3.1 Characterization of the adsorbents
The micrographs acquired from Scanning Electron Microscopy (SEM), presented in Fig. 2, show the surface morphology of the untreated (A) and of the plasma treated SB (B - J).
Figure 2 - Scanning electron micrographs of the raw and of the plasma treated SB with different exposure times and excitation powers at ×500 and ×3000 magnifications. A) Untreated SB; B) SB 80 W 2 min; C) SB 190 W 2 min; D) SB 300 W 2 min; E) SB 80 W 30 min; F) SB 190 W 30 min; G) SB 190 W 60 min; G) SB 300 W 2 min; G) SB 300 W 30 min; G) SB 300 W 2 min.
In Fig. 2.A, the micrograph reveals a fibrous and heterogeneous structure with few pores or with small pores that are not perceptible with the employed magnifications. Particulate material, rising from deteriorated regions of the fibers, is present in some surface regions, together with cracks. The structural complexity of the natural material is responsible for this heterogeneity and it was also reported by (Veiga et al. 2021; Liu et al. 2022a).
Upon the plasma treatment, morphological changes were detected with the modification degree being dependent on the plasma power and on the exposure time. The particulate and agglomerate content is reduced, indicating removal of the weakly connected material by the plasma. A clear improvement on the short-range surface smoothness is observed when comparing the higher magnification micrographs. Except for the fastest treatment (2 min), conducted at the lowest power (80 W), cracks are no longer evident after the treatment. However, a higher number of pores was identified, suggesting material removal is being favored from specific points of the biomass. Aside to this, there is rupture of the original continuous structure into large flakes of material, evidenced by the analysis of the inset (lower magnification of x500) micrographs. It is detected a general trend of increasing the flakes production with increasing the treatment intensity (time and/or power). According to Liu, Zhou, and Liu (Liu et al. 2022b), the power elevation in helium plasmas could enhance the surface roughness due to the corrosive effect of plasma, resulting in a more irregular morphology. In the present work, two trends in opposition were detected as one considers the surface morphology of the treated samples: a reduction of the surface defects in the low-range scale and an elevation in a larger-range one.
Thus, according to the above results plasma is removing material from the surface by chemical and physical reactions. In plasmas of SF6, it is observed the generation of SF5, SF4, SF3, SF2, SF, S, and F reactive groups (Resnik et al. 2018). The increment in the exposure time thus elevates the probability of material removal by some of the erosive species.
Another phenomenon explaining the same effect, but with other consequences, is the removal of material by physical routes, that is, by ion bombardment. The collision of fast ions with the material surface transfer energy that can cause emission of groups together with structural and chemical modifications. For low energy ions, dissipation of energy occurs by means of nuclear collisions, that provide displacement of atoms from their position in the structure and thus its weakening and breakage. On the other hand, for more energetic ions energy dissipation by electronic events (excitations, ionizations, free radical generations) is favored, producing active dangling bonds that may recombine by C bonds unsaturation and chain crosslinking.
Consequently, to understand the results obtained here it should be taken into account that fast ions will lose their energies primarily in electronic collision, producing crosslinkings and unsaturations, processes that improve the surface uniformity. When their energies are reduced, with increasing the penetration depth, nuclear events will be the main responsible for slowing down the ions, producing breakage of the SB polymeric backbones. With increasing treatment time and power this damage process is intensified flaking the overall SB structure. Thus, ion bombardment can contribute to the improvement of the surface uniformity due to dissipation of energy by electronic collision and to the structural rupture by deposition of energy from nuclear events. Moreover, ion bombardment favors the non-homogeneous removal of material. In the work of Man (Man et al. 2020), it was demonstrated that etching of SiO2 coated with an CHxF layer, in SF6/CH4 plasmas, is regulated, amongst others, by ion bombardment. The sputtering of groups from the CHxF top layer generates defect points where the reaction of neutral F can promptly erode the whole structure beneath, creating pores in the initially uniform layer. Thus, the observed pores in the treated SB investigated here (micrographs) are attributed to the simultaneous physical (sputtering) and chemical (etching) effect of the SF6 plasmas as demonstrated in the work (Man et al. 2020).
Therefore, all the modifications observed on the surface microstructure of the SB submitted to the plasma treatments, including pore formation, flaking and smoothening of the SB structure can be attributed to the physical and chemical effect of ion bombardment and to the etching caused by neutrals generated in SF6 plasmas.
The EDS results of the untreated and of the plasma treated SB samples are presented in Table 1. The main elements observed on the surfaces of the untreated SB were carbon, oxygen and nitrogen, in good agreement with the results of (Rocha et al. 2015) in which was evaluated the composition of the bagasse from 60 varieties of sugarcane produced in Brazil. Aside to these elements, the treated samples also presented fluorine and, in some cases, sulfur.
The untreated material is majorly composed of C with lower proportions of O and N. Hydrogen is also a component of the SB, but it is not detected by this methodology. According to Seah and collaborators (Seah et al. 2023) the structure of different woody and herbaceous biomasses is constituted by 6% of H. After the plasma treatment there is detection of small proportions of F in all the samples and of S (≤ 0.2%) only in the samples treated in the mild conditions (80 W, 2 and 30 min). Changes in the proportion of the other elements are also identified, but three major trends should be discussed here.
Table 1
Semi-quantitative analysis results of atomic proportions of C, N, O, Si, K, F and S on the untreated and plasma treated SB.
Sample
|
C (%)
|
N (%)
|
O (%)
|
Si (%)
|
K (%)
|
F (%)
|
S (%)
|
SB
|
86.3
|
1.1
|
11.2
|
0.4
|
0.6
|
-
|
-
|
SB 80 W 2 min
|
78.3
|
10.4
|
9.1
|
0.1
|
0.7
|
1.2
|
0.2
|
SB 80 W 30 min
|
79.3
|
11.3
|
8.5
|
0.1
|
0.3
|
0.4
|
0.1
|
SB 80 W 60 min
|
86.1
|
2.5
|
10.0
|
0.4
|
-
|
0.3
|
-
|
SB 190 W 2 min
|
85.6
|
2.4
|
10.9
|
0.3
|
0.4
|
0.4
|
-
|
SB 190 W 30 min
|
87.1
|
1.1
|
11.0
|
-
|
0.5
|
0.3
|
-
|
SB 190 W 60 min
|
87.4
|
1.3
|
9.0
|
1.4
|
0.7
|
0.2
|
-
|
SB 300 W 2 min
|
75.6
|
12.9
|
11.0
|
-
|
0.4
|
0.2
|
-
|
SB 300 W 30 min
|
88.0
|
-
|
11.7
|
0.1
|
0.1
|
0.1
|
-
|
SB 300 W 60 min
|
91.9
|
-
|
8.0
|
-
|
0.3
|
0.1
|
-
|
The first one is related with the C content reduction for the treatments with the lowest power (80 W), for 2 and 30 min, and with the highest power (300 W) for 2 min. The concomitant falls in the C and in the O proportions for the samples treated in plasmas of 80 W (2 and 30 min) are followed by compensatory rise in the N proportion. On the other hand, for the highest power treatments (300 W, 2 min), despite the reduction in the C proportion is similar to that observed in the previous discussed samples, that of O is not. Only an oscillation in the O proportion is detected in this case, indicating now N is replacing only C. So, different changing mechanisms are taking place in the low and high-power regime. A reduction in the carbon proportion rice husk derived hybrid silica/carbon biochar was also reported by (Mohammed et al. 2021) and is very consistent with the erosive nature of electronegative SF6 plasmas.
The second trend observed in the Table 1 is the elevation of the C proportion beyond that observed for the untreated SB, for the highest power and exposure time treatments (300 W, 30 and 60 min). Proportion of O is only barely influenced in this case. However, N is no longer incorporated. The inclusion of N, as well as O, is proposed to happen, majorly, from reactions of active dangling bonds, left on the material structure after treatment, with atmospheric groups when the sample is removed of the vacuum chamber. The absence of N indicates that the trapped radical content decreases for treatments in plasmas of 300 W (30 and 60 min). Pendant bonds, generated from H and N abstraction, are being consumed by chain crosslinking and by unsaturation of chemical bonds. Just N realize it enough to explains the elevation in the C proportion, but O is also being abstracted in some of these treatments.
The third aspect observed is related to the proportion of fluorine. Despite some oscillation, a general trend of decreasing F incorporation with increasing plasma excitation power and exposure time is detected. Fluorine inclusion occurs at low proportions (< 1.2%), showing that fluorination is not favored in the treatments conducted here.
Possible neutral species produced in pure low pressure SF6 plasmas are SF6, SF5, SF4, SF3, SF2, SF, S, F and F2. Positive (SF5+, SF4+, SF3+, SF2+, SF+, S+ and F+) and negative (SF6−, SF5−, SF4−, SF3−, SF2−, F− and F2−) ions may also be formed from the SF6 precursor. Nevertheless, the dissociation behavior and thus the concentration of radicals and ions depend on the electronic density and temperature, which, in turn, rely on the pressure and applied power. It was demonstrated in the work developed by Levko and co-workers (Levko et al. 2013), on the evaluation of the neutral and ionized species distribution on SF6 plasmas, using the one-dimensional fluid model, that an elevation in the plasma power provides an overall increment in density of electrons, positive and negative ions in the plasma. In all cases, the most abundant neutral fragment is SF5, followed by F and by SF4 – SF3. The highest densities of SF5 and F groups is explained by their formation route, due to direct electron impact, to be the most probable one by the reaction (Resnik et al. 2018)
$${SF}_{z}+ {e}^{-}\to {SF}_{z-1}+F+{e}^{-}$$
3
In the same work, it was shown that the most abundant charged species are SF5+ followed by F−, SF6−, SF3+, SF4+. Negligible densities of SF2 and SF2+ were observe. Using a radiofrequency (13.56 MHz, 10 mTorr, 900–1700 W) capacitively coupled SF6 plasma, Lallement and co-authors (Lallement et al. 2009) observed similar results concerning the most probable neutral and charged species as well as a trend of rise in the electron density with increasing plasma power, but a constancy in the electron temperature. Besides that, Amorim et al., (Amorim et al. 2020) stated that in low-pressure plasma, high electron temperatures generate an increase in fluoride concentration. This, combined with elevates electronegativity of SF6, leads to the generation of negative ions, by reactions such as:
$${SF}_{6}+ {e}^{-}\to {SF}_{6}^{-}$$
4
Thus, plasma is composed of negative and positive ionized groups from SF6 and also from O (reactor residual atmosphere), together with electrons. According to the previous discussion, an elevation in the concentration of ions and electrons is expected in SF6 plasma with increasing the excitation power. But it is important to mention here that negative ions will be attracted to the grounded upper electrode of the reactor and repelled of the negatively biased sample holder (lowermost electrode). This can be pointed as one of the reasons why there was a low fluorine incorporation on the SB structure. Only neutral F species, orders of magnitude more abundant than ions, that can diffuse to the region where the SB was accommodated, are prone to react with the organic structure to be incorporated by means of CFx groups. In the first stage of this reaction, F has to recombine with C or H from the SB, generating volatile groups, with low sticking probabilities that are emitted, explaining material removal and dangling bonds formation on the material structure (Resende et al. 2018; Mohammed et al. 2021)
This is another reason for the low F incorporation, that is, F is effectively acting as an eroding compound rather than a doping element. Fluorine inclusion would happen after the etching step. Ion bombardment of the bagasse with positive SF6 fragments (SF5+SF4+, SF3+) may also contribute to F incorporation, but still more with the atomic and molecular release and then with free-radical generation. The latter processes would enhance chain crosslinkings and bond unsaturation, explaining the low incorporation of atmospheric groups (N, O) observed in some samples as well as the reduction on the surface defects.
The energy deposited in the material structure by ion bombardment tends to increase with the rise in the power and in the exposure time. The probability of neutral reactive radicals reaching the sample surface also increases with exposure time.
Therefore, the low proportions of F detected in the samples studied here indicate that dangling bonds’ saturation by F is not an efficient process. The variation in the atomic proportions of O and N, together with that of F indicates the saturation of free radical is taking place during plasma treatment but also after it, when the sample is in contact with atmosphere. Furthermore, F should be concentrated on the topmost layers of the fibers, whereas EDS analysis is probing deeper untreated regions, promoting a mixed result of the treated and untreated regions.
Finally, sulfur was detected only in the samples prepared in plasmas of low power (80 W) for low (2 min) and moderate (30 min) exposure times. Figure 3 highlights the points where sulfur was detected. Figure 3. A and Fig. 3. B correspond of the material treated in plasmas of 80 W for 2 and 30 minutes, respectively. Interestingly, higher power levels (190 and 300 W) and longer exposure times (60 min) did not contribute to S incorporation. The optimization of the structural healing (crosslinking and unsaturation) in these conditions are pointed as the responsible for the lack of S and the low proportion of F detected in these cases.
With such results it is promptly observed that the highest F and S incorporation occurred for the lowest power plasma and for the lower exposure time (2 min). The proportion of C was the lowest in this sample while the N proportion was the highest indicating C and O are being replaced mainly by N. This inference is confirmed by the results of the sample prepared at the highest plasma power and exposure time where N was not detected and C proportion was the highest one. Based on that it is proposed that the plasma treatment is removing not only H, but also C and O (Si and K) from the biomass structure, generating active sites for atmospheric N and O incorporation, crosslinking, bonds unsaturation and F incorporation.
The analysis of FTIR spectra of the of the as-received and the modified SB under various time intervals and pressures (Figure S1, see material supplementary) shows the presence of aliphatic CH2 groups of the lignin is evident by the bands at 2865 and 2918 cm− 1 (Manyatshe et al. 2022). Contributions due to OH stretching vibrations are detected at 3450–3650 cm− 1 (Abdulhameed et al. 2021; Bai et al. 2022). C oxidized groups are identified by the contributions at 1740 (C = O) (Ordonez-Loza et al. 2021) and 670 (C-OH in cellulose) cm− 1. Peaks characteristics of lignin, normally found in 1616, 1586, 1508, (aromatic C = C) and 1234 (aromatic C-O) (Montero et al. 2018; Veiga et al. 2021; Dzoujo et al. 2022). The bands observed at 1740 and 1368, cm− 1 are attributed to the stretching of the COO- and CH = CH respectively. Those bands are related to hemicellulose and lignin compounds (Veiga et al. 2021; Sutthasupa et al. 2023). The bands around 1120 and 1149 cm− 1 are typical cellulose and hemicellulose peaks due to C-O and C-N stretching (Montero et al. 2018).
It is a consensus in the literature (Luz et al. 2007), that SB is composed of different proportions of cellulose (C₆H₁₀O₅)n, hemicellulose (C5H10O5) and lignin (C81H92O28), together with mineral contaminants. The organic fraction of this material is composed of structures formed by aromatic carbon rings to which hydroxyls, methyl, methylene and others components are attached (Rocha et al. 2015; ALVES MACEDO 2020). Comparing the spectra of the as-received and of the plasma-treated SB reveals a general preservation of the material’s chemical structure. However, it should be considered that whereas the plasma treatment is changing the topmost layers of the material, the infrared inspection is reaching untreated deep layers. But even so, some modifications in the infrared spectra of the samples are indicatives of the plasma induced changes. New bands are detected at 2100 cm− 1 (C≡C) and 1114 cm− 1 (C-O), the latter spectral band mainly appears in the sample spectrum subjected to a longer duration and higher power treatment.
Also, in the Fig. 4.A and 4.B is observed a peak in 1900 cm− 1 which is related to stretching C = C = C (Merck 2021). Aside to this, the treatment results in alterations in the intensities of several bands, including OH at 3580 cm− 1, CH2 at 2918 cm− 1, C = O at 1745 cm− 1 and C = C at 1586 cm− 1. Specifically, unsaturation of C bonds, proposed in the interpretations of elemental composition of the samples, is then corroborated by the rise of C≡C (2100 cm− 1) band and by the growth of C = C one (1586 cm− 1).
The peak ascribed to O-H stretching vibrations (3450–3650 cm− 1) exhibits heightened prominence in materials primarily treated with lower power. However, when the treatment power increases, its intensity decreases, probably because the plasma has a similar effect to heat treatment, that may lead to the release of part of the structural water contained in the plant material (Dzoujo et al. 2022).
The intensity of the band at 1700 cm− 1 (aromatic C = O), is significantly increased after plasma treatment, particularly when using power levels of 80 and 190 W. This effect may be due to the conversion of C-OH groups into C = O due to H abstraction (F or ion bombardment). The transformation of C-OH groups into C = O also corroborates the idea of dangling bond consumption by unsaturation of bonds.
After exposure to fluorine plasma, new peaks appeared in the range of 900 to 1300 cm− 1, which may be related to CFx groups (CF, CF2, and CF3) (Agopian et al. 2022; Zhou et al. 2023). The peaks observed at 700 − 600 cm− 1 correspond to the alkyl halides of CF (Mohammed et al. 2021). The rise of small bands around 1330 (CF2) and 600–700 cm− 1 (CF) in the spectra of treated samples suggest F incorporation in low proportions, what is in good accordance with the compositional results obtained by EDS.
In the spectra appears a peak in 1150 cm− 1 which can be attributed to the S = O symmetrical stretching of the sulfonate (Pavia et al. 2008).This result, combined with that of Table 1, confirms the presence of sulfur incorporation, especially in the SB samples with a power of 80 W.
Fluorine neutrals and charged species are majorly eroding and ion bombarding the material structure. It is also confirmed that dangling bonds generated by etching and sputtering are being consumed by unsaturation of C bonds and possibly by the counterpart process of crosslinking.
The determination of the zero-charge pH (pHpzc) plays a pivotal role in comprehending the electrostatic interaction dynamics between the adsorbate and the adsorbent. It is imperative for the charges on the adsorbent to be opposite those of the adsorbate, fostering a more robust interaction between the two entities, as emphasized by Alves Macedo (ALVES MACEDO 2020). In Fig. 4, the pHpzc graph is presented for both untreated SB and SB treated at 300W for 60 minutes. The point of intersection on the curve with the x-axis signifies an equilibrium between negative and positive charges on the adsorbate's surface.
Upon analysis, it is observed that the pHpzc of SB 300W 60 min is approximately 4.8. In contrast, the untreated SB exhibits a slightly higher difference pHpzc value, reaching 5.2. This suggests that the plasma treatment has not significantly altered the pHpzc. In conclusion, the surface of the adsorbent becomes positively charged at a pH < 4.8 and negatively charged at a pH > 4.8.