan, Q. Q. Wei and C. M. Lieber, Science, 2001, 291, 630–633.
104. G. H. Yu, A. Y. Cao and C. M. Lieber, Nat. Nanotechnol., 2007, 2, 372–377.
105. C. M. Hangarter and N. V. Myung, Chem. Mater., 2005, 17, 1320–1324.
106. E. M. Freer, O. Grachev, X. F. Duan, S. Martin and D. P. Stumbo, Nat. Nanotechnol., 2010, 5, 525–530.
107. Z. Y. Fan, J. C. Ho, Z. A. Jacobson, R. Yerushalmi, R. L. Alley, H. Razavi and A. Javey, Nano Lett., 2008, 8, 20–25.
108. A. C. Ford, J. C. Ho, Z. Y. Fan, O. Ergen, V. Altoe, S. Aloni, H. Razavi and A. Javey, Nano Res., 2008, 1, 32–39.
109. R. Yerushalmi, Z. A. Jacobson, J. C. Ho, Z. Y. Fan and A. Javey, Appl. Phys. Lett., 2007, 91, 203104.
110. S.-Y. Min, T.-S. Kim, B.-J. Kim, H. Cho, Y.-Y. Noh, H. Yang, J. H. Cho and T.-W. Lee, Nat. Commun., 2013, 4, 1773.
111. H. Cho, S.-Y. Min and T.-W. Lee, Macromol. Mater. Eng., 2013, 298, 475–486.
112. D. Li and Y. Xia, Adv. Mater., 2004, 16, 1151–1170.
113. A. L. Briseno, S. C. B. Mannsfeld, S. A. Jenekhe, Z. N. Bao and Y. N. Xia, Mater. Today, 2008, 11, 38–47.
114. N. L. Liu, Y. Zhou, N. Ai, C. Luo, J. B. Peng, J. Wang, J. Pei and Y. Cao, Langmuir, 2011, 27, 14710–14715.
115. X. J. Zhang, J. S. Jie, W. F. Zhang, C. Y. Zhang, L. B. Luo, Z. B. He, X. H. Zhang, W. J. Zhang, C. Lee and S. T. Lee, Adv. Mater., 2008, 20, 2427–2432.
116. Y. P. Zhang, X. D. Wang, Y. M. Wu, J. S. Jie, X. W. Zhang, Y. L. Xing, H. H. Wu, B. Zou, X. J. Zhang and X. H. Zhang, J. Mater. Chem., 2012, 22, 14357–14362.
117. L. Persano, C. Degdeviren, Y. W. Su, Y. H. Zhang, S. Girardo, D. Pisignano, Y. G. Huang and J. A. Rogers, Nat. Commun., 2013, 4, 1633.
118. A. R. Madaria, A. Kumar, F. N. Ishikawa and C. W. Zhou, Nano Res., 2010, 3, 564–573.
119. L. B. Hu, H. S. Kim, J.-Y. Lee, P. Peumans and Y. Cui, ACS Nano, 2010, 5, 2955–2963.
120. A. R. Rathmell and B. J. Wiley, Adv. Mater., 2011, 23, 4798–4803.
121. T. Takahashi, K. Takei, E. Adabi, Z. Y. Fan, A. M. Niknejad and A. Javey, ACS Nano, 2010, 10, 5855–5860.
122. M. C. Mcalpine, H. Ahmad, D. W. Wang and J. R. Heath, Nat. Mater., 2007, 6, 379–384.
123. D. K. Kim, Y. M. Lai, T. R. Vemulkar and C. R. Kagan, ACS Nano, 2011, 12, 10074–10083.
124. D. V. Talapin and C. B. Murray, Science, 2005, 310, 86–89.
125. F. Patolsky and C. M. Lieber, Mater. Today, 2005, 8, 20–28.
126. H. Kind, H. Q. Yan, B. Messer, M. Law and P. D. Yang, Adv. Mater., 2002, 14, 158–160.
127. C.-H. Lin, T.-T. Chen and Y.-F. Chen, Opt. Express, 2008, 16, 16916–16922.
128. D. Zhang, C. Li, S. Han, X. Liu, T. Tang, W. Jin and C. W. Zhou, Appl. Phys. A: Mater. Sci. Process., 2003, 77, 163–166.
129. L. Wang, M. Lu, P. Lv, J. S. Jie, T. X. Yan, Y. Q. Yu, C. Y. Wu, Y. Zhang, C. Xie, P. Jiang, Z. Wang and Z. H. Hu, Sci. Adv. Mater., 2012, 4, 332–336.
130. T. Y. Zhai, M. F. Ye, L. Li, X. S. Fang, M. Y. Liao, Y. F. Li, Y. Koide, Y. Bando and D. Golberg, Adv. Mater., 2010, 22, 4530–4533.
131. Z. T. You, X. S. Fang, M. Y. Liao, X. J. Xu, L. Li, B. D. Liu, Y. Koide, Y. Ma, J. N. Yao, Y. Bando and D. Golberg, ACS Nano, 2010, 4, 1596–1602.
132. Z. X. Wang, M. Safdar, C. Jiang and J. He, Nano Lett., 2012, 12, 4715–4721.
133. J. Svensson, N. Anttu, N. Vainorius, B. M. Borg and L. E. Wernersson, Nano Lett., 2013, 13, 1380–1385.
134. C. Y. Yan, N. D. Singh and P. S. Lee, Appl. Phys. Lett., 2010, 96, 053108.
135. C. Y. Yan and P. S. Lee, Sci. Adv. Mater., 2012, 4, 241–253.
136. S. Bai, W. W. Wu, Y. Qin, N. Y. Cui, D. J. Bayerl and X. D. Wang, Adv. Funct. Mater., 2011, 21, 4464–4469.
137. G. Chen, Z. Liu, B. Liang, G. Yu, Z. Xie, H. T Huang, B. Liu, X. F. Wang, D. Chen, M. Q. Zhu and G. Z. Shen, Adv. Funct. Mater., 2013, 23, 2681–2690.
138. R. R. He and P. D. Yang, Nat. Nanotechnol., 2006, 1, 42–46.
139. J. Zhou, Y. D. Gu, P. Fei, W. J. Mai, Y. F. Gao, R. S. Yang, G. Bao and Z. L. Wang, Nano Lett., 2008, 8, 3973–3977.
140. S. C. B. Mannsfeld, B. C.-K. Tee, C.-K. Tee, R. M. Stoltenberg, C. V. H.-H. Chen, S. Barman, B. V. O. Muir, A. N. Sokolov, C. Reese and Z. N. Bao, Nat. Mater., 2010, 9, 859–864.
141. K. Takei, T. Takahashi, J. C. Ho, H. Ko, A. G. Gillies, P. W. Leu, R. S. Fearing and A. Javey, Nat. Mater., 2010, 9, 821–826.
142. C. Wang, D. Hwang, Z. B. Zhu, K. Takei, J. Park, T. Chen, B. W. Ma and A. Javey, Nat. Mater., 2013, 12, 899–904.
143. M. C. McAlpine, R. S. Friedman, S. Jin, K.-H. Lin, W. U. Wang and C. M. Lieber, Nano Lett., 2003, 3, 1531–1535.
144. A. Nadarajah, R. C. Word, J. Meiss and R. Könenkamp, Nano Lett., 2008, 8, 534–537.
145. A. Javey, S. W. Nam, R. S. Friedman, H. Yan and C. M. Lieber, Nano Lett., 2007, 7, 773–777.
146. Y. G. Sun, H.-S. Kim, E. Menard, S. Kim, I. Adesida and J. A. Rogers, Small, 2006, 2(11), 1330–1334.
147. Q. Cao, H.-S. Kim, N. Pimparkar, J. P. Kulkarni, C. Wang, M. Shim, K. Roy, M. A. Alam and J. A. Rogers, Nature, 2008, 454, 495–500.
148. M. Armand and J. M. Tarascon, Nature, 2008, 451, 652–657.
149. H. Wu and Y. Cui, Nano Today, 2012, 7, 414–429.
150. A. S. Arico, P. Bruce, B. Scrosati, J.-M. Tarascon and W. van Schalkwijk, Nat. Mater., 2005, 4, 366–377.
151. S. Iijima, Nature, 1991, 354, 56–58.
152. B. Gao, C. Bower, J. D. Lorentzen, L. Fleming, A. Kleinhammes, X. P. Tang, L. E. McNeil, Y. Wu and O. Zhou, Chem. Phys. Lett., 2000, 327, 69–75.
153. S. W. Lee, N. Yabuuchi, B. M. Gallant, S. Chen, B.-S. Kim, P. T. Hammond and Y. Shao-Horn, Nat. Nanotechnol., 2010, 5, 531–537.
154. S. Y. Chew, S. H. Ng, J. Wang, P. Novák, F. Krumeich, S. L. Chou, J. Chen and H. K. Liu, Carbon, 2009, 47, 2976–2983.
155. J. Chen, A. I. Minett, Y. Liu, C. Lynam, P. Sherrell, C. Wang and G. G. Wallace, Adv. Mater., 2008, 20, 566–570.
156. B. J. Landi, M. J. Ganter, C. M. Schauerman, C. D. Cress and R. P. Raffaelle, J. Mater. Chem., 2008, 112, 7509–7515.
157. X. Li, J. Yang, Y. Hu, J. Wang, Y. Li, M. Cai, R. Li and X. Sun, J. Mater. Chem., 2012, 22, 18847–18853.
158. L. Hu, H. Wu, F. La Mantia, Y. Yang and Y. Cui, ACS Nano, 2010, 4, 5843–5848.
159. K. Wang, S. Luo, Y. Wu, X. F. He, F. Zhao, J. P. Wang, K. L. Jiang and S. S. Fan, Adv. Funct. Mater., 2013, 23, 846–853.
160. L. Hu, F. La Mantia, H. Wu, X. Xie, J. McDonough, M. Pasta and Y. Cui, Adv. Energy Mater., 2011, 1, 1012–1017.
161. V. L. Pushparaj, M. M. Shaijumon, A. Kumar, S. Murugesan, L. Ci, R. Vajtai, R. J. Linhardt, O. Nalamasu and P. M. Ajayan, Proc. Natl. Acad. Sci. U. S. A., 2007, 104, 13574–13577.
162. L. Hu and Y. Cui, Energy Environ. Sci., 2012, 5, 6423–6435.
163. L. Hu, J. W. Choi, Y. Yang, S. Jeong, F. La Mantia, L.-F. Cui and Y. Cui, Proc. Natl. Acad. Sci. U. S. A., 2009, 106, 21490–21494.
164. X. F. Wang, B. Liu, X. J. Hou, Q. F. Wang, D. Chen and G. Z. Shen, Nano Res., 2014, 7, 1073–1082.
165. C. K. Chan, X. F. Zhang and Y. Cui, Nano Lett., 2007, 8, 307–309.
166. M. Tian, W. Wang, S.-H. Lee, Y.-C. Lee and R. Yang, J. Power Sources, 2011, 196, 10207–10212.
167. B. Liu, J. Zhang, X. Wang, G. Chen, D. Chen, C. Zhou and G. Shen, Nano Lett., 2012, 12, 3005–3011.
168. L. Wang, B. Liu, S. Ran, H. Huang, X. Wang, B. Liang, D. Chen and G. Shen, J. Mater. Chem., 2012, 22, 23541–23546.
169. L. Gao, X. Wang, Z. Xie, W. Song, L. Wang, X. Wu, F. Qu, D. Chen and G. Shen, J. Mater. Chem. A, 2013, 1, 7167–7173.
170. W. Li, X. Wang, B. Liu, S. Luo, Z. Liu, X. Hou, Q. Xiang, D. Chen and G. Shen, Chem. – Eur. J., 2013, 19, 8650–8656.
171. W. Li, X. Wang, B. Liu, J. Xu, B. Liang, T. Luo, S. Luo, D. Chen and G. Shen, Nanoscale, 2013, 5, 10291–10299.
172. B. Liu, X. Wang, B. Liu, Q. Wang, D. Tan, W. Song, X. Hou, D. Chen and G. Shen, Nano Res., 2013, 6, 525–534.
173. X. Wang, Q. Xiang, B. Liu, L. Wang, T. Luo, D. Chen and G. Shen, Sci. Rep., 2013, 3, 2007.
174. J.-Y. Choi, D. J. Lee, Y. M. Lee, Y.-G. Lee, K. M. Kim, J.-K. Park and K. Y. Cho, Adv. Funct. Mater., 2013, 23, 2108–2114.
175. Y. Cui and C. M. Lieber, Science, 2001, 291, 851–853.
176. C. K. Chan, H. Peng, G. Liu, K. McIlwrath, X. F. Zhang, R. A. Huggins and Y. Cui, Nat. Nanotechnol., 2008, 3, 31–35.
177. S. Liu, Z. Wang, C. Yu, H. B. Wu, G. Wang, Q. Dong, J. Qiu, A. Eychmüller and X. W. Lou, Adv. Mater., 2013, 25, 3462–3467.
178. W. W. Li, X. F. Wang, B. Liu, S. J. Luo, Z. Liu, X. J. Hou, Q. Y. Xiang, D. Chen and G. Z. Shen, Chem. – Eur. J., 2013, 19, 8650–8656.
179. L. Lin, Z. Wu, S. Yuan and X. B. Zhang, Energy Environ. Sci., 2014, 7, 2101–2122.
180. G. Zhou, F. Li and H. M. Cheng, Energy Environ. Sci., 2014, 7, 1307–1338.
181. H. Gwon, J. Hong, H. Kim, D. H. Seo, S. Jeon and K. Kang, Energy Environ. Sci., 2014, 7, 538–551.
182. X. F. Wang, X. H. Lu, Y. X. Tong and G. Z. Shen, Adv. Mater., 2014, 26, 4763–4782.
183. S. Xu, Y. H. Zhang, J. Cho, J. Lee, X. Huang, L. Jia, J. A. Fan, Y. W. Su, J. Su, H. G. Zhang, H. Y. Cheng, B. W. Lu, C. J. Yu, C. Chuang, T. I. Kim, T. Song, K. Shigeta, S. Kang, C. Dagdeviren, I. Petrov, P. V. Braun, Y. G. Huang, U. Paik and J. A. Rogers, Nat. Commun., 2013, 4, 1543.
184. P. Simon and Y. Gogotsi, Nat. Mater., 2008, 7, 845–854.
185. Y. P. Zhai, Y. Q. Dou, D. Y. Zhao, P. F. Fulvio, R. T. Mayes and S. Dai, Adv. Mater., 2011, 23, 4828–4850.
186. Y. M. He, W. J. Chen, C. T. Gao, J. Y. Zhou, X. D. Li and E. Q. Xie, Nanoscale, 2013, 5, 8799–8820.
187. C. Niu, E. K. Sichel, R. Hoch, D. Moy and H. Tennent, Appl. Phys. Lett., 1997, 70, 1480–1482.
188. D. N. Futaba, K. Hata, T. Yamada, T. Hiraoka, Y. Hayamizu, Y. Kakudate, O. Tanaike, H. Hatori, M. Yumura and S. Iijima, Nat. Mater., 2006, 5, 987–994.
189. Z. Niu, W. Zhou, J. Chen, G. Feng, H. Li, W. Ma, J. Li, H. Dong, Y. Ren, D. Zhao and S. Xie, Energy Environ. Sci., 2011, 4, 1440–1446.
190. C. Zheng, W. Qian, C. Cui, Q. Zhang, Y. Jin, M. Zhao, P. Tan and F. Wei, Carbon, 2012, 50, 5167–5175.
191. G. Xu, C. Zheng, Q. Zhang, J. Huang, M. Zhao, J. Nie, X. Wang and F. Wei, Nano Res., 2011, 4, 870–881.
192. Q. Guo, X. Zhou, X. Li, S. Chen, A. Seema, A. Greiner and H. Hou, J. Mater. Chem., 2009, 19, 2810–2816.
193. Z. Chen, V. Augustyn, X. Jia, Q. Xiao, B. Dunn and Y. Lu, ACS Nano, 2012, 6, 4319–4327.
194. X. Zhang, W. Shi, J. Zhu, D. J. Kharistal, W. Zhao, B. S. Lalia, H. H. Hng and Q. Yan, ACS Nano, 2011, 5, 2013–2019.
195. Z. Chen, Y. Qin, D. Weng, Q. Xiao, Y. Peng, X. Wang, H. Li, F. Wei and Y. Lu, Adv. Funct. Mater., 2009, 19, 3420–3426.
196. Z. Niu, P. Luan, Q. Shao, H. Dong, J. Li, J. Chen, D. Zhao, L. Cai, W. Zhou, X. Chen and S. Xie, Energy Environ. Sci., 2012, 5, 8726–8733.
197. S. D. Perera, B. Patel, N. Nijem, K. Roodenko, O. Seitz, J. P. Ferraris, Y. J. Chabal and K. J. Balkus, Adv. Energy Mater., 2011, 1, 936–945.
198. C. Meng, C. Liu, L. Chen, C. Hu and S. Fan, Nano Lett., 2010, 10, 4025–4031.
199. L. Yang, S. Cheng, Y. Ding, X. Zhu, Z. L. Wang and M. Liu, Nano Lett., 2011, 12, 321–325.
200. L. Huang, D. Chen, Y. Ding, S. Feng, Z. L. Wang and M. Liu, Nano Lett., 2013, 13, 3135–3139.
201. L. Hu, H. Wu and Y. Cui, Appl. Phys. Lett., 2010, 96, 183502.
202. K. Yu Jin, C. Haegeun, H. Chi-Hwan and K. Woong, Nanotechnology, 2012, 23, 065401.
203. L. Nyholm, G. Nyström, A. Mihranyan and M. Strømme, Adv. Mater., 2011, 23, 3751–3769.
204. P. Karthika, N. Rajalakshmi and K. S. Dhathathreyan, ChemPhysChem, 2013, 14, 3822–3826.
205. S. Ju, J. Li, J. Liu, P.-C. Chen, Y.-g. Ha, F. Ishikawa, H. Chang, C. Zhou, A. Facchetti, D. B. Janes and T. J. Marks, Nano Lett., 2007, 8, 997–1004.
206. P.-C. Chen, G. Shen, S. Sukcharoenchoke and C. Zhou, Appl. Phys. Lett., 2009, 94, 043113.
207. J. Ge, G. Cheng and L. Chen, Nanoscale, 2011, 3, 3084–3088.
208. H. Y. Jung, M. B. Karimi, M. G. Hahm, P. M. Ajayan and Y. J. Jung, Sci. Rep., 2012, 2, 773.
209. K. Wang, H. Wu, Y. Meng, Y. Zhang and Z. Wei, Energy Environ. Sci., 2012, 5, 8384–8389.
210. P.-C. Chen, G. Shen, Y. Shi, H. Chen and C. Zhou, ACS Nano, 2010, 4, 4403–4411.
211. M. F. El-Kady, V. Strong, S. Dubin and R. B. Kaner, Science, 2012, 335, 1326–1330.
212. K. Jost, C. R. Perez, J. K. McDonough, V. Presser, M. Heon, G. Dion and Y. Gogotsi, Energy Environ. Sci., 2011, 4, 5060–5067.
213. J. Xu, Q. Wang, X. Wang, Q. Xiang, B. Liang, D. Chen and G. Shen, ACS Nano, 2013, 7, 5453–5462.
214. X. Lu, M. Yu, G. Wang, T. Zhai, S. Xie, Y. Ling, Y. Tong and Y. Li, Adv. Mater., 2012, 25, 267–272.
215. X. Lu, G. Wang, T. Zhai, M. Yu, S. Xie, Y. Ling, C. Liang, Y. Tong and Y. Li, Nano Lett., 2012, 12, 5376–5381.
216. Q. Wang, J. Xu, X. Wang, B. Liu, X. Hou, G. Yu, P. Wang, D. Chen and G. Shen, ChemElectroChem, 2014, 1, 559–564.
217. L. Hu, M. Pasta, F. L. Mantia, L. Cui, S. Jeong, H. D. Deshazer, J. W. Choi, S. M. Han and Y. Cui, Nano Lett., 2010, 10, 708–714.
218. M. Pasta, F. La Mantia, L. Hu, H. Deshazer and Y. Cui, Nano Res., 2010, 3, 452–458.
219. J. Xue, Y. Zhao, H. Cheng, C. Hu, Y. Hu, Y. Meng, H. Shao, Z. Zhang and L. Qu, Phys. Chem. Chem. Phys., 2013, 15, 8042–8045.
220. J. Xu, H. Wu, C. Xu, H. Huang, L. Lu, G. Ding, H. Wang, D. Liu, G. Shen, D. Li and X. Chen, Chem. – Eur. J., 2013, 19, 6451–6458.
221. K. Wang, P. Zhao, X. Zhou, H. Wu and Z. Wei, J. Mater. Chem., 2011, 21, 16373–16378.
222. V. T. Le, H. Kim, A. Ghosh, J. Kim, J. Chang, Q. A. Vu, D. T. Pham, J.-H. Lee, S.-W. Kim and Y. H. Lee, ACS Nano, 2013, 7, 5940–5947.
223. K. Wang, Q. Meng, Y. Zhang, Z. Wei and M. Miao, Adv. Mater., 2013, 25, 1494–1498.
224. J. Bae, M. K. Song, Y. J. Park, J. M. Kim, M. Liu and Z. L. Wang, Angew. Chem., Int. Ed., 2011, 50, 1683–1687.
225. Z. Yang, J. Deng, X. Chen, J. Ren and H. Peng, Angew. Chem., Int. Ed., 2013, 125, 13695–13699.
226. B. Liu, D. Tan, X. Wang, D. Chen and G. Shen, Small, 2013, 9, 1998–2004.
227. B. O'Regan and M. Grätzel, Nature, 1991, 353, 737–740.
228. J. Burschka, N. Pellet, S.-J. Moon, R. Humphry-Baker, P. Gao, M. K. Nazeeruddin and M. Gratzel, Nature, 2013, 499, 316–319.
229. M. Gratzel, Nature, 2001, 414, 338–344.
230. J. Akilavasan, K. Wijeratne, H. Moutinho, M. Al-Jassim, A. R. M. Alamoud, R. M. G. Rajapakse and J. Bandara, J. Mater. Chem. A, 2013, 1, 5377–5385.
231. M. Ye, X. Xin, C. Lin and Z. Lin, Nano Lett., 2011, 11, 3214–3220.
232. K. Zhu, N. R. Neale, A. Miedaner and A. J. Frank, Nano Lett., 2006, 7, 69–74.
233. C. Y. Jiang, X. W. Sun, K. W. Tan, G. Q. Lo, A. K. K. Kyaw and D. L. Kwong, Appl. Phys. Lett., 2008, 92, 143101.
234. S. Gubbala, V. Chakrapani, V. Kumar and M. K. Sunkara, Adv. Funct. Mater., 2008, 18, 2411–2418.
235. Z. Li, Y. Zhou, C. Bao, G. Xue, J. Zhang, J. Liu, T. Yu and Z. Zou, Nanoscale, 2012, 4, 3490–3494.
236. Y. Li, D.-K. Lee, J. Y. Kim, B. Kim, N.-G. Park, K. Kim, J.-H. Shin, I.-S. Choi and M. J. Ko, Energy Environ. Sci., 2012, 5, 8950–8957.
237. B.-L. Chen, H. Hu, Q.-D. Tai, N.-G. Zhang, F. Guo, B. Sebo, W. Liu, J.-K. Yuan, J.-B. Wang and X.-Z. Zhao, Electrochim. Acta, 2012, 59, 581–586.
238. Y.-Y. Kuo and C.-H. Chien, Electrochim. Acta, 2013, 91, 337–343.
239. J. Di, Z. Yong, Z. Yao, X. Liu, X. Shen, B. Sun, Z. Zhao, H. He and Q. Li, Small, 2013, 9, 148–155.
240. D. Kuang, J. r. m. Brillet, P. Chen, M. Takata, S. Uchida, H. Miura, K. Sumioka, S. M. Zakeeruddin and M. Grätzel, ACS Nano, 2008, 2, 1113–1116.
241. J.-Y. Liao, B.-X. Lei, H.-Y. Chen, D.-B. Kuang and C.-Y. Su, Energy Environ. Sci., 2012, 5, 5750–5757.
242. A. Vomiero, V. Galstyan, A. Braga, I. Concina, M. Brisotto, E. Bontempi and G. Sberveglieri, Energy Environ. Sci., 2011, 4, 3408–3413.
243. W.-Q. Wu, H.-S. Rao, Y.-F. Xu, Y.-F. Wang, C.-Y. Su and D.-B. Kuang, Sci. Rep., 2013, 3, 1352.
244. V. Galstyan, A. Vomiero, I. Concina, A. Braga, M. Brisotto, E. Bontempi, G. Faglia and G. Sberveglieri, Small, 2011, 7, 2437–2442.
245. D. Li, P.-C. Chang, C.-J. Chien and J. G. Lu, Chem. Mater., 2010, 22, 5707–5711.
246. P. Wang, S. M. Zakeeruddin, J. E. Moser, M. K. Nazeeruddin, T. Sekiguchi and M. Gratzel, Nat. Mater., 2003, 2, 402–407.
247. D.-W. Kim and Y.-K. Sun, J. Power Sources, 2001, 102, 41–45.
248. M. G. Kang, K. M. Kim, K. S. Ryu, S. H. Chang, N.-G. Park, J. S. Hong and K.-J. Kim, J. Electrochem. Soc., 2004, 151, E257–E260.
249. J.-Y. Park, J.-W. Lee, K. Park, T.-Y. Kim, S.-H. Yim, X. Zhao, H.-B. Gu and E. Jin, Polym. Bull., 2013, 70, 507–515.
250. E. Olsen, G. Hagen and S. Eric Lindquist, Sol. Energy Mater. Sol. Cells, 2000, 63, 267–273.
251. X. Xu, D. Huang, K. Cao, M. Wang, S. M. Zakeeruddin and M. Grätzel, Sci. Rep., 2013, 3, 1489.
252. J. Chen, K. Li, Y. Luo, X. Guo, D. Li, M. Deng, S. Huang and Q. Meng, Carbon, 2009, 47, 2704–2708.
253. K.-M. Lee, W.-H. Chiu, H.-Y. Wei, C.-W. Hu, V. Suryanarayanan, W.-F. Hsieh and K.-C. Ho, Thin Solid Films, 2010, 518, 1716–1721.
254. H. Sun, Y. Luo, Y. Zhang, D. Li, Z. Yu, K. Li and Q. Meng, J. Phys. Chem. C, 2010, 114, 11673–11679.
255. M. Wang, A. M. Anghel, B. t. Marsan, N.-L. Cevey Ha, N. Pootrakulchote, S. M. Zakeeruddin and M. Grätzel, J. Am. Chem. Soc., 2009, 131, 15976–15977.
256. B. Fan, X. Mei, K. Sun and J. Ouyang, Appl. Phys. Lett., 2008, 93, 143103.
257. J. G. Nam, Y. J. Park, B. S. Kim and J. S. Lee, Scr. Mater., 2010, 62, 148–150.
258. K. Suzuki, M. Yamaguchi, M. Kumagai and S. Yanagida, Chem. Lett., 2003, 28–29.
259. F. Malara, M. Manca, L. De Marco, P. Pareo and G. Gigli, ACS Appl. Mater. Interfaces, 2011, 3, 3625–3632.
260. F. Malara, M. Manca, M. Lanza, C. Hubner, E. Piperopoulos and G. Gigli, Energy Environ. Sci., 2012, 5, 8377–8383.
261. L.-Y. Chang, C.-P. Lee, K.-C. Huang, Y.-C. Wang, M.-H. Yeh, J.-J. Lin and K.-C. Ho, J. Mater. Chem., 2012, 22, 3185–3191.
262. K. S. Lee, W. J. Lee, N.-G. Park, S. O. Kim and J. H. Park, Chem. Commun., 2011, 47, 4264–4266.
263. C.-W. Kung, H.-W. Chen, C.-Y. Lin, K.-C. Huang, R. Vittal and K.-C. Ho, ACS Nano, 2012, 6, 7016–7025.
264. H.-W. Chen, C.-W. Kung, C.-M. Tseng, T.-C. Wei, N. Sakai, S. Morita, M. Ikegami, T. Miyasaka and K.-C. Ho, J. Mater. Chem. A, 2013, 1, 13759–13768.
265. X. Chen, Y. Hou, B. Zhang, X. H. Yang and H. G. Yang, Chem. Commun., 2013, 49, 5793–5795.
266. T. Chen, L. Qiu, Z. Yang and H. Peng, Chem. Soc. Rev., 2013, 42, 5031–5041.
267. W. Wang, Q. Zhao, H. Li, H. Wu, D. Zou and D. Yu, Adv. Funct. Mater., 2012, 22, 2775–2782.
268. J. Velten, Z. Kuanyshbekova, Ö. Göktepe, F. Göktepe and A. Zakhidov, Appl. Phys. Lett., 2013, 102, 203902.
269. L. Chen, Y. Zhou, H. Dai, Z. Li, T. Yu, J. Liu and Z. Zou, J. Mater. Chem. A, 2013, 1, 11790–11794.
270. S. Hou, X. Cai, H. Wu, Z. Lv, D. Wang, Y. Fu and D. Zou, J. Power Sources, 2012, 215, 164–169.
271. W. Guo, C. Xu, X. Wang, S. Wang, C. Pan, C. Lin and Z. L. Wang, J. Am. Chem. Soc., 2012, 134, 4437–4441.
272. Y. Chae, J. T. Park, J. K. Koh, J. H. Kim and E. Kim, Mater. Sci. Eng., B, 2013, 178, 1117–1123.
273. S. Zhang, C. Ji, Z. Bian, R. Liu, X. Xia, D. Yun, L. Zhang, C. Huang and A. Cao, Nano Lett., 2011, 11, 3383–3387.
274. Z. Lv, J. Yu, H. Wu, J. Shang, D. Wang, S. Hou, Y. Fu, K. Wu and D. Zou, Nanoscale, 2012, 4, 1248–1253.
275. M. J. Uddin, B. Davies, T. J. Dickens and O. I. Okoli, Sol. Energy Mater. Sol. Cells, 2013, 115, 166–171.
276. S. Pan, Z. Yang, P. Chen, X. Fang, G. Guan, Z. Zhang, J. Deng and H. Peng, J. Phys. Chem. C, 2014, 118, 16419–16425.
277. H. Sun, Z. Yang, X. Chen, L. Qiu, X. You, P. Chen and H. Peng, Angew. Chem., 2013, 125, 8434–8438.
278. Y. Fu, Z. Lv, S. Hou, H. Wu, D. Wang, C. Zhang, Z. Chu, X. Cai, X. Fan, Z. L. Wang and D. Zou, Energy Environ. Sci., 2011, 4, 3379–3383.
279. B. Weintraub, Y. Wei and Z. L. Wang, Angew. Chem., 2009, 121, 9143–9147.
280. T. Chen, L. Qiu, Z. Cai, F. Gong, Z. Yang, Z. Wang and H. Peng, Nano Lett., 2012, 12, 2568–2572.
281. Z. L. Wang and J. H. Song, Science, 2006, 312, 242–246.
282. G. Zhu, R. S. Yang, S. H. Wang and Z. L. Wang, Nano Lett., 2010, 10, 3151–3155.
283. S. Xu, Y. Qin, C. Xu, Y. G. Wei, R. S. Yang and Z. L. Wang, Nat. Nanotechnol., 2010, 5, 366–373.
284. X. F. Wang, B. Liu, R. Liu, Q. F. Wang, X. J. Hou, D. Chen, R. M. Wang and G. Z. Shen, Angew. Chem., Int. Ed., 2014, 53, 1849–1853.
285. L. Vj, J.-Y. Oh, A. P. Nayak, A. M. Katazenmeyer, K. H. Cilchrist, S. Grego, N. P. Kobayashi, S.-Y. Wang, A. A. Talin, N. K. Dhar and M. S. Islam, IEEE J. Sel. Top. Quantum Electron., 2011, 17, 1002–1032.
286. D. S. Tsai, W. C. Lien, D. H. Lien, K. M. Chen, M. L. Tsai, D. G. Senesky, Y. C. Yu, A. P. Pisano and J. H. He, Sci. Rep., 2013, 4, 2628.

---

Footnote

† Z. Liu and J. Xu contribute equally to this work. They are visiting students from Huazhong University of Science and Technology.

---