NC State
Zhang, H., Dou, C., Pal, L., and Hubbe, M. A. (2019). "Review of electrically conductive composites and films containing cellulosic fibers or nanocellulose," BioRes. 14(3), 7494-7542.


Strategic combinations of cellulosic materials with electrically conductive polymers or nanoconductors offer important potential advantages for technological advances, light-weight inexpensive products, applications of novel form factors, and more eco-friendly alternatives to certain forms of smart packaging and electronics. This review of the literature focuses on how such electrically conductive composite systems work, the roles that cellulosic materials can provide in such structures, processes by which electrically-conductive cellulose-based composites and films can be manufactured, and various potential applications that have been demonstrated. Several advantages of cellulose, such as ease of fabrication, compatibility with conductive agents, and sustainability, allow its integration with conductive agents in making conductive composites. Applications of electrically conducting cellulose-based composites for strain sensors, energy storage, solar cells, electrodes, supercapacitors, and smart packaging are discussed.

Download PDF

Full Article

Review of Electrically Conductive Composites and Films Containing Cellulosic Fibers or Nanocellulose

Hao Zhang,a,b Chang Dou,b Lokendra Pal,b and Martin A. Hubbe b,*

Strategic combinations of cellulosic materials with electrically conductive polymers or nanoconductors offer important potential advantages for technological advances, light-weight inexpensive products, applications of novel form factors, and more eco-friendly alternatives to certain forms of smart packaging and electronics. This review of the literature focuses on how such electrically conductive composite systems work, the roles that cellulosic materials can provide in such structures, processes by which electrically-conductive cellulose-based composites and films can be manufactured, and various potential applications that have been demonstrated. Several advantages of cellulose, such as ease of fabrication, compatibility with conductive agents, and sustainability, allow its integration with conductive agents in making conductive composites. Applications of electrically conducting cellulose-based composites for strain sensors, energy storage, solar cells, electrodes, supercapacitors, and smart packaging are discussed.

Keywords: Nanocellulose; Conductive material; Composite; Flexibility; Smart Packaging

Contact information: a: Department of Materials and Chemical Engineering, Henan University of Engineering; Zhengzhou, Henan 450000, PR China; b: Department of Forest Biomaterials, North Carolina State University, Box 8005, Raleigh, NC 27695-8005 USA;

* Corresponding author:


The pace of recent publications reveals intense interest in the production of thinner, cheaper, lighter, and more eco-friendly electro-conductive films and composites. The goal of this article is to review scholarly work leading towards an understanding of electrically conductive composites and films that contain various forms of cellulose, with an emphasis on nanocellulose. Cellulose itself is not a good conductor of electricity, but as a primary ingredient in electrically conductive composites it may serve as an eco-friendly, generally low-cost substrate. However, the combining of a cellulose-based non-conductor with conductive materials presents many challenges or potential advantages in terms of electrical connectivity, tolerance to forces and extension, interfacial adhesion, and chemical stability of the composite materials.

Motivations for the use of cellulosic materials in modern products, displays, and packaging often can be expressed in a negative sense: avoiding higher-cost materials, avoiding heavy structures, and avoiding usage of non-renewable materials, etc. (Siqueira et al. 2010; Abdul Khalil et al. 2012). If one is willing to accept that some classes of manufactured items will be discarded after a single usage (as in the case of a package), then it is important to explore the use of constituent materials that are inherently recyclable and also those that cause minimum harm if discharged to the environment. This pair of challenges will be kept as a subtheme in this article when examining the scholarly literature pertaining to various electrically conductive materials that have been paired with cellulosics in various studies. Another theme will be to consider the different potentially favorable attributes of cellulose in such applications. The list might include such attributes as fibrillar shape, flexibility, and various chemical details of cellulosic materials.

Certain earlier review articles should be mentioned, either to provide further background or as a means to cut down on the need for detailed discussion of subtopics that already have been well covered by others. Clerc et al. (1990) reviewed the AC electrical conductivity of disordered binary composite systems, with an emphasis on semi-random networks of electrical connectivity (percolation) and mathematical modeling. Roldughin and Vysotskii (2000) considered metal-filled films in their review of structures and mechanisms affecting conductivity. The reviews by Bauhofer and Kovacs (2009) and Min et al. (2010) deal with the use of carbon nanotubes in conductive composites. Svennersten et al. (2011) reviewed the topic of organic bioelectronics in nanomedicine. Tobjörk and Österbacka (2011) reviewed “paper electronics,” covering applications in which paper can be the substrate for printed electronics. Nyholm et al. (2011), in their review, covered flexible and paper-based energy storage devices. Irimia-Vladu (2014) reviewed the topic of “green electronics,” with some discussion of electrically conductive polymers and conductive paper. Agate et al. (2018) also reviewed recent advances in printed electronics, with particular emphasis on paper as a substrate. Qiu and Hu (2013) reviewed the topic of ‘smart” materials based on cellulose. Jabbour et al. (2013), Shi et al. (2015), Ummartyotin and Manuspiya (2015), and Du et al. (2017) covered the role of conductive cellulosic systems for use in batteries and other energy-related applications. Terzopoulou et al. (2015) as well as Rouf and Kokini (2016) reviewed the roles of graphene, its oxides, and other derivatives in electrically conductive composites containing polysaccharides. The general topic of cellulosic materials in composites and nanocomposites also has been reviewed (Hubbe et al. 2008; Siqueira et al. 2010; Abdul Khalil et al. 2012).

A key motivation for the use of cellulosic materials in various composites has been to reduce reliance on non-renewable materials such as synthetic plastics (Abdul Khalil et al. 2012). The wastes of conductive materials based on plastics and inorganic semiconductors are difficult to recycle or decompose; hence they are causing serious environmental problems (Irimia-Vladu 2014). More recently, bio-based materials, such as cellulose and chitin, are receiving a lot attention because of their compostable and cost-competitive characteristics. Cellulose is a polysaccharide of D-glucose. It represents the most abundant polymer in nature and has an annual production of over 150 million metric tons (Zhao and Li 2014). Due to its specific crystalline nature, outstanding mechanical properties, and high surface activity, cellulose and its derivatives are among the most promising biopolymers for many different applications, including but not limited to biomedical devices and wastewater treatment (Grishkewich et al. 2017).


Conductive Composites

Conduction within composite materials depends on the mobility of electrons (Last and Thouless 1971; Wosnitza 1996). As will be described in more detail in the next main section, electrons flow very easily in metals, but also within certain organic polymers such as polypyrrole, graphene, and carbon nanotubes. Most often, when dealing with electrically conducting composites, one of the phases conducts electricity well, whereas the other either is an insulator or a much poorer conductor. So in this review much attention also will be paid to gaps, defects, and complete lack of contact among the electrically conductive elements in the composite. The envisioned situation is illustrated schematically in Fig. 1, where a nano-scale gap space is represented in the lower-right quadrant.

Fig. 1. Schematic illustration of nanostructure within a hypothetical cellulose-based composite containing a conductive component such as carbon nanotubes, which are shown as either being stretched out or agglomerated into bunches

It is not always clear what is the most advantageous size of either the conductive particles or the structural particles (e.g. cellulose) in a conductive composite or film. A smaller particle size typically requires a higher energy to prepare the materials and greater costs if there is need for surface modification or the use of a compatibilizer (Lai et al. 2003; Qiu et al. 2005; Roy and Potiyaraj 2018). Also, depending on how a composite is constructed, there will be cases where higher surface areas may imply a greater requirement and cost of electrically conductive material. Shi et al. (2015) listed the following potential advantages of nanostructured conductive polymers for such applications: favorable accommodation of strains, large surface areas, short pathways for charge transport (maybe due to small size of the potential devices), and the availability of a variety of mechanisms of conduction. Though there are reasons to favor different sizes of particles in a composite, a key question will be what combination best meets the performance needs of the application while minimizing costs.

Though the main focus of this review is on composites, it will be important to bear in mind that there are some alternative strategies that can be suggested for future research. For instance, in any mixture of conductive and non-conductive materials there is an inherent risk of inefficient flow of current. Some of the conducting particles may fail to be in electrical contact with other conductors. Such problems often can be minimized by a layered structure, placing the conductive material in layers and then using adjacent layers of structural, binding, or insulating material for support and protection. Though such issues lie beyond the scope of the present review, it seems a promising topic of future research to compare the efficacy and practicality of using layered structures as an alternative to the composite mixtures that will be mainly considered in the present article.

Rule of Mixtures

One of the simplest, but not necessarily accurate ways to estimate the conductivity of a composite is to sum up the products of the proportional shares of the components and their respective electrical conductances (Du et al. 2006; Hubbe 2017). The prediction of such a “rule of mixtures” calculation is illustrated schematically in Fig. 2. The figure also shows hypothetical results for a system exhibiting positive or negative synergistic effects relative to the expected result based on the rule of mixtures assumption.

Fig. 2. Hypothetical plot illustrating what is meant by a “rule of mixtures” relationship, as well as positive and negative synergistic effects

Spitzig et al. (1993) observed a rule of mixtures relationship in the case of a heterogeneous amalgam of two metals having unequal conductance. Gaier et al. (2003) observed reasonable agreement when predicting the conductance of graphite-polymer composites, but only when testing broad, flat specimens. Narrow specimens showed deviations from the rule, which were attributed to there being too few paths of electrical connections, depending on the size of the specimen. The term “effective medium approximation” has been used when accounting for well-behaved systems in which the conductance properties of composites were predictable based on their composition (Clerc et al. 1990; Hsu and Wu 1993; Wang et al. 2015c; Xia et al. 2017). Similar to the rule of mixtures, the effective medium theory implies that the conductances of the constituents of a mixture are averaged in proportion to their content. Whatever terms and details are used for the calculations, there are some key limitations such approaches, since they do not account well for effects related to shapes of conductive particles, various details related to connectivity, and the statistics of completion of the conductive paths.

Percolation Theories

The word percolation, as an approach to predicting the electrical conductivity of composites, actually originated in efforts to predict the flow through granular media, as in the case of hot water flowing through coffee grounds (Hunt 2001). The concept of percolation perhaps can be best grasped by considering the experiment conducted by Last and Thouless (1971). These researchers randomly punched holes into conductive paper. They found that conductance went to zero when the proportion of holes approached 0.4. Subsequent researchers have used more sophisticated numerical approaches to perform the same general type of analysis, with options to consider three-dimensional paths of conductance and the effects of differently shaped conductive particles. Excellent reviews of the topic of percolation theory are provided by Kirkpatrick (1973), and Clerc et al. (1990). Hunt (2001) reviewed percolation in the case of fluid flow through granular media with random closure of passages for the fluid. A review article by Bauhofer and Kovacs (2009) provides a listing of “percolation threshold” values, i.e. the minimum content of the conductive component required to significantly increase the composite’s electrical conductance relative to that of the non-conductive phase.

The concept of percolation is illustrated in Fig. 3, which is patterned after examples shown by Zallen (1983). The case shown represents the kind of calculation in which electrical connections are systematically (usually randomly) removed from a regular grid, which can be regarded as a system of ordinary resistors. Elimination of links reduces the number of complete paths, until eventually the last complete connection has been lost. In the case shown, for instance, the removal of links completely eliminated left-to-right flow of current crossing the whole grid, and only one path remained for up-to-down current. It is typical for computations to be carried out with very large grids, so as to better represent real systems in which there are very high numbers of possible connections (Zallen 1983).

Fig. 3. Illustration of percolation concept, comparing a fully connected grid with a case in which about half of the connections have been eliminated, leading to reduced connections

The traditional approach inherent in percolation theory is an assumption that conductive elements are distributed in a purely random fashion (Last and Thouless 1971; Kirkpatrick 1973; Zallen 1983; Mamunya et al. 2002; Bauhofer and Kovacs 2009). For example, Last and Thouless (1971) achieved such an effect by punching random holes into a conductive paper sheet. The results of calculations based on percolation theory have shown quite good agreement with observed conductance measurements for systems that are sufficiently well removed from the critical range of content of the conductive component (Kirkpatrick 1973). Percolation theory predicts much more efficient completion of conductance paths with increasing length-to-width (aspect) ratio of the conductive particles (Kirkpatrick 1973; Mamunya et al. 2002; Li et al. 2007; Bao et al. 2013). Some progress also has been achieved in making predictions within the critical zone, for instance with the concept of a “Green’s function” (Kirkpatrick 1973), a “critical path analysis” (Hunt 2001), and a “magic ratio rule” (Geng et al. 2015) to account for connectivity. Roldughin and Vysotskii (2000) introduced the usage of fractal concepts in accounting for connectivity in percolation systems. However, Bauhofer and Kovacs (2009) found large divergence in data obtained from different studies. This can be taken as an indication that the assumption of purely random distribution may be incorrect in many or most cases when applying such calculations to predict experimental results.

Researchers have found it useful to summarize data by means of power laws, as shown in Eqs. 1 and 2 (Grassberger 1999; Mamunya et al. 2002; Bauhofer and Kovacs 2009; Min et al.2010),

where  is the electrical conductivity for filler volume contents equal to or larger than a critical value of  c c is the conductance of the conductive component,  c is the conductivity associated with the critical condition,  m is the conductivity of the matrix,  is the volume fraction of conductive particles, and F is the maximum packing of the conductive filler component. The exponent t was found empirically to lie within the range of 2.4 to 3.2 (Mamunya et al. 2002), which exceeds the theoretically expected value of 1.7. Min et al. (2010) reported that in a variety of structures the value of t can fall in the range between 1.6 and 2 according to various theoretical predictions. Experimentally determined values have been reported in the range of 0.7 and 3.1.

One of the challenges in such models involves how to decide whether or not adjacent conductive elements are in contact or not, since the conventional practice is to regard such contact points as on-off switches for current flow (Kirkpatrick 1973; Min et al. 2010). For example, Wang and Chatterjee (2003) assumed that full contact was established whenever the distance between the centers of rod-like particles did not exceed a predetermined value. However, what makes sense mathematically does not necessarily account properly for what actually happens. For instance, Du et al. (2006) observed a relatively large interfacial resistance at the junctions between carbon nanotubes that were in physical contact.

Hopping and Tunneling Mechanisms of Charge Flow

Since cellulosic materials generally don’t conduct electricity well, it is reasonable to expect there to be gaps in the networks of conductive material within cellulose-based conductive composites. To a first approximation, the cellulosic material might be regarded as a perfect resistor, and much attention has been directed towards electron transport related to the conductive substances. Two mechanisms to account for movement of electricity across such gaps are known as hopping and tunneling (Sheng 1980; Du et al. 2006; Min et al. 2010). Both of these mechanisms entail an activation energy that governs the probability of each unit transport of charge (Hunt 2001). Thus, the amount of current can be expected to follow an Arrhenius relationship,

i = A e-(E/RT) (3)

where A is the pre-exponential factor, E is the activation energy, R is the gas constant, and T is the absolute temperature.

Figure 4 envisions a situation in which a certain energy of activation is required for electrons to “hop” across a gap between conductive surfaces. As shown, the expected population of electrons becomes rarer and rarer within a gap of increasing size (occupied by an insulator or empty space) due to less favorable energy, as predicted by the Boltzmann distribution.

Fig. 4. Schematic illustration of the relationship between potential energy (vertical axis) and the expected population densities of electrons, as influenced by nano-scale gaps in a chain of electrically conductive particles

Kaiser et al. (2009) found that both mechanisms, hopping and tunneling, are favored by strong local electrical fields. Also, as suggested by Eq. 3, the conductance generally increases with increasing temperature. This is expected because of higher amounts of thermal energy, which increases the likelihood that a given electron will acquire sufficient potential energy to surpass the energy barrier.

The hopping concept applies most appropriately when electrically chargeable sites in a composite are small, in a quantum sense (Sheng 1980). Sato et al. (1991) concluded that hopping was the current-limiting factor for conduction within polypyrrole films at high temperature. Based on the observed dependence of current on temperature, Mavinakuli et al. (2010) invoked a hopping concept to account for conductivity within a composite of polypyrrole and silicon carbide. Roldughin and Vysotskii (2000) observed a frequency dependence of conductivity in metal-filled polymer films and proposed that this was consistent with a hopping or tunneling mechanism of transport of charge across gaps.

The tunneling concept, by contrast, pertains especially to systems comprised of relatively long conductive paths separated by short gaps (Sheng 1980). Min et al. (2010) described such gaps as providing either “contact resistance” or a potential barrier. One can imagine waves of electrical current sloshing back and forth within these pathways, such that the local electrical field varies and sometimes is sufficient to jump the gaps at a rate that can be estimated based on the Arrhenius equation.

The concept is illustrated in Fig. 5, using the analogy of a series of well-developed waves approaching a fixed barrier. This perspective is consistent with the frequency-dependent nature of conductance in such systems (Roldughin and Vysotskii 2000; Xia et al. 2017). Whereas hopping mechanisms are strongly associated with higher temperatures, the tunneling mechanism can be important for low to moderate temperature systems (Sheng 1980; Sato et al. 1991; Kaiser et al. 2009). A tunnel junction can be modeled as a capacitor in series with resistors (Sheng 1980), which suggests that the mechanism may be most effective for conduction of alternating current within favorable ranges of frequency.

Fig. 5. Analogy of a well-developed pattern of waves approaching a fixed barrier. In the envisioned example, kinetic energy is converted to sufficient potential energy (higher elevation) as the wave encounters the barrier so that some spill-over occurs.

Toker et al. (2003) studied systems that appeared to follow predictions of percolation theory, and yet the charge transport appeared to involve tunneling. The explanation given was that some of the nearest-neighbor distances were too far apart for tunneling to be effective. Wang et al. (2015c) likewise attempted to combine the concepts of percolation and tunneling-assisted conductance in the case of agglomerated graphene-containing nanocomposites.

Biased Distributions of Particles in Composites

Although conventional percolation models usually assume purely random distributions of conductive particles in a matrix, there are many reasons to expect otherwise in practice. For instance, the presence of repulsive forces between solid surfaces can be expected to have a dispersing effect, thus keeping conductive particles from contacting each other within the resulting composite, depending on the process by which the composite is formed. Attractive forces, on the other hand, can be expected to create a bias towards more agglomerated structures within a composite (Wang and Chatterjee 2003; Martin et al. 2004; Vigolo et al. 2005; Schmidt et al. 2007). For instance, Wang and Chatterjee (2003) suggested that such attractions can be included within simulation routines, leading to predictions of higher frequencies of contacts among rods within a homogeneous matrix. Martin et al. (2004) were able to achieve a bias towards connectivity among carbon nanotubes within an epoxy matrix by initially using charge-charge stabilization to favor a uniform distribution; this was followed by use of elevated temperatures and modest shearing forces to induce contacts favoring completion of electrical conduction paths.

High levels of agglomeration, leading to very non-uniform distributions, can result from the application of shear, especially in the case of carbon nanotubes, which are flexible and can have very high aspect ratio (Schmidt et al. 2007). Shearing of a suspension, even in cases where the fluid phase consists of a melted polymer, can be expected to bring about inter-particle collisions. As illustrated in Fig. 6, if the particles are sufficiently long and flexible, the resulting agglomerates may become entangled. Such effects have been demonstrated is studies of the forming process of wet-laid nonwoven fabrics (Shiffler 1988; Ramasubramanian et al. 2008; Hubbe and Koukoulas 2016). As noted by Du et al. (2005), a high degree of alignment of rod-like particles can be expected to yield low conductivity due to a lack of contact points. Their simulation study showed favorable results associated with a slight degree of alignment compared to a purely random distribution of orientations.

Fig. 6. Representation of process of entanglement of long, flexible suspended particles as a result of hydrodynamic shearing

It is important not to confuse a completely “random” distribution of particles with a completely “uniform” distribution. This point is illustrated in Figs. 7 and 8. In each case, the distribution shown in the left-hand frame was obtained with the help of a random number generator. In the case of Fig. 7, in the right-hand frame for the figure each rod-shaped particle was moved by about two pixels so that its center would be closer to that of its nearest neighbor. Despite the very small shifts involved, the overall effect is a decidedly more agglomerated appearance. Likewise, Fig. 8 shows the consequence of the same degree of shifting of locations, but this time in the direction of tending to fill in gaps in the structure. Again, though each component shift was tiny, the overall effect was a distribution that would appear to offer a higher frequency of connections, at least in the case of a two-dimensional structure filled with rod-like conductors placed in random directions.

Fig. 7. Graphic example of a two-dimensional distribution of equal rods distributed either according to a random number generator or by moving the center of each rod two pixels closer to its nearest neighbor

Fig. 8. Graphic example of a two-dimensional distribution of equal rods distributed either according to a random number generator or by moving the center of each rod two pixels farther away from its nearest neighbor

Though the images portrayed in Figs. 7 and 8 might tend to make one think that a more uniform dispersion of rod-like particles would imply more connections among the rods and hence high overall electrical conductivity in the case of conductive rods, it is important to keep in mind that most composites are inherently three-dimensional. Thus, what may appear to be connections in a two-dimensional view might be near misses, far misses, or possibly connections in the case of a realistic composite structure having three dimensions.

Hypotheses to be Considered

In light of the theoretical issues just discussed, there are two sets of hypotheses that appear worthy of special focus in this article. The first can be called a “weakest link” hypothesis. It is proposed that the overall conductivity within a typical cellulose-based composite will be governed to a large extent by defects and gaps in its structure. The second can be called an “optimized agglomeration” hypothesis. Here is it proposed that it is often an advantage if the conductive particles have a greater amount of interconnections than would be predicted for a randomly dispersed mixture. On the other hand, a severe level of agglomeration would be unfavorable for electrical connectivity throughout the whole volume of a structure.

Weakest link hypothesis

In light of the relatively high electrical conductivity of the substances to be considered in the next section, contrasting with the low conductivity of the cellulosic material, it is proposed here that the major determinants of composite conductivity will be related to defects, restrictions, and any other kind of gaps in the network of conductive material. A form of this hypothesis was proposed by Kaiser et al. (2009), who studied graphene monolayers and accounted for the net resistances based on Schottky barriers of conduction between adjacent flakes of graphene oxide.

A corollary to the first hypothesis is that the amount of conductive material needs to be at least great enough to provide a theoretical connectivity, i.e. probably greater than the predicted percolation threshold (except that the next hypothesis may call that requirement into question). Another corollary is that, given the much greater conductivity within conductive substances compared to the gaps or the cellulosic material, it makes sense to employ relatively long conductive particles, such as nanowires, nanotubes, etc. Indeed, many studies have shown advantages of using high aspect ratio conductive particles (Wang and Chatterjee 2003; Bryning et al. 2005; Li et al. 2007; Rosca and Hoa 2009; Bao et al. 2013; Vo et al. 2015). Yet a third corollary, which may require future research to validate, is that improving the conductivity of the cellulosic phase may be a promising approach toward achieving higher overall conductivity. As noted by Agate et al. (2018), cellulosic material can be made more conductive by suitable doping, as in the case of treatment with iodine or ammonium carbonate or acetate. Such an approach can contribute to conductivity by a proton-hopping mechanism (Rani et al. 2014).

Optimized agglomeration hypothesis

The second hypothesis that appears worthy of examination is that in order to achieve the most favorable conductivity results in a cellulose-based composites there needs to be an optimized level of agglomeration, i.e. not a uniformly dispersed system, but also not an over-agglomerated system. It is proposed that, by manipulating factors affecting particle contacts, one can achieve an optimal level of agglomeration of conductive elements within a composite. It is further proposed that the electrical conductivity in such a system can exceed the predictions of percolation theories based on randomly distributed particles.

Published articles have reported both negative and positive effects of agglomeration (or non-uniform distributions) of conductive particles on the conductivity of composites. The following authors observed negative effects of agglomeration on conductivity (Schmidt et al. 2007; Liao et al. 2008; Bao et al. 2013; Yang et al. 2014). For instance, Schmidt et al. (2007) observed increasing agglomeration of carbon nanotubes within a thermoplastic matrix with increasing time of shearing. Bao et al. (2013) explicitly included tangled agglomerates of carbon nanotubes in their excluded volume modeling and predicted a large effect of the dispersion state and the percolation threshold. By contrast, the following articles report positive relationships between the extent of agglomeration of conductive particles and conductivity of the composites (Martin et al. 2004; Bryning et al. 2005; Du et al. 2006; Hu et al. 2006; Li et al. 2007; Aguilar et al. 2010; Wang et al. 2015c). Thus, Bryning et al. (2005) achieved a very low conductivity threshold, and they attributed this to not only a high aspect ratio of the conductive elements, but also to a bias favoring contacts among them. The lowest percolation thresholds were observed when the carbon nanotubes were given an opportunity to reagglomerate.

Excessive agglomeration of conductive particles seems to be a likely explanation of some counterintuitive findings reported by Liao et al. (2008). Those authors observed that conductance values of composites leveled off and even decreased with increasing content of carbon nanotubes in a polypropylene matrix. In fact, the authors noticed “serious aggregation.” The relationship between increasing content of agitated fibers and increasing mechanically induced flocculation leading to the formation of clusters is well known in papermaking science (Kerekes and Schell 1992; Hubbe 2007).

It is clear that hydrodynamic shear can promote agglomeration of elongated particles in suspension. Schueler et al. (1997) reported that “light shearing” of carbon black particles dispersed in epoxy resin led to increased conductivity, which was attributed to optimal agglomeration. Vigolo et al. (2005) experimented with “sticky nanotubes.” They found that conditions favoring agglomeration gave a lower percolation threshold for electrical conductivity.

Tools to establish favorable contacts

Given the findings, as noted above, that increased agglomeration of conductive particles can both hurt and promote higher electrical conductance of composites, it follows that the details of agglomeration must be important. Dense entanglement of conductive particles, leaving most of the composite devoid of the particles, is to be avoided. On the other hand, it would be a great advantage to bias the system in favor of establishing contacts between fibers or chains of conductive particles that span the dimensions of the specimen. Some tools and principles that might be employed to bring about such contacts are included in Table 1, along with some selected references.

Table 1. Tools and Principles to Encourage Favorable Contacts to Promote Electrical Conductance in Composites

The last item listed in Table 1 merits further discussion. The term self-assembly, if it is favorable, implies that the particles may be able to organize themselves at a nano-scale level such as to achieve more efficient electrical conduction paths (Bryning et al. 2005). Conditions under which this is possible will include (a) the system must remain fluid-like long enough to allow significant diffusion of the conductive particles to take place, (b) the viscosity must be low enough for diffusion to take place (Schmidt et al. 2007), and (c) there needs to be a net attraction, e.g. van der Waals forces drawing the surfaces of the conducting particles towards each other.



Metal, as a highly conductive material, is widely applied in many electronic devices, for example, solar panels and touch screens. Positive features of metals include their generally outstanding electrical conductance, modest costs (in most cases), and ability to be formed into wires and nanowires (Elechiguerra et al. 2005; Hu et al. 2009; Song et al. 2015; Su et al. 2017; Lv et al. 2018). For example, Hu et al. (2009) employed silver wires (nanowires) so small as to be transparent. These were formed into nanopaper in combination with nanofibrillated cellulose, leading to conductive sheets exhibiting high flexibility. Su et al. (2017) carried such an approach an additional step further through the use of a conductive polymer (polydopamine) in combination with silver nanowires and nanocellulose. Lv et al. (2017) used a filtration method to apply a suspension of silver nanowires to a pre-formed mat of bacterial cellulose.

Metal can continue to function even after being exposed to high forces (Kawaguchi et al. 2017). In addition, the oxides of metals can work as conductive agents. For example, indium tin oxide has a very high transparency and it is applied in monitors and photoelectronic materials (Hasan Khondoker et al. 2012). The challenge of using metals in conductive materials is that they are hard to recycle. Elechiguerra et al. (2005) observed durability problems, which were attributed to the oxidation of silver nanowires. Additionally, the supply of certain metals requires intensive mining, high energy input, and the sources are limited.

Compared to pure metallic substrates, the use of cellulose in metal-containing composites can provide advantages with respect to cost, flexibility, and degradability. Many studies have been done to prepare conductive materials by adding metals into a nanocellulose matrix and to investigate the electrochemical properties. Table 2 provides highlights from a selection of relevant publications.

Table 2. Listing of Studies in Combining Metal Particles and Cellulosic Material in Electrically Conductive Composites or Films

Five of the articles listed in Table 2 stand out as examples in which metal particles were used in combination with conductive polymers to impart electrical conductance to cellulose-based composites (Sarma et al. 2002; Zhang et al. 2014b; Ummartyotin and Manuspiya 2015; Su et al. 2017; Wang et al. 2018a). Such approaches are consistent with the hypothesis proposed earlier, that conductance performance may be dominated by issues of poor connections among the conductive particles. In the cited cases the conductive polymers may serve as bridges between the more highly conductive metal particles or nanowires.

The bulk conductance of electricity within metals is due to the relatively free mobility of electrons within conduction bands (Mizutani 2001). As illustrated in Fig. 9, certain electrons are essentially shared among the metal as a whole rather than staying bound to atomic orbitals. As a consequence, these electrons respond freely to applied fields and can participate in bulk-scale currents.

Fig. 9. Representation of a conduction band, occupied by highly mobile electrons, within a metal

Resistance within the bulk phase of metals can be attributed, in part, to static imperfections and lattice vibrations, leading to scattering phenomena (Dugdale 2016). As noted by Greenwood (1958), the electrons may interact with scattering centers consisting of irregularities in the metal crystalline structure. Such scattering contributes to some electrical resistance even within the metal itself. However, due to the very high conductivity within metals relative to many other materials that may be present, the gaps between metal structures can be expected to have a greater contribution of resistivity of metal-containing composites than any features within the metals themselves.

Oxidation and corrosion can compromise the durability and performance of certain metals, especially when very thin, as in the case of nanowires. Elchiguerra et al. (2005) reported such problems in the case of silver nanowires.

Through magnetron sputtering, Hu et al. (2013) prepared a cellulosic conductive material by depositing indium oxide on the nanofibrillated cellulose. The soft conductive nanocellulose material, like a paper film, was found to achieve a 65% transmittance at the 550 nm wavelength of light with a low resistance of 12 Ω/sq. The conductivity of the paper remained constant after multiple twists. Using the magnetron sputtering technique, Yang et al. (2009) deposited ZnO as the conductive agent within in a nanocellulose paper with a thickness of 250 nm. The resistance of the material was as low as 4.62×10-4 Ω•cm. The average transmittance of the visible light was reported to be 93.7%.

As a highly conductive agent, silver has been applied in different studies for the preparation of cellulose-based conductive materials. Meulendijks et al. (2017) treated cellulose with oxidizing agents TEMPO and NaBr. The oxidized cellulose was mixed with SnCl2 and HCl and then centrifuged to obtain the cellulose matrix. After a silver deposition bath treatment, the cellulose-silver cellulosic material had a conductivity as high as 2.9×104 S/cm. Wang et al. (2018a) modified the surface of a cellulose substrate with dopamine via self-polymerization under weak base conditions. The materials were merged into the acetone and treated with UV light to obtain a porous structure. Once silver was deposited onto the porous structure, the composite was found to achieve a low resistance of only 0.2 mΩ•cm.

Recently there has been a lot of focus on using silver nanowire as the conductive agent in preparing cellulose-based conductive film, achieving high flexibility. Song et al. (2015) prepared a kind of nano-paper by integrating bamboo nanocellulose and silver nanowires using hydroxypropyl methylcellulose as the binding agent. With a thickness of 4.5 μm, the nano-paper had a conductivity of 500 S/cm. Surface modification of cellulose, such as oxidation with TEMPO, has the potential to improve the bonding strength between matrix and conductive agents significantly. Zhang et al. (2018) blended TEMPO-oxidized nanocellulose substrate with 0.2 wt% chitosan and 0.3 wt% nanocellulose. After magnetic field blending, pressure filtering, and low-temperature drying, they obtained a piece of 7.8 μm thick conductive material having a resistivity of 4.32 Ω/m2. It was suggested that chitosan could substantially increase the mechanical properties of the conducive material. When the cellulose:chitosan ratio was 7:3, the maximum strain of the film increased 112% compared to the control. Additionally, coating also has been used in preparing conductive nanocellulose-based material. Su et al. (2017) tested the coating of polydopamine on a nanocellulose surface, and then they attached silver nanowires onto the polydopamine. The multi-layer cellulosic product had a transparency of 90.9% of 550 nm light and a film resistance of 14.2 Ω/sq.

In addition to silver, other metals with high conductance can also be used to prepare cellulosic conductive materials. For example, Ueno et al. (2015) used the Kanigen method and nickel coating on recycled newspaper pulp. The conductive material had a conductivity of 0.28 S/m (0.0028 S/cm). On the other hand, cellulosic board containing 40% of such cellulosic conductive materials was reported to block an electric field of 30 dB at 1000 MHz. Root et al. (2018) treated the cellulosic substrate with copper coating. To be more specific, the cellulose matrix was treated with CuSOsolution, and the concentration of CuSO4 was 0.0028 mol. The film resistance of the conductive cellulosic materials was affected by the degree of copper electroplating, and the resistance was at the range between 16.5 and 369.3 Ω/sq. In other work, Cu was deposited on a cellulose surface, yielding a conductive material with high flexibility (Sakurai et al. 2018).

Conductive Polymers

Conductive polymers are typically macromolecules having highly conjugated π bonds (Svennersten et al. 2011). Because of the special electrochemical and photonic properties, low density, and corrosion resistance, some of the conductive polymers (e.g. polypyrrole and polyaniline) have been used to make organic light-emitting diodes (OLED) and sensors (Niu et al. 2016). The main challenge facing the wider application of conductive polymers in electronic devices is that they are difficult to manufacture. Cellulose, by contrast, has strong mechanical properties and is easy to form into films. It is proposed that mixing the conductive polymers with cellulose could ensure satisfactory mechanical quality. The supporting cellulose structure then could allow the conductive polymers to be in a form that is easy to manufacture. Many studies have been carried out in applying conductive polymers as conductive agent in preparing nanocellulose-based conductive materials (Wistrand et al. 2007). Table 3 lists various studies in which polymers such as polypyrrole and polyaniline were used in combination with cellulosic materials to produce electrically conductive composites.

Table 3. Listing of Studies in Combining Conductive Polymers and Cellulosic Material in Electrically Conductive Composites or Films


Important progress has been achieved in understanding the electrical conductivity of derivatives of polypyrrole, the structure of which is shown in Fig. 10 (Scott et al. 1984).

Fig. 10. Molecular structure of polypyrrole: (a) ordinary polypyrrole; (b) containing a polaron feature; (c) containing a bipolaron feature

In addition to showing a representation of pure polypyrrole, the figure also gives an example of a polaron and a bipolaron, which both contribute to electrical conductance (Brédas et al. 1984; Scott et al. 1984; Sato et al. 1991; Jenden et al. 1993). Polaron features contribute a single radical cation to the polypyrrole structure, whereas bipolarons contribute dual-cation features. Note that the “dot” shown in part B of the figure indicates the presence of an unpaired electron. In either case, a key role is played by ionized groups, including both electrons and positive sites, the latter of which can be regarded as “holes” where an electron is missing.

Brédas et al. (1984) reported that low oxidation levels of polypyrrole lead to the formation of polarons, whereas higher degrees of oxidation lead to a preponderance of biopolarons, which are associated with a high contribution to conductance. These features were confirmed by Jenden et al. (1993), who used FT-Raman spectrographic evidence.

Oxidized compounds related to polypyrrole were further studied by Hegewald et al. (2009), who found the best conductivity results after oxidation in a ferric chloride solution. The effects of oxidation are complex, however, since higher levels of oxidation increase the likelihood of defects in the conductive structure.

As noted by Roldughin and Vysotskii (2000), composites that contain conducting polymers tend to be highly disordered systems, in which there is likely to be a high frequency of defects that must be surmounted by quantum-mechanical tunneling and hopping mechanisms. According to Sato et al. (1991), the conduction at lower temperatures is mainly due to tunneling, as discussed earlier. Taking a related approach, Wang et al. (2018b) found that reductive treatment of graphene oxide decreased the frequency of oxygen-related defects in the structure and therefore yielded easier transport of electric current. Perhaps associated with such redox transitions, one of the negative features of polypyrrole-based conductive materials is poor durability, which may be due either to chemical instability or to effects of swelling and shrinkage in the course of charging and discharging cycles (Shi et al. 2015).

In 1977, in work that later led to a Nobel prize, it was found that the conductance of certain organic polymers can be increased by doping with electrolytic substances (Shirakawa et al. 1977). Doping of the polypyrrole films with p-toluene sulfonate and dodecyl sulfate contributes to conductivity (Jenden et al. 1993; Ramelow et al. 2001; Mavinakuli et al. 2010), which is dependent on charged sites. The sensitivity of conductance to doping treatment was studied further by Ruangchuay et al. (2004), who used such effects as the basis for a unique sensing system for acetone vapor. Tezuka et al. (1995) studied the reverse problem of “undoping,” including the electrochemical reduction of oxidized polypyrrole.

Polypyrrole applications in cellulose composites

Polypyrrole has been widely used in cellulose-based conductive materials (Mihranyan et al. 2008). Usually the polymerization process of pyrrole monomers has involved the use of Fe3+ as an oxidative agent in an acidic environment (Fedorkova et al. 2010; Lei et al. 2013). Nyström et al. (2010) homogenized 2 wt% cellulose hydrogel with pyrrole monomer in aqueous solution and added FeCl3 solution and 37 wt% HCl to prepare the polypyrrole. The surface-coating of cellulosic matrix was completed with polypyrrole by in-situ polymerization. The surface density of the conductive material was 0.011g/m2 with a conductivity of 1.5 S/m (0.0015 S/cm). Lay et al. (2016) fabricated polypyrrole/cellulose composite by in-situ polymerization. The cellulosic substrate was bleached softwood pulp that was oxidized with FeCl3. It was found that when the content of polypyrrole reached 20%, the conductivity of the composites was 5.2 S/m (0.052 S/cm).

The preparation methods affect the microstructure, mechanical strength, and conductivity of polypyrrole/cellulose composites (Wang et al. 2015b). Saravanan et al. (2017) improved the conductivity of polypyrrole/cellulose composite by adding a mixture of styrene-triazine sulfonate sodium salt and 9,10-anthraquinone-2-sulfonate sodium salt. The results showed a conductivity ranging from 3 KΩ/sq to 26 KΩ/sq with different contents of the mixture of styrene-triazine sulfonate sodium salt and 9,10-anthraquinone-2-sulfonate sodium salt. Other oxidizing agents can be added to change the material structure and to improve the performance of conductive polymer. For example, TEMPO oxidizes the hydroxyl groups of cellulose into carboxyl groups. Such treatment was applied to promote the formation of hydrogen bonds between cellulose and pyrrole, thus improving the stability of the composites (Wu et al. 2014a). Wu et al. (2014b) preprocessed the cellulose by attaching a layer of N-vinylpyrrolidone. This layer of polymer helped the attachment of polypyrrole on the surface of cellulose material, resulting in a conductivity of 36.9 S/cm. Shi et al. (2014) introduced supercritical CO2 as a drying method to process the polypyrrole/cellulose composite from hydrogel into an aerogel. The final material had a low density of 0.41 to 0.53 g/cm-3 and a conductivity of 0.08 S/cm.

Cellulosic materials that are in non-fibrous form can also be used as the base material and to integrate with polypyrrole. Zhang et al. (2017) treated filter paper as the matrix and melted the paraffin-based ink on the surface of the filter paper. Then the ink formed several channels into the filter paper, which were conductive. With the conductive channels, the polymerization of pyrrole monomers was able to be improved, and finally to form a kind of conductive network inside the cellulosic substrate. Several studies have shown that pH is another reason that cellulose/polypyrrole conductive composite changes conductivity. For example, Li et al. (2010) found that the conductivity of cellulose/polypyrrole composite decreased under basic conditions but increased under acidic conditions.

Fig. 11. Schematic of the reactions between cellulose oxidized with TEMPO and pyrrole monomers. Figure based on one shown by Lay et al. (2016), redrawn.


Polyaniline, another widely used conductive polymer, has been applied in conductive composites such as electrodes, biosensors, and antistatic coatings (Mi et al. 2008). The structure of polyaniline is shown in Fig. 12, which is based on Alonso et al. (2018). In the figure, the term “X” indicates a negative ion, such as a chloride ion. Most of the polyaniline has been applied on non-renewable materials, including synthetic rubber, plastics, and fabrics. Other researchers have evaluated cellulose as a potential substitute for petroleum-based base materials (Petersson and Oksman 2006).

Fig. 12. Molecular structure of polyaniline, in its electrically conductive form (figure based on Alonso et al. 2018)

The electrical properties of cellulose/polyaniline composites are greatly affected by the content of polyaniline. Luong et al. (2013) prepared nanocellulose/polyaniline composites by in-situpolymerization and obtained flexible conductive film through vacuum filtration. It was found that when the content of polyaniline was over 4.75%, the conductivity reached 2.6×10-5 S/cm, indicating potential applications such as soft electrodes and antistatic coatings. Through in situ polymerization of aniline within bleached sulfate pulp, Sharifi et al. (2018) treated the kraft pulp as the raw materials and obtained a cellulose/polyaniline film with a density of 60 g/m2. Results obtained with films dried under 800 KPa and 65 °C for 60 min indicated that both polymerization time and the mass ratio of cellulose/aniline affected the conductivity of cellulose/polyaniline films. A maximal conductivity (1.49 S/m or 0.0149 S/cm) was observed after 8 hours of polymerization when the cellulose/polyaniline mass ratio was 1:2. Mo et al. (2009) investigated a chemical oxidation method to prepare the cellulose/polyaniline composite. In their study, cellulose was first mixed with a HCl solution containing aniline for activation; then 1 mol/L ammonium persulfate was added into the mixture for the polymerization of aniline monomers. It was found that the conductivity of cellulose/polyaniline composite ranged from 1.0×10-5 S/cm to 3.0×10-2 S/cm and was closely related to cellulose activation time and aniline dosage during preparation.

Cellulosic materials from different sources have been applied to prepare the cellulose/polyaniline conductive composites. For example, sodium carboxymethyl cellulose and polyaniline will form into hydrogels with semi-interpenetrating polymer network structure using glycerol diglycidyl ether as the crosslinking agent (Li et al. 2017). With higher dosages of sodium carboxymethyl cellulose and glycerol diglycidyl ether, the conductivity of the composite hydrogels increased first and then decreased. Bacterial cellulose was studied by Alonso et al.(2018) in preparing a cellulose/polyaniline conductive film. A conductivity of 1.4×10-1 S/cm was reported.

Other conductive polymers

Other than polypyrrole and polyaniline, polybenzazole is also used in making conductive composites. Polybenzazole has higher heat stability and redox activity than polypyrrole and polyaniline. Cai et al. (2018) prepared a bacterial cellulose/polybenzazole composite by electrospinning. The composite achieved a conductivity of 4.6×10-2 S/cm, which was significantly higher than the control with only bacterial cellulose (1.5×10-8 S/cm).

In addition, polythiophene also has been used as the conductive agent in preparing cellulose-based composites. Lay et al. (2017) prepared cellulose nanofibers by introducing poly-3,4-ethylenedioxythiophene and polystyrene sulfonate as conductive agents. The electrical conductivity and the specific capacitance of the nanopapers were 2.58 S/cm and 6.21 F/g, respectively. Hafez et al. (2017) fabricated composite films by adding wood microfibers to poly(3,4-ethylenedioxythiophene)/poly(styrene sulfonate). The mechanical strength of the composite was improved with higher fiber content. At a wood fiber content of 75%, the resistivity of the composites was the lowest (340 Ω/sq). Zhao et al. (2017) assembled cellulose and poly(3,4-ethylenedioxythiophene)/poly(styrene sulfonate) in an ionic liquid and fabricated a flexible and conductive composite matrix. By incorporating multiwalled carbon nanotubes, the composite matrix could achieve higher flexibility and excellent electrochemical properties, including a low resistance of 0.45 Ω and a high specific capacitance of 485 F/g at 1 A/g.

Carbon-based Conductive Agents

The two widely-studied forms of carbon having high promise for imparting electrical conductivity to composites are carbon nanotubes and graphene. These materials, in their pure form, share a structural similarity. The structure can be described as either tubular or flat layers of hybridized sp2 carbons (Zhang et al. 2013a; Saeed and Ibrahim 2013; Terzopoulou et al. 2015). The molecular structure of graphene and an example of a graphene oxide are illustrated in Fig. 13, which is based on structures shown by Chua and Pumera (2014).

Fig. 13. Molecular structures of graphene (left) and a graphene oxide (right)

Kaiser et al. (2009) described the mechanism of current flow as two-dimensional variable-range hopping in parallel with tunneling driven by the electrical field. Their research suggested that highly conductive graphene regions were joined by disordered regions, across which the current flow was wholly dependent on charge-carrier hopping and tunneling, which was sensitive to local electric fields. A much-cited article by Bockrath et al. (1999) shows the current within carbon nanotubes rising steadily with increasing applied voltage, but only after exceeding a limiting value of voltage. Dependencies on temperature and applied voltage were consistent with a tunneling mechanism. Within the conducting nanotubes the current flow was consistent with the concept of a Luttinger liquid, describing transport of electricity in one dimension. Yao et al. (2000) studied contributions to electrical resistivity within carbon nanotubes at high electrical field strength and concluded that a main contribution to resistance is scattering of electrons associated with emission of light or phonons. However, the major impediments to current flow were associated with poor transfer of current at the ends of nanotubes, even when the nanotubes are directly connected with metal conductors at those ends.

Carbon nanotubes

Studies in which carbon nanotubes were incorporated into electrically conductive composites with cellulosic material are listed in Table 4.

Table 4. Listing of Studies in Combining Carbon Nanotubes and Cellulosic Material in Electrically Conductive Composites or Films

Notes: CNT = Carbon nanotubes, unspecified; SWCNT = Single walled carbon nanotubes; MWCNT = multi-walled carbon nanotubes

Carbon-based conductive agents have very high conductivity, usually higher than that of polymer conductive agents. Carbon nanotubes (CNTs) are allotropes of carbon with a cylindrical nanostructure. There are mainly two types of carbon nanotubes, Single-Wall CNTs (SWCNTs) and Multi-Wall CNTs (MWCNTs) (Lijima 1991). Besides excellent conductivity, CNTs also have outstanding physical properties. A study by Salvetat et al. (1999) has shown that the Young’s moduli and bending strength of CNTs can be approximately 1 TPa and 100 GPa, respectively, and the conductivity and current capacity can be 10S/m (10S/cm) and 10A/cm2, respectively. CNTs are stable and retain their electrical properties when stretched or bent. As a result of these characteristics, the surface-modified CNTs have the promising potential to improve the conductivity of chemical and biological polymers with the formation of a conductive network.

A negative aspect of individual carbon nanotubes is their high variability in performance, which can be attributed to differences in chirality and molecular structure (Hu et al. 2004). Some tubes act as metal-like conductors, whereas others behave as semiconductors (Bockrath et al. 1999). However, as shown by Hu et al. (2004), reliable and promising performance can be achieved by the use of networks of carbon nanotubes, wherein the performance is dictated by the ensemble averaging of properties of the individual tubes.

Developers of technology related to carbon nanotubes can choose between single-walled and multi-walled tube products, as well as mixtures thereof. According to Min et al. (2010), in the case of MWCNTs the current is carried only by the outermost wall. Though this might be regarded as a penalty in terms of efficiency, one also needs to take into consideration such factors as the length and stiffness of the tubes, which may be key to performance in different applications. Rosca and Hoa (2009) used MWCNTs in epoxy composites to produce “bulky paper” sheets.

Hamedi et al. (2014) dispersed 43 wt% of CNTs directly into the nanocellulose. Then the conductive fiber and nanopaper were made through molecular self-assembly of cellulose. Lee et al. (2016) described a method to mix 0.5 to 30.0% of multi-wall CNTs with 5 wt% of cellulose and prepare the conductive fiber through solvent-based solution spinning. The reported conductivity can be as high as 2.7 S/cm when there is 30 wt% of multi-wall CNTs. Cellulose-based aerogel was also introduced as conductive material. Qi et al. (2013) prepared an aerogel with a mixture of cellulose and CNTs by the flash freezing/lyophilization process. Through adjusting the content of cellulose and CNTs, the Young’s modulus could be tuned to reach about 90 MPa with a conductivity of 2.2×10-2 S/cm.

During the preparation of conductive material, surface modification of cellulose has the potential to improve the compatibility and the uniformity of dispersion of CNTs in the modified matrix. Koga et al. (2013) fabricated transparent, conductive, and printable composites by mixing TEMPO-oxidized nanocellulose and CNTs, indicating a substantial improvement of CNTs dispersion in the TEMPO-oxidized nanocellulose. In addition to TEMPO-oxidization, surface reactant was also found to contribute to the dispersion of CNTs. Huang et al. (2015) reported that CNTs were uniformly dispersed in the presence of cetyltrimethylammonium bromide in a NaOH/urea aqueous solution. The compressed nanocomposite film reached a relatively high electrical conductivity (7.2 S/m or 0.072 S/cm) when there was 5 wt% CNTs. Imai et al. (2010) found that CNTs disperse more uniformly in the cellulose material previously mixed with anionic surfactant, resulting in a high conductivity of 672 S/m (6.72 S/cm). Oppositely, Yoon et al. (2006) dispersed multi-wall CNTs into the cationic cetyltrimethylammonium bromide first and dipped the bacterial cellulose into the CNTs/cetyltrimethylammonium bromide solution. After extraction and drying, the conductivity of the CNTs-incorporated cellulose was 1.4×10-1 S/cm at a multi-wall CNTs content of 9.6 wt %.


Graphene, with its high performance in conductivity and mechanics, has recently attracted great interest in the applications of soft conductive materials and supercapacitors (Zhu et al. 2011; Wang et al. 2014). Two-dimensional nano-scale graphene has extremely high surface-area-to-volume ratio and flexibility. However, the strong π-π interaction among graphene particles promotes coalescence, decreasing the surface-area-to-volume ratio (Zhang et al. 2012). The hydrophilicity and nanostructure of nanocellulose not only can promote the dispersion of nano graphene, but it also can improve the orientation of polymer and enhance the mechanical property of composites. Studies in which graphene or graphene oxide-type materials were incorporated into electrically conductive composites with cellulosic material are listed in Table 5.

Table 5. Listing of Studies in Combining Graphene or Graphene Oxide-related Substances and Cellulosic Material in Electrically Conductive Composites or Films