State-ofthe-Art of Cellulose Nanocrystals and Optimal Method for their Dispersion for Construction-Related Applications

In this paper, we reviewed the existing literature on the fabrication of nanocomposites based on cellulose and cellulose nanocrystals (CNCs), and analyzed their dispersion mechanism with respect to their use in the field of construction. First, the existing literature on CNC-based nanocomposites that exhibit the physical and chemical properties of nanocellulose and CNCs was reviewed. Next, keeping the use of these nanocomposites in the field of construction in mind, we determined the optimal mechanical method for their dispersion as an alternative to the currently used harmful chemical techniques. To this end, we evaluated the dispersibility of colloidal CNCs using two dispersion methods: magnetic stirring (for stirring times of 60 min, 120 min, and 180 min) and high-pressure dispersion (at pressures of 345 × 105 Pa, 1035 × 105 Pa, and 1587 × 105 Pa, and one to three dispersion passes). The optimal dispersion conditions were determined by analyzing the size and zeta potential of the CNC particles. It was found that the difference in the average diameter was reduced by approximately 76% at 1587 × 105 Pa during high-pressure dispersion.

Recently, Cao et al. [16] explored novel cellulose cement composites with improved mechanical performance based on cellulose nanocrystals (CNCs).CNCs are environmentally friendly because they can be produced from abundantly available natural sources and are biodegradable in nature.Scanning electron microscopy (SEM) analysis has revealed that CNCs dispersed in an unhydrated cement core act as water channels and enhance hydration in the core, thus improving the strength of the cement.
Alain [17] studied the Young's modulus of nanocellulose with the density of crystalline cellulose of approximately 1.5-1.6 g/cm 3 .Further, it is known that CNCs are stronger than steel (Young's modulus of 200-220 GPa, density of approximately 8 g/cm 3 .)While the process for producing CNCs is a complex one, the final product suggests improvements in the cement composites by improving mechanical properties.Calcium silicate hydrate (CSH) gel formation was improved in CNC cement mortar, with the compressive strength of the mortar being 42-45% higher than that of conventional cement mortar.The formation of the CSH gel improved the strength of the cement by improving its hydration.In order to quantify the performance of cement composites based on CNCs, their degree of hydration was measured using an isothermal calorimeter and a thermogravimetric analyzer [18,19].
Studies on the application of cement to conventional CNCs have been carried out only in some strength studies using optimal mixing conditions.Ultrasonic methods are primarily used to disperse CNCs [16,[18][19][20][21].However, ultrasonic dispersion techniques have certain limitations with respect to the production of cement paste and mortar, because they can only produce small amounts of CNC suspensions (10-250 mL by Sonics and Processor USA).When attempting to produce large amounts of CNC suspensions, as is the case when making concrete mixtures, the ultrasonic dispersion equipment may exhibit problems caused by prolonged use.
Therefore, methods should be developed that allow one to prepare CNC-based concrete on a large scale while causing fewer equipment problems.In this study, we attempted to determine the optimal CNC dispersion conditions based on an evaluation of the previously reported dispersion data with the aim of using CNCs as a construction material.

Review of Existing Research
We reviewed previous studies on cellulose and cellulose nanomaterials that are relevant to the field of construction.Table 1 summarizes the review of the existing literature on CNCs.

Authors Synopsis
Cao et al. [13] Investigated the effect of addition of CNCs on cement paste.Highest ball-on-three-ball (B3B) flexural strength (strength increased by ~30%) was observed after the addition of 0.2 vol% CNCs and longer curing times.At higher CNC concentrations, an agglomeration of NCs occurred, resulting in a decrease in strength.

Siqueira et al. [22]
Use of sisal fibers as a source of CNCs as well as cellulose whiskers and microfibrillated cellulose (MFC).CNCs, cellulose whiskers, and MFC were obtained from sisal fibers and modified with n-octadecyl isocyanate.Efficiency of chemical transformation was found to be highly dependent on the nature of the nanoparticles, with explanations relating to specific areas, peeling ability, and solvent dispersion.Surface modification with n-octadecyl isocyanate improved the dispersion of nanoparticles in organic solvents.It is possible to form nanocomposite films on a wide range of polymer matrices by casting/evaporation techniques.
Heux et al. [23] Electrostatic repulsion between cellulose microcrystals (CMCs) is inefficient in nonpolar organic solvents, while strong hydrogen bonding between MCs induces the rapid aggregation of colloidal suspension.As many applications require a good dispersion of the filler phase, this type of aggregation limits the range of solvents that can be used for forming CMC suspensions.The chemical modification of microcrystalline surfaces is an effective approach for improving their dispersion.
Pei et al. [24] Unmodified and silylated CNCs (SCNCs) were prepared and used to solution-cast nanocomposite films of poly-L-lactic acid (PLLA).The effects of surface silylation on the morphology, nonisothermal, and isothermal crystallization behaviors, and mechanical properties of nanostructured composites were investigated.Unmodified CNCs formed agglomerates, whereas SCNCs were well dispersed in PLLA.The tensile modulus and tensile strength of PLLA/SCNC nanocomposite films containing only 1 wt% SCNCs were more than 20% higher than those of pure PLLA; this was owing to the crystallinity effects of SCNCs and their good dispersion.

Authors Synopsis
Kloser and Gray [25] The aqueous suspension of polyethylene oxide-grafted nanocrystalline cellulose (NCC) was prepared to obtain stereos instead of quenching stabilization.NCC suspension, which was prepared by sulfuric acid hydrolysis using a two-step process, was desulfurized with NaOH and functionalized with epoxy-terminated polyethylene oxide under alkaline conditions.

Habibi and
Dufresne [26] Cellulose and starch nanocrystals obtained from the acid hydrolysis of ramie fibers and waxy cornstarch granules can be used to graft polycaprolactone (PCL) rings with varying molecular weights onto surfaces through isocyanate-mediated reactions.It was confirmed that when PCL rings are grafted onto the surfaces of cellulose nanoparticles, the modulus decreases, while the fracture strain becomes higher than that of unmodified nanoparticles (i.e., modified nanoparticles exhibit excellent ductility).This can be ascribed to the entanglement and crystallization of the rings, namely, to the phenomenon of crystal reinforcement.
The addition of CNWs increased the modulus of nanocomposites, resulting in a significantly higher elastic modulus.A high elastic modulus is essential for shape-memory effect.
Dagnon.[28] Studied the production of cellulose from seaweeds.Nanocomposites were prepared using polyglutaric styrene, co-butadiene, and polybutadiene.Seaweed CNWs were used as highly elastic filler in the nanocomposites.In addition, the tensile dynamic modulus was determined through dynamic mechanical analysis, and interactions between the internal network and the nanomatrix were analyzed.
Espino-Pérez et al. [29] Analyzed the compatibility of CNWs and poly (lactic) acid (PLA).PLA/CNWs bionanocomposites were synthesized and their thermal, mechanical, and barrier properties were evaluated.The surfaces of CNWs were modified by in situ grafting them onto PLA; this enhanced the tensile strength and increased the heat resistance of PLA.Further, surface grafting at low CNW-ICN(Cellulose nanowhiskers were grafted by n-octadecyl-isocyanate) concentrations led to improved compatibility.
Landry et al. [30] Investigated change in the performance of films with the incorporation of NCCs.Effects on properties such as film morphology, rigidity, and heat resistance were analyzed in order to determine the suitability of NCCs as a reinforcement material.
Boufi et al. [31] Mechanical performance and transparency of NC-reinforced polymer nanocomposites formed using CNCs and NFCs as reinforcement materials, and polymers with various aspect ratios were analyzed.The reinforcement effect in formed nanocomposite films was evaluated.

CNC-Based Nanocomposites
In the case of nanostructured organic composites, the surface area between the filler material and the organic matrix must be large.The dispersibility of the filler is the most important factor affecting the available surface area.If the dispersion is poor, the filler will form aggregates owing to van der Waals bonding; this will reduce the surface area in contact with the organic material and result in poor-quality nanocomposites.In addition, strong bonding at the interface between the filler and the organic material is essential.Otherwise, any external force acting on the nanocomposite will not be transferred to the filler, resulting in poor mechanical performance.CNCs can be used as reinforcing materials in multiphase composite materials based on polymers, as they lead to significant improvements in the mechanical strength, even at very low volume fractions.In addition, CNCs have been employed in biomimetic foams, reinforced paper, and flexible panels for flat panel electronics and displays [32].Table 2 lists the various technologies that currently use nanocellulose materials.

Technology Description
Method for replicating biological structures on surfaces of CNCs (Aalto University and University of Eastern Finland) The surface by using a method to mimic the biological structure on the surface of nano-sized cellulose crystals adsorbs the virus and destroys the function of the virus.The results of this study are expected to be useful for the development of antiviral ointment and surface.[33] Transparent paper (University of Maryland, South China University of Technology, and University of Nebraska-Lincoln) Developed transparent paper that can have high optical transparency (96%) and high degree of fogging (60%) simultaneously using nanocrystalline cellulose has very effective properties as a solar cell substrate [34].
Cellulose nanofibers-based filter (Gyeongsang National University) Improvement in wet strength due to the hydrophilic characteristics of cellulose nanofiber prepared by the mechanical treatment and formation of a cellulose nanofiber filter by layering the characteristics of the adsorption of a metal ion by a high anionic functional group of carboxylated cellulose nanofibers [35].

Supercapacitors and batteries
A novel nanocellulose-based composite was fabricated for producing flexible energy storage devices.The device, which had a simple structure, used nanocellulose paper with good conductivity at room temperature as the separator, and was flexible because it employed multiwalled carbon nanotubes-based electrodes [36].
Organic light-emitting diode (OLED) devices based on nanocellulose (Kyoto University) OLEDs that show low thermal expansion coefficients (21 ppm/K) were recently developed using nanocellulose substrates.Various resins were used to produce transparent and flexible devices [37].
Energy recovery device 80% of the currently used solar panels are made of silicon, and have long lifetimes and good driving power.However, it is expensive to produce high-purity silicon, and the process is complicated, making mass production difficult.Nanocellulose is attracting attention as a novel flexible substrate for solar energy applications [38,39].

Nanoporous cellulose membrane
A silver nanoparticles-doped bacterial cellulose (AgNP@BC) nanoporous membrane containing AgNPs with diameters of ~8.1 nm was successfully fabricated without employing any other reductants or capping or dispersing agents.The BC hydrogel with a 3D network not only acts as a stable scaffold, but also as a reductant for the synthesis of AgNPs.The as-prepared membrane exhibited high efficiency during the continuous catalytic decolorization of two typical organic dyes (rhodamine 6G and methyl orange), owing to its distinct nanoporous structure [40].

CNC films
Cross-polarized reflected micrographs showing the effect of surface anchoring on planar orientation, CNC film dried in a water vapor-saturated environment assisted by orbital shear with coverslip (A) off and (B) on during drying.Uniform photonic properties were achieved over orders of magnitude with greater length scales by understanding the effects of initial concentration, orbital shear, surface anchoring, and drying conditions on the microstructure and photonic properties of fabricated films.In addition, the fabricated biomimetic films exhibited a double-peak spectrum similar to that of the chiral nematic photonic structure that was observed in Lomaptera beetles [41].

Use of CNCs in Construction Applications
The use of nanomaterials in existing construction materials can improve their durability-related properties such as the strength and salt resistance by improving the cement microstructure.Studies have shown that the powdered nanosilica (NS, ~40 nm) affects the production of calcium silicate hydrate (C-S-H) [42,43].Methods for optimizing the dispersion of CNCs were investigated for producing CNC/cement composites that are suitable for use in construction-related applications.In these composites, a nanoscale matrix is formed within the concrete owing to the well-dispersed CNCs.The flexural performances of the thus-produced concrete samples were evaluated based on their moduli of elasticity.It may also be possible to improve the tensile performance of the thus-formed concrete samples by combining them with existing fiber-reinforced concrete [18][19][20][21].

Test Methods
In this study, various nanocomposites consisting of conventional nanocellulose and CNCs were investigated, and the suitability of CNCs for use in the field of construction was evaluated.The optimal conditions for dispersing CNCs in powder form through magnetic stirring and using high-pressure dispersion to prepare colloidal samples for use in cement-base composites were also analyzed.
Nanomaterials experience van der Waals attraction, which is the primary reason for the aggregation of their particles.As the particle size decreases, the effect of the van der Waals attraction becomes more pronounced compared to that of the other forces.Sulfuric acid has been used to control the agglomeration of CNCs [11].However, the direct dispersion of CNCs can be harmful to humans.Further, it can also cause the CNCs to chemically react with the cement.Cao et al. [20] dispersed CNCs using sonication.They observed that the transparency of the dispersion varied with the sonication time.For a quantitative evaluation, they performed rheological measurements on freeze-dried CNC suspensions at 0 minutes, one minute, five minutes, 15 minutes, and 30 minutes.The thus-dispersed CNCs could move freely, in contrast to coagulated CNC suspensions.The total shear stress decreased with an increase in the ultrasonic treatment time.This type of mechanical dispersion allows the inherent properties of CNCs to be combined with those of cement, in contrast to the case for dispersion using chemical additives such as sulfuric acid.In addition, it was found that sonication is safe with respect to humans.Therefore, in this study, CNCs were mechanically dispersed (except for ultrasound dispersion) using a magnetic stirrer and a high-pressure dispersing machine.Figure 1 shows the raw CNCs, which were in the form of a powder, and a CNC suspension prepared using distilled water as the solvent.The physical properties of the CNCs and the dispersion methods used and the corresponding parameters are listed in Tables 3 and 4, respectively.In the case of magnetic stirring, the dispersibility was measured based on the stirring time, whereas for high-pressure dispersion, the dispersibility was measured based on the pressure and number of dispersion passes.shows the raw CNCs, which were in the form of a powder, and a CNC suspension prepared using distilled water as the solvent.The physical properties of the CNCs and the dispersion methods used and the corresponding parameters are listed in Tables 3 and 4, respectively.In the case of magnetic stirring, the dispersibility was measured based on the stirring time, whereas for high-pressure dispersion, the dispersibility was measured based on the pressure and number of dispersion passes.High-pressure dispersion involves applying a pressure to the sample to be dispersed using a hydraulic system and a pump.The material is passed through a nanocell and is dispersed via the mechanisms of high shear, impact, and cavitation.This process is shown schematically in Figure 2. The high-pressure disperser that was used in this study was operated at pressures ranging from 206 × 10 5 Pa to 2757 × 10 5 Pa and at a flow rate of up to 300 mL/min.Generally, a pressure of 1587 × 10 5 Pa is used to disperse nanomaterials through high-pressure dispersion.In this study, high-pressure dispersion was performed at pressures of 345 × 10 5 Pa, 1035 × 10 5 Pa, and 1587 × 10 5 Pa, with one to three dispersion passes.The CNCs used in this study were very soft and in powder form; hence, the maximum pressure was limited to 1587 × 10 5 Pa, given the possibility of their ionization in water at  High-pressure dispersion involves applying a pressure to the sample to be dispersed using a hydraulic system and a pump.The material is passed through a nanocell and is dispersed via the mechanisms of high shear, impact, and cavitation.This process is shown schematically in Figure 2. The high-pressure disperser that was used in this study was operated at pressures ranging from 206 × 10 5 Pa to 2757 × 10 5 Pa and at a flow rate of up to 300 mL/min.Generally, a pressure of 1587 × 10 5 Pa is used to disperse nanomaterials through high-pressure dispersion.In this study, high-pressure dispersion was performed at pressures of 345 × 10 5 Pa, 1035 × 10 5 Pa, and 1587 × 10 5 Pa, with one to three dispersion passes.The CNCs used in this study were very soft and in powder form; hence, the maximum pressure was limited to 1587 × 10 5 Pa, given the possibility of their ionization in water at higher pressures.

Results and Discussion
The results of the particle size analysis are shown in Figure 3 and Table 5.The particle sizes of three samples (dispersion times of 60 minutes, 120 minutes, and 180 minutes) were measured over a period of 60 minutes.In the case of magnetic stirring, the mean diameter of the particles was 299 nm for 60 minutes, 277 nm for 120 minutes, and 267 nm for 180 minutes.The average diameter decreased with the increase in the dispersion time.In addition, the deviation from the mean decreased by approximately 20% for a dispersion time of 180 minutes, and the distribution was also close to the average value.Thus, a dispersion time of 180 minutes was considered to be optimal.
Transmission electron microscopy (TEM) imaging indicated that the CNCs were bar-shaped and had formed clusters.Based on these results, we concluded that CNC dispersion using magnetic stirring is not suitable, as the dispersion time that is needed is very high, even though it results in a high degree of dispersion with the diameter of the CNC clusters decreasing with time.Based on the results of this experiment, the authors will evaluate the dispersibility according to the time and energy usage through comparison with the direct energy dispersion method: ultrasonic dispersion.

Results and Discussion
The results of the particle size analysis are shown in Figure 3 and Table 5.The particle sizes of three samples (dispersion times of 60 minutes, 120 minutes, and 180 minutes) were measured over a period of 60 minutes.In the case of magnetic stirring, the mean diameter of the particles was 299 nm for 60 minutes, 277 nm for 120 minutes, and 267 nm for 180 minutes.The average diameter decreased with the increase in the dispersion time.In addition, the deviation from the mean decreased by approximately 20% for a dispersion time of 180 minutes, and the distribution was also close to the average value.Thus, a dispersion time of 180 minutes was considered to be optimal.The zeta potential can be used as a measure of dispersibility, with a higher absolute value indicating higher dispersibility.In general, when the zeta potential is measured, it is seen that particles have a positive charge in the case of acidic solutions, exhibit an isoelectric point at a pH of 9, and have a negative charge in alkaline solutions.The lowest average particle size is observed in the pH region where the absolute value of the zeta potential is high, with the dispersibility of the particles improving owing to the electrostatic repulsive force experienced by them.On the other hand, near the isoelectric point, the particles form agglomerates, resulting in a higher average particle size, and the system becomes unstable (Otsuka Electronics Co. Ltd., Seongnam, Korea).Figure 4 shows the zeta potentials of alumina particles dispersed using the same equipment.In the case of the CNC samples, the highest zeta potential value was −48.83 mV, which was observed for a dispersion time of 180 minutes; this was approximately 19% higher than that of the sample dispersed for 120 minutes (see Figure 5).Transmission electron microscopy (TEM) imaging indicated that the CNCs were bar-shaped and had formed clusters.Based on these results, we concluded that CNC dispersion using magnetic stirring is not suitable, as the dispersion time that is needed is very high, even though it results in a high degree of dispersion with the diameter of the CNC clusters decreasing with time.Based on the results of this experiment, the authors will evaluate the dispersibility according to the time and energy usage through comparison with the direct energy dispersion method: ultrasonic dispersion.
The zeta potential can be used as a measure of dispersibility, with a higher absolute value indicating higher dispersibility.In general, when the zeta potential is measured, it is seen that particles have a positive charge in the case of acidic solutions, exhibit an isoelectric point at a pH of 9, and have a negative charge in alkaline solutions.The lowest average particle size is observed in the pH region where the absolute value of the zeta potential is high, with the dispersibility of the particles improving owing to the electrostatic repulsive force experienced by them.On the other hand, near the isoelectric point, the particles form agglomerates, resulting in a higher average particle size, and the system becomes unstable (Otsuka Electronics Co. Ltd., Seongnam, Korea).Figure 4 shows the zeta potentials of alumina particles dispersed using the same equipment.In the case of the CNC samples, the highest zeta potential value was −48.83 mV, which was observed for a dispersion time of 180 minutes; this was approximately 19% higher than that of the sample dispersed for 120 minutes (see Figure 5).Figure 6 shows the particle size distributions for the different numbers of dispersion passes at 345 × 10 5 Pa.It can be seen that under these conditions, the range of the particle diameters is 69-102 nm, and the spread of the diameters is not significant.However, the mean diameter was the lowest after three passes (79 nm).Further, while there was a decrease in the particle size after the second pass, the average diameter did not change significantly between passes two and three.Figure 6 shows the particle size distributions for the different numbers of dispersion passes at 345 × 10 5 Pa.It can be seen that under these conditions, the range of the particle diameters is 69-102 nm, and the spread of the diameters is not significant.However, the mean diameter was the lowest after three passes (79 nm).Further, while there was a decrease in the particle size after the second pass, the average diameter did not change significantly between passes two and three.Figure 6 shows the particle size distributions for the different numbers of dispersion passes at 345 × 10 5 Pa.It can be seen that under these conditions, the range of the particle diameters is 69-102 nm, and the spread of the diameters is not significant.However, the mean diameter was the lowest after three passes (79 nm).Further, while there was a decrease in the particle size after the second pass, the average diameter did not change significantly between passes two and three.As shown in Figure 7, the particle size distribution at 1035 × 10 5 Pa became wider after the second pass; however, the average diameter was not significantly different from the overall average.For a pressure of 345 × 10 5 Pa, the average diameter was approximately 76 nm after three passes (the lowest value), with the difference between the second and third passes not being significant.At 1035 × 10 5 Pa, the range of the diameters was 68-98 nm.In the case of the highest pressure (1587 × 10 5 Pa) used in this experiment (as shown in Figure 8), the diameter distribution after three passes was the widest, with the range of the diameters being 56-86 nm; this was approximately 16% lower than that at 1035 × 10 5 Pa.In addition, the standard deviation of the mean diameter was smaller than those for the 345 × 10 5 Pa and 1035 × 10 5 Pa samples.As shown in Figure 7, the particle size distribution at 1035 × 10 5 Pa became wider after the second pass; however, the average diameter was not significantly different from the overall average.For a pressure of 345 × 10 5 Pa, the average diameter was approximately 76 nm after three passes (the lowest value), with the difference between the second and third passes not being significant.At 1035 × 10 5 Pa, the range of the diameters was 68-98 nm.In the case of the highest pressure (1587 × 10 5 Pa) used in this experiment (as shown in Figure 8), the diameter distribution after three passes was the widest, with the range of the diameters being 56-86 nm; this was approximately 16% lower than that at 1035 × 10 5 Pa.In addition, the standard deviation of the mean diameter was smaller than those for the 345 × 10 5 Pa and 1035 × 10 5 Pa samples.
A comparison of the high-pressure and magnetic dispersion methods revealed that the deviation in the particle size was smaller in the case of the former.In addition, as shown in Table 6, most of the particles had diameters smaller than 100 nm, indicating high dispersibility.Among these samples, the particles treated at 1587 × 10 5 Pa showed the lowest diameter (67 nm), with the deviation in the diameters also being the smallest.Hence, it can be concluded that high-pressure dispersion is the superior method, with the highest pressure yielding the best results.The zeta potential data shown in Figure 9 indicates that the highest potential (−52.02 mV) was observed for the sample processed at 1587 × 10 5 Pa.
pressure of 345 × 10 5 Pa, the average diameter was approximately 76 nm after three passes (the lowest value), with the difference between the second and third passes not being significant.At 1035 × 10 5 Pa, the range of the diameters was 68-98 nm.In the case of the highest pressure (1587 × 10 5 Pa) used in this experiment (as shown in Figure 8), the diameter distribution after three passes was the widest, with the range of the diameters being 56-86 nm; this was approximately 16% lower than that at 1035 × 10 5 Pa.In addition, the standard deviation of the mean diameter was smaller than those for the 345 × 10 5 Pa and 1035 × 10 5 Pa samples.A comparison of the high-pressure and magnetic dispersion methods revealed that the deviation in the particle size was smaller in the case of the former.In addition, as shown in Table 6, most of the particles had diameters smaller than 100 nm, indicating high dispersibility.Among these samples, the particles treated at 1587 × 10 5 Pa showed the lowest diameter (67 nm), with the deviation in the diameters also being the smallest.Hence, it can be concluded that high-pressure dispersion is the superior method, with the highest pressure yielding the best results.The zeta potential data shown    A comparison of the high-pressure and magnetic dispersion methods revealed that the deviation in the particle size was smaller in the case of the former.In addition, as shown in Table 6, most of the particles had diameters smaller than 100 nm, indicating high dispersibility.Among these samples, the particles treated at 1587 × 10 5 Pa showed the lowest diameter (67 nm), with the deviation in the diameters also being the smallest.Hence, it can be concluded that high-pressure dispersion is the superior method, with the highest pressure yielding the best results.The zeta potential data shown was the smallest after the third pass in general, with the 1587 × 10 5 Pa-three pass combination resulting in the highest degree of dispersion.However, it should be noted that the particle diameters after each pass were not significantly different from the average diameter in the final dispersion.
(3) High-pressure dispersion can be used for mass-producing concrete and is advantageous, as it is relatively simple and consumes less time than magnetic stirring.Further, the difference in the average diameter is reduced by approximately 76% in the case of a pressure of 1587 × 10 5 Pa.Hence, the optimal way of dispersing CNCs is to subject them to high-pressure dispersion at 1587 × 10 5 Pa for three passes.
(4) CNCs are expected to find increasing use as a reinforcing nanomaterial that can improve material properties such as the elastic modulus, tensile strength, bending strength, fracture energy, and impact resistance, when dispersed in concrete using the most efficient dispersion method.

Figure 3 .Table 5 .
Figure 3. CNC particle size during magnetic stirring.Size distributions with average diameters are shown for dispersion times of (a) 60 min, (b) 120 min, and (c) 180 min.(d) TEM image of particles dispersed for 180 min.

Figure 4 .
Figure 4. Dependence of particle size and zeta potential of alumina particles on pH (adapted from Otsuka Electronics Korea Co. Ltd.).

Figure 5 .
Figure 5. Zeta potential of CNC samples as a function of the magnetic dispersion time.

Figure 4 .
Figure 4. Dependence of particle size and zeta potential of alumina particles on pH (adapted from Otsuka Electronics Korea Co. Ltd.).

Figure 4 .
Figure 4. Dependence of particle size and zeta potential of alumina particles on pH (adapted from Otsuka Electronics Korea Co. Ltd.).

Figure 5 .
Figure 5. Zeta potential of CNC samples as a function of the magnetic dispersion time.

Figure 5 .
Figure 5. Zeta potential of CNC samples as a function of the magnetic dispersion time.

Figure 6 .
Figure 6.Results of particle size analysis after high-pressure dispersion at 345 × 10 5 Pa.Diameter distributions after (a) one dispersion pass as a function of time, (b) two dispersion passes as a function of time, and (c) three dispersion passes as a function of time.

Figure 6 .
Figure 6.Results of particle size analysis after high-pressure dispersion at 345 × 10 5 Pa.Diameter distributions after (a) one dispersion pass as a function of time, (b) two dispersion passes as a function of time, and (c) three dispersion passes as a function of time.

Figure 7 .
Figure 7. Results of particle size analysis during high-pressure dispersion at 1035 × 10 5 Pa.(a) Overall particle size as function of time.Diameter distribution after (b) one dispersion pass as a function of time, (c) two dispersion passes as a function of time.

Figure 8 .
Figure 8. Results of particle size analysis during high-pressure dispersion at 1587 × 10 5 Pa.Diameter distributions after (a) one dispersion pass as a function of time, (b) two dispersion passes as a function of time, and (c) three dispersion passes as a function of time.

Figure 7 .
Figure 7. Results of particle size analysis during high-pressure dispersion at 1035 × 10 5 Pa.(a) Overall particle size as function of time.Diameter distribution after (b) one dispersion pass as a function of time, (c) two dispersion passes as a function of time.

Figure 7 .
Figure 7. Results of particle size analysis during high-pressure dispersion at 1035 × 10 5 Pa.(a) Overall particle size as function of time.Diameter distribution after (b) one dispersion pass as a function of time, (c) two dispersion passes as a function of time.

Figure 8 .
Figure 8. Results of particle size analysis during high-pressure dispersion at 1587 × 10 5 Pa.Diameter distributions after (a) one dispersion pass as a function of time, (b) two dispersion passes as a function of time, and (c) three dispersion passes as a function of time.

Figure 8 .
Figure 8. Results of particle size analysis during high-pressure dispersion at 1587 × 10 5 Pa.Diameter distributions after (a) one dispersion pass as a function of time, (b) two dispersion passes as a function of time, and (c) three dispersion passes as a function of time.

Table 1 .
Review of existing literature on cellulose nanocrystals (CNCs).

Table 2 .
Use of nanocellulose in the industry.

Table 3 .
Physical properties of CNCs used.

Table 4 .
Methods used to disperse CNCs and corresponding parameters.

Table 3 .
Physical properties of CNCs used.

Table 4 .
Methods used to disperse CNCs and corresponding parameters.