MCC yield and elemental compositions
The MCC yield obtained by this study from extraction of CCW was 83%, which was higher than that of the previous reports, which also used hydrochloric acid hydrolysis. For example, MCC from sugarcane bagasse with 55% yield , MCC from brown algae with 69.5% yield , and MCC from date seeds with 12.5% yield . The high cellulose component in CCW might be responsible for the improvement of the MCC yield. This shows that CCW has outstanding potential in recovering valuable cellulose material. Elemental compositions of samples are shown in Table S1. The percentage of carbon, hydrogen, and oxygen elements were determined in MCC, CMCC, TCC, and the CCW sample. The sulfur and nitrogen contents were not detected in any analyzed samples. The results indicated that all samples contained the typical characteristic of cellulose. As indicated in Table S1, the percentage of elemental compositions of MCC was similar to CMCC, implying successful extraction of MCC from CCW.
Surface characterization by morphological analysis
The macro photographs and SEM images of CCW, TCC, and MCC are shown in Fig. 1. As shown in Fig. 1a and b, the TCC has a bright white color compared to cotton waste which containes many dyed colors, implying the effectiveness of the pretreatment process in removing dyed fabric color. The CCW was observed as a ribbon-like shape with a smooth and uniform surface (Fig. 1d). After the pretreatment process, the fibers of TCC became swollen and appeared on a rough surface (Fig. 1e). The findings indicated that pretreatment could effectively eliminate the impurity components in CCW, such as hemicellulose, wax, and pectin . Then, TCC was subject to hydrochloric acid and resulted in a fine and white powder (Fig. 1c), which were observed as short cylindrical microcrystal morphology (Fig. 1f). This is because hydrolysis reaction formed hydronium ions (H3O+) to attack the amorphous regions and broke the β-1,4-glycosidic bonds between two anhydroglucose units from long cellulose chains, resulting in the formation of smaller fragments .
Crystalline speciation identification
The XRD spectra of MCC, TCC, CCW, and CMCC are presented in Fig. 2. There are four similar characteristic peaks by all samples at 14.8, 16.6, 22.6, and 34.5o, corresponding to the crystal planes of (1–10), (110), (200), and (004) of cellulose I structure . The XRD diffraction displays that MCC remains intact cellulose I structure during chemical treatments. The result was in agreement with another work reported by Shao et al. , who prepared MCC using hydrochloric acid hydrolysis of the corncob. The authors reported that corncob-MCC possessed cellulose I and did not convert it into cellulose II following the chemical treatment process . As shown in Fig. 2, the characteristic peak at around 22.6o was sharper and narrower from CCW to MCC, implying that the obtained MCC grasped the highly ordered crystals and an increase of CrI . This is because hydrochloric acid hydrolysis could remove the amorphous region of CCW and break the glycosidic linkages, resulting in releasing crystallites and increasing the CrI value .
The CrI of MCC (95%) obtained by this work is higher than that of oil palm MCC (87%) , giant reed MCC (79%)  and pomelo peel MCC (41%) . The finding can be related to the low amorphous content of CCW which is easily eliminated by chemical treatments. In addition, the CrI of MCC is found higher than that of CMCC, indicating the extraction process is acceptable and the obtained MCC is a suitable cellulose source for CNC production.
Spherical CNC characterization
Acid to cellulose ratio effect on morphological CNC
Figure 3 displays the TEM images of isolated CNC with various acid to cellulose ratios (CNC30, CNC40, and CNC50). Rod-like CNC was observed mainly in CNC30 (Fig. 3a) and CNC40 (Fig. 3b). The average length and width of CNC30 were 53 ± 19 nm and 6.6 ± 1.3 nm, respectively; the average length and width of CNC40 were 49 ± 13 nm and 4.3 ± 2.0 nm, respectively. However, when the acid to cellulose ratio further increased to 50 mL g− 1, the CNC morphology turned to be spherical CNC (CN50), with a small diameter of 14 ± 4 nm (Fig. 3c). Spherical CNC is a spherical or corn-like shaped nanoparticle . These results are consistent with previous study done by Xu et al. , in which when the acid to cellulose ratio increased from 9.5 to 10.5 mL g− 1, the morphology of CNC became spherical CNC with an average diameter of 3.4 nm. The morphology of spherical CNC obtained in this research is similar to those of spherical CNC reported in other literature [10, 16]. The crystalline structure of spherical CNC is illustrated in HR-TEM results (Fig. 3d). Additionally, selected area diffraction (SAED) patterns and crystalline lattices of spherical CNC were also examined, as shown in Fig. 3c-e. The inset image in Fig. 3c displays the SAED pattern obtained from the blue circle. The SAED pattern indicated that spherical CNC was a single crystal, and the weak ring pattern was observed due to the appearance of residual amorphous regions. Figure 3e shows crystalline lattice of spherical CNC, which collected from the yellow dashed square in Fig. 3d and produced by an inverse fast fourier transform of a masked fast fourier transform (inset image in Fig. 3e). As shown in Fig. 3e, the (200) plane was confirmed with lattice-spacing of 3.9 Å, corresponding to the cellulose I structure .
The increase of acid to cellulose ratio investigated by this work is found to be successful in forming spherical CNC. The formation mechanism of different CNC morphologies could be derived from the effects of acid to cellulose ratio related to H3O+ produced from the acid hydrolysis process and could attack glycosidic linkages and disintegrate the amorphous regions in the cellulose matrix . Besides, the crystalline regions could retain essentially intact and thus the rod-like CNC was obtained. Furthermore, increasing the acid to cellulose ratio can result in more active hydronium ions in a given amount of cellulose, which tends to hydrolyze the semi-crystalline or amorphous regions faster and more intensely. The result is shorter rods obtained with higher crystallinity and higher-ordered structure. Then, these shorter rods tend to self-assembly via interfacial hydrogen bond and ultimately form the spherical CNC .
Table 1 shows the spherical CNC (CNC50) diameter in this study (14 nm) was comparable to other works using sulfuric acid hydrolysis process; such as spherical CNC obtained from baby diaper waste (10–20 nm)  and sago seed shells (10–15 nm) . Furthermore, the nanoparticle diameter of CNC50 is also found much smaller than that of previous researches where the spherical CNC diameter of 30–40 and 40 nm using ultrasound-assisted sulfuric acid hydrolysis from empty fruit bunch fiber  and compound enzymatic hydrolysis from eucalyptus pulp , respectively. In general, the spherical with smaller size has remarkable potential in applications, such as drug delivery  and cellulose uptake .
CNC yield and ZP
The yield, sulfur compositions, and ZP of CNC samples are shown in Table S2. When the acid to cellulose ratio increased from 30:1 to 50:1 mL g− 1, the yield of CNC significantly decreased from 52 to 31%. This is because the increase in acid to cellulose ratio means decreasing the amount of cellulose of the hydrolysis process, resulting in decreased CNC yield. Also, the increase of acid to cellulose ratio could promote the degree of cellulose hydrolysis. A similar observation was also reported in the previous study of Xu et al.  who indicated that the yield of CNC markedly dropped off when the acid to pulp ratio increased from 8.5:1–10.5:1 mL g− 1.
Although the yield of spherical CNC obtained was not as high as expected, this result reflects that the spherical CNC can successfully be prepared by increasing the acid to cellulose ratio without ultrasound during the hydrolysis process. In previous research, the spherical CNC was prepared from cotton-derived cellulose by Xiong et al.  using ultrasound during sulfuric acid hydrolysis of waste cotton fabric with 21.5% yield and Meyabadi et al.  applying enzymatic hydrolysis and sonication treatment of waste cotton fiber with 20% yield. The limitations of ultrasound treatment and enzymatic hydrolysis in spherical CNC preparation are low yields. Therefore, the spherical CNC from CCW with a yield of 31% obtained in this study can consider as a prospective method for yield improvement of spherical CNC.
The acid to cellulose ratio significantly influenced the sulfate groups on the CNC surface (Table S2). The increase of acid to cellulose ratio could intensify the degree of esterification of the hydroxyl groups and the substitution of these hydroxyl groups by sulfate groups, resulting in grafting more negatively charged sulfate groups on the CNC surface. If nanocellulose has a highly negative charged surface, it can avoid agglomeration and make a stable suspension due to electrostatic repulsion generated between the nanoparticles .
The stability of CNC suspension was determined via ZP analysis. ZP is related to the charge present that nanoparticles have. If the absolute ZP value of CNC suspension is higher than 30 mV, the nanoparticles have enough charge to repulse each other and resist forming agglomeration . Furthermore, the ZP values obtained are proportional to the sulfate group’s density (Table S2), and an absolute ZP value of CNC50 (46.8 mV) was higher than that of CNC30 and CNC40. This could suggest that CNC50 with spherical nanocellulose disperses more stable in the aqueous phase than rod-like CNC of CNC30 and CNC40.
Moreover, the obtained ZP of spherical nanocellulose was much higher than compared to that reported by Naduparambath et al.  preparing spherical nanocellulose with an absolute ZP value of 37.8 mV using sulfuric acid of sago seed shells. Xu and Chen  reported that the spherical CNC obtained from composite enzymolysis of pulp fibers having an absolute ZP value of 28.2 mV. This means that the spherical nanocellulose prepared from sulfuric acid exhibits better dispersion than the enzymatic hydrolysis method. The nanocellulose should generally possess high ZP, which resists the agglomeration and maintains stable dispersibility in the solution.
Figure 4 shows FTIR spectra of CCW, TCC, MCC, CNC30, CNC40, and CNC50. Peaks at 3340 and 2900 cm− 1 correspond to O-H stretching vibration and C-H asymmetric stretching vibration . The peaks around 1505 and 1596 cm− 1 were assigned to lignin aromatic skeletal vibration C=C bonds. The C-O-C stretching band at 1260 cm− 1 could present the lignin ether linkages. However, these bands were not observed in the FTIR analysis, indicating that the lignin composition does not appear in all samples, and the result is consistent with the published works [12, 16]. The peak at 1429 cm− 1 is related to the bending vibration of -CH2-. The peaks observed at 1160 and 1110 cm− 1 are assigned to the bending vibration of C-C and C-O-C within the anhydroglucose ring, respectively , indicating that the degradation of cellulose was unlikely to attack glucose rings, whereas the degradation of cellulose should be at β-1,4-glycosidic ring linkage . Some small peaks at 1052 and 896 cm− 1 correspond to C-O-C pyranose ring vibration and C-H vibration in cellulose, respectively [40, 41]. These bands found in all spectra are related to cellulose I structure, showing that the cellulose components remain during pretreatment and acid hydrolysis. The result agrees with Dungani et al.  who extracted CNC from oil palm fronds via sulfuric acid.
Although the cellulose structure did not change, the peak intensity of the obtained cellulose samples markedly changed. The intensity of O-H vibration around 3340 cm− 1 for MCC, CNC30, CNC40, and CNC50 was lower than that of CCW and TCC. This is because H3O+ generated in acid hydrolysis reaction weakened intermolecular hydrogen bonds and broke the cellulose chains, which might be responsible for decreasing hydroxyl groups on the cellulose surface. For CNC samples, the depletion of their surface hydroxyl groups is also due to the contribution of esterification, which replaces these hydroxyl groups with sulfate groups . Observing the intensities of C-O-C (around 1052 cm− 1) and C-H (around 896 cm− 1) vibrations in MCC, these intensities were higher than that of CCW and TCC. The result indicated that the non-cellulosic components could be removed effectively in the chemical treatment process and enhance the percentage of cellulose . However, the intensity of these infrared bands in CNC samples showed a noticeable decrease and was proportional to the increase in the acid to cellulose ratio. The findings show that the acid to cellulose ratio affects the magnitude of the hydrolysis reaction on the cellulose structure, resulting in a decrease in the amount of cellulose, consistent with yield results.
CNC crystalline speciation identification
The XRD patterns of rod-like CNC (CNC30 and CNC40) and spherical CNC (CNC50) are shown in Fig. 5. The CrI % and crystallite size are also presented in the inset of Fig. 5. The characteristic peaks were displayed at around 14.7, 16.6, 22.7, and 34.5o, corresponding to the (1–10), (110), (200), and (400) crystal planes, respectively. These planes indicate that rod-like CNC and spherical CNC have the characteristics of cellulose I. The findings are consistent with other researches where diffraction peaks of rod-like CNC and spherical CNC are similar [9, 16, 20]. The results also indicate that the structure characteristics of cellulose I remain intact with increasing ratio of sulfuric acid to cellulose, consistent with FTIR analysis. The CrI of CNC samples did not improve in comparison with that of MCC after sulfuric acid hydrolysis. The findings show that the hydrolysis reaction in the CNC synthesis not only degrades the amorphous regions but also destroys some crystalline domains . Moreover, the inset (Fig. 5) shows that the CrI values of CNC is stable when the acid to cellulose ratio increased from 30:1 to 40:1 mL g− 1. The finding was similar to Xu et al.  who indicated that the CrI values of CNC were not significant different as the acid to cellulose ratio increased from 8.5 to 9.5 mL g− 1. However, when the ratio was further increased to 50:1 mL g− 1, the CrI of CNC50 appeared to increase. The results could be derived from the different morphologies of CNC, since CNC50 with spherical nanocellulose can possess a more organized and crystalline structure which might be responsible for the higher crystallinity index of CNC50 .
Crystallite size is another crucial parameter relating to the crystal structure of cellulose. The previous reports mentioned that the increase of crystallite size could result from the increase of CrI [12, 15]. In this study, the CrI of CNC50 is higher than that of CNC30 and CNC40 and consequently larger crystallite size (as shown in inserted table). The crystallite sizes obtained were similar to Maciel et al.  preparing CNC samples with a crystallite size of 7.44 nm through sulfuric acid hydrolysis of industrial textile cotton waste. It is worth mentioning that CNC50 with spherical nanocellulose possesses a high CrI, which could improve the mechanical properties (rigidity, tensile strength) of nanocellulose-based composites .
The CrI of spherical nanocellulose (CNC50) is approximately 95%, greater than those of results published by others (Table 1). For example, Tong et al.  used compound enzymatic hydrolysis to prepare spherical nanocellulose with CrI of 91.2%. The results indicated that the spherical CNC with high CrI could be manufactured from CCW via the adequate sulfuric acid to cellulose ratio.
Thermal analysis in this study focused on the thermal degradation behavior on CNC samples with different morphologies, including rod-like CNC30 and spherical CNC50. The result of the thermal properties of CNC obtained in Thermogravimetry (TG) and Derivative Thermogravimetry (DTG) curves are displayed in Fig. 6 and Table S3. Thermal analysis of spherical CNC and rod-like CNC was investigated at different heating rates of 5, 10, 20, and 40 °C min− 1. The TG curves of spherical CNC (Fig. 6a) and rod-like CNC (Fig. 6c) showed an initial slight weight loss of only 2%, found in the temperature range of 100–120 °C due to the evaporation of absorbed water or low molecule compounds.
For spherical CNC, the degradation of cellulose is displayed in two main pyrolysis regions (Fig. 6a and b). The first stage in low temperature of 153–235 °C corresponded to the amorphous degradation regions, which were more accessible and sulfated . The second one at high temperature range of 260–448 °C was attributed to un-sulfated crystalline regions . This degradation process was also observed by Naduparambath et al.  who prepared spherical nanocellulose using sulfuric acid hydrolysis of sago seed shells. A similar thermal degradation behavior was observed for rod-like CNC, which also included two degradation stages (Fig. 6c and d). The first and second stages occurred between 150 and 260 °C and 262–450 °C, respectively.
The thermal stability of spherical CNC is expected to be higher than that of rod-like CNC. Onset decomposition temperature (Tonset), which is contributed by the thermal stability of a material, was higher for spherical CNC than the rod-like CNC (Table S3). This is because the CrI of spherical CNC was higher than that of rod-like CNC (as shown in XRD results) and consequently higher thermal stability. The results indicated that the CrI contributes positively to the initial thermostability improvement of spherical CNC .
However, the spherical CNC exhibited slightly lower thermal stability than the rod-like CNC in the main degradation regions (Table S3). The finding can be related to the presence of sulfate groups on the CNC surface. The previous studies suggested that the higher sulfate groups might be responsible for the lower thermal stability . Also, the Ea of spherical CNC is lower by introducing the negatively charged sulfate groups on the CNC outer surface. Another factor affecting spherical CNC thermal stability could suggest that increasing the acid to cellulose ratio could generate a higher H3O+ concentration to access the crystalline domain. As a result, part of the crystalline regions can be dissolved, resulting in its susceptibility to disintegration at high temperatures . In short, both the CrI and the sulfate group are essential factors for determining the thermal stability of spherical CNC.
The Ea is directly associated with the thermal stability of CNC. Hence, the investigation of Ea values could provide a better insight into the thermal degradation of spherical CNC and rod-like CNC. The correlation between log β and 1000/T using the FWO method at different conversion (α) values for spherical CNC (CNC50) and rod-like CNC (CNC30) is shown in Fig. 7. The lines illustrated nearly parallel in all samples, confirming approximate Ea values in various conversions . Table S4 indicates a strong correlation with the high value of R2 and in agreement with the previous research .
Both spherical CNC and rod-like CNC (CNC30) showed an increase in Ea values in the conversion range from 0.1 to 0.4, followed by decreased Ea values in the conversion range from 0.5 to 0.8 (Fig. 8). The Ea values in this study were superior to those reported by Henrique et al. , who prepared CNC from Kraft pulp using sulfuric acid hydrolysis with the Ea values range from 156 to 269 kJ mol− 1. The Ea values of spherical CNC were higher than that of rod-like CNC until conversion closer to 0.35. However, in the conversion ratio from 0.4 to 0.8, the Ea values of spherical CNC were lower than rod-like CNC. These findings demonstrated that the various reactions appeared simultaneously with the CNC degradation .
The CrI was considered as an influential factor on CNC’s Ea values . Observing in conversion ratio of 0.1 to 0.35, the rod-like CNC has lower Ea values than spherical CNC. This is because the low crystallinity preferred the dehydration reactions, resulting in reduced Ea values . However, after conversion higher than 0.4, the Ea values of spherical CNC were generally lower than that of rod-like CNC. This is because the more negatively charged sulfate groups are introduced into the spherical CNC surface. The results indicated that the more the number of sulfate groups in spherical CNCs, the lesser the Ea needed to be supplied to the degradation process .
Rights and permissions
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated in a credit line to the data.