Corncob and giant cane cut-up were obtained from the experimental farm ‘A. Menozzi' of the University of Milan (Landriano, Italy) (Calvo et al., 2018). These lignocellulosic wastes were first air-dried at 60°C (Memmert UF110plus, Schwabach, Germany) and then stored at 23.0 ± 0.5°C in a desiccator containing calcium chloride for 2 days before further processing. Garlic stalks (Allium sativum) were directly placed in a desiccator without any pretreatment. All lignocellulosic agro-wastes were then subjected to grinding using an M20 Universal mill (IKA^®-Werke GmbH & Co., Staufen, Germany) provided with a double-walled grinding chamber for internal cooling, at a fixed speed of 20,000 rpm for 10 min. To select the smaller fractions of milled agro-wastes, a 70-mesh metal sieve (particle size retention: >210 μm) was used. The dry-sieved powder was kept at 23.0 ± 0.5°C in a desiccator for 2 days before chemical extraction. All the adopted chemicals (VWR International S.r.l., Milan, Italy) were of reagent grade and used without further purification. For each lignocellulosic feedstock, the initial composition was drawn from the literature and is reported in Table 1.

The recovery of cellulose from lignocellulosic agro-wastes followed the extraction protocol used in previous works (Mondal and Haque, 2007; Mondal et al., 2015) with slight modifications intended to optimize each purification step (Figure 1). In the beginning, 2 g of lignocellulosic agro-wastes powder were dispersed into a 20% NaOH solution (w/v) at a solid-to-liquid ratio (S/L) of 1:100 (g/mL). The obtained mixture was kept under stirring at 300 rpm for 120 min (garlic stalks, and giant cane cut-up) or 180 min (corncob), at room temperature. The longer contact time adopted for corncob arose from the high abundancy of hemicellulose in the biomass (Table 1), thus requiring more severe conditions.

At the end of this first stage, the solid/liquid separation was achieved by means of a Büchner filtration setup using grade 230 Whatman^® filter paper (particles retention: 20–30 μm, Merck KGaA, Darmstadt, Germany). The solid residue was then washed thoroughly with 2% (v/v) acetic acid and hot distilled water (70°C) until the effluent became clear. The subsequent removal of the organic compounds was performed using pure xylene (S/L = 1:25 g/mL) under stirring at 500 rpm for 20 min at room temperature, followed by a washing step with technical-grade ethanol and subsequent filtration. As the last step, a 1% (w/v) NaClO₂ solution (S/L = 1:50 g/mL) buffered at pH 4 was used at 90°C and 500 rpm for 90 min (garlic stalks and giant cane cut-up) and 60 min (corncob) as both bleaching agent and to allow the removal of lignin residues possibly still present in the biomass.

At last, the so obtained cellulose was separated by filtering and washing using technical-grade ethanol and distilled water, and then air-dried at 60°C overnight before undergoing further characterization.

BC was used in this study for comparison purposes, that is, as a reference substrate to be compared with the cellulose extracted from the three different lignocellulosic biomasses. The production of BC followed the same procedure as described in our previous work (Rovera et al., 2018). Briefly, BC was synthesized by static fermentation (7 days at 30°C in a Hestrin and Schramm culture medium) using the Komagataeibacter sucrofermentans DSM 15973 strain (Leibniz Institute DSMZ-German Collection of Microorganisms and Cell Cultures, Braunschweig, Germany). BC pellicles were washed with deionized water and boiled in an alkaline solution (NaOH, 1M) for 30 min to remove the residual bacterial cells. After washing several times with distilled water, the cellulosic material was homogenized for 15 min with an Ultra-turrax^® T25 Basic homogenizer (Ika-Werke, Stanfen, Germany) at 12,000 rpm and finally freeze-dried at −55°C and 0.63 mbar for 24 h using an ALPHA 1-2 LDplus freeze dryer (Martin Christ, Osterode am Harz, Germany).

The conversion of macro-sized cellulose to cellulose nanocrystals was performed by acid hydrolysis, using both sulfuric and hydrochloric acid. Specifically, 12 g of freeze-dried cellulose were added to 88 mL of distilled water. Then, 1 g of the stock dispersion was added to the acid solution (15 g of sulfuric acid – H₂SO₄, 58% v/v or hydrochloric acid – HCl 33% v/v). To promote the penetration of the acid inside the cellulosic network, cellulose-based aqueous dispersions were mixed by Ultra-turrax at 8,000 rpm for 5 min before hydrolysis. The hydrolytic process with sulfuric acid was performed under magnetic stirring (400 rpm) at 55 ± 1°C for 120 min, while that involving hydrochloric acid was performed under gentle magnetic stirring at 60 ± 1°C for 240 min. The hydrolysis reaction was terminated by adding 14 mL of cold distilled water, followed by centrifugation at 10,000 rpm for 15 min up to four times utilizing a Frontier™ 5000 Series Multi centrifuge (Ohaus, Parsippany, USA). Between each centrifugation cycle, CNCs were collected at the bottom of the centrifuge tubes as pellets, while the acidic supernatant was discarded and replaced with fresh distilled water, followed by pellet redispersion at 3,000 rpm for 3 min. Finally, pH was adjusted to neutrality using a dialysis tube against deionized water, (MW cut off = 12,000) (Sigma-Aldrich, Milan, Italy). For comparison purposes, nanocrystals from bacterial cellulose (BCNCs) as a pure counterpart were also obtained using the same hydrolytic procedure as described for agro-waste feedstocks.

For all the samples, the yield of both the cellulose extraction (Y_C, Equation 1) and hydrolysis (Y_CNCs, Equation 2) processes was determined gravimetrically with an HS43S-MC halogen-lamp moisture-content analyzer (Mettler Toledo, Greifensee, Switzerland) set at 105°C.

In particular, m_{C.R., i} represents the mass of dry cellulosic residue extracted from the interest biomass (i = garlic stalk, corncob, and giant cane cut-up), m_{0, i} stands for the initial amount of dry biomass undergoing the extraction process (2 g), and m_{CNCs, i} is the dry mass of cellulose nanocrystals retrieved from the acid hydrolysis process of the i-th raw material.

Solid-state 13C Cross-Polarization Magic Angle Spinning (CP-MAS) spectra were collected at 125.76 MHz on an Avance 500 MHz NMR Spectrometer (Bruker Italia s.r.l., Milan, Italy), operating at a magnetic field of 11.7 T and equipped with a 4 mm MAS probe, spinning the sample at the magic angle up to 15 kHz, which, with the addition of high power 1H decoupling capability, allows the decrease or the elimination of homo and heteronuclear anisotropies. The chemical shifts were recorded relative to the glycine standard that was previously acquired (C=O signal: 176.03 ppm, relative to tetramethylsilane reference). All the samples were prepared by packing them in Zirconia (ZrO₂) rotors, closed with Kel-F caps (50 μL internal volume) and the MAS rate was optimized to 10 kHz, after having run some experiments in the range of 2–15 kHz. Cross-polarization (CP) spectra, under Hartmann–Hahn conditions, were recorded with a variable spin-lock sequence (ramp CP-MAS) and a Contact time (CT) of 1 ms, optimized into a range between 1 and 5 ms. The following acquisition conditions were also applied: repetition time (d1) = 2 s, 1H 90° pulse length (p1) = 3.4 μs, spectral width (sw) = 240 ppm (30303.03 Hz) and acquisition Time (AQ) = 46 ms.

The crystallinity index (Cr.I.) related to the percentage of crystalline domains existing in the cellulose fibrils of different cellulose-based samples was calculated considering the region of the C4 13C CPMAS signals associated with the carbon C4 of the glucose units in the cellulose polymer chain, by comparing the area of C4 peak, in the range 86–92 ppm (crystalline domain), with that of the up-field C4 signal, in the range 80–86 ppm, related to the less ordered (amorphous) domains in the cellulose fibrils (Focher et al., 2001; Vismara et al., 2009). The software used for the comparison of interest areas was “MestReNova 14.1.2” (Mestrelab Research S.L., Santiago de Compostela, Spain), and the Global Spectral Deconvolution (GSD) analysis algorithm (Focher et al., 2001), able to apply an automatic multiplet deconvolution to the whole spectrum and fully characterize each fit peak (chemical shift, peak width, area, intensity, etc.) was launched. The Cr.I. (%) was finally calculated according to Equation (3), as follows:

Where x represents the C4 crystalline area below the interest peak, while y stands for the C4 para-crystalline one. For clarity purposes, the GSD-derived plot obtained from the NMR spectrum of the giant cane cut-up derived cellulose is reported in Figure 2, with the insert pointing to the regions (i.e., C4 crystalline and C4 para-crystalline) addressed for the Cr.I. calculation.

Solid-state Fourier Transform Infrared Spectroscopy (FT-IR) analysis was performed using a Spectrum 100 instrument (Perkin Elmer Inc., Waltham, MA), coupled with an Attenuated Total Reflectance (ATR) accessory and a Ge/Ge crystal, fixed at an incident angle of 45°. All spectra were collected at a resolution of 4 cm⁻¹ over a wavenumber range between 4,000 cm⁻¹ and 800 cm⁻¹ and resulting from an average of 180 scans. A background scan of clean Ge crystal in air was acquired before each scan.

Basic information about the morphological features of cellulose was collected using a Nikon Eclipse ME600 Inspection optical microscope (Nikon Instruments, Sesto Fiorentino, Italy), equipped with a Nikon DS-Fi3 digital camera (Nikon Instruments, Sesto Fiorentino, Italy) using magnifications of 10 × , 20 × , and 50 × . Microscopic images were captured by NIS-Element software (Nikon Instruments, Sesto Fiorentino, Italy).

In-depth morphological and size features of CNCs were determined using a Tosca™ 400 AFM (Anton Paar Italia Srl, Rivoli, Italy) in contact resonance amplitude imaging (CRAI) mode. Toward this goal, an Arrow-FMR-10 Force Modulation probe (Nanoworld, Neuchâtel, Switzerland) featuring a rectangular cantilever with a triangular free end and a tetrahedral tip with a typical height of 10–15 μm was used. Additionally, the tip radius of curvature is ~10 nm. The cantilever has a spring constant and resonance frequency of 2.8 N m⁻¹ and 75 kHz, respectively. For the analyses, approximately 100 μl of diluted CNC water dispersion (~0.1 mg/mL) were dropped onto a mica disc substrate (Ted Pella Inc., Redding, California). Dimensional calculations and image editing were conducted via Tosca Analysis Software (version 7.30, Anton Paar, Graz, Austria). The mean values reported for each CNCs dimension were calculated over at least five images, from which at least ten measurements were taken.

The size distribution of CNCs was measured by photon correlation spectroscopy experiments using a dynamic light scattering (DLS) Nanotrac Flex In-situ Analyzer (MicroTrac GmbH, Krefeld, German). Tests were carried out at 25°C, with a stabilization time of 60 s, and using viscosity (0.8872 cP) and refractive index (1.33) of water as reference values. The software used the non-negative least squares algorithm to retrieve the size distribution.

The ζ-potential of CNCs aqueous solutions was investigated by electrophoretic light scattering (ELS) analysis, using a Litesizer^TM 500 (Anton Paar, Rivoli, Italy) system at neutral pH. For each sample, at least three replicates were performed.

Transmission electron microscopy (TEM) images of BCNCs were captured using a LEO 912 AB energy-filtering transmission electron microscope (EF-TEM, Carl Zeiss, Oberkochen, Germany) operating at 80 kV. Digital images were recorded with a ProScan 1K Slow-Scan CCD camera (Proscan, Scheuring, Germany). Samples for TEM analyses were prepared by drop-casting a few mL of a 0.2% wt. BCNCs dispersion onto Formvar-coated Cu grids (400-mesh) and letting the samples rest for 24 h at room temperature to allow water evaporation.

Data were analyzed using Statgraphics Plus 4.0 software (STSC, Rockville, MD, USA), and a one-way analysis of variance was used to check for differences among samples. The significance level (p) was fixed at 0.05. Unless otherwise stated, all the analyses were performed in triplicate.

Cellulose was extracted from the pristine biomasses through a cascaded chemical procedure, as schematically depicted in Figure 1. Each step of the proposed protocol had the main goal of sequentially removing all the molecules other than cellulose (i.e., hemicellulose, lignin, ash, and organic compounds). In total, this procedure took approximately 6 h. The thermogravimetric measurements revealed that the ultimate process yields were as high as 35.73 ± 0.95% DW_b, 37.15 ± 2.75% DW_b, and 39.10 ± 1.69% DW_b for garlic stalk, corncob, and giant cane cut-up samples, respectively (Table 2). These results clearly demonstrate the possibility of efficiently recovering cellulose from all the investigated matrices by a simple and fast chemical procedure, with ultimate yield values close to the theoretical ones (Table 1).

Information on the quality and chemical structure of cellulose extracted from garlic stalk, giant cane cut-up, and corncob was acquired by both FT-IR and 13C CP MAS NMR. FT-IR spectra are displayed in Figure 3. At first glance, it can be seen clearly that the individual patterns are very similar, suggesting an almost identical chemical composition of the biomass extracted from the three lignocellulosic wastes. More specifically, according to the assignments by Halib et al. (2012) and Atykyan et al. (2020), the characteristic absorption peaks of cellulose at ~3,400 cm⁻¹ and ~2,900 cm⁻¹ are attributed to the O-H and C-H stretching vibrations, respectively, while the absorption peak at 1,650 cm⁻¹ was attributed to the O-H deformational vibration of the absorbed water. The peak assigned to the C-H and C-O vibrations of the polysaccharide ring of cellulose is centered between 1,300 and 1,400 cm⁻¹, whereas the peak at 1,460 cm⁻¹ corresponds to CH₂ bending. The peaks associated with the vibration of C-O-C in the pyranose ring are centered at 1,178 and 1,060 cm⁻¹ (Halib et al., 2012; Atykyan et al., 2020). The subtle peak at ~900 cm⁻¹ was attributed to the β-1,4 bond vibrations. Finally, it is interesting to note the peak at 1,670 cm⁻¹, which has been ascribed to the C=O stretching vibration of the carbonyl and acetyl groups of the xylan components of hemicellulose (Fiore et al., 2014). This observation seems to suggest a possible contamination of the main cellulosic residue (Fiore et al., 2014). For comparison purposes, the FT-IR spectra of BC and filter paper were reported in Figure 3. Also in this case, there is a clear correspondence of the different spectra obtained from the lignocellulosic biomass and the spectra of BC and commercial cellulose. The different shape of the peak centered at 3,350 cm⁻¹ for the BC spectrum (thinner than the bumpy band of plant cellulose) is thought to be due both to the higher crystallinity of BC compared to plant cellulose (Azeredo et al., 2019) and to the different crystalline structure between the triclinic I_α crystalline pattern of BC and the monoclinic I_β pattern of plant cellulose. At the same time, this different shape has also been ascribed to a more ordered and pure structure of BC (Seoane et al., 2017). Therefore, from the comparison of the FTIR spectra reported in Figure 3, it can be inferred that the cellulose extracted form the agri-food waste has not the same degree of purity of BC, due to possible residues of other plant components such as hemicellulose and lignin.

More detailed information on the purity of the biomass obtained after the extraction procedure has been obtained through 13C CP MAS NMR. In full agreement with the findings of Park et al. (2009), Figure 4 highlights a full correspondence between the NMR spectra obtained from the agri-food biomass and the NMR reference spectrum of pure cellulose (BC). This correspondence was particularly clear in the correspondence between the C4 and C6 crystalline regions, C4 para-crystalline region, and C2,3,5 fractions (Focher et al., 2001), especially for corncob and giant cane cut-up samples. The presence of a subtle peak at 175.70 ppm for the garlic stalk-derived cellulose may indicate the presence of some organic compounds not yet completely removed after the purification step (Park et al., 2009).

These results confirm that the biomass obtained after the extraction procedure was represented by high-purity cellulose, due to the absence of unpredictable signals coming from other fibrous components (e.g., hemicellulose, and lignin). This statement was supported by some evidence in the literature. Huang et al. (2017) observed that, in lignin samples, the characteristic peaks revealing the presence of such a molecule within NMR spectra stood within the aromatic region (160–103 ppm), and in the oxygenated and non-oxygenated regions (86–50 ppm). In a study by Zheng et al. (2015), different patterns with a single sharped peak in the region of 100–60 ppm were detected in hemicellulosic samples. All of the aforementioned regions were characterized by the presence of an up-field wing which might be attributed to the less ordered domains in the cellulose fibrils (Vismara et al., 2009).

¹³C CP MAS NMR spectra also allowed the calculation of the “crystallinity index” (Cr.I.) by means of Equation 3. As reported in Table 2, the agro-wastes-derived cellulose exhibited a larger number of crystalline domains over the amorphous ones. Specifically, Cr.I. average values were 73.84, 82.82, and 75.04% for garlic stalk, corncob, and giant cane cut-up, respectively (Table 2). These values were in line with those reported in the literature. For example, the Cr.I. of cotton linter-derived cellulose was ~72.2% (Park et al., 2009), while Bahloul et al. (2021) calculated a Cr.I. of ~62% after the extraction of cellulose from eggplant. By comparison, the crystallinity of bacterial cellulose (as a pure counterpart) was ~70–80% (Vazquez et al., 2013) and 71.2% (Watanabe et al., 1998).

A preliminary morphological assessment of the cellulose extracted from garlic stalk, corncob, and giant cane cut-up was performed by optical microscopy analysis (Figure 5). Cellulose from all three lignocellulosic wastes was represented by a fibrous morphology, as confirmed by previous similar works (Luan et al., 2012; Vanderghem et al., 2012; Arantes et al., 2017; Wu and Cheng, 2017). Moreover, the overall length of the cellulose fibers was up to 1 mm, with a width varying from ~10 to ~15 μm.

Although low environmental impact approaches (e.g., enzymatic hydrolysis) have recently been proposed (Rovera et al., 2018, 2020), acid hydrolysis remains the most commonly used, efficient, and faster method to synthesize CNCs (Habibi et al., 2010). In this work, the cellulosic substrate obtained from garlic stalk, corncob, and giant cane cut-up was subjected to acid hydrolysis to isolate the crystalline domains from the amorphous regions of the cellulosic backbone, eventually obtaining CNCs. By using a 58 % (v/v) sulfuric acid solution, the ultimate yield was as high as 41.98 ± 2.65 % DW_C.R., 40.87 ± 3.15 % DW_C.R., and 42.50 ± 0.81 % DW_C.R. for CNCs derived from garlic stalk, corncob, and giant cane cut-up, respectively (Table 2). These results are in line with those found by Reddy and Rhim (2018), who reported a CNCs yield of ~39.5% from garlic skin utilizing 45% (v/v) of sulfuric acid for 2 h at 60°C. Silvério et al. (2013), working with cellulose extracted from corncob, reported a yield of 57 and 50% by using a 9.17 M of sulfuric acid solution at 45°C, for 30 and 60 min, respectively. The same authors obtained a 46% yield for CNCs generated via a 90 min contact with sulfuric acid. Instead, using a 33% (v/v) hydrochloric acid solution, the yield of hydrolysis was significantly higher for all the investigated substrates (50.56 ± 3.10 % DW_C.R., 49.73 ± 1.45 % DW_C.R., and 53.24 ± 2.65 % DW_C.R. for CNCs derived from garlic stalk, corncob, and giant cane cut-up, respectively).

The stability of the CNCs water suspensions was evaluated by measuring the ζ-potential by DLS analysis. For the CNC derived from all the three agro-wastes, a mean value of ζ > −30 mV was measured when sulfuric acid was used, whereas ζ < −15 mV was obtained as an average for CNCs derived from hydrochloric acid. According to a study by Clogston and Patri (2011), ζ-potential values of > |30| mV are deemed the minimum requirement to produce a stable colloidal dispersion. The greater stability of the CNCs suspension obtained using sulfuric acid can be explained by the esterification of the hydroxyl groups on the C6 of the cellulosic backbone to form sulfate groups (i.e., –SO4-2), hence negative charges that were responsible for intermolecular repulsion forces. In contrast, ζ-potential values < |30| mV pertain to less stable colloidal dispersions, which most likely will undergo aggregation, flocculation, and phase separation over time. In this specific case, the absence of a net charge on the CNCs surface will bring to flocculation due to the tendency of CNCs to aggregate by H-bonding.

A morphological investigation of CNCs was first performed by AFM (Figure 6). Figures 6a–d, in particular, shows the AFM images of CNCs obtained after acid hydrolysis of garlic stalk (a and b), corncob (c), and giant cane cut-up (d) extracted by means of sulfuric acid. Nanocrystals generated from a sulfuric acid-driven hydrolysis process showed a typical needle-like shape, with an apparent averaged size (length) according to the following order: garlic stalks>corncob>giant cane. CNCs from giant cane cut-up extracted using hydrochloric acid (Figures 6e, f) showed a less defined shape, with a frayed morphology and a generally larger size (300–450 nm) than CNCs obtained using sulfuric acid. Similar results were observed in the case of garlic stalk and corncob.

The different sizes of CNCs treated with the two acids were confirmed by the particle size distribution measurements (Figure 7). Although DLS is not the most suitable analytical technique to determine the absolute size distribution of non-spherical particles (Rovera et al., 2018), it has been demonstrated that it is useful to monitor the relative difference in the size of particles obtained according to different treatments/setups. Figure 7A shows the bimodal size distribution of the nano-sized CNCs generated from sulfuric acid-treated cellulose. The position and intensity of the two main peaks were matrix-dependent. In particular, in CNCs derived from giant cane cut-up and corncob agro-wastes, the most intense peak was centered at ~90 and 120 nm, respectively, while the less intense one was centered at around 450 nm for both types of CNCs. Similarly, the size distribution of CNCs from garlic stalks showed a more intense peak at ~190 nm, while a less intense peak was observed at 750 nm. When the hydrolysis was carried out using hydrochloric acid, a general shift toward larger sizes was observed for the CNCs, though with an identical bimodal distribution (Figure 7B). More specifically, in CNCs derived from corncob and giant cane cut-up agro-wastes, the largest peak in intensity was centered at ~900–1,200 nm, while the less intense one was centered at ~260 nm. CNCs from garlic stalks exhibited the largest peak in intensity centered at ~1.8 μm, while a less intense peak was centered at ~630 nm (Figure 7B). These results thus seem to confirm our previous observations with AFM (Figure 6).

Based on the findings arising from both AFM and DLS analyses, it can be said that the hydrolytic process of cellulose from three different wastes provided non-homogeneous size distributions of the CNCs population, that is, two main size domains were detected when both sulfuric and hydrochloric acid were used. However, it seemed that the process mediated by sulfuric acid was more efficient in providing CNCs with smaller sizes. Our findings are in line with those reported in previous works. Silvério et al. (2013) reported a size distribution centered at ~290 and ~195 nm for CNCs extracted by acid hydrolysis (sulfuric acid 9.17 M) at 45°C, for 30 or 90 min. Similarly, Barana et al. (2016) reported an overall length of CNCs from Arundo donax of ~200 nm (60% H₂SO₄ at 55°C for 1 or 2 h). Kallel et al. (2016) found an average length of ~480 nm for CNCs extracted from garlic stalks by using a 65% (v/v) H₂SO₄ solution (45°C for 40 min).

With the goal of gathering further information on the quality of CNCs extracted from lignocellulosic wastes, a comparison with CNCs extracted from bacterial cellulose (BC) was performed, in consideration of the fact that BC is inherently devoid of contaminating molecules, such as lignin, pectins, hemicelluloses, etc., typical of cellulose of plant origin; that is, BC has a high purity degree (Gorgieva and Trček, 2019).

Acid hydrolysis of BC yielded 51.61 ± 1.51% and 69.83 ± 3.94% when sulfuric acid and hydrochloric acid were used, respectively. This means that, compared with the extraction of CNCs from cellulose from agri-food wastes, a + 23.3% in the final yield was obtained for the sulfuric acid-mediated hydrolysis, whereas an average 1.4-fold increase was achieved with hydrochloric acid. The yield increase in the case of BC can be ascribed to the different ratios between amorphous and crystalline domains. In particular, the degree of crystallinity of BC is higher than that of plant-based cellulose (85–90% for bacterial cellulose vs. 65–75% for plant-based cellulose) (Sacui et al., 2014).

Additional morphological and dimensional information on BCNCs was obtained by TEM, AFM, and DLS analyses. BCNCs with a needle-like shape were obtained when both HCl (Figures 8a, c) and H₂SO₄ (Figure 8b) were used, in contrast to what was observed for CNCs obtained from the lignocellulosic wastes (in this case, HCl hydrolysis yielded irregularly shaped nanocrystals, see Figures 6c, d). Apparently, BCNCs were wider than the plant-derived counterpart, thus displaying average length and width of 390 ± 35 and 18 ± 3 nm, respectively. The DLS analysis confirmed a bimodal size distribution for BCNCs obtained via sulfuric acid, in agreement with Rollini et al. (2020), who found two main dimensional families (~240 nm and ~1 μm) of BCNCs obtained using a 65% (v/v) sulfuric acid solution at 55 ± 1°C for 2 h. Conversely, only one main peak centered at ~2 μm was detected for BCNCs obtained using HCl. Irrespective of the acid used, a shift toward higher dimensions was observed for BCNCs, confirming what has been seen through TEM and AFM (Figure 8). Finally, the DLS pattern for BCNCs obtained using HCl (i.e., one main narrow peak) is typical of monodisperse systems, hence suggesting a better efficiency for HCl-mediated hydrolysis of BC over H₂SO₄ hydrolysis. Instead, Arserim-Uçar et al. (2021) detected a size distribution of BCNCs from 920 nm to ~1 μm when a 4 N HCl (reagent grade 37% v/v) acid solution at 95 °C and 105 °C for 240 min was adopted.