Cellulose-Based Nanomaterials for Energy Applications

Abstract

Cellulose is the most abundant natural polymer on earth, providing a sustainable green resource that is renewable, degradable, biocompatible and cost effective. Recently, nanocellulose-based mesoporous structure, flexible thin films, fibers, and networks are increasingly developed and used in photovoltaic devices, energy storage systems, mechanical energy harvesters, and catalysts components, showing tremendous materials science value and application potential in many energy-related fields. In this review article, we review the most recent advancements of processing, integration and application of cellulose nanomaterials in the areas of solar energy harvesting, energy storage, and mechanical energy harvesting. For solar energy harvesting, promising applications of cellulose-based nanostructures for both solar cells and photoelectrochemical electrodes development are reviewed, and their morphology-related merits are discussed. For energy storage, our discussion is primarily focused on the applications of cellulose-based nanomateriales in lithium ion batteries, including electrodes (e.g. active materials, binders and structural support), electrolytes, and separators. Applications of cellulose nanomaterials in supercapacitors are also overviewed briefly. For mechanical energy harvesting, we review the most recent technology evolution of cellulose-based triboelectric nanogenerators, from fundamental property tuning to practical implementations. At last, the future research potential and opportunities of cellulose nanomaterials as a new energy material are commented.

Graphical Abstract

This article reviews the most recent advancements of processing, integration and application of cellulose nanomaterials in the areas of solar energy harvesting, energy storage, and mechanical energy harvesting; underlining cellulose nanomaterials as a new energy material with tremendous materials science value and application potential in many energy-related fields.

1. Introduction

Invention and utilization of paper marked an important milestone in human civilization. Such a revolutionary use of natural fibers brought in an affordable, abundant and renewable substance to human race, which significantly facilitated the heritage and evolution of knowledge, technology and humanity. Today, as our technology entering a new era of information technology, materials innovation is experiencing an unprecedented pace of progressing. Machines, vehicles, computers, display panels, memory devices, batteries, solar cells, and numerous types of sensors are built everyday using advanced ceramics, metals and plastics. The explosive advancement of modern technology is bringing revolutionary changes to the human society in recent decades, meanwhile, however it also leads to a more and more serious environmental issue on electronic and plastic wastes. While paper is still mostly playing its traditional role currently, it is maybe about the time that the plants-based materials stand out again to pay the debt that human civilization owns to the environment and make our technology evolution more sustainable.

Cellulose, as the most abundant natural polymer on earth, provides a sustainable green resource that is renewable, degradable, biocompatible and cost effective.^[1–3] It is directly generated from natural plants such as woods, cotton and hemp or from microbes such as bacteria, algae and fungi. Among them, wood-origin cellulose is the most common one, and is usually used in the form of wood pulp.^[4] Chemically, cellulose consists of linear chains of repeating β-D-glucopyranose units covalently linked through β-1,4 glycosidic bonds.^[2] A large number of hydrogen bonds exist intra- and inter-molecularly and yields different configuration of cellulose structures. The nanoscale structures of cellulose (e.g. nanofibrils or nanocrystals), offer a unique combinations of properties including flexible surface chemistry, transparency, low thermal expansion, high elasticity and anisotropy. While most study on cellulose was centered around cellulose-to-fuel conversion, contemporary research recently began to address the functionalization and engineering of cellulose nanostructures for electronic and energy applications. With its unique structural and chemical features, cellulose can be assembled into various nanoscale configurations to meet the design needs of the energy devices. Currently, nanocellulose-based mesoporous structure, flexible thin films, fibers, and networks are developed and used in photovoltaic (PV) devices, energy storage systems, mechanical energy harvesters, and catalysts components. Inclusion of nanocellulose in those energy-related devices largely raises the portion of eco-friendly materials and is very promising in addressing the relevant environmental concerns. Furthermore, cellulose manifests itself in the low cost and large scale promises. As the cellulose-related materials become more and more popular, many quality review articles have been published.^{[2, 5–7]} Most of them, however, are still in the area of cellulose molecule processing for chemical fuel production. A few most recent ones begin to address contemporary applications such as cellulose-based electronics ^[8] and electronics from wood-based materials.^[9] To address the emergent applications of cellulose nanomaterials in energy-related areas, this review article discusses the most current advancements of processing, integration and application of cellulose nanomaterials in the areas of solar energy harvesting, mechanical energy harvesting, and energy storage.

2. Cellulose nanostructures for solar energy harvesting

Solar energy harvesting devices typically prefer high surface area and good charge transport properties so that photons can be effectively absorbed and converted into electrical energy. Three dimensional (3D) mesoporous structures composed of cellulose nanofibrils (CNFs) offer a very large surface area, and comparable mechanical properties (e.g. Young’s modulus and tensile strength) as other fibrous materials (e.g., carbon fibers and glass fibers).^[10] Therefore, CNFs are often considered as an attractive template/matrix for processing functional 3D nanostructures for photoelectrode development.^[11] In addition, 3D CNF network or compacted CNF films could bring unique optical properties that are beneficial to photon absorption. In this section, promising applications of CNF-based nanostructures for both solar cells and photoelectrochemical (PEC) electrodes development are reviewed, and their morphology-related merits are discussed.

2.1 Nanocellulose-based paper substrate for solar cell development

Since cellulose intrinsically is an insulator, the most common application of CNF in solar cell development is the transparent substrate. In order to maximize the photon absorption, both high optical transparency and high optical haze (the percentage of light diffusely scattered from a transparent surface) are desired for solar cell substrates. Regular paper substrates offer good optical haze but typically have low transmittance.^{[12, 13]} Therefore, organic solar cells built on paper substrates usually exhibited a fairly low power conversion efficiency (PCE, <1.4%).^[14–16] As for nanocellulose (e.g. CNF and CNC), because their sizes are much less than the wavelength of visible light, paper made from nanocellulose could be highly transparent with significant scattering along the light transport direction.^[17] Thus, nanocellulose paper could be an ideal substrate for supporting the photoactive components of solar cells.

Hu et al. fabricated a carboxymethylated nanocellulose paper and studied its optical properties and subsequent organic solar cell application.^[17] The nanocellulose-paper (~40 μm thick) was fabricated by compressing the carboxymethylated cellulose pulp in a sheet-former at 70 °C and 1 bar. The as-fabricated nanocellulose paper was highly transparent to a closely-contacted object, but was hazy when the object was ~2 inches away due to significant light scattering of the porous cellulose film (Fig. 1a). Angular reflection and transmittance revealed that the nanocellulose paper did not have strong angle-dependent reflection and transmission, which is drastically different from regular transparent substrates. This nanocellulose paper was then applied as the substrate for organic bulk heterojunction solar cells that were composed of poly(3,4-ethylenedioxythiophene):poly(styrene sulfonate) (PEDOT:PSS), poly(3-hexylthiophene) (P3HT) and [6,6]-phenyl-C₆₁-butyric acid methyl ester (PCBM). Although the optical property of the nanocellulose paper substrate was superior, the photovoltaic performance of the solar cell did not excel due to the large sheet resistance of ITO coating and other fabrication-related issues. As shown in Fig. 1b, the short circuit current (J_sc) and V_oc were ~2.4 mA cm⁻² and of ~0.4 V, respectively, revealing a PCE of only ~0.2%.

Figure 1. Solar cells built on nanocellulose-based paper substrate.

(a) Photographs of a transparent nanocellulose paper made from carboxymethylation of softwood cellulose fiber showing the scattering effect. (b) I-V curve of an organic solar cell built on the nanocellulose paper. Inset is a photo of the organic solar cell. (Reprint with permission from ref [17], Copyright 2013 Royal Society of Chemistry). (c) Schematic illustration and photograph of an organic solar cell built on CNC substrate. (d) I-V curve of the organic solar cell. (Reprint with permission from ref [18], Copyright 2013 Creative Commons Attribution 4.0 International License.) (e) Photograph showing the high transparency and high haze of a nanocellulose paper obtained from TEMPO oxidization. (f) Photovoltaic performance of organic solar cells built with and without this nanocellulose paper. (Reprint with permission from ref [13], Copyright 2013 American Chemical Society.)

Zhou et al. demonstrated organic solar cells on CNC substrates which yielded higher efficiencies and excellent rectification behaviors (Figs. 1c and d).^[18] The CNC substrates offered similar optical properties as the CNF papers, while the better performance was mainly attributed to the utilization of highly-conductive Ag electrode and better fabrication strategy. However, the low transmittance of the Ag electrode capped the efficiency at ~2.7%. Zhou et al. further fabricated such organic solar cells on recyclable CNC substrates and the efficiency reached almost 4.0% by using polyethylenimine (PEI)-modified Ag electrode. The dry lamination transfer process of PEDOT:PSS to the CNC substrate could avoid the damage of cellulose from typical aqueous-based processes, and thus led to very low reverse saturation current and good diode rectification of the device.^[19]

Most recently, Fang et al. further improved the nanocellulose paper using a tetramethylpiperidine-1-oxy (TEMPO)-oxidized cellulose as the building block. This paper had a high fragment content and fewer hollow structures, which led to a higher packing density and thus exhibited excellent optical transparency (~96%) and ultrahigh haze (~60%) (Fig. 1e).^[13] By simply laminating a piece of such paper on an ITO glass, the haze effect enhanced the incident angle independency of the solar cell device. When the angle of the incident light was larger than 7°, the photocurrents of the nanocellulose paper-attached solar cell started to excel. Over ~15% photocurrent improvement was observed within the incident angle range of 60~87° (Fig. 1f), which was a result of the reduced light reflection and broadened angular distribution of incident light due to the haze effect of the nanocellulose paper.^[13]

In general, nanocellulose paper can offer enhanced photon absorption at oblique incidence, therefore they are superior in collecting ambient light with greater effectiveness compared to regular glass substrates. The optical properties are directly related to the fiber length, packing density and processing techniques. With further material and processing optimization, nanocellulose papers may take a dominating role in fabricating high efficiency solar cells.

2.2 CNF-templated mesoporous structure as solar cell electrodes

The very large surface area offered by the CNF network is a general structural merit to achieve high photocurrent density in solar cell design. Nevertheless, due to the intrinsic insulating property of CNF, using CNF networks as the scaffold for solar cell electrode development was less popular compared to substrate renovation. The mostly practiced strategy is to use CNF networks as a template to fabricate mesoporous structure of a functional material that is photoactive and conductive. TiO₂ was frequently used to coat the CNF template, followed by high temperature calcination to remove the CNF template, forming a hollow nanofiber network with a very high porosity.^[20] Due to the excellent hydrophilicity of CNFs, aqueous solution-based approaches have been found very effective in producing a uniform coating with tens of nanometer thickness (Fig. 2a). Such a cross-linked TiO₂ network yielded improved electron transport properties than mesoporous films assembled from TiO₂ nanoparticles. Therefore, the CNF-templated TiO₂ nanofibers could provide structural advantages in photon interaction and charge transport. In a typical demonstration, the nanofibers were directly casted on a transparent conducting glass substrate with designed thickness to serve as the photoelectrode of a dye-sensitized solar cell (DSSC). It was expected that the high surface area could largely improve the loading of dye molecules and thus yields high photocurrent density and high incident photon-to-current conversion efficiency. Compared to nanoparticle-based DSSC with the same thickness, the nanofiber-based devices, however, showed a slightly lower photocurrent density but higher open circuit voltage (V_oc, Fig. 2b). The relatively low roughness factor of the fiber network was believed to be responsible for the lower current density, which was not odd in many fiber- or nanowire-based electrodes. The higher V_oc of fiber-based DSSC was attributed to the higher photoelectron density (10¹⁷ cm⁻³) compared to the nanoparticle electrode (Fig. 2c). The higher electron density at a constant V_oc is a result of the lowering of the conduction band.^{[20, 21]} By comparing the V_oc difference at the same photoelectron density, the recombination lifetime of electrons (τ_rec) in the 6 μm-thick fibrous electrode was ~3 times higher than that of the nanoparticles-based photoelectrode (Fig. 2d). The TiO₂ hollow fibers also exhibited 2–3 times shorter electron collection time (τ_n) compared to the TiO₂ nanoparticles networks with the same film thickness (Fig. 2e). Although no specific reason was given by the literature, better charge transport property of the nanocellulose-templated TiO₂ could be attributed to the residue carbon in the TiO₂ lattice, which facilitated electron transport. These merits are beneficial to solar cell performance. Further improving the surface area density (i.e. the nanocellulose packing density) but without losing the porosity could eventually raise the overall PV performance of nanocellulose-templated solar cells.

Figure 2. Solar cell electrodes from CNF-templated TiO₂ mesoporous structure.

(a) SEM image of TiO₂ fibers template by CNF networks. (b) J-V characteristics of dye-sensitized solar cells based on fibrous photoanodes. (c) Open circuit voltage at different charge density levels for DSCs made of TiO₂ nanoparticles or fibers. (d) Electron recombination lifetime τ_rec and (e) electron collection time τ_n extracted from transient open circuit photovoltage. (Reprint with permission from ref [20], Copyright 2010 American Chemical Society.)

2.3 Nanocellulose in photoelectrochemical (PEC) cell development

Photoelectrochemical (PEC) water splitting is a promising strategy for directly converting solar energy into hydrogen fuels.^[22] It shares much similarity with solar cell development, including rapid charge generation and separation, large surface area for redox reactions, and broadband light absorption.^{[11, 23]} Therefore, PEC cells can also benefit from the structural merits offered by nanocellulose-based 3D structures.^{[11, 24–27]} In addition, the hydrophilic mesopores within the cellulose film could serve as an ideal host matrix for the embedment of nanoparticle catalysts with minimized agglomeration. As a representative example, Ke et al. immobilized photoactive CdS nanoparticles in mesoporous nanocellulose films by sequentially immersing a nanocellulose film in CdCl₂ solution and Na₂S solution.^[25] Pt nanoparticles were then photochemically deposited on the CdS/cellulose nanocomposite film for photocatalytic H₂ evolution. The mesoscale pores in the cellulose matrix offered large light scattering, high-density and uniform CdS coating, and good charge and mass transport. Thus, a high H₂ production rate of 1.323 mmol g⁻¹ h⁻¹ was achieved, evidencing a promising role of porous nanocellulose films in photocatalyst loading.

The mesoporous cellulose films were also used as a supporting framework for encapsulating photosynthetic materials,^{[26, 28]} which could support the photoactive materials, organize the guest redox couples, and protect them by providing a stable microenvironment.^[29] Lee et al. reported an integrated artificial photosynthetic system in which a mesoporous lignocellulosic composite was used as a support to immobilize light-harvesting pigments.^[26] The porphyrins were embedded in the lignocellulosic matrix and interacted with lights to perform photochemical reactions, regenerate pyridine nucleotide cofactors, and synthesize enzymes. Furthermore, the presence of redox-active lignin component on the mesoporous lignocellulosic composite surfaces could further enhance the photosynthesis activity.

Taking the advantage of the excellent hydrophilicity of cellulose, Li et al. developed a capillary PEC system using CNF-templated TiO₂ nanostructures that can uniquely perform PEC water splitting reactions outside of the electrolyte body.^[11] A mesoporous CNF film was used to template the fabrication of anatase TiO₂ nanofibrous 3D structure via atomic layer deposition (ALD) followed by high-temperature annealing (Fig. 3a). A CNF strip was then attached to the TiO₂ nanostructure, which acted as the photoanode and was placed outside of the electrolyte facing a light source. The electrolyte was supplied through the nano/micro-channels in the CNF film driven by the capillary force (Fig. 3b). By comparing to regular in-electrolyte PEC setup, the capillary PEC exhibited a superior performance with a photocurrent density (J_ph) of 0.87 mA cm⁻², which was 107% higher than that of the in-electrolyte set up. The PEC performance enhancement of the capillary design was attributed to the less light scattering from minimal electrolyte coating and higher interfacial charge transport kinetics.

Figure 3. PEC electrodes based on CNF-templated TiO₂ nanostructure.

(a) SEM image of CNF film coated with TiO₂. (b) Schematic illustration and photograph of the capillary PEC water splitting setup. (c) Photograph of the CNF films before and after TiO₂ coating, and after annealing in vacuum and oxygen atmosphere. (d) Performance comparison of capillary PEC and normal PEC using CNF-templated TiO₂ photoanodes. (Reprint with permission from ref [11], Copyright 2014 WILEY-VCH Verlag GmbH & Co. KGaA.) (e) Schematic synthesis procedure (left) and SEM image (right) of CNF-templated 3D TiO₂ fiber-nanorods heterostructure. (Reprint with permission from ref [24], Copyright 2014 IOP Publishing Ltd.)

Li et al. further showed that the CNF template was able to modulate the chemical composition of the TiO₂ coating.^[11] By annealing the TiO₂-coated CNF structures in vacuum instead of in O₂ atmosphere, the final structure appeared black while the fibrous morphology was still well preserved (Fig. 3c). The black color was a result of appreciable amount of preserved carbon elements in the TiO₂ lattices, suggesting significant visible light absorption. This “black” cellulose-templated TiO₂ nanostructure exhibited appreciable photoactivity under visible light illumination and further improved the PEC performance of the capillary design. The enhancements J_ph and PEC efficiency (η) were as large as ~767 % and ~800% under visible light illumination (Fig. 3d). This comparison clearly shows that the CNF-enabled capillary PEC is able to provide significant additional enhancement of the PEC performance to those achieved by other strategies, such as optimizations of light absorption, surface/interface properties and charge transport.

In addition to templating low-temperature ALD film coating, mesoporous CNF was also used as templates for high temperature nanorods (NRs) growth. Li et al. adapted the surface-reaction-limited pulsed chemical vapor deposition (SPCVD) technique ^[30–32] for synthesizing high density TiO₂ NR branches within mesoporous CNF templates.^[24] The CNF framework was well preserved under high deposition temperature (~600 °C) by first introducing a ZnO overcoating. The ZnO layer interacted with TiCl₄ vapor and was quickly converted into polycrystalline TiO₂ thin films following the Kirkendall effect,^[33] seeding the growth of TiO₂ NRs while the CNF framework was burned off (Fig. 3e). This 3D heterostructure offers long optical paths, high quality charge transport channels, and a super large electrolyte-semiconductor interface area. When used as a PEC photoanode, its efficiency could be more than 200% higher than those of CNF-templated TiO₂ nanofibrous 3D structure.

3. Cellulose Nanostructures for Energy Storage Applications

Lithium-ion batteries (LIBs) are broadly used as an energy storage device in automobiles and portable electronic devices. Rapid expansion of these industries generates a growing need for LIBs with greater safety, lower manufacture cost, higher energy density and power density. Environmental concerns also demand lithium-ion batteries to use environmental-friendly materials. To manufacture LIBs with these desirable properties, the development of new materials and innovative structural design for LIBs seems imminently inevitable. Cellulose, with its intrinsic chemical and physical properties as well as advantages arising from its microstructures, starts to draw a considerable amount of attention from researchers.^[34] For example, the good solubility of cellulose in water allows the utilization of aqueous electrolyte in place of hazardous organic solvents and the safety concern of lithium-ion batteries can therefore be greatly alleviated. The nanostructured cellulose and assemblies can bring in some unique advantages and even another dimension in designing battery electrodes. To date, various forms of cellulose, including wood CNFs, wood CNCs, and bacterial nanocellulose, have been applied in all components of LIBs. In this section, we will primarily focus our discussion on the applications of cellulose-based nanomateriales in LIBs, including electrodes (e.g. active materials, binders and structural support), electrolytes, and separators. Contemporary applications of cellulose nanomaterials in supercapacitors, which share much similarity in materials requirements, will also be overviewed briefly.

3.1. Paper-based LIBs

Conventional LIB electrodes employ lithium intercalation materials in both cathodes and anodes. These electrode materials are usually coated onto a metal foil serving as a current collector and structural support. The lithium ion storage materials, or the active materials, are used in conjunction with binders and electronic conductivity-enhancing materials (e.g. carbon black). In the form of paper sheets, cellulose can be used as the substrates for LIB electrodes in place of the metal foil current collectors. In this regard, the technique of coating the current collector and electrode materials onto cellulose paper sheets is critical. Hu et al. first exercised a simple Meyer rod coating method to coat carbon nanotubes (CNTs) and silver nanowires onto commercial Xerox papers and achieved sheet resistance as low as 1 ohm per square.^[35] Compared with plastics, the authors demonstrated that cellulose paper substrates could dramatically improve the film adhesion, and thereby simplify the coating process and lower the cost. It was also pointed out that the porous structure of cellulose paper could lead to more conformal coating of the conductive materials as it provides strong capillary force which enabled large contacting surface area between flexible nanotubes/nanowires and cellulose paper after the solvent was absorbed and dried out.

Since cellulose paper sheets can be used as both separators and electrode substrates, there has been a new research frontier where LIBs are completely built on paper sheets.^[36] Such paper-based LIBs are thin and flexible in nature and therefore can unlock a number of applications where conventional LIBs would fail to deliver success. These applications include batteries for bendable electronics and those where irregular shaped batteries are needed to fill the void spaces such as in the chassis of electric vehicles. Hu et al. later demonstrated a lamination process that could integrate all components of a LIB into a single paper sheet.^[37] The process began with assembling a CNT current collector layer and active material layer on a stainless steel, followed by transferring this double-layer structure to a cellulose paper coated with polyvinylidene fluoride (PVDF) as the glue. Figure 4 illustrates the lamination process. The CNT layer was coated by a doctor blade method using an aqueous ink containing CNTs and dodecylbenzenesulfonate (SDBS); while the active materials layer was coated by a slurry containing Li₄Ti₅O₁₂ (LTO, anode material)/LiCoO₂ (LCO, cathode material), with active carbon and PVDF as binder. The anode and cathode double layer structures could be peeled off and glued to both sides of the cellulose paper, which served as a separator membrane as well as a mechanical support. Compared to directly coating the double layer structure onto the cellulose paper, the lamination process could provide excellent mechanical integrity of the electrode materials and prevented leaking through the paper separator. Similar configuration can be applied to zinc-air batteries too. Hilder et al. successfully fabricated flexible zinc-air batteries by printing zinc/carbon/polymer composite anode on one side and poly(3.4-ethylenedioxythiophene) on the other side of a cellulose paper.^[38]

Figure 4. A thin, flexible paper-based Li-ion battery.

(a) Schematic fabrication process of free-standing double layer thin films. (b) (Left) 5 in. × 5 in. LTO/CNT double layer film coated on SS substrate; (middle) the double layer film can be easily separated from the SS substrate by immersing in water; (right) the final free-standing film after drying. (c) Schematic of the lamination process. (d) Schematic of the final paper Li-ion battery structure. (e) Photograph of the Li-ion paper battery before encapsulation for measurement. (Reprint with permission from ref [37], Copyright 2010 American Chemical Society.)

3.2. Binders

An important application of cellulose in LIBs is the binder in both cathodes and anodes for electrode materials including layered oxides,^[39] lithium iron phosphate (LFP),^{[40, 41]} sulfur,^{[42, 43]} TiO₂(B),^[44] graphite,^[45–47] and silicon.^[47–52] Although binders are not the most essential component in LIBs, they can greatly affect the properties of electrodes, particularly their mechanical and electrochemical performances.^{[43, 49, 53]} Current LIB technology uses PVDF as binders and requires the use of volatile and toxic solvents for processing, such as N-methyl-pyrrolidone (NMP). Both PVDF and NMP are expensive materials and difficult to recycle. The use of organic solvents can lead to hazardous events such as explosion. Therefore, fluorine-free, aqueous binder replacement is desired to address the economic, environmental, and safety concerns, paving the way toward safer and greener LIBs. In this regard, water-soluble carbonxymethyl cellulose (CMC), a derivative of linear polymeric cellulose with substitutional anionic carboxymethyl groups in various degrees, is frequently used. Typically, sodium CMC is often applied in conjunction with styrene-butadiene rubber by a weight ratio.^[54]

Studies have shown that CMC may not only be a safer, cheaper, and more environmental friendly alternative to PVDF as binders, but it also enhances the performance of the electrodes. Despite its high stiffness and a small strain at fracture, CMC could yield better bonding with the electrode materials and the electrode substrates. This can be intuitively attributed to the large number of carboxylmethyl groups on cellulose chains. Munao et al. compared the performance of PVDF and CMC when using a nano-porous silicon material as the cathode.^[49] They found that the sample with CMC had a much longer cycle life than that with PVDF, where both have similar initial capacity (1153 mAh/g for 35 cycles). This agrees with previous findings by Li et al. where a specific capacity about 1100 mAh/g for 70 cycles was obtained.^[52] Figure 5a schematically illustrates the proposed cycling failure mechanism of both binder materials. Compared with PVDF, CMC was much more adherent to the Si particle surfaces and the structure of the composite electrode materials thus had a much better stability upon cycling. The specific bonding between CMC and Si was proposed to be the carboxylic groups on CMC and the OH groups on the surface oxide of Si. Liu et al. reported that styrene-butadiene rubber (SBR), together with sodium CMC, can provide great adhesion between the electrode materials and the copper substrate. The cycle life of the LIB was thus improved nearly 10 times for Si-based cathode and 100 times for graphite-based cathode.^[50] Significant improvement in the cyclability of the specific capacity was also discovered when sulfur was used as the cathode material.^[42]

Figure 5. Cellulose-based materials as binders in LIBs.

(a) Proposed capacity fading mechanisms for Si composite LIB electrodes when using carbonxymethyl cellulose and PVDF as binder. (Reprint with permission from ref [49], Copyright 2010 Elsevier B.V..) (b) Production process of LIB electrodes with natural cellulose/ionic liquid binders. (Reprint with permission from ref [55], Copyright 2011 Elsevier B.V..)

Natural cellulose without any modification can also be used as binders. However, ionic liquids have to be used as solvents in this case as natural cellulose is not soluble in water. Jeong et al. prepared natural cellulose in 1-ethyl-3-methylimidazolium acetate (EMIMAc) and mixed active materials (graphite and LFP, respectively for anode and cathode) into the solution.^[55] After coating the slurry on metal foils, the electrodes were immediately immersed in water to extract EMIMAc (Fig. 5b). Water is a good solvent for EMIMAc but not for cellulose. In this phase inversion process, EMIMAc is completely recyclable. Unlike most above-mentioned work where a single electrode was tested against a lithium foil counter electrode, full coin cells were tested and a maximum capacity of 140 mAh/g (LFP) was obtained at the 55^th cycle and 85% of that capacity was retained up to 120 cycles.

3.3. LIB separators

Cellulose paper sheets are widely used in commercial alkaline batteries thanks to their excellent wettability, low cost and weight, high porosity, and good mechanical properties. Their use in LIBs are very limited primarily because of the lack of a thermal shutdown mechanism that closes the pores and terminates the ionic flow at high temperature.^[56] The hydrophilic nature of cellulose papers is another concern as too much water can cause degradation of lithium salts. However, recent work by Jabbour et al. and Leijonmarck et al. suggested that prolonged thermal treatment of cellulose papers can largely alleviate this problem.^{[57, 58]} It was estimated that the cost of the separator can be over 20% of the total cost of a LIB, and the majority of which comes from the manufacturing process.^[59] The low cost of cellulose has driven continuous efforts on using cellulose papers and cellulose composite papers for LIB separators as a cheap alternative to existing polyethylene(PE)/polypropylene(PP) membranes especially for low-power applications.^[60–63]

Zhang et al. for the first time reported the use of a commercial rice paper as a separator in LIBs. The cellulose fibers in the rice papers were about 5–40 μm in diameter. The performance of such separators was compared to a commercial PP/PE/PP Celgard membrane and a lower ionic resistance was discovered.^[64] Kim et al. prepared cellulose nanofiber papers with pore size tuned by incorporating SiO₂ nanoparticles and non-conductive spacer particles. Such cellulose composite papers were prepared by vacuum filtering a suspension consisting of cellulose powders and SiO₂ nanoparticles and subsequently drying between two filter papers under 1 MPa in a 80 °C oven.^{[65, 66]} They found that 5 wt.% SiO₂ content yielded the highest ionic conduction due to a balanced combination of nanoporous structure and the separator thickness. Figure 6 shows the SEM images and an illustration of the composite separator structure. Increasing content of SiO₂ nanoparticles increased the pore size of the separator membrane and thus the ionic conductivity. Thus, a significantly lower MacMullin number could be achieved compared to the commercial PP/PE/PP separators. Xu et al. prepared cellulose/polydopamine (CPD) membrane simply using a paper-making machine instead of a vacuum filtration system.^[67] A complete lithium cobalt oxide/graphite cell was fabricated using CPD separator, and better cycling stability and rate capability was observed compared to commercial PP separator and pristine cellulose separator. Alternating-current impedance of the cell with CPD separator showed only a minor variation of 9 Ω after 100 cycles indicating an excellent interfacial stability. Zhang et al. developed a nonwoven cellulose composite membrane that contains cellulose, sodium alginate, flame retardant and silica using a papermaking machine as well, and a complete LiFePO₄/lithium cell was fabricated and tested using such separators.^[68] In addition to good mechanical strength, improved electrolyte uptake and ionic conductivity, such composite separators showed stable cyclability and thermal dimensional stability at 120 °C owing to the addition of the flame retardant. The addition of sodium alginate and silica also enhanced the bonding between cellulose fibers and thus the tensile strength of the separators.

Figure 6. Cellulose-based paper sheet as LIB separator.

(a) Molecular and crystalline structure of cellulose fibers. (b–d) SEM images of the Si-cellulose nanofibers papers with 0%, 1 wt.%, and 5 wt.% SiO₂ content, respectively. (e) Schematic illustrations of the structure of the composite paper. (Reprint with permission from ref [65], Copyright 2013 Elsevier B.V.)

3.4. LIB electrolyte

In current LIBs, mixtures of ethylene carbonate and a linear carbonate are most commonly used as electrolytes. Such polar electrolytes raised safety concerns as they are volatile and flammable. Water and ionic liquids have also been used as electrolytes. In general, liquid electrolytes can suffer from leaking problems which threatens the life time of LIBs. Researchers have adapted polymer electrolytes where liquid electrolytes are trapped inside a polymer matrix to prevent leakage and allow the development of flexible LIBs. Cellulose and cellulose derivatives are studied for this purpose. In the simplest case, cellulose can be used as filler materials to reinforce the mechanical strength of polymer electrolytes. Nair et al. used a natural cellulose hand-sheet in a methacrylic-based thermos-set gel-polymer electrolyte.^[69] To enhance its compatibility with the polymer matrix, cellulose was modified by grating poly(ethylene glycol) methyl ether methacrylate under UV. The cellulose composite exhibited excellent mechanical properties and preserved the good electrochemical performance. Azizi Samir et al. conducted a series of study on the tunicin cellulose whisker-reinforced cross-linked poly(oxyethylene) (POE) electrolyte. They were able to obtain higher ionic conductivity over a wide temperature range and better mechanical properties compared to POE without tunicin cellulose whiskers. The reinforcing effect was more prominent at high temperature.^[70–73] In addition, the cellulose filler could provide a thermal stabilization effect to the polymer electrolyte above the melting point of poly(oxyethylene).^[71] Willgert et al. developed poly(ethylene glycol) (PEG) electrolyte reinforced with CNF paper. The CNF paper was made from never-dried softwood dissolving pulp and was modified by acryloyl chloride and propionyl chloride via reactions with the OH groups on cellulose, before their polymerization with the PEG methacrylate oligomers.^[74] Therefore, covalent bonds were created between the PEG and cellulose phases, which were beneficial to the ionic conductivity when the composite was swelled with liquid electrolyte. An ionic conductivity of 5×10⁻⁵ S cm⁻¹ and an elastic modulus around 400 MPa at 25 °C were obtained.

Cellulose and cellulose derivatives alone can be used as the polymer matrix for electrolytes. Lee et al. reported a composite polymer gels made of cellulose triacetate (CTA), N-methyl-N′-propylpyrrolidinium bis(trifluoromethanesulfonyl)imide, and lithium bis(trifluoromethanesulfonylimide with a ratio of 1:3:1.5.^[75] They demonstrated a significant enhancement in ionic conductivity owing to the strong interaction between lithium ions and the carbonyl group of CTA. Other cellulose derivatives have been used as well, with similar motivations of solvation and complexation of lithium ions to improve the ionic conductivity.^{[76, 77]} Grafting anionic side chains to the polymer backbones is another means of improving the ionic conductivity as the translational mobility of negative charges can be reduced so that positive lithium ions can have a significant mobility.^{[78, 79]} Paracha et al. grafted lithium salts of 2-acrylamido-2-methylpropane sulphonic acid on ethyl cellulose, and then copolymerized methyl methacrylate (MMA) with a crosslinking agent to form a semi-interpenetrating network for LIB electrolyte.^[80] 1.33 mS/cm was obtained from conductivity tests and the mechanical strength of the gels could be controlled by the concentration of MMA and ethyl cellulose.

3.5. As-grown template for complex electrode structures

The fibrous form factor of cellulose nanofibers also contributes to the development of novel LIBs as it allows cellulose nanofibers to be used as templates to grow tubular nanostructures of electrode materials. Guo et al. took the advantage of the morphological hierarchy of filter papers made from natural cellulose, and applied a surface sol-gel process and subsequent calcination process to prepare nanotubular hollow SiO_x.^[81] The authors argued that the nanotubes possessed firm framework and prevented the electrodes from pulverizing in the charge-discharge process while having sufficient voids to accommodate the volume change. Wang et al. reported a templated growth of TiO₂ nanotubes from cellulose filter papers. They first had Ti(OBu)₄ sufficiently adsorbed into filter papers in ethanol and then calcinated the paper at 500 °C.^[82] Compared with other templated growth of TiO₂ nanotubes, this method is cheaper, simpler, and more capable of being integrated into the fabrication of LIBs. Cellulose can also be easily converted to carbon which can be used as an anode material. Wang et al. reported that they prepared carbon aerogel from bacterials nanocellulose aerogels, which was obtained by freeze-drying bacterial nanocellulose hydrogels.^[83] After calcination, the 3D open network of bacterials nanocellulose hydrogels was preserved and carbon aerogels composed of about 20 nm nanofibers were produced. A much better capacity retention behavior than mesoporous carbon was observed thanks to the higher surface area of the aerogel electrode.

3.6. Supercapacitors

Similar to LIBs, cellulose papers can also be used as substrates for supercapacitors, which give flexibility to the devices compared to conventional metal substrates in order to meet various design and power needs of electronic gadgets. Gui et al. developed cellulose paper/CNTs/MnO₂/CNTs supercapacitor electrodes (Fig. 7a).^[84] The cellulose fibers in the paper not only provided a high surface area, but also acted as an interior electrolyte reservoir. The devices had dual ion diffusion and electron transfer pathways so they work as both double-layer capacitor and pseudocapacitors. Comparisons were made between electrodes with cellulose paper replaced by Al₂O₃-coated polyester textile, showing cellulose paper-based ones had better ionic and electron conductivity and a better capacitance retention up to 50k cycles. Surface area is a critical factor for supercapacitors, where cellulose paper-based supercapacitors can find its unique advantages. Bacterial nanocellulose (BCN), typically obtained by bacterial synthesis from low molecular weight sugars and alcohols, have drawn people’s attention as BCN fibers are much smaller (20–100 nm) and have better mechanical strength, superior chemical stability, can be made thinner.^{[4, 84–87]} For example, Kim et al. made BNC papers from cultured gluconacetobacter xylinum and prepared CNT/BNC composite papers (Fig. 7b).^[88] The supercapacitor showed an excellent specific capacity of about 20 mF/cm² with a reduction of smaller than 0.5% over 5k cycles at a current density of 10 A/g. They were also tested against bending and the performance was retained through 200 bending cycles. In addition to CNTs, graphene and conducting polymers are often used to form the composite too.^[89–91] A new advance in this area was made when Liew et al. reported their polypyrrole/cellulose nanocrystals composite supercapacitors.^[92] Unlike other laboratory models, where very thin films were used and thus their area specific capacitance was too small for commercial applications, over 100 μm-thick films were fabricated, which yielded a high capacitance of 240 F/g. The highest area specific capacitance of 1.54 F/cm² was demonstrated before electrolyte diffusion reaches its limitation in thicker electrodes.

Figure 7. Cellulose-based supercapacitors.

(a) Schematic structure and SEM image of a paper/ CNTs/MnO₂/CNTs supercapacitor electrode. Red arrows indicate location of CNTs. (Reprint with permission from ref [84], Copyright 2013 American Chemical Society.) (b) SEM image (top) and a cross-sectional image (bottom) of a CNT-coated bacterial nanocellulose (BNC) paper supercapacitor electrode. Inset is a photo of the CNT-coated BNC paper. (Reprint with permission from ref [88], Copyright 2012 American Chemical Society.) (c) Supercapacitor assembled using cellulose/CNTs composite paper units. Top left is a schematic illustration showing the structure of the nanocomposite paper units. Top right are photos of the nanocomposite units. Bottom is a cross-sectional SEM image of the nanocomposite paper showing MWNT protruding from the cellulose-RTIL thin films. (Reprint with permission from ref [93], Copyright 2007 The National Academy of Sciences of the USA.)

Rather than using pre-made cellulose papers, Pushparaj et al. developed cellulose/CNTs composite paper supercapacitors by embedding cellulose solutions into vertically-aligned multi-walled carbon nanotube (MWNT) arrays grown on Si substrates (Fig. 7c).^[93] Unmodified plant cellulose in an ionic liquid (1-butyl,3-methylimidazolium chloride) was infiltrated into the MWNT to form a uniform film of cellulose and the ionic liquid was removed by solidification of the film on dry ice and immersing it in ethanol for extraction. This thin film was then peeled off from the Si substrate and exhibited excellent mechanical flexibility and integrity during operation. A very intriguing feature of such nanocomposite paper-based supercapacitors was the choice of electrolyte. In addition to water and ionic liquid electrolytes, the device could operate with a suite of electrolytes based on body fluids, such as sweat or blood, allowing body implant or use under special circumstances.^[94]

Cellulose can be used as binders and separators in supercapacitors just like in LIBs. Used in conventional active carbon-based double layer capacitors, Varzi et al. demonstrated that cellulose fiber binders enabled superior performance retention under float tests at high voltage compared to PVDF, a particularly important merit of double layer capacitor electrodes with ionic liquids as electrolytes. Their work demonstrated that the improvement was more significant at the negative electrode.^[95]

4. Cellulose Nanomaterials for Nanogenerator Developments

Mechanical energy from the environment (e.g. ambient vibrations or activities of human body) ^{[96, 97]} represents a unique and sustainable energy source that perfectly suits for powering portable electronics and unattended devices.^[97–99] In 2012, the triboelectric nanogenerator (TENG) was first reported ^[100] as a new type of technology that converts environment mechanical energy into electrical energy.^[13] Compared to other technologies, TENG is advantageous in terms of high efficiency, high power density, light weight, low cost, and great manufacturability.^{[97, 101–103]} TENG operates based on the coupling effects of contact electrification and electrostatic induction. The working principle requires two dissimilar surfaces to be oppositely charged upon contact. Positive (polyamides,^[104] metals ^[97]) and negative (fluorinated ethylene propylene (FEP)^{[105, 106]}, polytetrafluoroethylene (PTFE) ^[104] and polyethylene terephthalate (PET)^{[104, 107–109]}) triboelectric materials are typically used in pairs in TENGs due to their opposite tendencies to gain or lose charges upon contact. Cellulose, the most abundant natural polymer on Earth, can also be a good candidate as a triboelectric material, bringing unprecedented natural degradability, recyclability and biocompatibility to TENG systems. As a result, CNF thin films are increasingly explored in TENG developments in recent a couple of years. Here, we review the most recent technology evolvement of CNF-based TENGs, from fundamental property tuning to practical implementations of CNF materials in TENG systems. Cellulose-based piezoelectric nanogenerators (PENGs) are also briefly reviewed as an alternative for mechanical energy harvesting.

4.1. Cellulose nanostructure-based TENGs

Natural cellulose is a slightly triboelectric positive material. In order to effectively generate triboelectric output, it was typically paired with strong triboelectric negative materials to fabricate a TENG device. The first CNF-based TENG was demonstrated by Yao et al. by paring a CNF mesoporous film with a FEP film.^[110] The CNF film surface exhibited a clear fibrous feature and a surface roughness of ~300 nm, offering a large surface area for contact and electrostatic charge generation (Fig. 8a). The as-prepared film also showed excellent transparency and flexibility and the entire TENG device is flexible and transparent (inset of Fig. 8b). The average triboelectric output under regular mechanical reached ~5 V (Fig. 8b), which were comparable to the performance of TENGs made of typical synthetic polymer pairs such as Kapton-PET ^[100] and PTFE-polyamide.^[104] External load matching test showed the optimal power was 0.56 mW at ~1 MΩ.

Figure 8. TENG fabricated from CNF films.

(a) Atomic force microscopy (AFM) topography image and roughness of a CNF film surface. (b) Voltage output of a TENG fabricated from a CNF-FEP pair. Inset is a photograph of the TENG. (c) Manufacturing process of the triboelectric fiberboard: a CNF-FEP TENG is embedded into a fiber mat and the entire piece is then cold-pressed and dried to obtain the final fiberboard. The bottom right photograph shows the fiberboard lighting up LED lights when stepped on. (Reprint with permission from ref [110], Copyright 2016 Elsevier Ltd.)

Successful development of TENG using CNF film opened a possibility of creating power-harvesting boards using wood-extracted fibers. CNF-based TENG was integrated with a fiberboard made from recycled cardboards as a practical demonstration. Figure 8c shows the large-scale manufacturing process of the triboelectric fiberboard. This board was able to generate electricity as high as ~10–30 V and ~10–90 μA, when a person of average weight repeatedly stepped on this TENG fiberboard. This amount of electric energy from one single step can light up 35 green LEDs connected in series (panel 6 in Fig. 8c), demonstrating the promise of powering commercial electronics from foot traffic. This research fully exploited the merits of mechanical flexibility and nanostructured surface features contributed by the NC films.

Following this work, cellulose nanomaterials were more and more frequently used in TENG developments. Kim et al ^[111] regenerated bacteria cellulose to obtain a uniform, flexible and transparent cellulose film. Since the chemistry of bacteria cellulose is nearly the same as wood cellulose, their dielectric and triboelectric polarization are generally the same, and thus the triboelectric output remained at the same level as the previous report. When the cellulose film was assembled with Cu to form a vertical contact-separation mode TENG, the typical voltage output from a 5 × 5 cm² TENG under a compressing force of 16.8N at 1 Hz was ~13 V (Fig. 9a), where a curved cellulose film was found favorable for higher output. Uddin et al ^[112] also fabricated pure-cellulose based TENG by assembling two flexible cellulose films made from hydrolyzed microcrystalline cellulose with a PI (polyimide)-Al-PET (polyethylene terephthalate)-Al-PI corrugated-core structure (inset of Fig. 9b). The TENG showed a maximum output voltage of 142–153 V and output current of 3.2–3.9 μA at a frequency of 3 Hz and a series of applied forces ranging from 3 to 10.1 N. This TENG device also exhibited a H₂-gas responsive output: when the TENG was exposed to H₂ gas, the voltage output decreased remarkably, from ~142 V to ~41 V in 0.1 vol % H₂ (Fig. 9b). This force-independent output response was possibly owing to the strong interaction between cellulose fibrils and H₂ molecules, showing a promising application potential of self-powered gas sensing.

Figure 9. TENGs developed based on cellulose films.

(a) Schematic structure of a BNC-Cu TENG and the voltage output from the TENGs with different active surface area. (Reprint with permission from ref [111], Copyright 2017 Elsevier Ltd.) (b) Photograph of a microcrystalline cellulose film-polyimide TENG and its voltage output under various compression force when exposed to H₂ and air. (Reprint with permission from ref [112], Copyright 2017 Elsevier B.V..) (c) Schematic synthesis process and SEM image of PDMS/CNC composite. Center image shows the film adhered to a glove surface, demonstrating its good flexibility. (d) The voltage output of TENGs fabricated from the PDMS/CNC films with different CNC concentrations. (Reprint with permission from ref [114], Copyright 2017 The Royal Society of Chemistry.)

In addition to pure cellulose films, cellulose composites were also developed for TENG applications. Chandrasekhar et al. fabricated polydimethylsiloxane (PDMS)/cellulose composite films with improved flexibility and mechanical integrality.^[113] By paring with Al film, the TENG generated a voltage output of 28 V at a force of 32.16 N. A similar cellulose composite film was created by mixing cellulose nanocrystal flakes in a PDMS matrix (Fig. 9c ).^[114] Under a compressing force of 20–40 N at a frequency of 10 Hz, the TENGs assembled from the composite film against Al showed voltage output of 150–320V (Fig. 9d), which was 4–10 times higher than its pure PDMS counterpart. This output was also much higher than the other cellulose composite films, which might be related to the internal polarization from the cellulose nanocrystal flakes.

4.2. Polarity tuning of cellulose nanomaterials

Above pioneer work on cellulose TENG opened a new exciting material engineering direction for TENG development using nature and recyclable materials. However, all of them still included synthetic polymers and/or other electrode materials to pair with cellulose for effective charge separation. The weak polarization of nature cellulose limits their capability of generating surface charge, which keeps their performance low compared to those synthetic polymers. Inclusion of synthetic polymers would also reduce the recyclability and eco-friendliness, as well as raise the cost. In order to realize all-cellulose TENGs, the electric property of CNF fibers needs to be tuned. Thus, both TENG active layers can be made from CNF thin films. Taking advantage of the abundant hydroxyl groups on cellulose molecules, CNF can be functionalized via simple wet-chemistry reactions.^[115–117] Their electrical properties could then be drastically tuned by the new chemical functional groups. For example, nitro groups (-NO₃) can largely raise the electron affinity of cellulose molecules; while methyl groups (-CH₃) can improve the cellulose’s capability of donating electrons.

Recently, Yao et al. employed wet-chemical reaction approaches to introduce nitro groups and methyl groups, to rationally tune the triboelectric polarity of CNF for TENG application.^[118] Nitro-CNF was achieved by treating CNFs with nitration acid mixture comprising HNO₃, H₂SO₄ and water. Methyl-CNF was received by reacting mercerized CNF with dimethyl sulfate. The as-prepared nitro- and methyl-CNF films were both flexible and transparent showing the chemical functionalization imposed negligible influences to the film’s other physical properties. Triboelectric figure of merit (FOM) characterization was performed on these CNF thin films by quantifying the surface charge density upon contacting standard Ga-In eutectic liquid metal electrode. Measurements revealed the charge densities from pristine NC, nitro-NC, methyl-NC and FEP were −13.3, 85.8, −62.5 and 120.9 μC m⁻², respectively (Fig. 10). The negative sign indicates that electrons were transferred from the testing film to the electrode upon contact. FEP is the most tribo-negative materials and was used as a reference here (FOM=1). It can be seen that, after the methylation and nitration, the charge density of CNF exhibited 5–10 times enhancement toward the same and opposite charge transfer directions, respectively. The charge density of nitro-CNF raised to nearly 80% of FEP. Corresponding FOMs of nitro-CNF and methyl-CNF were calculated to be 0.51 and 0.28, respectively, which are comparable to other commonly used synthetic polymers such as Kapton (0.60), PVDF (0.45) and polyethylene (0.43).

Figure 10. Surface charge property of triboelectric films.

(a – d) Current flow measured during one contact-separation event when electric contact was made between Ga-In eutectic liquid metal and the triboelectric film of CNF (a), FEP (b), nitro-CNF, and methyl-CNF. Inset in (a) is the setup that can move vertically for making the contacts. Inset in (c) and (d) are the molecular structure of nitro-CNF and methyl-CNF, respectively. (Reprint with permission from ref [118], Copyright 2017 WILEY-VCH Verlag GmbH & Co. KGaA.)

The remarkable triboelectric polarity tuning allowed the creation of all-CNF-based TENGs with both triboelectric functional layers being CNF materials. TENGs with the vertical contact mode were assembled with different pairs of pristine and functionalized CNF and their voltage and current outputs were characterized. The two chemically functionalized CNF films both showed enhanced output when paired with the pristine CNF film. Because nitro- and methyl-functionalization shifted the triboelectric polarization toward opposite directions, when these two materials were paired together, the output was further raised to an average voltage and current value of 8 V and 9 μA respectively. It is important to notice that output from the nitro-CNF-methyl-CNF pair was approaching that of the FEP-CNF pair (9 V and 10.5 μA); whereas the FEP-methyl-CNF pair showed an even higher output (12 V and 10.5 μA) due to the more positive methyl functionalization. Since this polarization tuning was based on molecular-level functionalization, the enhanced output was fairly stable. Long-term test of the nitro-CNF-methyl-CNF TENG revealed no output signal decay after 5×10⁵ cycles. These experiments evidenced the great potential for replacing commercial synthetic polymers by renewable and biodegradable CNF based films with comparable or even superior triboelectric performance.

4.3. Flexible, transparent and conductive CNF paper

Most recently, Yang et al. presented a development of flexible and nature degradable transparent electrodes based on CNF paper.^[119]. In their work, an aluminum-doped zinc oxide (AZO) thin film was grown by ALD on a CNF paper as the transparent conductive oxide (TCO) layer. While the deposition temperature was only 150°C to sustain the structural and chemical integrity of the CNF paper, the AZO film showed excellent crystallinity, conformality and a very low surface roughness (~2 nm, Fig. 11a). By optimizing the Al concentration, the highest resistivity of the AZO-coated CNF paper was found to be 2.7×10⁻³ Ω.cm at 3.7 at.% Al. Meanwhile, the AZO-CNF papers showed a high transmittance of >84% in the visible range of 400–760 nm, which made them acceptable for TCOs applications like displayers and solar cell windows. The AZO film with 3.7 at.% Al also exhibited the best resistivity and transparency combination, which was at the same level of the reported values of ITO and graphene (Fig. 11b) ^{[120, 121]}

Figure 11. Transparent conductive AZO-CNF papers.

(a) Left is AFM image of an AZO-coated CNF surface. Right are optical images of an as-fabricated CNF paper from the CNF gel (top) and an AZO-coated CNF paper (bottom). (b) Transmittance (at 550nm) and sheet resistance of AZO-CNF in comparison with other typical flexible transparent electrode materials. (Reprint with permission from ref [119], Copyright 2017 Creative Commons Attribution 4.0 International License.)

The relative open structure of the CNF paper was believed to be able to yield strong binding of the oxide layer. As evidenced by the cross-sectional energy disperse spectrum (EDS) mapping (inset of Fig. 12a), appreciable amount of Zn signal was detected inside the CNF film area. This was attributed to the porous structure of the CNF paper which facilitated precursor infiltration. Growth of AZO a certain distance into the CNF paper could enhance the binding strength between the AZO layer and the CNF surface (Fig. 12a) and improve its mechanical integrity. As a result, the sheet resistance of the AZO-CNF papers exhibited an excellent stability within a range of 135 to 180 Ω·sq⁻¹ under compressive strain from −0.057% to −0.457% (Fig. 12b). This stable strain region of the AZO-CNF film was comparable to typical ITO-based flexible transparent films, suggesting its good potential to be used as a competitive new type of flexible transparent substrate.

Figure 12. AZO-CNF papers for TENG development.

(a) A cross-sectional SEM image of the AZO-CNF paper. Inset is the corresponding EDS mapping of the Zn elements showing substantial Zn signal within the CNF paper. (b) Change of sheet resistance and resistivity as a function of bending radii when the AZO film was under compressive strains. (c) A series of photos showing the process of dissolving an AZO-CNF paper in city water at 70 °C. The AZO-CNF paper is completely dissolved after 60-minute stirring. (d) Schematic image and photograph of the all-transparent TENG made from a pair of AZO-CNF papers. (e) The long-term stability of the AZO-CNF TENG. (Reprint with permission from ref [119], Copyright 2017 Creative Commons Attribution 4.0 International License.)

One unique merit of AZO-CNF paper that distinguishes it from other synthetic flexible polymer substrates is its excellent degradability and recyclability. As shown in Fig. 12c, one piece of AZO-CNF paper was placed into a warm city water (70 °C). After 60-minute stirring, the AZO-CNF paper was completely dissolved into the water. The final solution was clear and no suspension can be observed. It became a CNF gel with a small amount of Zn and Al ions coming from the AZO coating. Using AZO as the TCO component ensures completely dissolving of all the film components. This solution can thus be further concentrated and filtered to produce CNF film again, suggesting an excellent recyclability.

A pair of AZO-CNF paper with chemically tuned polarity was then used to fabricate a nature degradable TENG device (Fig. 12d). Because no more metal contact was used in this TENG assembly, the entire device was also completely water degradable. When the TENG was under a constant 10 Hz impact, the maximum voltage was from 1~6 V under a series of external load resistance (Fig. 12e). The maximum power of ~0.5 mW was obtained at 1 MΩ load resistance, comparable to other cellulose-based TENG devices. This is the first report of flexible, transparent and metal-free TENG that are completely naturally degradable. With a reasonably high electrical output and operational lifespan, it offers a unique solution for developing green, biocompatible and self-degradable powering units for self-sustainable electronic systems.

4.4. Cellulose-based PENGs

PENG is another widely applied principle for mechanical energy harvesting using piezoelectric nanomaterials.^[122] Crystalline cellulose is a well-known polymeric piezoelectric material, and its piezoelectric polarity arises from the alignment of the glucose units through glycosidic linkage linearly along the C₂ monoclinic lattice, a non-centrosymmetric crystallographic structure. While the piezoelectric property has been observed over half century ago from natural wood,^[123] isolated cellulose was not investigated as a piezoelectric material until 2006 when Jaehwan Kim et al. published their pioneer work on a cellulose paper (which was termed as EAPap, standing for cellulose electro-active paper) demonstrating an actuation phenomenon.^[124] This cellulose paper was a film of regenerated cellulose from xanthate cellulose solution, and morphologically consists of ordered and disordered regions. This paper exhibited appreciable mechanical deflection under relatively small electrical potential, whereas this response was also very sensitive to humidity. Oriented CNC film was expected to generate more significant piezoelectricity due to their highly ordered crystal structure. Levente et al. fabricated an ordered CNC film by a combination of shear and electric fields.^[125] AFM measurements revealed the piezoelectric constant d₂₅ of such a highly ordered CNC film to be 2.1Å/V.

However, so far there is no work reporting directly using crystalline cellulose for PENG development, possibly because of its relatively weak piezoelectric effect and high processing requirements. Instead, cellulose-based composites were successfully explored as a promising PENG structure. Zhang et al. reported a piezoelectric composite paper based on ferroelectric BaTiO₃ (BTO) nanoparticles and bacterial cellulose.^[126] This composite paper exhibited remarkably enhanced piezoelectric output with a peak voltage of 14 V and a peak current density of 190 nAcm⁻². The maximum power density reached 0.64 μW cm⁻². This output was nearly ten times higher than that of piezoelectric composite paper made from the same BTO nanoparticles in a PDMS matrix. The enhancement was attributed to the extremely uniform distribution of BTO nanoparticles in the cellulose matrix in contrast to more agglomerations in PDMS media. This advantage was intuitive since cellulose offers preferable hydrophilicity and strong surface charge to stabilize nanoparticle suspension during composite processing. It should be noted that this output power was still much lower than cellulose-based TENG devices. It may be promising to integrate both bulk-related piezoelectric effect and surface-related triboelectric effect in cellulose nanostructures to further improve mechanical energy harvesting capability.

5. Conclusion and Remarks

In general, this article reviews the most recent developments and contemporary applications of cellulose nanomaterials in energy-related areas, including lithium ion batteries, supercapacitors, solar cells, photoelectrochemical electrodes, and nanogenerators. Compared to the conventional application of cellulose in bio-fuel generation, these applications emphasize the polymeric material nature of cellulose. As we addressed in the introduction part, cellulose possesses many unique features and advantages as a polymer and presents a unique combination of chemical, structural, dielectric, and optical properties that can be rarely found in other polymer systems. In this regard, using cellulose nanomaterial as a functional component in modern energy devices and systems may lead to a new paradigm of materials innovation. Therefore, we can highlight the studies reviewed in this article as functional cellulose nanomaterials development. Although it just emerged in the most recent a few years, unprecedented application potentials have been seen in various directions. It is foreseeable that functional cellulose nanomaterials are going to take a more and more important role in the family of energy materials. Certainly, challenges also exist and require substantial research efforts to overcome.

First, synthesis of cellulose nanomaterials with well-controlled degree-of-polymerization. This is the fundamental requirement for us to obtain reliable understanding of the property and structure relationships in cellulose nanomaterials and predict their functionality.

Second, molecular-level understanding of how cellulose molecules arrange into different phases, i.e. how the molecular chains organize and interact with each other in short- and long-range ordered fashion. Although many cellulose polymorphs have been found, how they can be controlled and thus tune the macroscopic property is mostly still a mystery. This understanding will essentially uncover the unprecedented application potentials of cellulose nanomaterials as an essential functional component in many modern energy and electronic systems.

Third, new approaches to functionalization, modification, and assembly of cellulose nanomaterials in different length scale. This capability, which is currently lacking, is an essential step leading the fundamental and conceptual studies of cellulose nanomaterials toward practical applications.

To conclude, study of functional cellulose nanomaterials is just at the beginning. As reviewed in this article, it possesses tremendous materials science value and application potential in many energy-related fields. With advanced synthesis, processing and characterization techniques, and the talents from different scientific disciplines, cellulose nanomaterials may soon become a powerful multifunctional material platform that nature offered to us to bring our energy materials technology to a new horizon.