1. Introduction
Cellulose molecules are biosynthesized by enzymes, deposited in a continuous fashion, and aggregated to form microfibrils, which are long threadlike bundles of molecules stabilized laterally by intermolecular hydrogen bonds and hydrophobic interactions. Thus, the molecular chains in a single microfibril can be elongated during biosynthesis, and are therefore remarkably uniform in size and shape. This extended chain conformation and fibrillar morphology result in a significant load-carrying capability. Depending on their biological origin, the diameters of microfibril range from about 2 to 20 nm, and their length can reach several tens of microns. As they are devoid of chain folding and contain only a small number of defects, each microfibril can be considered as a string of cellulose crystals, linked along the microfibril by amorphous domains. Microfibril has a modulus close to that of a perfect crystal of native cellulose (estimated to be around 150 GPa), and strength in the order of 10 GPa (Abe et al. 2007; Habibi et al. 2010; Iwamoto et al. 2007; Klemm et al. 2011). In addition, microfibril has piezoelectric properties equivalent to quartz, and can be manipulated to produce photonic structures (Postek et al. 2013). This microfibril is so-called cellulose nanofibril (CNF).
CNF can be found in the structure of the cell wall of all plants, as well as some fungi, animals, and bacteria (Saxena and brown 2005). Particularly, in the cell wall of lignocellulosic biomass, CNF aggregates further to form cellulose fibers due to its self-assembly properties, and becomes embedded in a polymer matrix composed of hemicellulose and lignin. Therefore, lignocellulosic biomass is itself a composite material with a complex cascading hierarchical structure. Various techniques have been used to isolate CNF from lignocellulosic biomass, by disrupting and extracting matrix polymers from the cell wall structure through mechanical and chemical treatments, acid hydrolysis, and enzymatic treatments. Cellulose nanocrystals (CNCs), which are rod-like cellulose nanoparticles with a high crystallinity, can be prepared by the further acid or enzymatic hydrolysis of the amorphous region of cellulosic materials or CNFs.
Several review papers about the preparation and potential applications of CNF and CNC have already been published, and are briefly summarized as follows. Siró and Plackett (2010) summarized the lignocellulose feedstocks for nanocellulose preparation, and its progress with a particular focus on CNF and its application for the fabrication of bio-nanocomposites. Mechanical defibrillation such as refining, high-pressure homogenization, grinding, and cryocrushing, with and without chemical and enzymatic pretreatments such as alkaline, oxidation, and cellulase treatments, were introduced for the production of CNF from various lignocellulosic feedstocks such as potato, soybean stock, wheat straw, hemp fiber, sugar beet, kraft, and sulfite wood pulp. They also summarized the morphological characteristics of CNF obtained from different feedstocks, and the mechanical, optical, and barrier properties of CNF-based films. In addition, they also introduced the preparation and characterization of nanocomposites with different polymer matrices such as polylactide, polyethylene, polypropylene, polycaprolactone, poly(styrene-co-butyl acrylate), ethylene vinyl alcohol copolymers, polyurethanes, polyvinyl alcohol, starch, amylopectine, polyethylene oxide, and chitosan. The modification of CNF was also introduced, including its acetylation, silylation, application of coupling agents, and grafting. Kalia et al. (2014) also summarized several nanofibrillation technologies, such as high-pressure homogenizer, microfluidization, grinding, and ultrasonification, for CNF in terms of their advantages and disadvantages. Pretreatments to facilitate the release of CNF from the cell wall structure were introduced, focusing on mechanical refining, TEMPO-mediated oxidation, enzymatic pretreatments, steam explosion, and delignification using NaClO2/acetic acid. Habibi et al. (2010) reviewed the chemistry, self-assembly, and applications of CNC. They specially focused on the chemical modifications of CNC, including noncovalent surface chemical modifications, (2,2,6,6-Tetramethylpiperidin-1-yl)oxyl (TEMPO)-mediated oxidation, cationization, esterification, silylation, and polymer grafting. In addition, they introduced the applications of CNC in nanocomposite materials, focusing on nanocomposite processing such as casting-evaporation and sol-gel processing. Charreau et al. (2013) summarized the comprehensive review of patent trends on nanocellulose, with the aim to provide researchers with patent information that may help them visualize the evolution of nanocellulose technology.
In this paper, we present a literature review of previously published research and review papers on the preparation methods of nanocellulose, targeting the research activities of our interest.
2. Scale-up production facilities of nanocellulose
Since 2014, TAPPI (Technical Association of the Pulp and Paper Industry) has reported the commercial entities for the mass production of nanocellulose at the TAPPI Nano Conferences (www. tappinano.org). Since Innventia opened the world’s first pilot plant for the production of CNC in 2011, with a capacity of 100 kg/day, many companies and institutes have started to commercialize nanocellulose for market development. In 2014, Borregaard and Inventia started 1,000 kg/day and 100 kg/day demonstration plants, respectively, for CNF production. In 2015, Paperlogic announced a facility for CNF production with a capacity of 2,000 kg/day. Other facilities at the University of Maine (1,000 kg/day), American Process (500 kg/day), Nippon Paper (150 kg/day), Oji paper (100 kg/day), and Innventia (100 kg/ day) were currently reported as commercial scale plants. For CNC production, CelluForce started with a 1,000 kg/day demonstration plant in 2012. Currently, American Process (500 kg/day), Holmen (100 kg/day), Alberta Innovates (20 kg/day), US Forest Products Lab. (10 kg/day), Blue Goose Biorefineries (10 kg/day), and FPInnovations (3 kg/day), etc. have reported commercial scale plants for CNC.
In particular, Japanese Economy, the Trade and Industry Ministry, and the National Institute of Advanced Industrial Science and Technology organized a Nanocellulose Forum in 2014 (https://unit.aist.go.jp/rpd-mc/ncf/eng/index.html). In this consortium, more than 176 companies were involved in the exploration of nanocellulose applications, among them Nippon Paper Industries, Oji Holdings Corp., Toyota Auto Body Co., Mitsubishi Motors Corp., Mitsui Chemicals Inc., and Daicel Corp., Kao Corp., etc.
3. Production of nanocellulose
3.1 Mechanical defibrillation for CNF production
High-pressure homogenizers (HPH), disk-mills (DM), ultrasonicators, ball-mills, etc., have been introduced as devices for mechanical defibrillation to prepare CNF (Bhatnagar and Sain 2005; Chen et al. 2011; Chun et al. 2011; Lee et al. 2010; Kalia et al. 2014; Zhang et al. 2015). A high-pressure homogenizer is a device that efficiently converts liquid fluid pressure into shear and impact forces for defibrillation, particle size reduction and dispersion, deagglomeration, emulsification, cell disruption, etc. For CNF production, a combination of crushing, shearing, and cavitational forces due to hitting materials among themselves, and extreme velocity changes in the material stream, mainly works to defibrillate cellulosic materials into the nanoscale. (Sun-Young Lee et al. 2009; Siró and Plackett 2010; Klemm et al. 2011). Turbak et al. (1983) are known as the first researchers to introduce HPH for the production of CNF from wood pulp, with the diameter and length of the obtained CNF between 10-100 nm and in the micro-scale, respectively. Ren et al. (2014) reported the effect of chemical and HPH treatment conditions on the morphology of cellulose nanoparticles. CNF and CNC with different morphologies were prepared from microcrystalline cellulose (MCC) through combined acid hydrolysis and HPH treatments. The dimensions of the CNC and CNF were reduced by increasing the acid-to-MCC ratio and number of HPH. Davoudpour et al. (2015) optimized the high-pressure homogenization parameters for CNF production using response surface methodology (RSM). RSM was used to determine the effects of pressure, and the number of cycles in HPH on the isolated yield, crystallinity, and diameter of kenaf bast-CNF. They reported the optimized experimental conditions to be a pressure of 56 MPa, 44 cycles, and a 0.1 wt% fiber suspension concentration, yielding an 89.9% CNF with a crystallinity of 56.5%, and a diameter of 8 nm.
A DM can be used to prepare nanoscale fibers or particles by grinding, cutting, defibrillating, pulverizing, and rubbing the materials at an adjustable clearance between the upper and lower disks. The DM method has been widely used for CNF production. Generally, the concentration of the feedstock suspension for efficient DM is known to be about 1-2%, which is higher than that for high-pressure homogenizing (0.3-0.5%). In addition, the energy consumption per one cycle operation of DM is known to be 620 kJ/kg, which is remarkably lower than the 3,940 kJ/kg needed for high-pressure homogenizing (Spence et al. 2011). Another advantage of DM is its availability for mass production. The combination of DM and HPH treatments has been tried for CNF production. For example, Yano’s research group produced CNF with various degrees of fibrillation from kraft pulp using several DM passes, and subsequently passing it through HPH. (Iwamoto et al. 2005; Nakagaito and Yano 2004; Nakagaito and Yano 2005; Yano et al. 2005) In more detail, a 3% concentration pulp fiber slurry was passed 2, 4, 8, 16, and 30 times through DM with a disk gap of 100 μm, and the last portion was subsequently passed through an HPH up to 30 times. Abe et al. (2007) introduced the DM method for obtaining CNF with a uniform morphology. They reported that a uniform diameter of 15 nm could be obtained by DM from Radiata pine after delignification with a sodium chlorite-acetic acid method, and the removal of hemicellulose by an alkali treatment with 6 wt% potassium hydroxide. Iwamoto et al. (2007) have also produced CNF with a diameter of 20-50 nm and a length of over 1 μm from Pinus radiate pulp fiber by DM. However, they reported that the excessively increasing number of passes led to a decrease in the degree of polymerization, and reduced the aspect ratio of CNF.
An ultrasonic wave can produce microbubbles in water. These bubbles can be grown and imploded to generate acoustic cavitations. The cavitation energy can facilitate the defibrillation of cellulosic materials into the nano-scale (Chen et al. 2011; Cheng et al. 2009; Zhao et al. 2007). Chen et al. (2011) prepared CNF from poplar wood by ultrasonicating the wood at 400-1,200 W for 30 min, after the chemical process, to eliminate the lignin and hemicellulose. CNF with a diameter of 5-20 nm was produced as the output power increased to a value greater than 1,000 W. Xie et al. (2016) reported the isolation method from bamboo using microwave liquefaction combined with a chemical treatment and ultrasonication. Microwave liquefaction using a mix of glycerol and methanol at a ratio of 2/1 (w/w) was capable of eliminating all the lignin, making the resultant product easy to purify by a chemical treatment using an acidified NaClO2 solution. Thereafter, ultrasonication was applied to the purified residues, generating the presence of elementary fibrils, nano-sized fibril bundles, and aggregated fibril bundles.
Ball-milling (BM) is method that has been widely used to reduce the size of materials into the nano-scale through impact and friction in the dry state, because of its simple operation and relatively low-priced equipment, which consists of a hollow cylindrical shell partially filled with balls. (Koch 1997). BM has been used to defibrillate lignocellulosic materials into nanofibrils, in the wet state, by reducing the decrystallization effect due to severe operation conditions (Avolio et al. 2012; Ouajai shanks 2006; Phanthong et al. 2016). Zhang et al. (2015) investigated the effect of ballmilling conditions, including the ball-to-pulp mass ratio, ball size and mass, milling time, and alkali pretreatment on the characteristics of CNF from bleached softwood kraft pulp. They found that the crystal structure, crystallinity, and crystallite size of the resulting CNF were not sensitive to the differences in the milling conditions, and that the careful selection of the ball size was important to produce CNF instead of particles. The milling time and ball-to-pulp mass ratio were also found to be important, showing that a ballto-pulp mass ratio that was too low produced damage to the nanofibers in a short milling time. An alkali pretreatment was effective for defibrillation.
3.2 Pretreatments to facilitate mechanical defibrillation
Lignocellulosic biomass is difficult to be isolated because of its biomass recalcitrant characteristics, meaning that it has a complex and hierarchical structure (Himmel et al. 2007). In general, a pretreatment is necessary prior to mechanical defibrillation to improve the efficiency of CNF production from lignocellulosic biomass. The main role of the pretreatment is to disturb the tightened structure of lignocellulosic biomass by partially, or completely, removing the lignin and hemicellulose (Lee et al. 2010), and to reduce the energy consumption of mechanical defibrillation (Siró and Plackett 2010). Various pretreatments such as hot-compressed water (HCW), ozone, steam, and chemical and enzyme treatments have been introduced. Chang et al. (2012) reported the effect of a HCW treatment before DM on the defibrillation efficiency of bamboo. Some hemicellulose and lignin were partially removed by the HCW treatment, enhancing the defibrillation efficiency. Jang et al. (2013) carried out an ozone and steam pretreatment on Korean White Pine for CNF production by DM. When compared with the non-pretreated sample, the pretreated samples had a fine and uniform morphology, with a large specific surface area, for the short DM time. Henriksson et al. (2007) treated cellulosic wood fiber pulps with hydrochloric acid and endoglucanase as a pretreatment before the HPH treatment, showing that the CNF obtained with the enzyme pretreatment had a higher average molar mass and a larger aspect ratio than the CNFs resulted from the acidic pretreatment. Pääkkö et al. (2007) carried out an enzymatic hydrolysis combined with mechanical shearing and HPH for the production of CNF production from bleached sulphite pulp, leading to a controlled fibrillation down to the nanoscale, and a network of long and highly entangled CNFs. They fractionated two groups of CNFs, i.e., one with lateral dimensions of 5-6 nm, and another one with lateral dimensions of about 10-20 nm.
Recently, Isogai et al. (2011) published a review paper titled, “TEMPO (2,2,6,6-tetramethylpiperidine-1-oxyl radical)-oxidized cellulose nanofibers.” By TEMPO oxidation as a pretreatment before mechanical defibrillation, cellulosic materials can be efficiently converted to CNFs with diameters of 3-4 nm, and several microns in length. Carboxylate groups on the C6 position of each CNF surface could be selectively formed without any changes to the original crystallinity of the material, and the electrostatic repulsion and/or osmotic effects improved the defibrillation efficiency to produce completely individualized CNFs by gentle mechanical disintegration. They insisted that TEMPO-oxidized CNFs have potential applications as environmentally friendly and new biobased nanomaterials in high-tech fields.
3.3 Hydrolysis treatments for CNC production
CNC can be produced by acid hydrolysis using sulfuric acid, hydrochloric acid and phosphoric acid, and enzymatic hydrolysis. The amorphous region of cellulose is preferentially decomposed by hydrolysis because the cellulose molecular chains in the crystalline region have a higher hydrolytic resistance than those in the amorphous region. CNC has a rod-shape morphology with a diameter of 4-10 nm, a length of 100-400 nm, and high crystallinity. The CNC production yield by hydrolysis with hydrochloric and phosphoric acid is known to be higher than that obtained using sulfuric acid. Yu et al. (2013) prepared CNC from microcrystalline cellulose (MCC) using 64% sulfuric acid and 6 M hydrochloric acid, and compared the production yield and dimensional characterization of CNC. They found that the length and width of the CNC obtained with sulfuric acid and hydrochloric acid were about 239 and 16 nm, and 258 and 16 nm, respectively. In addition, the CNC production yield by sulfuric acid and hydrochloric acid were found to be 30.2 and 93.8%, respectively. Rodriguez et al. (2006) hydrolyzed sisal fiber with sulfuric acid into CNC with a diameter of 3-5 nm, and a length of 100-500 nm. A production yield of approximately 30% was obtained, and the dimensional characteristics were depended on the hydrolysis conditions. Espinosa et al. (2013) conducted a hydrolysis with 10.7 M phosphoric acid for 90 min at 100°C. The CNC production yield was 76-80%, and their diameter and length were 31± 14 and 316±127 nm, respectively.
On the other hand, CNC can also be prepared by cellulase enzymatic hydrolysis using cellulases. Among cellulases, endoglucanase is known to preferably hydrolyze the amorphous regions of cellulose (Filson et al. 2009; Liu et al. 2009; Siqueira et al. 2010). Enzymatic hydrolysis is less likely to reduce the degree of polymerization of cellulose than strong acid hydrolysis, and it can also easily control the dimensions of CNC by adjusting the reaction conditions (George et al. 2011). Teixeira et al. (2015) hydrolyzed commercial CNF with endoglucanase from P. horikoshii and β-glucosidase from P. furiosus for 72 h at 85°C. As a result, the average diameter and length were decreased from about 20 to 8 nm, and from about 3 um to 600 nm, respectively. Xu et al. (2013) hydrolyzed flax fiber with endoglucanase from Aspergillus oryzae to CNC with a diameter of 5-7 nm, and a length of about 200 nm.


