Morphological And Thermal Analysis
The morphology and composition of the LFs, LSFs, and LCFs were observed using SEM and EDS. As shown in Fig. S1, LFs had an average diameter of 158.5 ± 15.9 µm and exhibited a smooth surface. After chemical stabilization, the fibrous diameter decreased to 53.9 ± 9.6 µm. Pores and splits were generated on the LSF-0 surface (Fig. 2a). With Ni(acac)2 addition, the inner and exterior morphologies of the LSFs were clearly altered (Figs. 2b–d). These changes can be explained by material interactions. Firstly, ether bond cleavage occurs between the lignin matrix and PEG side chains under concentrated acidic conditions. These residues at low molecular weight wash out easily, forming defects on the LSF surface (Lin et al., 2012). Secondly, Ni cations, which are also electrophiles in this system, coordinate different nucleophiles and promote ether bond cleavage, yielding a rough fibrous surface (Fig. 3). Additionally, the deposited Ni ions can be a graphitic catalyst in subsequent carbonization(Malik et al., 2021).
LCF-0 with an average diameter of 64.68 ± 4.6 µm was obtained, as shown in Figs. 2e–h. The fiber diameter was reduced to 48.9 ± 5.6 µm after carbonization in LCF-0.4. Notably, a fibrous melt surface was observed from LCF-0.4 and LCF-0.6. With confirmed increased Ni loading from EDS (Fig. S2), the melt surface might be caused by Ni-assisted catalytic graphitization. This graphitization is explained in the following sections.
Table 1
Decomposition temperatures and mass loss analysis for the LSF samples.
Sample | Extrapolated onset decomposition temperature (Tod, °C) | Extrapolated endset decomposition temperature (Ted, °C) | Mass loss at decomposition (%) | Mass loss at carbonization and graphitization (%) | Total mass loss (%) |
LSF-0 | 312.7 | 471.7 | 39.4 | 14.1 | 53.5 |
LSF-0.2 | 309.1 | 468.4 | 41.2 | 13.3 | 54.5 |
LSF-0.4 | 308.4 | 467.9 | 43.9 | 12.5 | 56.4 |
LSF-0.6 | 289.5 | 463.5 | 45.7 | 12.4 | 58.1 |
Simultaneous TG–DSC measurements were performed under N2 atmosphere to investigate the decomposition and melting behavior of LSFs during carbonization (Fig. 4 and Table 1). All the LSFs exhibited one main weight-loss stage in the TG curve, which corresponds to the decomposition of unstable small-molecule components. The onset temperatures of decomposition decreased from 312.7°C to 289.5°C with increasing Ni content. Such decomposition afforded a mass loss of 39.4% for LSF-0, which increased to 45.7% for LSF-0.6. This increase proves that Ni assists in ether bond cleavage and small-fragment formation. From the DSC curve above the extrapolated Ted, LSF-0 exhibited a broad exothermic region with an exothermic peak at 582.7°C, a typical feature in biomass carbonization (Yoo et al., 2018). Meanwhile for LSF-0.2, the DSC heat flow changes from an exothermic to endothermic reaction at 660.6°C. This temperature decreased to 491.4°C and 472.0°C for LSF-0.4 and LSF-0.6, respectively, indicating the developing ordered carbon structure (graphitic and turbostatic carbons) (Gutiérrez-Pardo et al., 2015; Inagaki and Kang, 2014; Yoo et al., 2018). According to the proposed mechanism for catalytic graphitization, the endothermic reaction suggests dissolution of sp3 C atoms in Ni and precipitation of sp2 C in saturated C/Ni solution, affording the melt surface observed in the SEM images.
Structural Determination
To understand the carbon structure of the LCF samples, we performed Raman spectroscopy on the fibers. As shown in Fig. 5a, all LCF spectra had both D-band at 1350 cm–1 and G-band at ~ 1600 cm–1 after deconvolution, where the G-band refers to the in-plane vibration of graphitic carbon atoms, and D-band corresponds to disordered sp2-hybridized graphitic carbon atoms (Zhengshuai Sun, 2022-06-20 ). The integrated area ratio of ID/IG (R) can be calculated to gauge the degree of disorder at the surface structure of the LCFs (Gong et al., 2017). As shown in Table 1, LCF-0 exhibited the highest R value of 8.16 among the prepared samples. With increasing Ni loading, R decreased to 2.57 for LCF-0.4. This sample exhibited the largest La plane size of 7.48 nm among the LCF samples. When the Ni(acac)2 amount was further increased, the R and La values of LCF-0.6 increased to 3.71 and 5.18 nm, respectively, indicating a smaller graphitic region than LCF-0.4. This difference may be due to Ni being a graphitization catalyst as well as an etching reagent during chemical stabilization, generating a porous surface on the LSFs. Thus, high Ni concentrations yield aggressive etching, causing discontinuous graphitic carbon regions after carbonization. This phenomenon indicates that biomass graphitization could not be improved by simply increasing the amount of Ni ions (Gai et al., 2021; Zhengshuai Sun, 2022-06-20 ).
XRD characterization was conducted to elucidate the carbon structure changes from those of the bulk LCF samples. As shown in Fig. 5b, no diffraction peaks were observed for LCF-0 in the entire scan range, indicating that only amorphous carbon was generated without Ni(acac)2. Meanwhile, broad XRD peaks from the LCF samples at 24° and 44° were identified, which correspond to the (002) and (101) crystal planes, respectively.(Inagaki and Kang, 2014) This is because Ni ions can only be adsorbed on the surface of the LSFs during chemical stabilization and cannot access the insides. The average Lc values for LCF-0.2, LCF-0.4, and LCF-0.6, based on the (002) peaks, were 0.26, 0.41, and 0.39 nm, respectively (Table 2).
Table 2
Calculated Raman spectroscopy and XRD results for the LCFs.
Samples | Raman | | XRD |
R = ID/IG | La (nm) | Lc (nm) | d002 (nm) |
LCF-0 | 8.16 | 2.36 | | – | – |
LCF-0.2 | 3.74 | 5.14 | | 0.26 | 0.37 |
LCF-0.4 | 2.57 | 7.48 | | 0.41 | 0.35 |
LCF-0.6 | 3.71 | 5.18 | | 0.39 | 0.37 |
High-resolution TEM clearly shows nanostructures in the LCFs (Fig. 5). Consistent with the XRD and Raman spectroscopy results, ordered microstructures were difficult to find in LCF-0 and LCF-0.2 owing to low Ni content on the fibrous surface. For LCF-0.4 and LCF-0.6, distinct lattice fringes embedded in the amorphous regions were identified. LCF-0.4 exhibited large ordered layers. The measured spacing between the crystal layers was 0.34 nm, proving Ni-assisted graphitic structure generation. Thus, the LCFs comprised mainly amorphous carbon in the bulk covered by crystallized carbon layers on the surface. Such a structure would be beneficial for electrolyte ion transport and accumulation in supercapacitors.
Table 3
Textural properties of the LCFs with varying Ni(acac)2 content.
Samples | SBET a (m2 g–1) | Wp b (nm) | Vtotal c (cm3 g–1) | Sinternal d (m2 g–1) | Sexternal e (m2 g–1) |
LCF-0 | 350.95 | 1.62 | 0.13 | 316.67 | 34.28 |
LCF-0.2 | 103.52 | 1.65 | 0.04 | 93.96 | 9.57 |
LCF-0.4 | 89.92 | 1.61 | 0.03 | 83.74 | 6.19 |
LCF-0.6 | 269.56 | 1.67 | 0.11 | 234.48 | 35.08 |
a SBET: Specific surface area computed using the BET model.
b Wp: Adsorption average pore width.
c Vtotal: Total pore volume.
d Sinternal: Surface area of micropores and mesopores using the t–plot method.
e Sexternal: Surface area of macropores using the t–plot method.
N2 adsorption/desorption curves (Fig. S3) exhibit typical type-I behavior. The specific surface area and pore information for the LCFs are listed in Table 3. The specific surface area of LCF-0 was 350.95 m2 g− 1, with 316.67 m2 g− 1 from micro/mesopores and 34.28 m2 g− 1 from macropores, according to the t-plot method. The specific surface area decreased with Ni(acac)2 content and reached a minimum value of 89.92 m2 g− 1 for LCF-0.4. Moreover, the total volume and internal/external surface areas exhibited a similar trend. This can be explained by Ni-induced melting on the fiber surface, which filled up several pores and decreased the specific surface area. Thus, the total pore volume decreased from 0.13 m3 g− 1 for LCF-0 to 0.03 m3 g− 1 for LCF-0.4. Interestingly, the surface area and total pore volume increased to 269.56 m2 g− 1 and 0.11 m3 g− 1 for LCF-0.6, respectively. These may be due to the influence of Ni ions on etching rather than filling the surface at high Ni concentrations in the Ni-treated LCF samples. This feature is expected to influence supercapacitor performance, as discussed in the next section (Kim et al., 2021; Li et al., 2021b; Xia et al., 2020; Xu et al., 2021).
Electrochemical Performance Of The Lcf-based Supercapacitors
EIS, CV, and GCD experiments using a two-electrode system with 1.2 M TEMABF4/PC electrolyte were performed to determine the electrochemical performance of the LCF-based electrodes. As shown in Fig. 6a, Nyquist plots at the high-frequency region show the electrode resistance (Re) of the supercapacitors at the x-axis intercept, which indicates the resistance of the electrode materials and their contacts. The equivalent series resistance (Rs) was obtained from the intersection of the straight line in the low-frequency region and Zre axis. Rs is the sum of the electrode and electrolyte resistances (Mei et al., 2017; Vicentini et al., 2021). Without graphitic structures on its surface, LCF-0 had very large Re (39.74 Ω) and Rs (326.39 Ω) even when CB was added in the electrode recipe. When Ni(acac)2 was used in chemical thermostabilization, Re and Rs decreased remarkably and achieved the lowest values of 0.9 and 1.2 Ω for LCF-0.4, respectively. The low resistances suggest that a low-resistance network was established on the surface of LCF-0.4 owing to Ni catalytic graphitization. Additionally, the straight line in the low-frequency region for LCF-0.4 exhibited the highest slope among the LCF samples, indicating rapid diffusion of electrolyte ions during the charge–discharge process.
Table 4
Summarized supercapacitor performance using different LCF electrodes.
Sample | CCV a (F g–1) | CGCD b (F g–1) | Rs c (Ω) | E d (Wh kg–1) | P e (kW kg–1) |
LCF-0 | 1.7 | 2.30 | 6.3 | 0.7 | 13.4 |
LCF-0.2 | 52.6 | 37.4 | 2.0 | 11.7 | 282.4 |
LCF-0.4 | 75.0 | 50.2 | 1.2 | 15.7 | 644.9 |
LCF-0.6 | 66.4 | 42.5 | 21.0 | 13.2 | 202.6 |
a CCV: Gravimetric specific capacitance calculated from CV profiles.
b CGCD: Specific capacitance calculated from GCD profiles.
c Rs: Equivalent series resistance from Nyquist plots.
d E: Energy density.
e P: Power density.
CV curves of the supercapacitors with LCFs exhibited quasi-rectangular shapes within 0–3 V, indicating typical supercapacitor behavior (Fig. 6b). Even though LCF-0 had a large surface area and well-developed porous structure, the capacitance of the assembled supercapacitor was only 1.7 F g− 1 at a scan rate of 0.05 V s− 1. This value increased considerably for the Ni-catalyzed LCF electrodes, and a high capacitance of up to 75.0 F g− 1 was obtained for LCF-0.4.
GCD measurements were conducted to further evaluate the electrochemical performance of the LCF electrodes (Fig. 6c). Very short discharge time of 0.1 s was observed when LCF-0 was used as the electrode. The discharge time extended to 65 s and specific capacitance of 50.2 F g− 1 was obtained when graphitic LCF electrodes at the same current density (0.5 A g− 1) were employed. These findings are consistent with the CV results. Moreover, LCF-0.4 exhibited excellent electrochemical stability of 100% after 1500 cycles. Notably, the normalized capacitance of the LCF-0.4 electrode outperformed that of most reported biomass-derived supercapacitors at 0.5 A g− 1 (Fig. 7a) (Chmiola et al., 2006; Du et al., 2021; Wang et al., 2017). Thus, the assembled supercapacitor using LCF-0.4 exhibited the best performance. It had an energy density of 15.7 Wh kg− 1 and power density of 644.9 kW kg− 1, which are comparable to those of previously reported biomass-derived supercapacitors (Fig. 7b) (Dai et al., 2019; Poochai et al., 2021; Schlee et al., 2019; Sevilla et al., 2019; Zheng et al., 2019; Zou et al., 2019). Thus, we infer that the Ni-catalyzed graphitization induced conductive network had more influence on electrolyte ion accumulation than on the surface area and pore size distribution, affording high electrochemical performance.