Oxidative exfoliation of spent cathode carbon:A two-in-one strategy for its decontamination and high-valued application

Runze Chen, Yuran Chen, Xuemin Liang, Yapeng Kong,*, Yangyang Fan, Quan Liu, Zhenyu Yang,Feiying Tang, Johnny Muya Chabu, Maru Dessie Walle, Liqiang Wang,*

1 Henan Province Industrial Technology Research Institute of Resources and Materials, School of Material Science and Engineering, Zhengzhou University, Zhengzhou 450001, China

2 College of Chemical Engineering, Zhengzhou University, Zhengzhou 450001, China

3 College of Chemical Engineering, Xiangtan University, Xiangtan 411105, China

4 Foshan Green Intelligent Manufacturing Research Institute of Xiangtan University, Foshan 528010, China

5 Department of Chemistry, Faculty of Science, University of Lubumbashi, Lubumbashi, BP 1825, Democratic Republic of the Congo

6 College of Science, Bahir Dar University, Bahir Dar 79, Ethiopia

Keywords:

ABSTRACT

1. Introduction

According to the 2021 report from the International Aluminum Institute, the world primary aluminum output of 2020 reached as high as 65 million metric tons[1].However,the primary aluminum production is commonly accompanied by the emission of a large amount of waste which causes a series of environmental and ecological issues [2,3]. Among the various wastes, the spent cathode carbon (SCC) has received particular attention. The SCC is generated during the overhauling of the aluminum reduction cell [4],and it is reported that about 22 kg of SCC would be discharged per ton of primary aluminum production [5]. The SCC has been classified as hazardous solid waste since it contains such a large amount of fluorides and cyanides that exceed the discharge standard. Specifically, SCC contains about 20%–40% (mass) fluorides(e.g., NaF, CaF2, and NaAlF6) and 0.01%–0.02% (mass) cyanides [6].If not handled properly, the hazardous constituents can easily be transferred into the atmosphere, soil, and groundwater, leading to severe environmental and health crises [7]. SCC comprises carbon resources (～30%–70% (mass)) [8]. While, cathode carbon(CC) itself is mainly composed of graphitized carbon, whereas the graphitization degree has been further improved after long-term(mostly 5–8 years) use under high temperature [9,10]. Such constituents make the SCC a potential candidate for high-value graphite resources. However, the SCC cannot be utilized directly due to the embedded hazardous substances [11]. Thus several strategies have been proposed for removing the contaminants in the SCC for recycling purposes. Current methods for SCC treatment include physical separation [12], chemical leaching [13], and high-temperature processing [14]. For instance, Zhang and coworkers [15] found that ultrasound-assisted caustic leaching can efficiently remove the fluorides in the SCC, and they were able to achieve 94.72% of carbon recovery. Comparatively, the hightemperature method relies on the combustion process to convert the hazardous materials inside of the SCC from solid phases into gases, thus removing them. For example, Koppala and co-workers[8] used high-temperature roasting to remove fluorides, and their optimal removal rate was 88% under 1499 °C for 3.33 h. Similarly,Liu and co-workers [16] adopted microwave high-temperature roasting to remove fluorides in SCC. Although these methods exhibited certain perspectives in eliminating the fluorides and cyanides in SCC, it remains challenging to remove those hazardous substances completely, since most of them were encapsulated by thick carbon layers after serving for several years. Consequently,the disposal of SCC has yet still relied on being passively openpiled or simply buried, which is a waste of resources and associated with high running costs and secondary pollution issues[6,17]. On the other hand, graphene oxide (GO) or GO-like carbon structures have received wide attention in diverse fields due to their unique chemico-physical properties [18]. Notably, due to its relatively large surface area, rich pore structure, strong chemical stability, and abundant oxygen-containing functional groups, GO has been extensively explored for water purification [19–22]. For example,Wang et al.[23]found that polyacrylamide(PAM)grafted GO could adsorb radionuclides in radio-active water. Yang et al.[24] demonstrated that GO was an effective absorbent for Cu2+because there existed a strong electrostatic interaction between Cu2+and oxygen atoms within GO.Moreover,Zr/Fe-MOFs/GO composites were found capable of efficiently removing the tetracycline hydrochloride in water [25]. Notably, GO is generally fabricated from graphite by the chemical exfoliation method, whereby the‘‘thick”carbon can be engineered into a single-or few-layer carbon structure[26].Therefore the carbon presents in the SCC with a high graphitization degree can find suitable utilization if converted into GO-like structure. Meanwhile, the embedded fluorides and cyanides would be well exposed and thus easier to be removed.More importantly, it could promote recycling utilization and enhance the economic value of SCC.

Considering the above hypothesis, in this paper, we adopted a modified Hummers method to exfoliate SCC into SCC-GO for recycling purposes and minimize its ecological and environmental impact.Intriguingly,after the exfoliation,the originally embedded hazardous substances were sufficiently exposed and could be almost entirely removed(i.e.,99.9%fluorides removal rate)by continuous acid and water leaching. Moreover, structural characterization reveals that the as-produced SCC-GO exhibits a defected structure with enhanced specific surface areas and plenty of oxygen-contained functional groups,making it an excellent absorbent for the removal of organic pollutants in water (Fig. 1(a)).

2. Experimental

2.1. Preparation of SCC-GO using SCC

SCC was obtained from Henan Zhongfu Industrial Co., ltd. The element composition of SCC is summarized in Table 1. As shown in Table 1, the SCC is mainly comprised of 41.7% C, 31.4% F, 24%Na, and 21.6% O (mass fraction). The SCC was crushed into SCC powder of specific size ranges, i.e., >125 μm, 74 to 125 μm,and < 74 μm, respectively. A modified Hummers method was exploited to exfoliate SCC [27]. Specifically, in an ice bath, 2 g of SCC powder,1.2 g of NaNO3was added to 48 ml 98%H2SO4under continuous stirring. Then 30 min later, 8 g KMnO4was gradually added to the above mixture solution at room temperature. After stirring for 2 h, 240 ml DI water was added dropwise, and the obtained mixture was heated to 90 °C for another 1 h. Subsequently, the mixture solution was cooled to 60 °C, followed by the addition of 6 ml 30% H2O2. The suspension was later washed with HCl (0.5 mol?L–1) and deionized (DI) water three times,respectively. Finally, the gained suspension was freeze-dried to obtain the SCC-GO.

2.2. Characterization

The morphology of the as-prepared sample was characterized by transmission electron microscopy (TEM, Titan G2 60–300 with spherical aberration correction, FEI, USA), field-emission scanning electron microscopy (FE-SEM, HeliosNanoLab 600i Dual Beam FIB/FE-SEM, FEI, USA), and atomic force microscopy (AFM, SPM-9600, Shimadazu, Japan). X-ray fluorescence (XRF, XRF-1800 Shimadzu,Japan)was used for elemental analysis.The content of fluorides and cyanides in SCC and SCC-GO was measured by an ion activity meter (PXSJ-216F, Shanghai Inesa Scientific Instrument Co.,ltd.,China)based on the ion selective electrode method(CETC,1996).The crystal phases were examined by powder X-ray diffraction(XRD,Bruker D8 Advance,GER).N2adsorption/desorption was measured by static N2physisorption at -196 °C with AutosorbiQ analyzer (NOVAe, Quantachrome, USA). The surface area was calculated using the multipoint Brunauer-Emmett-Teller (BET)method implemented on a Micromeritics ASAP 2020 Plus 2.00.Raman spectroscopy (LabRAM HR Evolution, HORIBA JY, France)was applied to measure the degree of defects and graphitization.The X-ray photoelectron spectrometer (XPS, ESCALAB Xi+,ThermoFischer,USA)was applied for elemental analysis.The functional groups of SCC-GO were investigated by Fourier transform mid-infrared (FT-IR, Nicolet iS50, USA).

2.3. Adsorption investigation

The adsorption isotherm was conducted by the following procedures, 30 mg of GO was added into 40 ml of methylene blue (MB)solution(pH 6) with concentrations ranging from 50-1000 mg?L–1,and the mixture solution was stirred for 90 min at room temperature. The mixture solution was centrifuged, and the supernatant was analyzed by a UV–Vis spectrophotometer. The equilibrium concentration (ce) and equilibrium adsorption capacity (qe) of MB was attained by the calibration curve of MB. The effect of adsorption temperature was carried out to study the endothermic/exothermic characteristics of the adsorption process.Two temperature points were adopted,i.e.,25°C and 50°C.SCC-GO(30 mg)as SCC of different sizes were added into 40 ml MB solution(50 mg?L–1)to evaluate the size effect of SCC on the adsorption capacity of the gained SCC-GO. The pH influence of solution on MB adsorption was conducted at 25 °C, with the pH value ranging from 2 to 12,adjusted by dilute HCl or NaOH aqueous solution.

2.4. Theoretical calculations

Theoretical calculations were carried out on the Material Studio,using the plane wave basis set with an energy cutoff of 299.3 eV and the generalized gradient approximation parameterized by Perdew, Burke, and Ernzerhof (GGA-PBE) for exchange–correlation functional.The model structures are fully optimized by the convergence tolerance of 10–4eV?atom-1. Use the Grimme method for DFT-D correction. A graphene monolayer was exploited for simulating the carbon structure using a supercell consisting of 8 × 8 graphene unit cells, where surface carbon vacancy, edge carbon vacancy, and vacancy modified by O-contained functional groups were introduced. The model structures are optimized by the modules of CASTEP.

The adsorption energy(Ead)was calculated by using the following equation:where Esub+suris the total energy of surface adsorbed with the substrate;Esuris the energy of surface without substrate and Esubis the energy of the substrate.

Fig.1. (a)Scheme diagram of the formation mechanism and application of SCC-GO.(b,c)SEM images of SCC:(d)SEM(inset presents the photographs of Tyndall effect of SCCGO solution), (e) TEM, and (f) AFM images of SCC-GO. (g) Sectional height of the selected area within SCC-GO.

Table 1 Elemental analysis results of raw spent cathode carbon (SCC)

3. Results and Discussion

3.1. Structure and morphology

A modified Hummers method was exploited to upgrade SCC into SCC-GO[28].For typical Hummers’strategies,oxidizing agents and strong acid were introduced into the aqueous, whereby graphite was exfoliated into single or few-layer,partially oxidized carbon structures. SEM, TEM, and AFM characterization were carried out to provide a deep insight into the structural transformation of SCC to SCC-GO. SEM images of SCC showed that it possessed a compact structure (Fig. 1(b) and (c)). By contrast, SCC-GO portrayed a loose structure composed of layer-like carbon sheets(Fig. 1(d)) and exhibited a pronounced Tyndall effect (Inset in Fig.1(d)),revealing the formation of hydrophilic nanoscale carbon.TEM images of SCC-GO showed that SCC-GO possessed a wrinkled,sheet-like structure (Fig. 1(e)). Most of the SCC-GO possessed a thickness of 1.25-1.5 nm as revealed by the AFM (Fig. 1(f) and(g)), demonstrating that it contains single- or few-layers carbon[29]. Notably, the gained SCC-GO showed a Brunauer-Emmett-Teller (BET) surface area of 164.42 m2?g-1, much larger than that of the SCC (18.70 m2?g-1) (Table S1).

Table 2 Removal rate of fluorides by our proposed strategy

The fluorides content in SCC and SCC-GO was measured by ion activity meter based on the ion selective electrode method (CETC,1996). As listed in Tables 1 and 2, the fluorides content in SCC-GO was four orders of magnitude smaller than that in SCC, suggesting that over 99.9% of the fluorides in SCC can be removed by the oxidative exfoliation process.Meanwhile,only a neglected amount of cyanide ions could be detected after oxidative exfoliation treatment (see Table S2). XRD was further exploited to investigate the structural changes of SCC before and after chemical oxidation. As shown in Fig. 2(a), there was a strong peak at 26.5° and a slight peak at 54.5°,corresponding to(002)and(004)planes of graphite,respectively, revealing that the SCC was highly graphitized [30].Besides, characteristic peaks belonging to CaF2, NaAlF6, and NaF could be clearly observed. By contrast, SCC-GO possessed GO-like XRD profiles with a distinctive peak at 9.98° belonging to the diffraction peak of (002) of graphene layers [31–33]. Notably, the peaks belonging to CaF2, NaAlF6, and NaF disappeared in the XRD pattern of SCC-GO, suggesting the successful removal of fluorides after chemical exfoliation and the acid/aqueous leaching, which coincides well with the result of the fluorides content measurement [13]. Fig. 2(b) and (c) represent the XPS spectra of SCC and SCC-GO. Peaks indexed to the binding energy of Na or F could be clearly seen in the survey spectra of SCC but disappeared in SCCGO. These findings agree well with the XRD results. The deconvoluted C 1s of SCC and SCC-GO are shown in Fig.2(d)and(e).Compared with SCC, there were more oxidized functional species in SCC-GO. Specially, SCC-GO showed peaks at 284.8, 286.92, and 288.32 eV, corresponding to C=C/C—C (66.25%), C—O—C/C–OH(24.50%), and O–C=O (7.87%), respectively, indicating the formation of O-containing groups on the surface of SCC-GO after the oxidative exfoliation (Fig. 2(e)) [34,35]. We subsequently investigated the level of defects in the SCC before and after chemical exfoliation by Raman spectra.SCC had a weak peak at 1368 cm-1and a strong one at 1582 cm-1, which corresponded to the defectinduced D band and the first-order scattering of the E2gmode for sp2carbon lattice, respectively (Fig. 2(f)) [36]. The ID/IGis often exploited for describing the defect degree in graphitic materials.The ID/IGfor SCC and SCC-GO was 0.2571 and 0.9466,respectively,suggesting that SCC-GO has more defects (Fig. 2(f)) [37].

FT-IR further confirmed the existence of O-containing groups in SCC-GO. As shown in Fig. 3(a), no distinctive spectrum indexed to O-containing groups could be observed for SCC.By contrast,in the FT-IR spectra of SCC-GO, peaks at around 3000 – 3600 cm-1could be indexed to the stretching vibrations of the hydroxyl (—OH)group, while the one at ～1720 cm-1to the stretching vibrations of carboxyl C=O groups [38], and the one at 1051 cm-1to the bending vibration of C—O (Fig. 3(b)) [39]. The above results demonstrated that we were able to convert SCC into SCC-GO with a few layers, defected structure, enhanced specific surface areas,and a large amount of O-containing groups.

3.2. Adsorption investigation

Fig.2. (a)XRD patterns of SCC and SCC-GO.Survey XPS spectra of(b)SCC and(c)SCC-GO.C 1s XPS spectra of(d)SCC and(e)SCC-GO.(f)Raman spectra of SCC and SCC-GO.

Fig. 3. FT-IR spectra of (a) SCC and (b) SCC-GO.

SCC-GO owns large specific surface areas and plenty of oxygencontained functional groups and defects.Such characteristics make SCC-GO an excellent candidate for pollutants remediation in water.Methylene blue was taken as a pollutant model to investigate the ability of SCC-GO to remove organic contaminants. MB solution exhibited a strong UV–Vis absorbance centered at 665 nm (Fig. 4(a)). After incubation with SCC-GO, the solution became colorless,accompanied by the remarkable attenuation of the absorbance peaks(Fig. 4(a)).As shown in Fig.4(b), the equilibrium adsorption capacity increases rapidly when the cevalue is at a low stage(<37 mg?L–1), while it increases much slower at a high cevalue(37 mg?L–1). 99.5% of MB was removed at an initial concentration of 125 mg?L–1and could maintain a 90% removal even when the initial MB concentration was up to 250 mg?L–1. The maximum value of equilibrium adsorption capacity was calculated to be 347.2 mg?g-1at an initial concentration of 800 mg?L–1.The adsorption process was then investigated, following Langmuir and Freundlich model [40,41]. The Freundlich isotherm (Eq. (2)) is often exploited to describe the adsorption process on heterogeneous surfaces, which is not restricted to the monolayer [42]. Whereas the Langmuir model (Eq. (3)) is based on the assumption that adsorption took place on homogeneous surfaces and represents monolayer adsorption [43]. The maximum adsorption capacity (qm)and Freundlich constant(KF)were calculated by fitting the adsorption results into the corresponding equations.

The most correlations (R2) for the Langmuir model and Freundlich model of MB adsorption by GO was 0.941 and 0.999(Fig. 4(c) and (d)), respectively, indicating the Freundlich model was more suitable for studying the adsorption process than Langmuir model did. And the result suggests that the adsorption process is more likely to be a multi-layer adsorption. Specifically,the 1/n value of the Freundlich model was calculated to be 0.2369, much smaller than 1, suggesting the adsorption process proceeded smoothly [44]. The maximum adsorption capacity (qm)was calculated to be 347.2 mg?g-1according to the Langmuir model, which agreed well with the experiment results(355.9 mg?g-1), and its qmcould rival that of the recently welldesigned carbon-based absorbents (Table 3). Notably, SCC only exhibited a neglected qmof 14.6 mg?g-1(Fig. S3(a)).

Table 3 Maximum adsorption capacity (qm) of various absorbents

Fig. 5(a) shows the relationship between the adsorption capacity and temperature. The adsorption capacity decreased remarkably when the solution temperature increased to 50 °C, proving the adsorption process an exothermic reaction. The size effect of SCC on the absorbance efficiency of the derived SCC-GO was also investigated.As shown in Fig.5(b),SCC-GO derived from SCC of different sizes exhibited a similar MB remove efficiency, i.e., 99.88%,99.87%, and 99.67% for SCC < 74 μm, 74 μm < SCC < 125 μm, and SCC > 125 μm, respectively. It reveals that the grain size of SCC exhibits neglected influences on the adsorption ability of the derived SCC-GO. We investigated the influences of pH value on the adsorption of MB. As is shown in Fig. 5(c), the adsorption capacity increased with pH value, which rose from 265 mg?g-1to 332 mg?g-1with pH changing from 2 to 12. This situation can be attributed to that SCC-GO possesses plenty of O-containing groups such as –OH and –COOH. These groups can interact with MB strongly via electrostatic interaction, which is highly dependent on the pH value. Specifically, SCC-GO had a higher negative ζpotential when the pH value increased,promoting the electrostatic interaction between MB cations(Fig.5(d))[44].By contrast,methyl orange (MO), an anionic dye, was exploited to evaluate the absorbance capacity of SCC-GO.Its adsorption process was also fitted by Langmuir model and Freundlich model (Fig. S4(c) and (d)). The equilibrium adsorption capacity of MO was calculated to be 26.06 mg?g-1, much lower than that of MB (347.20 mg?g-1). The results suggest that the electrostatic interaction plays a critical role in the adsorption process of MB onto SCC-GO.

Theoretical calculation was performed to further understand the origin of the excellent adsorption capability of SCC-GO towards MB (Fig. 6(a)). To simplify the calculation, we only considered the influences of a particular functional group,i.e.,–OH and–COOH,or defected structures,i.e.,edges or defects.As shown in Fig.6(b),(c),and(d),compared with perfect graphene(G-1)structures,the ones possessing defected structures, i.e., vacancy graphene (VG) and edge graphene(EG),exhibited a stronger adsorption ability,represented by the relatively low adsorption energy (–0.3 eV for G1,–1.1 eV for VG-1, and –1.2 for EG-1). Moreover, carbon structures attached by O-containing groups exhibited much lower adsorption energy. Those with –COOH showed stronger adsorption strength than those with –OH, while oxygen-containing groups located at the edge favored the adsorption of MB more than that on the vacancy (Fig. 6(e)-(i)). Note that compared with SCC, the SCC-GO is smaller in size and thus possesses a higher content of edge structures.Also,it possesses a higher level of defects and more oxygencontaining functional groups. These results can explain why SCCGO demonstrated an excellent adsorption ability.

Fig. 4. (a) UV–Vis spectra of MB solution before and after adsorption by SCC-GO, and the inset presents the corresponding photographs. (b) Adsorption isotherm of MB adsorption by SCC-GO. Langmuir (c) and Freundlich (d) isotherms for the adsorption of MB by SCC-GO.

Fig. 5. The influence of (a) temperature, (b) SCC grain size, and (c) pH on the adsorption of MB over SCC-GO. (d) The zeta potential of SCC-GO under different pH values.

Fig.6. (a)The molecule structure of methylene blue.Calculated adsorption energies of methylene blue on:(b)perfect graphene(G-1),(c)vacancy graphene(VG-1),(d)edge of graphene (EG-1), (e) vacancy graphene with –OH (VG-2), (f) vacancy graphene with –COOH (VG-3), (g) edge of graphene with –OH (EG-2), (h) edge of graphene with –COOH (EG-3). (i) Summary diagram of adsorption energies of methylene blue on calculated models.

4. Conclusions

In conclusion, the oxidative exfoliation is demonstrated to be an efficient strategy for converting SCC into valuable chemicals.Oxidative exfoliation treatment can fully expose the hazardous substances (e.g., fluorides, cyanides) that are originally embedded within carbon layers, facilitating their subsequent removal;meanwhile, after treatment, the gained samples possess fewlayers carbon structures and plenty of O-containing groups on their surfaces, making it an excellent candidate for adsorbing contaminants in water. As a result, the maximum adsorption capacity of SCC-GO towards methylene blue can reach 347 mg?g-1owing to its relatively large specific surface areas and plenty of O-containing functional groups. This work provides an efficient recycling route and utilization of hazardous carbonaceous wastes.

Declaration of Competing Interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Acknowledgements

This work was supported by the National Natural Science Foundation of China (22008221); Startup Research Fund of Zhengzhou University(32211716);Key Scientific Research Projects of Colleges and Universities in Henan Province(21A530005);Guangdong Basic and Applied Basic Research Foundation (2021A1515110789);Hunan Provincial Natural Science Foundation of China(2022JJ40431);Zhengzhou Collaborative Innovation Major Project.

Supplementary Material

Supplementary data to this article can be found online at https://doi.org/10.1016/j.cjche.2022.10.020.