Structure–color mechanism of iridescent cellulose nanocrystal films

Abstract

Chirality and repulsion interactions among sulfate cellulose nanocrystals (CNCs) have vital impact on the formation of a cholesteric liquid crystal (CLC) phase in a suspension or solidified film. In this work, a facile sonication treatment was applied to change the structure and repulsion interactions of CNCs and consequently tune the chiroptical properties of the resultant films. The results show that increasing the sonication energy either by improving the input power or prolonging the aging time resulted in the reduction of particle size and surface charge density, thereby increasing the cholesteric pitch and red-shifting the reflective wavelength of the iridescent films. The optical properties of the film followed the regulation of Bragg reflection and thin-film interference. However, an over-energy input would result in the multi-dispersion of the CNCs according to the level of the surface charge density, thus leading to the formation of polydomain CLC instead of planar CLC because of multi-distributed intra-axial drive forces. Hence, a schematic model was built up to describe the structure transition, as well as the color variation and to correlate the mesoscopic behavior of CNCs and the microscopic interactions of electrostatic repulsions, hydrogen bonding affinity and chirality. Hence, we provide some meaningful information on building up a hierarchical organization assembled from charged rigid biological rods, and help to recognize the structure–color mechanism of solidified films of polysaccharide nanocrystals.

---

Introduction

Cellulose nanocrystals have attracted significant attention because of their unique properties such as high stiffness and strength, high specific surface area, low coefficient of thermal expansion, optical transparency and self-assembly behavior. The low cost, renewability and recyclability, and chemical reactivity, allowing its chemical and physical properties to be tailored, make CNC attractive for various applications.¹ CNCs are typically extracted from cellulosic biomass using strong acid hydrolysis,² strong oxidation,³ or ionic liquids.⁴ The acid hydrolysis of cellulose is a well-known process involving the preferential digestion of the amorphous domains and cleavage of the nanofibril bundles, therefore breaking down the hierarchical structure of the raw materials into crystalline nanocrystals. Strong acids such as sulfuric acid⁵ or hydrochloric acid⁶ are generally used; however, hydrochloric acid CNCs have minimal surface charge and limited aqueous dispersibility, whereas sulfuric acid CNCs provide highly stable aqueous suspensions because of the esterification of surface hydroxyl groups resulting in negatively charged sulfate groups.^7–9 When the homogeneous suspension of sulfate CNCs was concentrated, nematic liquid crystals or CLCs were formed after a slow self-assembling process.^10,11 The aqueous suspension of the sulfate CNCs is one kind of lyotropic liquid crystal. Another interesting property was evident when an aqueous CNCs suspension was evaporated, and the chiral nematic phase was retained in the produced solid semitranslucent CNC films. CNC rods were parallelly aligned to each other and to the plane of layers, each layer being rotated slightly with respect to the layers above or below it, thereby producing helical pseudo-layers. Inspired by the fascinating chiroptical properties of CNC mesogens, thin coatings, pigments, printing ink, and colorful films with tunable optical properties were designed,¹² and porous chiral photonic crystals were also fabricated by using CNCs as templates for the deposition of inorganic nanoparticles.^13,14 Research in CNC mesogens spontaneously arising from aqueous suspensions have attracted significant interest and is becoming a novel scientific topic. Moreover, chirality and nematic phases are vital to optically functionalize the CNCs based nanomaterials or templated phonotic crystals.

In addition to acidic hydrolysis and oxidation, mechanical treatments, such as sonication¹⁵ and homogenization^16–19 are commonly used to disintegrate and disperse polysaccharide nanocrystals in suspensions, e.g., sonication was thought to break up CNCs side-by-side, resulting in decrease in viscosity,⁹ and control the red-shift color of CNC films by increasing the input sonication energy.²⁰

The cholesteric characters were described as being dependent on the polydispersity, physical dimension, surface charge, and the ionic strength of the CNC particles.¹⁶ In previous reports,⁷ it was thought that the smaller average length of CNC rods would lead to a smaller pitch. However, this turns out not to be true in the case of the sonicated CNCs because the smaller sonicated particle size would lead to a larger pitch. Furthermore, based on our research experiments, CNC films processed by sonication didn't show uniform color as compared with freshly- prepared films. To the best of our knowledge, the liquid crystal properties of CNC suspensions have had investigated in detail,^6,7,9–12 but the color mechanisms of the CNC films are still vague. In this work, we investigated the optical properties of CNC films through turning structure and properties of CNCs subjected to different energy levels of sonication. We hope to discover the mystery of color and establish the structure–color mechanism so that it can be applied to many other natural polymeric nanocrystalline optical materials.

Experimental section

Materials

Microcrystalline cellulose powder (MCC, column chromatography) and sulfuric acid were purchased from Sinopharm Chemical Reagent Co., Ltd, Shanghai, China. Regenerated cellulose dialysis tubing having a molecular weight cut-off of 8000–10 000 was supplied by Nanjing Wanqing Chemical Glassware Instrument Co., China.

Preparation of CNC suspension and film

Aqueous CNC suspensions were prepared from MCC by sulfuric acid (64%) hydrolysis at 50 °C for 2 h, and then it was diluted five times to stop the hydrolysis reaction. The diluted suspension was poured into the dialysis tubing and dialyzed against deionized water for several days until the pH reached around 5.0. The freshly prepared CNC suspension had a solid content around 0.5 wt%. Mechanical sonication was performed on an ultrasonic processor (Sonics Vibra-Cell 1200 W, 19–21 kHz, HN98-IIID, Shanghai, China) with a 20 mm diameter probe and a 25% ultrasonic power ratio. Typically, 50 mL of a 0.5 wt% CNC suspension was placed in a 100 mL beaker and sonicated at 300 W. Ultrasonic disruptors were typically operated in a pulsed mode, in which the duration of the on and off intervals were regulated at 10 and 20 s, respectively. In this work, the sonications were carried out in an ice bath to prevent desulfation on the surface of the nanocrystallites resulting from the heating of the suspension, and prolonged for 10, 20, 30, 40, 50, 60 and 70 min.

Aliquots of 0.5 wt% CNC suspension (50 mL) were sonicated with increasing energy inputs and were allowed to stabilize in a sealed plastic box (3.9 × 3.9 × 4.7 cm³) over a period of 72 h at ambient conditions. Then, the treated suspension, in the open plastic box, was left undisturbed and slowly evaporated to obtain solidified CNC films in an oven at 30 °C. All the images of CNC films were taken normal to the films, which were placed against a black background in order to obtain the reflected iridescence under diffuse incident light.

Characterization

Zeta potential, polydispersity index (PDI), and Z-average size were measured using a Zetasizer Nano ZS90 (Malvern Instruments, UK). The CNC suspension was added to folded capillary cells to measure zeta potential values, which were calculated using the Smoluchowski equation. The surface charge density of the CNCs was determined by conductometric titration. Approximate 1 g of a 5 wt% CNC suspension was transferred to a three-neck round bottom flask followed by the addition of 200 mL of deionized water and 4 mL of 0.1 N HCl, which ensured an excess of H⁺ in the suspension. The acidic suspension was titrated with 0.01 N NaOH.

Several drops of the diluted suspension (0.1 wt%) were spin-coated onto silicon wafer chips and freshly cleaned using a modified RCA protocol. First, the chips were immersed in a solution of NH₄OH : H₂O₂ : H₂O (1 : 1 : 5) at 80 °C for 5 min, and then rinsed repeatedly with water and then immersed in a solution of HCl : H₂O₂ : H₂O (1 : 1 : 5) at 80 °C for 5 min. Finally, the chips were dried under an N₂ stream. The AFM imaging of the CNCs was performed with an Asylum Research MFP-3D atomic force microscope (Santa Barbara, CA). Samples on the chips were scanned in the AC tapping mode with Tap 300 standard silicon probes (tip radius < 10 nm, spring constant of 37 N m⁻¹, resonant frequency of 300 kHz) (Budget Sensor, USA) under a 1 Hz scan rate and 1024 pixels × 1024 pixels image resolution. A polarized optical microscope (POM) equipped with a Nikon (Tokyo, Japan) MDA502AA E400 was used to determine the cholesteric pitch of the liquid crystalline phase in the CNC films, which were placed directly on glass slides. Scanning electron microscopy (SEM) was performed using a Hitachi S-3600N VP SEM (Hitaschi, Japan) to investigate the morphology of the liquid crystal films. The cross-section of the CNC film was coated with gold for SEM observation at 20 kV.

UV-vis spectra were recorded on a Perkin-Elmer ultraviolet-visible spectrometer (Lambda 3B model). The optical reflectance of the sample films was measured over the wavelength region from 200 to 800 nm under 90° incident illuminations.

Results and discussion

Structure and properties of CNCs

Z-average size, PDI, Zeta-potential, and surface charge density of the CNCs measured using Zeta-sizer are presented in Table 1. The freshly prepared CNCs in this work had an Z-average size of about 176 nm, while the particle size of the CNCs sonicated only for 10 min was greatly reduced to 95 nm, and then the down-size effects slowed down and tended to reach an equilibrium after treatment for 50–70 min. The effect of sonication treatment on particle size and morphology was also investigated using AFM. As shown in Fig. 1, the freshly prepared CNCs and the sonicated CNCs in an ice bath for 70 min is needle shaped with a mean length of 159 and 81 nm, and a mean diameter of 50 and 28 nm, respectively. It is interesting that the Z-average size of the CNCs determined by the Zeta-sizer is in considerable accord with the length of nanocrystals measured by AFM. Sonication treatments markedly shortened the mean particle length and width. This may be because of the fact that CNCs bearing defects or cracks produced by sulphuric acid hydrolysis, were further subjected to powerful sonication and were broken into scaled down nanocrystals. As an indicator of electrostatic repulsion, surface charge density was reduced from 0.67 e nm⁻² for the untreated CNCs to 0.37 e nm⁻² for the CNCs treated for 70 min, owing to the increase in the total surface area of the particles.

Table 1 Parameters of CNCs in the suspension or in solidified cholesteric liquid crystal films with different sonication times. (Zeta-potential, Z-average size, and PDI were measured by Zeta-sizer; length or diameter of CNC were measured by AFM; and pitch and wavelength at peak was measured by POM and UV-vis, respectively.)

Sonication time (min)	Zeta potential (mV)	Z-average size (nm)	PDI	Length of CNC (nm)	Diameter of CNC (nm)	Surface charge density (e nm^−2)	Half pitch (μm)	Maximum reflection wavelength (nm)
0	−50.6 ± 0.2	176.2	0.427	159 ± 8	50 ± 2	0.67	0.554	328.46
10	−51.0 ± 0.5	95.6	0.519	122 ± 15	38 ± 4	0.51	0.756	442.03
20	−52.2 ± 0.2	76.2	0.517	120 ± 18	36 ± 5	0.48	0.877	507.25
30	−52.7 ± 0.3	74.1	0.501	119 ± 20	35 ± 5	0.47	0.965	538.66
40	−53.0 ± 0.3	73.1	0.481	114 ± 19	34 ± 6	0.45	1.010	550.73
50	−53.6 ± 0.4	69.1	0.490	106 ± 24	33 ± 7	0.44	1.068	594.22
60	−54.0 ± 0.3	68.7	0.496	87 ± 25	30 ± 6	0.39	1.205	664.31
70	−54.1 ± 0.4	67.7	0.549	81 ± 28	28 ± 8	0.37	1.260	671.57

---

Fig. 1 (a) AFM micrographs of the freshly prepared CNCs and (b) CNCs sonicated for 70 min, and (c and d) the corresponding statistical distribution of particle length.

The pronounced zeta potential values of the freshly prepared CNCs and sonicated CNCs were over −50 mV and showed no distinguishing differences as listed in Table 1. The zeta potential was thought to be an important parameter of suspended particles as it could influence both particle stability, as well as particle mucoadhesion.²¹ In the case of the sonicated CNCs, the absolute zeta potential value was still at around −60 mV, indicating that the surface charge was strong enough to produce repellent interactions among the nanocrystals, thus leading to good dispersion in the suspension.²² Furthermore, as shown in Fig. 1b, either large or small nanocrystals were exhibited in the sonicated suspension, thus the size distribution was not narrowed but widened as indicated by PDI. Therefore, bath sonication dispersed the CNCs completely, but endowed the nanocrystals with a less uniformly distributed particle size and lower surface charge density than untreated ones.

Structural color varied with processing temperature

Fig. 2 shows the POM micrographs of CNC films dried at a temperature of 30 (a), 45 (b), 50 (c), and 70 °C (d), respectively. The perfect planar texture in Fig. 2a indicates the perfect cholesteric LC characteristics of the free-standing film dried at 30 °C. Although the planar texture was also observed in Fig. 2b, the planar mesophases were not continuous because of existing defects. When the casting temperature was 50 °C, the phase without any fingerprint textures shifted to nematic mesogen as shown in Fig. 2c. On further increase of the casting temperature to 70 °C, the LC phase absolutely disappeared and the color of the film faded away (Fig. 2d). In brief, with the increase in processing temperature, the liquid crystal structure of the CNC films shifted from a cholesteric phase to a nematic phase, till no LC phase. It is well known that sulfuric acid treated CNCs are normally associated with covalent-bound surface sulfate ester groups (typically 0.5–2%). These functional groups play an important role in forming long range ordered chiral liquid crystals, which is also true for many other cellulose derivatives bearing ethyl, hydroxypropyl, acetyl, and other groups. However, the sulfuric acid treated CNCs are sensitive to heat. At temperatures above 40–45 °C, the nanocrystals underwent slow desulfation; at 70 °C the helical and chiral LC domains of CNCs were severely destroyed because of deep desulfation, which leads to the dissociation of the negatively charged sulfate groups from the surface of the CNCs.^23,24 On the other hand, the level of order of the nematic LC was related to the evaporation ratio of water from the suspension because slow drying at a relatively low temperatures would be beneficial for the organization of the CNCs into perfect cholesteric patterns, otherwise defects existed in the mesophases of the fast drying samples. In this work, we preferred a temperature of 30 °C as the optimum evaporation condition for casting CNC films.

Fig. 2 POM images of the CNC films processed at (a) 30 °C, (b) 45 °C, (c) 50 °C and (d) 70 °C, respectively. (sonication at 300 W for 10 min).

Structural color of CNC films with varied thickness

Fig. 3 presents the reflection spectra of the CNC films with thickness increasing from 42, to 63 and then from 82 to 96 μm. It is worth noting that with the increasing thickness of the films (or number of layers), the peak reflectivity rather than the wavelength at the maximum peak increased rapidly; moreover the bandwidth decreased gradually. The optical characteristic of the CNC chiral nematic liquid film was thought as multilayer interference, which is qualitatively understood in terms that many thin layers stacked periodically.²⁵ When light is incident on a film, the light passes through air (refractive index, n = 1.0) → CNC (n = 1.5) → air (n = 1.0) as shown in the inset of Fig. 3. Multilayered thin film interference was applicable for the reflection of the chiral nematic CNC films, and the bandwidth in this case can be estimated from the difference in the wavelengths at minimum reflectivity in the nearest oscillation. Supposing that these wavelengths reflected from (N + 1) interfaces, the destructive interfere was expressed as follows:²⁶where is the optical path length of one layer and m is an integer, which is obtained under the assumption of only one-time reflection at each interface. According to equation established by Kinoshita,²⁶ the bandwidth was estimated as follows:

Fig. 3 Reflection spectra (90° incidence and detection angles) of the CNC films with thickness of 42, 63, 82, 96 μm. (sonication at 300 W for 30 min and casting at 30 °C).

For large N the bandwidth is inversely proportional to the number of layers. Therefore, it is necessary to pile up nematic layers to obtain a high reflectivity for CNC multilayers having a small refractive-index difference, which inevitably decreases the bandwidth. To reach maximum reflectivity, a CNC film with a thickness of 96 μm was selected as the target for color–structure observations in this work.

Relationship between color and structure for the sonication treated CNC film

Fig. 4 shows images of CNC films cast from suspensions sonicated for 0, 10, 20, 30, 40, and 50 min, respectively. The films exhibit reflected iridescence in naked view under diffuse incident light, and the color shifted from blue to red with the increasing sonication time. However, after 30 min, the color of the solidified film didn't appear as uniform as the previous one, and especially the central zone of the sonicated films over 50 min was bleak and broad. The corresponding UV-vis reflection spectra of the CNC films at a 90° incident angle are shown in Fig. 5. With an extended sonication time from 0 to 70 min, the wavelengths at the maximum reflectivity of the films red-shifted from 328 to 671 nm; moreover, the bandwidth of reflection was broadened. It is interesting that with prolonged sonication time from 0 to 30 min, the maximum reflectivity gradually increased, but then decreased after 30 min until the reflection peak appeared vague for the films with sonication for 60 or 70 min. It is indicated that in this work proper sonication (300 W, <30 min) was beneficial for producing uniform and bright colored films, whereas ultrahigh sonication energy input (high sonication power and prolonged sonication time) would result counterproductive effects.

Fig. 4 Images of the CNC films cast from the suspensions sonicated for (a) 0, (b) 10, (c) 20, (d) 30, (e) 40, and (f) 50 min, respectively.

Fig. 5 Reflection spectra (90° detection angle) of the CNC films cast from the suspensions sonicated for 0, 10, 20, 30, 40, 50, 60, and 70 min, respectively. (Thickness of the films is 96 μm).

POM micrographs of iridescent films prepared at varying sonication times are shown in Fig. 6. Under polarized light, free-standing films exhibit a typical planar cholesteric texture and the reflective color shifted from violet (a), blue (b), cyan (c), green (d), yellow (e), orange (f), fuchsine (g), to kermesina (h) with prolonged sonication. The corresponding half cholesteric pitch, defined as the distance required for CNC rods to make a 180° rotation, which is the distance between two neighboring planar textures, dramatically increased from 0.554, 0.756, 0.877, 0.965, 1.010, 1.068, 1.205, to 1.260 μm as summarized in Table 1. The liquid crystal films self-assembled from smaller crystals presents larger pitch as observed by POM.

Fig. 6 POM micrographs of the CNC films cast from the suspensions sonicated for (a) 0, (b) 10, (c) 20, (d) 30, (e) 40, (f) 50, (g) 60, and (h) 70 min, respectively.

Supposing that the observed pitch is uniformly distributed in a birefringent CNC film, the reflection followed Bragg's law:²⁷

λ = (n_‖ − n_⊥)P sin θ (3)

where P is the pitch of CLCs, n_‖ is the refractive index of the polarized light parallel (extraordinary) to the optical axis and n_⊥ is for the polarizations perpendicular (ordinary) to the optical axis, and θ is the angle of incidence. The refractive index is a function of the molecular composition, orientation and packing within a liquid crystal. In a cholesteric mesogen, the refractive index difference ( n_‖ − n_⊥) leads to the birefringence. Fig. 7 demonstrates the maximum reflection wavelength as a function of the cholesteric pitch of the CNC films with incident angle at a 90°. The result shows that there is a linear relationship between the maximum reflection wavelength and cholesteric pitch. The value of  n_‖ − n_⊥ obtained from the slope was 0.29, which is higher than reported value of 0.05 for the cholesteric CNC colloidal suspension.²⁸ It is worth noting that the curve of the CNC film treated by prolonged sonication over 30 min deviated from its linear route because the Bragg reflection only accounts for the optical properties of the planar cholesterics consisting of a large number of the birefringent layers. In this case,  n_‖ − n_⊥ exhibited a decreasing trend because polydomain mesophases were observed in those over-sonicated CNC films as discussed in the following section. However, the structure–color of a CNC film with planar mesophases originates from thin-film interference and Bragg reflection.

Fig. 7 Dependence of the maximum reflective wavelength on the cholesteric pitch of the CNC films.

Polydomain mesophases of CNC films

It is well known that a high applied power or long treatment time would deliver high input sonication energy. Fig. 8 presents the reflective spectra of the CNC films produced under sonication at 520 W for 10, 30, 120 min, respectively. The maximum reflective wavelength was red-shifted and the bandwidth was broadened, thereby resulting in elevated reflectivity. Under a relatively high sonication power (520 W) the color of the film red-shifted in a shorter sonication time than that under low power (300 W). However, we found that under treatment at 520 W for 30 min, the resultant solid film was not uniform and the centre of the film appeared faint, and the film treated by over-energy input (520 W for 120 min) appeared smooth, transparent, and colorless, as shown in the inset of Fig 8. Therefore, it is essential to control the ultrasonic power and aging time to obtain red-shifted iridescent films.

Fig. 8 Reflection spectra (90° detection angle) of CNC films cast from the suspensions sonicated at 520 W for 10, 30, and 120 min. (The inset shows photographs of solidified CNC films under corresponding sonication treatment. Thickness of the films is 96 μm.)

Fig. 9 presents POM micrographs of the gloomy regions of the film produced under the sonication at 520 W for 10, 60, 120, and 150 min, respectively. Interestingly, the fingerprint region in Fig. 9a–d was composed of a focal conic texture with a random distribution of the helical axes. The amount of polydomains with the interval distance (P/2) varying from 0.70 to 2.08 μm in Fig. 9d is obviously more than those in Fig. 9b. This is evidence to address the nonlinear relationship between pitch–wavelength as shown in Fig. 7, while sonication time was over 30 min. Owing to the ultrahigh energy input, the cholesteric mesophases was very weak (Fig. 9e) and the birefringence with a fingerprint texture was intermittently exhibited in the transparent CNC films (Fig. 9f). Moreover, the incident light was weakly and randomly reflected and scattered in all directions, thus leading to the low reflectivity, broadened bandwidth, mixed color, and even no iridescence of the film.

Fig. 9 POM images of the CNC films cast from the suspensions sonicated at 520 W for (a and b) 10, (c and d) 60, (e) 120, and (f) 150 min, respectively.

Mechanism of structure–color correlation

Fig. 10 shows the SEM micrographs of the fractured surface of the CNC films, which were cast from the suspension sonicated for 0, 10, 30, 50, 60, 70 min, respectively. The CNC films in Fig. 10a–d exhibit fingerprint lines with a regular adjacent space of about 1–2 μm at an oblique angle. Obviously, the cholesteric axis was perpendicular to the film surface and parallel to the lines. The inset image of Fig. 10b is a high magnification image of the layered structure, in which the adjacent layer distance agreed with the P/2 value from the optical microscopy. A striking feature of the continuous “hill” and “hole” pattern resulted from a spontaneous twist deformation of the left-handed helicoidal hierarchy, which have been described in previous work.²⁵ Interestingly, when ultrasonication was over 10 min, the perfect pattern was not observed on the fracture surface, and even disappeared in Fig. 10e and f, indicating the transition from the perfect and unique-orientated planar phase to the polydomain mesophase as evidenced in Fig. 9.

Fig. 10 SEM micrographs of the cross-section of the CNC films cast from the suspensions sonicated for (a) 0, (b) 10, (c) 30, (d) 50, (e) 60, (f) 70 min, respectively. (The inset of (b) shows high resolution micrograph of solidified CNC films.)

A great number of plant cell walls are composed of cellulose microfibrils associated with the matrix components of xylan, and are typically designed as a helicoidal pattern that is analogous to a cholesteric order.^29–31 Cellulose chains in ramie has the formal geometric characteristics of a helical form with seven cellobiose residues per turn, radius r = 1.5830 Å, and an helix angle of 7°51′.^32,33 Similar to several other biological macromolecules, cellulose and its derivatives have been reported to have chiroptical properties in solution or in bulk because the cellulose backbone is chiral. However, strong hydrogen bonds originating from the –OH groups hanging on the cellulose chains are easily formed to aggregate the molecules. Because of this the HCl hydrolyzed CNCs did not show liquid crystal phases in the aqueous suspension. Therefore, it is crucial to create a method for weakening the hydrogen bonding attraction and strengthen the repulsion. To the best of our knowledge, several effective methods were adopted to obtain cellulosic CLCs. The first approach is to screen out the affinity by changing the polarity of the solvent; the second approach is to block the hydrogen bonds by increasing hydrophobicity by grafting or derivation, e.g., hydroxypropyl cellulose, ethyl cellulose, etc. The pitch height increased significantly with a reduction of the substitution degree for butyric esters of (2-hydroxypropyl) cellulose, Ethyl-cyanoethyl cellulose, e.g., increase of DS from 1.90–2.96 resulted in a significant decrease of P from 641 to 275 nm.³⁴ The third approach is to maintain the handness of cellulose, but improve electrostatic repulsions by increasing surface charges, e.g., sodium cellulose sulfate, sulfate hydrolyzed CNCs, etc. According to Chilaya,³⁵ the cholesteric pitch and the twist angle θ is related by:

where θ is the angle between neighboring nematic planes, and r is the unknown parameter indicating separation between adjacent nematic planes. Obviously, increasing the layer distance or weakening the twisting force or angle is beneficial for increasing the pitch. Therefore, as mentioned above, increasing the DS of cellulose derivatives, weakening the hydrogen bonding by solvents, or properly increasing ionic strength by the introduction of salts will increase the twisting angle θ, thereby reducing the pitch value.

In this work, the sulfuric acid CNCs have negatively charged surfaces because of the esterification of hydroxyl groups by sulfate ions, leading to stable nanocrystal aqueous suspensions. With increasing sonication energy, the CNCs were broken into smaller sized nanocrystals by mechanical sonication disruption, thus increasing the total number and the specific surface area of the particles. However, sonication did not affect the total surface charges, which were determined by the hydrolysis of CNCs by sulfuric acid. This means that the surface charge density of the small nanocrystals was effectively reduced. As shown in Scheme 1, under low energy input sonication, the repulsion between particles was reduced and the intra-axial force at constant interaxial spacing decreased, which consequently leads to the reduction of the twisting angle θ. Moreover, the increased nanocrystals spontaneously entered into macroscopic mesogens, thereby increasing the number of nematic layers. Hence, the pitch of the sonicated CNC mesogen increased and the maximum reflective wavelength red-shifted. However in the case of freshly prepared CNCs,⁷ the small-sized crystals, which were hydrolysed for longer time under high acidic concentrations, obviously had a higher surface charge density than those large-sized CNCs, accordingly leading to a smaller pitch. Thus, it is the polar opposite of the CLCs assembled from small-sized CNCs in this work. However, under high energy input sonication, the size or charge of polydispersed nanocrystals may join the different macroscopic mesogens according to the level of intra-axial force, thereby slowly self-assembling into polydomain mesophases with widely-distributed chiral pitches. Sonication would further force those nanocrystalline mesogens to be suitably stacked and positioned, thus forming CLCs with focal conic or polydomain textures. When the mean surface charge density was greatly reduced and the repulsion could not block the strong hydrogen bonding affinity, the long-range ordered mesophases were completely collapsed and the iridescence originating from birefringence was almost lost as evidenced in Fig. 1d and 9f. On the basis of Scheme 1, the electrostatic repulsion resulted from the surface charge, and the helical nature of cellulose played vital roles in building up the structure–color hierarchical organizations of CNCs.

Scheme 1 Cholesteric phase transition of CNCs induced by inputting sonication energy.

Conclusion

CNCs prepared by sulfuric acid hydrolysis retained the helicoids or handness character of cellulose chains, but blocked the hydrogen bonding affinity by electrostatic repulsions originating from the negative surface charge. Sonication treatment was an effective way to generate sub-scale level crystals with a diminishing surface charge density, thus leading to the increment of the cholesteric pitch and the maximum reflective wavelength with increasing input sonication energy. The structure–color of CNC films followed the regulation of Bragg reflection (birefringence) and thin-film interference. However, there is deviation from the linear relationship between pitch and wavelength because poly-domain hierarchical mesogen instead of planar phase was self-organized in the iridescent film cast from high-level sonication treated suspension. As a result, the structure–color of CNCs was controlled by the electrostatic repulsion; and the exact mechanism was thought to be applicable for many other polysaccharide nanocrystals. In order to obtain a uniform iridescent CNC film, short-range positional order and long-range orientation should take synergic effects. However, the chiral nematic transition during the solidification process was not clear and needs to be addressed in future work.

Acknowledgements

The authors are grateful to National Natural Science Foundation of China (no. 51103073 & 51473077), Natural Science Foundation of Jiangsu Province (no. BK2011828), Scientific Research Foundation for the Returned Overseas Chinese Scholars, and Qing Lan Project and Six Talented Peak Program of Jiangsu Province and the Priority Academic Program Development of Jiangsu Higher Education Institutions for financial support.

References

1. M. Grunert and W. T. Winter, Nanocomposites of Cellulose Acetate Butyrate Reinforced with Cellulose Nanocrystals, J. Polym. Environ., 2002, 10, 27–30.
2. S. Shafiei-Sabet, W. Y. Hamad and S. G. Hatzikiriakos, Rheology of Nanocrystalline Cellulose Aqueous Suspensions, Langmuir, 2012, 28, 17124–17133.
3. Y. Habibi, H. Chanzy and M. R. Vignon, TEMPO-mediated Surface Oxidation of Cellulose Whiskers, Cellulose, 2006, 13, 679–687.
4. H. Song, J. Zhang, Y. Niu and Z. Wang, Phase Transition and Rheological Behaviors of Concentrated Cellulose/ionic Liquid Solutions, J. Phys. Chem. B, 2010, 114, 6006–6013.
5. R. F. Nickerson and J. A. Habrle, Cellulose Intercrystalline Structure, Ind. Eng. Chem., 1947, 39, 1507–1512.
6. J. Araki, M. Wada, S. Kuga and T. Okano, Influence of Surface Charge on Viscosity Behavior of Cellulose Microcrystal Suspension, J. Wood Sci, 1999, 45, 258–261.
7. S. B. Candanedo, M. Roman and D. G. Gray, Effect of Conditions on the Properties and Behavior of Wood Cellulose Nanocrystal Suspension, Biomacromolecules, 2005, 6, 1048–1054.
8. B. G. Ranby, Aqueous Colloidal Solutions of Cellulose Micelles, Acta Chem. Scand., 1949, 3, 649–650.
9. R. H. Marchessault, F. F. Morehead and M. J. Koch, Some Hydrodynamic Properties of Neutral Suspensions of Cellulose Crystallites as Related to Size and Shape, J. Colloid Sci., 1961, 16, 327–344.
10. J. F. Revol, L. Godbout, X. M. Dong and D. G. Gray, Chiral Nematic Suspensions of Cellulose Crystallites: Phase Separation and Magnetic Field Orientation., Liq. Cryst., 1994, 16, 127–134.
11. E. E. Urena-Benavides, G. Ao, V. A. Davis and C. L. Kitchens, Rheology and Phase Behavior of Lyotropic Cellulose Nanocrystal Suspensions, Macromolecules, 2011, 44, 8990–8998.
12. E. D. Cranston and D. G. Gray, Morphological and Optical Characterization of Polyelectrolyte Multilayers Incorporating Nanocrystalline Cellulose, Biomacromolecules, 2006, 7, 2522–2530.
13. H. Qi, K. E. Shopsowitz, W. Y. Hamad and M. J. MacLachlan, Chiral Nematic Assemblies of Silver Nanoparticles in Mesoporous Silica Thin Films, J. Am. Chem. Soc., 2011, 133, 3728–3731.
14. K. E. Shopsowitz, H. Qi, W. Y. Hamad and M. J. MacLachlan, Free-standing Mesoporous Silica Films with Tunable Chiral Nematic Structures, Nature, 2010, 468, 422–425.
15. S. Shafiei-Sabet, W. Y. Hamad and S. G. Hatzikiriakos, Rheology of Nanocrystalline Cellulose Aqueous Suspensions, Langmuir, 2012, 28, 17124–17133.
16. J. Han, C. Zhou, Y. Wu, F. Liu and Q. Wu, Self-assembling Behavior of Cellulose Nanoparticles during Freeze-drying: Effect of Suspension Concentration, Particle Size, Crystal Structure, and Surface Charge, Biomacromolecules, 2013, 14, 1529–1540.
17. D. Liu, P. R. Chang, M. Chen and Q. Wu, Chitosan Colloidal Suspension Composed of Mechanically Disassembled Nanofibers, J. Colloid Interface Sci., 2011, 354, 637–643.
18. D. Liu, Q. Wu, H. Chen and P. R. Chang, Transitional Properties of Starch Colloid with Particle Size Reduction from Micro-to Nanometer, J. Colloid Interface Sci., 2009, 339, 117–124.
19. D. Liu, X. Chen, Y. Yue, M. Chen and Q. Wu, Structure and Rheology of Nanocrystalline Cellulose, Carbohydr. Polym., 2011, 84, 316–322.
20. S. Beck, J. Bouchard and R. Berry, Controlling the Reflection Wavelength of Iridescent Solid Films of Nanocrystalline Cellulose, Biomacromolecules, 2011, 12, 167–172.
21. M. N. V. Ravi Kumar, U. Bakowsky and C. M. Lehr, Preparation and Characterization of Cationic PLGA Nanospheres as DNA Carriers, Biomaterials, 2004, 25, 1771–1777.
22. S. Feng and G. Huang, Effects of Emulsifiers on the Controlled Release of Paclitaxel (Taxol) from Nanospheres of Biodegradable Polymers, J. Controlled Release, 2001, 71, 53–69.
23. X. M. Dong and D. G. Gray, Induced Phase Separation in Cellulose Nanocrystal Suspensions Containing Ionic Dye Species, Langmuir, 1997, 13, 2404–2409.
24. X. M. Dong, J. F. Revol and D. G. Gray, Effect of Microcrystallite Preparation Conditions on the Formation of Colloid Crystals of Cellulose, Cellulose, 1998, 5, 19–32.
25. J. Majoinen, E. Kontturi, O. Ikkala and D. G. Gray, SEM Imaging of Chiral Nematic Films Cast From Cellulose Nanocrystal Suspensions, Cellulose, 2012, 19, 1599–1605.
26. S. Kinoshita, S. Yoshioka and J. Miyazaki, Physics of Structural Colors, Rep. Prog. Phys., 2008, 71, 076401.
27. H. de Vries, Rotatory Power and Other Optical Properties of Certain Liquid Crystals, Acta Crystallogr., 1951, 4, 219–226.
28. G. S. Chilaya and L. N. Lisetski, Cholesteric Liquid Crystals: Physical Properties and Molecular-statistical Theories, Mol. Cryst. Liq. Cryst., 1986, 140, 243–286.
29. A. C. Neville, Biology of Fibrous Composites; Development beyond the Cell Membrane, Cambridge: The University Press, 1993.
30. R. H. Marchessault, F. F. Morehead and N. M. Walter, Liquid Crystal Systems From Fibrillar Polysaccharides, Nature, 1959, 184, 632–633.
31. J. F. Revol, H. Bradford, J. Giasson, R. H. Marchessault and D. G. Gray, Helicoidal Self-Ordering of Cellulose Microfibrils in Aqueous Suspension, Int. J. Biol. Macromol., 1992, 14, 170–172.
32. A. Viswanathan and S. G. Shenouda, The Helical Structure of Cellulose I, J. Appl. Polym. Sci., 1971, 15, 519–535.
33. Y. Nishiyama, J. Sugiyama, H. Chanzy and P. Langan, Crystal Structure and Hydrogen Bonding System in Cellulose Iα From Synchrotron X-ray and Neutron Fiber Diffraction, J. Am. Chem. Soc., 2003, 125, 14300–14306.
34. H. Hou, A. Reuning, J. H. Wendorff and A. Greiner, Tuning of the Pitch Height of Thermotropic Cellulose Esters, Macromol. Chem. Phys., 2000, 201, 2050–2054.
35. G. Chilaya, Induction of Chirality in Nematic Phases, Rev. Phys. Appl., 1981, 16, 193–208.

---