Use of Chènevotte, a Valuable Co-Product of Industrial Hemp Fiber, as Adsorbent for Pollutant Removal. Part I: Chemical, Microscopic, Spectroscopic and Thermogravimetric Characterization of Raw and Modified Samples

2.2. Hemp Shive Microstructure Analysis

Figure 3a shows the typical cross-section of a hemp chènevotte revealed by nano-CT with some remaining pith on the bottom and its woody part (xylem) made of woody fibers, vessels and rays. The hemp shive microstructure and ultrastructure have been recently investigated by Jiang et al. [46] using SEM and computed tomography. The results showed distinctive microstructures. The provided knowledge is used in the present work to characterize the influence of the treatments on the microstructure of the hemp shives. X-ray computed nanotomography images of a sample in the three main planes, namely L-R (Longitudinal-Radial), R-T (Radial-Tangential) and L-T (Longitudinal-Tangential) planes are presented in Figure 3b–d, respectively. In the raw shives, the vessels are isolated or grouped by two or three, rarely by more, and then they deform one another. The vessels have a quite thin cell wall and a diameter of approximately 50 to 150 µm (Figure 3b). They are surrounded by relatively thick-walled woody fibers with diameters of only a few µm and an irregularly polygonal section with a rounded cavity. The rays (made of parenchyma cells) are oriented in the radial direction. They are relatively thin, generally comprising one cell in width. Longitudinal sections through the hemp shives (Figure 3b,d) reveal the inner surface of vessels containing perforations (pits), which allow movement of moisture from cell to cell along the stem. These perforations connect smaller cells with the vessels. They are essential components in the water-transport system of higher plants. The pit membrane, which lies in the center of each pit, allows water to pass between xylem conduits but limits the spread of embolism and vascular pathogens in the xylem.

The influence of the tested treatments on the shive microstructure is clearly visible on nano-CT scans presented in Figure 4, in particular when treated with KOH and Na₂CO₃. The removal of the hemicelluloses deconstructs partially the woody fiber walls that leads to the fiber collapse and a general disorganization of the cells within the tissue, directly observed through the dealignment of the rays. The meso and microstructures of the shives are severely affected when treated with Na₂CO₃ in particular. The porosity level decreases from 75% to 57% due to the pronounced fiber collapse.

The surfaces of hemp shives were also examined with a scanning electron microscope (Figure 5). SEM images of the raw sample (SHI-R) showed the presence of impurities on the surface, with a compact and unfibrillated structure. The data analysis also indicated that each treatment significantly changed the morphology and roughness of the material surface. The strongest impact occurred after KOH treatment and degradation of the structure was observed after acid treatment.

Elemental analysis of the surfaces was also performed using a probe for energy-dispersive X-ray (EDX) spectroscopy and the electron beam excitation (Figure 5). The elements that composed the surface of SHI-R were mainly carbon, oxygen, potassium and calcium, whereas carbon and oxygen were the main components of all treated samples. Potassium and calcium are known to be essential for plant metabolism and various physiological processes related to growth [47,48,49,50,51,52]. Other residual elements that were observed include magnesium, aluminum, silicon, phosphorus, sulfur, chlorine and copper that might have come from soil [53]. Moreover, sodium was also found on the surface of hemp shives after Na₂CO₃ treatment.

2.3. Spectroscopic and Thermogravimetric Analysis

Besides the bulk sample characterization, the samples’ surface chemistry (determined up to 10 nm in depth by using ATR-FTIR) was studied given that it is one of the most essential factors affecting the biosorption of heavy metal ions. ATR-FTIR spectra of raw and treated hemp samples in powder form are shown in Figure 6. The main absorption bands observed in each spectrum and their assignment to chemical group vibrations and components are summarized in Table 2. The main changes determined in infrared spectra of hemp shives that were ascribed to the chemical treatments were found in two regions, i.e., 1800–800 and 3300–2700 cm⁻¹. One important difference in the spectra is the band at 1734 cm⁻¹ (SHI-R), characteristic of the stretching of unconjugated C=O groups present in hemicelluloses. This band fully disappeared after KOH treatment (SHI-OH sample) (Table 2) and was very weak after Na₂CO₃ treatment, while the H₃PO₄ treatment generated a more pronounced band in comparison to raw shives. Several authors have demonstrated a relationship between the removal of hemicelluloses by alkaline treatments such as KOH and NaOH using infrared spectroscopic measurements and the decrease/disappearance of band at 1734 cm⁻¹ [17,34,54,55,56,57,58,59,60,61]. In SHI-W and SHI-C samples, the C=O stretching at 1734 and 1733 cm⁻¹, respectively, appears less evident if the flakes are analyzed. Mirmohamadsadeghi et al. [60] indicated that the decrease of the band intensity at 1734 cm⁻¹ is due to the hydrolysis of the ester bonds between acetyl groups and hemicellulose by KOH and Na₂CO₃ treatments. The intensity of the peaks at 1600–1650 cm⁻¹, which corresponds to water adsorbed in polysaccharides, increases slightly after KOH and Na₂CO₃ treatments, which can be ascribed to the reaction with hydroxyl groups present on polysaccharides to form water molecules, as reported before [17,60].

In all ATR-FTIR spectra, the broad band around 3348 cm⁻¹ indicates the existence of hydroxyl groups primarily present in cellulose and hemicelluloses. The other two typical bands assigned to cellulose are observed at 899 cm⁻¹ (glycosidic bond symmetric ring-stretching mode) and 1327 cm⁻¹ (C-O stretching). The last band distinguishes between amorphous and crystalline cellulose [62]. After Na₂CO₃ and KOH treatments, these two bands are more pronounced, which can be explained by the fact that cellulose structure becomes more exposed due to the removal of hemicelluloses and lignin [61,62]. Additional confirmation of this statement is provided by the increased cellulose content in the Na₂CO₃ and KOH-treated samples (Table 1). The region between 1200 and 1700 cm⁻¹, which corresponds to the lignin component, is also affected by the treatments. Moreover, for treated samples, the signal at 1236 cm⁻¹ (assigned to the C-O aryl group of lignin) appeared pronounced, though this is less evident if the flakes are analyzed. The intensity of the band at 1506 cm⁻¹ suggests that none of the chemical treatments was successful in the complete removal of lignin from hemp fiber bundles [17,42,63], which is in accordance with the data listed in Table 1. The peak at 1422 cm⁻¹ could be attributed to pectins [57,64]. However, this band is also characteristic of C-H bonds present in all organic molecules (CH₂ symmetric bending).

The intensity of some particular bands in infrared spectra was proposed to calculate a crystallinity index [57,65,66]. The ratios of the peaks at 1421 and 893, 1375 and 2898 and 1375 and 660 cm⁻¹ were used to measure relative cellulose crystallinity (Table 3). The ratio of the peaks at 1375 and 660 cm⁻¹ is the most relevant descriptor according to Richter et al. [67] because there is no ambiguity in the assignment of the bands, unlike other bands, such as the one at 1424 cm⁻¹, which can be assigned to almost any component present in hemp. However, it is difficult to correlate this ratio with the crystallinity index obtained with XRD analysis [57] (as a technique for sample bulk characterization), since the results obtained by ATR-FTIR spectroscopy referred to the samples’ surface chemistry (around 10 nm in depth).

The cross-polarization magic angle spinning (CPMAS) nuclear magnetic resonance (NMR) spectra for untreated and treated shives are depicted in Figure 7. These spectra show the C1–C6 peaks of the glucopyranose unit present in the cellulose structure in the range 50–110 ppm. These peaks are characteristics of disordered cellulose. The spectra for SHI-R also show peaks that are ascribed to the presence of residual lignin and/or impurities, which is in agreement with the shives chemical composition. After alkaline treatment, these bands disappeared, in particular the bands at 22 and 170 ppm attributed to impurities (soluble extracts) and/or lignin.

Untreated and treated samples were studied using thermogravimetric analysis to obtain information relevant for application purposes such as thermal stability and the final residue. In addition, a previous study has shown that the study of the decomposition mechanism can be used to highlight structural changes better than other observable experimental method [68]. The knowledge of the hemp shive degradation process may be useful to evidence any structural changes provoked by the applicative use of these materials such as the adsorption of contaminants. All TG curves (Figure 8) showed an initial 4–6% mass loss that was mainly related to the evaporation of physically adsorbed and weakly bound water on hemp shives. This mass loss occurred at a temperature lower than 120 °C for all samples except for SHI-OH where, instead, the process ended at about 135 °C. According to the literature, the thermal degradation of hemp shives, after water evaporation, is generally a multi-stage process associated with the decomposition of their components, which are mainly hemicelluloses, cellulose and lignin, along with other minor substances depending on the physical and/or chemical treatment to which the samples were subjected [69]. Frequently, as in this case, the various stages cannot be highlighted in the TG curves because of their partial or complete overlapping in the same temperature range. Therefore, it is necessary to refer to the DTG curves (Figure 8) obtained by taking the derivative of the TG curves. The analysis of the DTG curve of SHI-R showed a two-stage thermal decomposition process. A first mass loss in the range of 260 to 380 °C with a maximum slope at 318.1 °C can be associated with the simultaneous degradation of hemicelluloses, cellulose and lignin [70]. It follows a second mass loss, partially convoluted with the first one, evidenced by a different slope in the range of 380 to 600 °C, which is likely associated initially with the end of the cellulose decomposition and then with the main process of thermal decomposition of the lignin polymer structure. Lignin thermal decomposition occurred in a wide temperature range, the process being very slow at the beginning [69,71]. Moreover, the lack of evident decomposition in the range 200–260 °C suggested the absence or negligible presence of pectin [72]. At 690 °C, a residue of 1.8% was obtained. SHI-W decomposition began at about 270 °C and showed a two-step process convoluted as already reported in the literature: the main step, which ends at 411 °C and had a maximum rate at 370.9 °C, ascribable mainly to cellulose decomposition, and the other step, evidenced in the DTG plot by the shoulder at about 307 °C due to the hemicellulose decomposition [69,73]. After 411 °C, a linear decrease of mass from 19.9% down to 13.7% at 690 °C was observed, likely due to the slow decomposition of lignin. As a result, in comparison with SHI-R, the thermal decomposition of SHI-W began at a higher temperature and was distributed in a shorter range of temperature than SHI-R but was less complete. The DTG curves of washed alkaline-treated samples, SHI-OH and SHI-C, were almost symmetric. The disappearance of the shoulder associated with hemicellulose decomposition could indicate a decrease of the hemicellulose amount or changes in the hemicellulose structure, likely due to the removal of side groups by means of the alkaline washing [73]. In both samples, the decomposition had an onset and a maximum rate at temperatures higher than SHI-R: T_onset = 266.7 and 263.6 °C and T_max rate = 331.9 and 335.6 °C for SHI-OH and SHI-C, respectively. Moreover, these chemically treated samples had the highest formation of carbonaceous residue (20.3% for SHI-OH and 17.2% for SHI-C) promoted by potassium and sodium ion presence, thus changing the decomposition mechanism of cellulose and lignin [73]. The orthophosphoric acid treatment of the hemp shives, instead, gave rise to a decrease in thermal stability. SHI-H thermal decomposition began at 239.3 °C, and it was likely due to the degradation of hemicelluloses as evidenced by the shoulder in the DTG curve at 265.6 °C. After the maximum rate of SHI-H decomposition at 293.2 °C, there was a change in the TG curve slope between 332.4 and 493.8 °C, which can be associated, as in the case of SHI-R, to the lignin decomposition with a final residue of 21.8% at 690 °C. All the indications obtained from the TG analysis are in good agreement with the composition of the hemp samples reported in Table 1 and evidence that all treatments give rise to an increase in the thermal stability of hemp shives, with respect to the untreated sample, with the exception of the sample treated with orthophosphoric acid.

3.2. Treatment Procedures

Raw hemp shives were either simply washed with water over 2 days at room temperature or treated with a chemical reagent for 4 h at 40 °C. The reagents used were 30% (v/v) H₃PO₄, 1 M KOH and 1 M Na₂CO₃. These conditions were selected according to those already used on site by the industrial partner for other materials. After each treatment, the samples were washed extensively with water until a neutral pH was obtained and then dried in an oven at 80 °C until a constant mass was obtained. The denotation of the samples is the following: SHI-R for raw shives and SHI-W, SHI-H, SHI-OH and SHI-C for raw shives treated with H₂O, H₃PO₄, KOH and Na₂CO₃, respectively.

N.M.-C. and G.C. thank Eurochanvre (Arc-les-Gray, France) for the gift of the materials, Région Bourgogne Franche-Comté (France) and FEDER (Fonds Européen de Développement Régional) for financial support and the PEA²t Platform (Chrono-environnement, Université Bourgogne Franche-Comté, France), which manages and maintains the analytical equipment used in this study. X.G. and V.P. are grateful to MIFHySTO technological platform (FEMTO-ST, France) for the use of X-ray nanotomography. A.R.L.R. acknowledges the support from Base-UIDB/50020/2020 and Programmatic-UIDP/50020/2020 Funding of LSRE-LCM, funded by national funds through FCT/MCTES (PIDDAC). The doctoral student Chiara Mongioví also thanks the Région Bourgogne Franche-Comté (France) for awarding her a research grant.