The method of cultivation of Pennisetum hydridum using municipal sludge sewage was the same as that in our previous work (Jiang et al., 2019). Stalks of Pennisetum hydridum were harvested after 6 months. The Pennisetum hydridum stalks were ground and sifted after equilibrium moisture was achieved. The powder of Pennisetum hydridum that passed a 40-mesh sieve but was retained on a 60-mesh sieve was collected and put into a storage jar to be kept at room temperature.

Lignin of the powder of Pennisetum hydridum was removed by the traditional sodium chlorite method. The steps were as follows. An amount of 10 g Pennisetum hydridum powder was loaded into a 500 ml beaker, and 325 ml distilled water was added at a ratio of 1:32.5. The mixture was placed in a thermostat water bath at 75°C for 4 h, an amount of 3 g sodium chlorite was added every 1 h, and then glacial acetic acid was added to adjust the pH to 4.5. The above steps were repeated four times until the sample became white. The sample was placed in an extractor and cleaned with distilled water, repeatedly, until the filtrate pH was 7. The obtained delignified Pennisetum hydridum fiber was dried in an oven at 50°C.

Lignin and pentosane (hemicellulose) of the original Pennisetum hydridum and delignified sample were determined according to GB/T 2677.8-1994 and GB/T 2677.9-1994, respectively. The cellulose content was determined by the nitrate method.

An amount of 3 g delignified Pennisetum hydridum fiber was mixed with sulfuric acid (30 ml) with a certain mass fraction (40, 55, and 65% wt), and the reaction was carried out in a Thermostat Ultrasonic Cleaner (SB-5200 DTD, Xinzhi Biological Technology Co., Ltd., Ningbo, China) at 40°C, 55°C, 70°C, and room temperature (RT, 25°C) for a certain period (50 min, 120 min). The frequency and power of the ultrasound were 40 Hz and 50 kW. The reaction product was centrifuged using a centrifuge (LL5-2A, Beijing Medical Centrifuge Factory, China) for 15 min at 5,310 × g, and the supernatant was removed. The supernatant was repeatedly centrifuged several times until the pH of the solution was about 6, and the supernatant became turbid. A certain amount of dried solid samples were taken to obtain CNCs.

2 ml CNC sample solution was taken, diluted 100 times using deionized water, dispersed with ultrasonic treatment, and dropped into the sample pool. The particle size and zeta potential were measured using the Malvern laser particle analyzer (Nano ZS90, Malvern Instruments Co., Ltd., United Kingdom). The reported results were the average of two measurements.

The obtained CNC suspension was dried to a constant weight at 50°C for at least 8 h in an oven and weighed after drying. The yield of CNCs was calculated as follows: Y = m 2 m 1 × 100 % , (1) where Y is the yield of CNCs, %; m₂ is the weight of CNCs after drying, g; and m₁ is the weight of the original Pennisetum hydridum, g.

The Pennisetum hydridum fiber was fixed on the observation table with copper adhesive conductive tape. A small amount of powder was sprinkled on the conductive adhesive and plated with gold. SEM imaging (XL-30-ESEM, FEI Corporation, Holland) was conducted using a gaseous secondary electron detector (GSED) at a temperature of about 23°C, accelerating voltage of 10–20 kV, and pressure of 800 Pa in the sample chamber. SEM images at different scales of 200, 100, and 10 µm were obtained, and their contrast and brightness were adjusted.

The CNC solution was dispersed evenly by ultrasonic treatment (power 600 W, shaking time 2 s, interval time 1 s, and duration of shaking 10 min), 10 μL solution was absorbed and dipped onto a carbon film copper mesh (200 mesh), and the excess liquid was absorbed by a clean filter paper. After a few minutes, a 10 μL 3% phosphotungstate stain (pH 7) was adsorbed and dropped onto the copper net for negative staining for 5 min, the missing liquid was absorbed by a clean filter paper, and then the sample was dried naturally. The sample was observed using TEM under 80 kV acceleration voltage (Tecnai 12, FEI Corporation, Holland). 50 nanofibers were measured for the morphological analysis.

The amounts of 1 mg CNC powder and 100 mg KBr were mixed and ground, and the mixture was pressed into a high-transparency tablet with a diameter of 13 mm by the tablet pressing method and then tested by FTIR (Vertex 70, Bruker Corporation, Germany) in the range of 4,000–500 cm⁻¹, with an accumulation of 64 scans and a resolution of 4 cm⁻¹.

The crystal structure of CNCs was analyzed by XRD (Ultima IV, Rigaku Corporation, Japan). The analyzing conditions were Cu-Kα source, 40 kV, and λ = 0.154. The scanning range was 5°–40°, with a scanning step width of 0.020° per scan. The crystallinity (CrI) was calculated by the Segal empirical formula (Chen et al., 2018) as CrI,% = I 002 - I am I 002 × 100 , (2) where I₀₀₂ is the maximum diffraction intensity of the cellulose (002) crystal plane and I_am is the diffraction intensity of the non-crystalline zone between the 002 peak and the 101 peak.

An amount of 10 mg sample was taken and tested by a thermogravimetric analyzer (STA449 F3, NETZSCH Corporation, Germany). The analysis conditions were nitrogen atmosphere, a room temperature of 600°C, and a heating rate of 10°C/min.

The basic chemical components of Pennisetum hydridum before and after delignification are shown in Table 1. The cellulose content of Pennisetum hydridum after delignification reached 50.9%, which showed Pennisetum hydridum could be a good material for the production of nanocellulose. The lignin content of Pennisetum hydridum after delignification became very low. The hemicellulose content of Pennisetum hydridum after delignification became 24.1%.

The flowchart of the CNC production process used in this work is presented in Figure 1. The preparation method of nanocellulose is mainly to decompose large cellulose molecules into smaller nanocellulose molecules through physical crushing and chemical deconstruction (Du et al., 2018), which is known as the top-down approach (Nechyporchuk et al., 2016). Sulfuric acid hydrolysis is the earliest inorganic acid hydrolysis method of CNCs. Later, there have been studies to prepare CNCs by hydrolysis of hydrochloric acid, phosphoric acid, and other inorganic acids (Kontturi et al., 2016; Espinosa et al., 2013; Vanderfleet et al., 2018).

Three treatment variables (treatment temperature, sulfuric acid mass fraction, and treatment time) were set in this study. After sulfuric acid hydrolysis for a certain period, the amorphous structure of cellulose was decomposed to obtain cellulose with a smaller size (Liu et al., 2017). As shown in Tables 2, 3, 4, the acid-hydrolyzed products under all conditions reached the nanometer size, and the mean particle size of CNCs was the smallest under RT, 65% wt H₂SO₄, and 120 min (Figure 2A; Table 4). When the other treatment conditions were the same, the longer the acid hydrolysis time, the smaller the particle size of the CNC. In general, the CNC prepared at room temperature had the smallest particle size compared to that at other temperatures when the other treatment conditions were the same. The particle size of the CNC was highest under 40°C and 65% H₂SO₄, as shown in Table 2. This may be caused by overreaction resulting in the partial flocculation of glucose generated by decomposition and carbonization of cellulose.

Different colors of production are shown due to the difference in the mass fraction of sulfuric acid (Figure 2C). When the temperature reaches 70°C and the mass fraction of sulfuric acid ≥ 55%, the product carbonizes and the color turns black (Figure 2B), so the measurement results of the product obtained at 70°C are not listed.

The zeta potential is used to measure the strength of repulsion or attraction between particles, which is an important parameter of the characterization of colloid system stability (Qin, 2018). Usually, the greater the zeta potential, the greater the stability of the colloid system. In most of the colloid systems, 30 mV is called the stability threshold. The zeta potential is higher than the threshold; strong electrostatic repulsion can prevent particles from getting close, thus increasing their stability (Qin, 2018). If the zeta potential is between −15 and 15 mV, the gel will condense. Due to the addition of sulfuric acid, the resulting nanocrystals have sulfuric acid ester groups on the surface, which is negatively charged (Xu et al., 2016). It can be seen from Table 2 that the zeta potential of CNCs obtained under all conditions is negative, and the absolute value is above 15 mV, which indicates that the prepared CNCs have good stability.

As shown in Table 4, the CNC yield was higher when the mass fraction of sulfuric acid was 65%, and the CNC yield increased with the increase of treatment time under the same mass fraction of sulfuric acid. When the temperature rises to 40°C, the yield of CNCs increases first and then decreases with the increase of sulfuric acid mass fraction. When the temperature is 55°C, the yield of CNCs increases first and then decreases with the increase of sulfuric acid. This is because of the different results produced by the interaction of temperature, sulfuric acid mass fraction, and time (Tang et al., 2011).

Although CNCs can be prepared from various cellulose sources and hydrolysis conditions, the conditions are not ideal, the yield and charge content may be very low, or the agglomeration may be caused by large CNC particles (Vanderfleet et al., 2018). It is generally concluded from previous studies that, by increasing time, temperature, and acid concentration, hydrolysis would be more intense, CNCs with a higher sulfate content and smaller size would be produced, and the yield would be higher (Beck-candanedo et al., 2005). However, if the intensity of hydrolysis exceeds a certain point, the crystallinity of CNC would be damaged, cellulose would be degraded into sugar, and furfural would be formed further (Wang, 2014), which would also lead to the decrease of CNC yield and the increase of by-products (Agarwal, 2015). In this study, CNCs were not obtained when hydrolyzed at 70°C with a higher concentration of sulfuric acid, which indicated that hydrolysis was too severe and the experimental conditions should be adjusted.

From images of SEM, it can be seen that the length and width of the Pennisetum hydridum fiber were within 100 μm and about 10 μm, respectively (Figure 3). The average particle size of CNCs was around 400 nm under different acid hydrolysis conditions (Tables 2, 3, 4), and the difference was not big. Therefore, a few samples were observed by TEM (Figures 2A,B: RT, 40% wt, and 120 min; Figure 2C: RT, 55% wt, and 120 min; Figure 2C: RT, 65% wt, and 120 min). From Figure 4, it can be seen that the fiber length of CNCs was within 500 nm and the diameter was within 10 nm. Under the conditions of sulfuric acid concentration of 40 and 55% wt, there were some fiber bundles formed between celluloses, and it was not easy to observe a single CNC. It can be seen that CNCs at the concentration of 65% wt sulfuric acid had the smallest length, were relatively uniform, and had the best dispersion. This might be due to the degradation of more amorphous cellulose of fiber as the acid concentration increases. Mukherjee (1953) was the first person to use TEM to obtain images of CNCs from sulfuric acid hydrolysis and observed rod-like particles of about 200 nm length and 10–20 nm diameter, which were consistent with the TEM images of CNCs prepared in this study (Figure 4).

The chemical structure analysis of the delignified/bleached Pennisetum hydridum fiber and CNCs prepared at different temperatures was carried out by FTIR (Figure 5). The absorption bands and the corresponding structure assignments from the infrared spectra are based on literature values (Chen et al., 2013). From Figure 5, it can be seen that there was a major peak near 3,390 cm⁻¹ with the presence of -OH and C-H absorption peaks near 2,918 cm⁻¹ and a C-O absorption peak near 1,060 cm⁻¹. These main characteristic peaks did not change significantly, indicating that the CNC still had the basic structure of cellulose after acid hydrolysis.

The appearance of the asymmetrical S=O vibration associated with the C-O-SO₃ group in the CNC samples was approximately 1,250 cm⁻¹ and the appearance of the symmetrical C-O-SO₃ vibration was approximately 833 cm⁻¹ (Gu et al., 2013). There were some bands between 750 and 1,000 cm⁻¹ and other bands around 1,350 and 1,175 cm⁻¹, which indicated the presence of sulfonates in the CNC samples (Morais et al., 2013).

The main feature of the CNC samples was the appearance of the band at 1,730 cm⁻¹ related to carbonyl groups (C=O) from hemicellulose or esterification of cellulose (Liu et al., 2017). The band of CNCs prepared at 55°C at 1,163 cm⁻¹ assigned to C-O in lignin and xylan disappeared, which demonstrated that the lignin or hemicellulose was reduced during the acid hydrolysis procedure with the increasing reaction temperature.

The relative intensity of the band at 1,100 cm⁻¹ related to crystalline cellulose was increased in the CNC samples; however, the relative intensity of the band at 900 cm⁻¹ related to amorphous cellulose was decreased. It demonstrated that the ratio of crystalline to amorphous cellulose significantly increased after acid hydrolysis.

Cellulose is composed of the crystalline zone and amorphous zone. The percentage of the crystalline zone as a whole of cellulose is crystallinity. Above cellulose’s microproperties, it is very important to study the crystallinity of CNCs. As shown in Figure 6, the 002 crystal plane diffraction peak of cellulose I appeared near 22°. The diffraction angle overlapped each other between 101 and 10Ī crystal planes of cellulose I forming the broad diffraction peak during 14.5–17°. It is not hard to find that the peak patterns of the CNC and delignified Pennisetum hydridum fiber were the same after sulfuric acid hydrolysis. They retained the crystal structure of cellulose I and have been changed. This is because the amorphous region of cellulose is damaged during sulfuric acid hydrolysis, which is consistent with the results of FTIR. The CrI of CNCs increased with the increase of the mass fraction of sulfuric acid due to the hydrolysis of amorphous cellulose. Compared with that of the Pennisetum hydridum fiber, the crystallinity of the CNC prepared with the mass fraction of sulfuric acid of 40, 55, and 65% wt increased by 6.79, 18.21, and 18.34%, respectively, and the maximum crystallinity reached 62.59% when H₂SO₄ was 65% wt.

During the heating process, the structure of cellulose will be changed and the crystalline zone will be destroyed and even carbonized, finally. The relationship between the sample quality and the temperature change was tested by a thermal analyzer. As shown in Figure 7, when the temperature was below 200°C, that stage was mainly the process of water evaporation. TG curves of the CNC and Pennisetum hydridum fiber showed a small weight loss ratio, but the CNC (−6.06%) had more weight loss than the Pennisetum hydridum fiber (−4.02%) because of its better absorption. The stage of 250–400°C was the process of mass weight loss, and the structure of cellulose was decomposed into various volatile substances. The CNC first exhibited a degradation peak (291.1°C); this is because the presence of sulfur ester groups accelerated the pyrolysis process. When the temperature increased above 400°C, the degradation of cellulose was finished, and the residual masses of the CNC and Pennisetum hydridum fiber were equal. The CNC exhibited lower thermal stability compared to the Pennisetum hydridum fiber. This is consistent with the research results of Xu (2016), in which the thermal stability of nanocellulose obtained from palm sheath is significantly lower than that of raw materials. The main reason for this might be linked to disruption of its crystalline structure and the introduction of sulfate groups during sulfuric acid hydrolysis. Moreover, smaller fiber dimensions, leading to higher surface areas exposed to heat, also have negative effects on thermal stability (Widsten et al., 2014).

Thermal stability is also an important parameter that is often considered in applications of materials. Kargarzadeh (2012) studied the influence of hydrolysis conditions on the thermal stability of CNCs obtained from kenaf bast fibers using sulfuric acid. A continuous and progressive decrease in the thermal stability of the nanoparticles occurred as the hydrolysis time increased, probably because of the high sulfation rate demonstrated by the zeta potential measurements (Kargarzadeh et al., 2012). Although reducing the hydrolysis time could improve thermal stability, the size of CNCs produced was larger and the colloid stability was lower. The experimental results of this study also indicated that the size of the CNC was smaller when the hydrolysis time was longer because of degradation of the amorphous region.