Cobalt oxide nanoparticles embedded in flexible carbon nanofibers: attractive material for supercapacitor electrodes and CO₂ adsorption

Abstract

Introducing flexibility and high porosity into carbon nanofibers (CNFs) is one of the critical challenges for the next generation of multifunctional energy storage and CO₂ adsorption materials. Herein, we developed an efficient strategy for the controllable fabrication of a flexible and mechanically robust Co₃O₄ nanoparticles (NPs) doped CNFs (CNFs-Co) hybrid membrane via electrospinning and subsequent carbonization treatment. The quantitative pore size distribution and fractal analysis revealed that the CNFs-Co possessed a tunable porous structure with high surface area of 483 m² g⁻¹. Therefore, it exhibited exceptional performance in CO₂ capture, i.e. a high CO₂ adsorption capacity of 5.4 mmol g⁻¹ at 1 bar and room temperature. Electrochemical measurements performed on CNFs-Co for supercapacitor applications demonstrated very high capacitance of up to ∼911 F g⁻¹ at 5 mV s⁻¹ (76% capacitance retention after 1000 cycles) in 1 M H₂SO₄ solution. The successful synthesis of this hybrid membrane may also provide new insights towards the development of materials for various multifunctional applications.

---

1. Introduction

Ordered carbon constituents have attracted significant attention owing to their high surface area, tunable porous structure, and admirable chemical and physical properties. They are commonly used in catalyst supports, gas storage, separation, and in energy storage materials,^1–4 although, a large number of the carbon constituents only have a small fraction of precise active positions with extremely hydrophobic planes, which inhibits their practical applications. Therefore, various efforts have been made to improve the surface properties of the carbon constituents.^2,5

Carbon materials such as graphene, carbon nanotubes,⁶ mesoporous carbon⁷ and CNFs, are among the most common materials for energy storage and adsorption. Excellent conductivity, high chemical stability, tunable structural flexibility, and cost-effectiveness were the key reasons for the selection of CNFs for the current work focusing on the synthesis of a flexible membrane. Meanwhile, a versatile and industrially feasible method, i.e. electrospinning, has achieved remarkable interest for the fabrication of CNFs in both academia and in industry.⁸ Significantly, the high mechanical strength and flexibility, good electrical conductivity, excellent porosity and surface area to volume ratio of cobalt oxide doped CNFs could be further tuned and enhanced when produced by electrospinning.^9,10

The reported literature shows that RuO₂ is considered as an ideal material for electrochemical capacitors, owing to its exceptional electron/proton conductivity with high rate capability, and huge theoretical specific capacitance.^11,12 Nevertheless, its toxicity and high cost hinder its large-scale application in supercapacitors therefore, substitute electrode materials with outstanding electrochemical performance are being targeted for advanced supercapacitors.^13,14 Recently, NiO, CoO, MnO₂,^15,16 etc. possessing attractive supercapacitive properties have been reported as substitute electrode materials for RuO₂.¹⁷ Moreover, Co₃O₄ membranes have been found to exhibit substantial capacitance up to 3560 F g⁻¹, having high redox reactivity.¹⁸ However, similar to other metal oxides, Co₃O₄ still suffers from poor electrical conductivity and limited electrochemical stability during cycles. To address these issues, nanostructured Co₃O₄ ¹⁹ has been assembled with a carbon matrix to form composites in various forms for supercapacitors.

CNFs have been turned into flexible membranes to extend their range of applications, such as in numerous portable electronic facilities and adsorption materials.^20–22 However, it is hard to manufacture them at a large scale due to the complicated fabrication process, and these integrated membranes are generally brittle.^2,23 Although many previous approaches have been proposed to predict the flexible properties of CNFs with highly-idealized structures,²⁴ a predictive understanding of mechanical properties and rupture behavior, based on the internal structure of CNFs have proven challenging.

Additionally, an extensive reliance on fossil fuels for energy production has ignited concerns over global warming, the most severe environmental issue of this time. Porous solid sorbents are prospective candidates for the purpose of CO₂ capture including porous carbon,⁵ metal–organic frameworks (MOFs),²⁵ porous polymer networks (PPNs),²⁶ covalent organic polymers (COPs), zeolites,^27,28 and amine functionalized materials.^29–31 Among these sorbents, carbon based components such as CNFs have many advantages i.e., their high thermal, mechanical, and chemical stabilities, low production cost, hydrophobic properties, and ease of modification with other NPs. However, CNFs are generally brittle in nature, offering poor mechanical characteristics, high flow rate and lower efficiency for CO₂ adsorption. Recently, three-dimensional (3D) metal oxide based flexible CNFs composites with interrelated paths have attracted growing attention,⁵ since they can display lower diffusive resistance to mass and ion transport from micropores, good flexibility and a high surface area for active site dispersion.

CNFs have been used for the controlled analysis established for a superficial mechanism of elasticity with the help of metal oxides. Different hybrid materials have thus been made from Co₃O₄,³² with different carbon materials. Co₃O₄/graphene and Co₃O₄/carbon nanotubes (CNTs) have been synthesized with improved conductivities of electrodes, and a large surface area between the active materials and electrolytes for efficient chemical reactions and CO₂ adsorption. Apart from graphene and CNTs, electrospun CNFs have been widely applied as an electrode material for various types of energy storage applications.

In this contribution, we developed a facile in situ technique for the synthesis of flexible CNFs membranes derived from hierarchical porous CNFs embedded with Co₃O₄ NPs. CNFs-Co hybrid membranes were prepared by altering the relative quantity of cobalt acetylacetonate (Co(acac)₂) in the precursor solution. The PAN nanofibers were carbonized at a high temperature in order to increase the electrical conductivity of the final CNFs matrix. The stepwise process of CNFs formation and FE-SEM (field emission electron microscopy) images of the pure, oxidized and doped nanofiber membranes are represented in Fig. 1. The resulting CNFs-Co exhibited a large surface area and high flexibility with an exceptional CO₂ capturing capacity of 5.4–6.9 mmol g⁻¹ at 25 °C and 0 °C (1 bar), also offering a reasonable specific capacitance of 911 F g⁻¹ (in 1 M H₂SO₄), showing good enough potential to be used in supercapacitors as well.

Fig. 1 (a) Schematic diagram showing the synthetic steps of CNFs-Co. (b) Digital images of the electrospun Co doped membrane, pre-oxidized membrane, and CNFs-Co-2. FE-SEM images of (c) Co NFs-2, (d) pre-oxidized Co NFs-2, and (e) CNFs-Co-2 and cross section (inset) of CNFs-Co-2.

2. Experimental section

2.1 Materials

Polyacrylonitrile (PAN, M_w = 90 000) was supplied by Spectrum Chemicals & Laboratory Products Co., Ltd, USA. Dimethylformamide (DMF), polyvinyl alcohol (PVA, M_w = 86 000), sulfuric acid (H₂SO₄) and cobalt acetylacetonate (Co(acac)₂) were procured from Shanghai Chemical Reagents Co., Ltd., China. All chemicals were used as received without further purification.

2.2 Synthesis of Co nanoparticles/CNFs

2.2.1 Fabrication of precursor fiber. PAN concentration in solutions was made stable at 10 wt%, and the total amount of Co(acac)₂ was adjusted to 0, 2, 3, and 4 wt%. Typically, a suitable quantity of salt, i.e., 2, 3, and 4 wt% was dissolved in DMF, after stirring for 20 min and PAN powder i.e. 5 g was added gradually to form a solution (total weight of solution = 50 g). A reddish sticky solution was obtained after 5 h vigorous stirring at room temperature. The obtained solution was loaded into syringes that were attached to a high voltage power supply (25 kV). The flow rate of solution was maintained at 1 mL h⁻¹ using a syringe pump (SOF Nanotechnology Co., Ltd., China). A grounded metallic roller covered with copper mesh collected the resultant fibers 20 cm away from the needle tip. The temperature and humidity of the electrospinning compartment were ∼25 °C and 50 ± 2%, respectively. Pure PAN fibers were also prepared in a similar manner for comparison.

2.2.2 Synthesis of CNFs-Co. The temperature for drying precursor fibers was 70 °C for 1 h under vacuum to remove the solvent, and then the material was stabilized in an oven at 280 °C for 2 h with external tension. Afterwards, the obtained nanofibrous membranes were annealed at 850 °C for 2 h with a heating rate of 2 °C min⁻¹ under N₂ flow in a tube furnace. The resulting membranes were smooth and flexible, and the carbonized final sample was denoted as CNFs-Co. Another three samples, CNFs-Co-2, CNFs-Co-3 and CNFs, were prepared by using 1 g, 1.5 g and 0 g of Co(acac)₂ for comparison, in the precursor solution, respectively.

2.3 Structural characterization

The morphology of the as-synthesised samples was studied using FE-SEM (S-4800, Hitachi Ltd., Japan) operating at 2 kV and 5 kV. Raman spectra were examined using a micro-Raman spectroscopy system (in Via-Reflex, Renishaw, Co., UK). High resolution transmission electron microscopy (HR-TEM, JEM-2100F, JEOL Ltd., Japan) was selected to complete the characterization of nanostructures, with selected area electron diffraction (SAED). The phase structures of the samples were characterized using X-ray diffraction (XRD) (D/Max-2550 PC Rigaku Co., Japan, Cu Kα, λ = 1.5406 Å). The Brunauer–Emmett–Teller (BET) surface area and porous structure were analyzed by N₂ adsorption–desorption isotherms using an automatic adsorption system (ASAP 2020). X-ray photoelectron spectroscopy (XPS) was carried out using a Thermo Scientific ESCALAB 250Xi X-ray source to obtain the elemental composition.

2.4 Electrochemical measurements

CNFs-Co was used as the working electrode in electrochemical tests, namely cyclic voltammetry (CV), chronopotentiometry (CP) and electrochemical impedance spectroscopy (EIS), on a CHI660E electrochemical workstation (Chenhua, Shanghai). Galvanic charge/discharge measurements were carried out in a three-electrode cell with a working electrode, a platinum plate counter electrode and a saturated calomel electrode (SCE) as the reference electrode at room temperature. 1 M H₂SO₄ aqueous solution was used as an electrolyte. The cycling performance test was carried out with a CT2001D tester (China). A PVA/H₂SO₄ electrolyte was sandwiched by assembling a coin cell (CR 2025) with CNFs-Co electrodes and the electrolyte (PVA/H₂SO₄). The specific capacitance of these self-supported CNFs-Co can be calculated using the following equation (C_s can be calculated from the CV curves):

C_s = Q/mΔU, (1)

where Q is the average of the absolute values of charges during the charge/discharge processes, and ΔU is the potential window.³³ The enclosed area S in the CV loop can be obtained using the equation:If V is the potential scan rate, the enclosed area in the CV curve is proportional to the average charge Q at a certain scan rate according to the equation.¹⁵ So eqn (1) can be replaced by:

C_s = S/2VmΔU, (3)

It can be seen that for electrodes undergoing similar testing parameters (the same mass, potential window, and scan rate), where m is the mass of the two electrodes. For cyclic voltammetry, the specific capacitance was taken as,^36,37

In this equation V_h is the high voltage cut-off, V_I is the low voltage cut-off, and v is the scan rate.

2.5 CO₂ adsorption measurement

A Micromeritics ASAP 2020 apparatus was used for CO₂ adsorption measurements. First, the CO₂ adsorption capacities of the CNFs-Co samples (around 0.02139 g) were measured under different parameters. Outgassing of the samples was carried out in ASAP 2020 by heating and evacuation. The chamber was then back-filled with CO₂ and the temperature swung from 25 °C to a high temperature, while recording the mass lost, which was subsequently gained upon cooling to 25 °C. Finally, the CO₂ adsorption at 25 °C was determined by controlling the flow rates of both N₂ and CO₂.

3. Results and discussion

3.1 Structure and morphology

Fig. 2 presents FE-SEM images of the as-synthesized CNFs-Co with different Co₃O₄ loadings, which indicate that the resultant nanofibers were randomly oriented as three-dimensional (3D) open-cell nonwoven membrane geometry (see S1 in ESI†). The fiber diameters of CNFs-Co-4, CNFs-Co-3 and CNFs-Co-2 were 674, 743 and 796 nm, respectively, which clearly showed that with the increase of the salt content, the diameter of the CNFs-Co membrane changes. It has been reported that increasing the conductivity of spinning solutions obtained by introducing salt would lead to an increase in the net charge density, which results in a decrease in the fiber diameter. The high-magnification SEM images of surface (Fig. 2b–d, insets) and a cross-sectional image of Co₃O₄ doped CNFs (Fig. 1e) clearly show that the Co₃O₄ NPs are not only dispersed uniformly on the surface of the CNFs (Fig. 2b–d), but are also embedded into the CNF framework, as evident from the cross-section of the Co₃O₄ doped CNFs. The uniform dispersion of electroactive NPs (e.g. Co₃O₄) on and/or embedded into 3D porous hierarchical CNFs, provides them with a large surface accessible by electrolyte via the inherent porous diffusion channels and an excellent electrical contact with the substrate, resulting in sufficient faradaic reactions for efficient energy storage.³⁸

Fig. 2 FE-SEM images of CNFs with different salt contents of (a) 0, (b) 2, (c) 3, and (d) 4 wt%. (e) Digital images presenting the robust flexibility of the CNFs-Co-4 membranes.

Although previous reports have proposed to predict the flexible properties of CNFs with highly-idealized structures, the poor mechanical properties and rupture behavior of CNFs are still big challenges to overcome. Here, we use organized analysis to establish an apparent mechanism of elasticity for the nanostructures of CNFs membranes, that has been revealed in Fig. 2e. The Co₃O₄ nanoparticles and the carbon layers are uniformly distributed in the carbon matrix, forming an organized composite nanostructure as explained previously (see movie 1 in ESI†). The external stress on the CNFs membranes leads to bending distortion of the fibers.³⁴ For pure CNFs, the pressure is concentrated on the bending area, resulting in rapid extended cracks across the fiber matrix, and finally causing the fracture of fibers (for pure CNFs, see movie 2 in ESI†).

Nitrogen adsorption–desorption measurements at 77 K were conducted to investigate the porous textures of the as-prepared CNFs. Fig. 3a shows that the isotherms of all samples demonstrate typical adsorption/desorption behavior, including micropore filling, monolayer adsorption, multilayer adsorption and capillary condensation.^35,36 For pure CNFs, the majority of nitrogen adsorption occurred at a low relative pressure (P/P₀ < 0.1), and a plateau appeared at a middle relative pressure, demonstrating a typical microporous structure with few mesopores. A slight decrease of the adsorption plateau could be attributed to the distortion of micropores. To further investigate the formation of the porous structure of the samples in detail, the specific surface area, total pore volume, average pore diameter and pore size distribution were determined using the BET surface area method (Fig. 3). Table 1 shows the specific surface areas of conventional CNFs samples for CNFs-Co-2, CNFs-Co-3, and CNFs-Co-4, i.e. ∼131, 416, 457 and 483 m² g⁻¹, respectively, indicating the major contributing role of Co NPs on determining the specific surface area.

Fig. 3 (a) N₂ adsorption–desorption isotherms, and (b) 2D-NLDFT pore size distribution curves of relevant CNFs-Co materials.

Table 1 The summary of pore structure parameters of relevant pure CNFs and doped CNFs

Samples	SBET^a (m^2 g^−1)	Smeso^b (m^2 g^−1)	Vtotal^c (cm^3 g^−1)	Vmeso^d (cm^3 g^−1)	PVFmeso^e (%)	Dav^f (nm)
a Total surface area was calculated by the BET method.b Mesopore surface area was calculated by the Barrett, Joyner, and Halenda (BJH) method.c Total pore volume was estimated was calculated at P/P0 = 0.99.d Vmeso was calculated by the BJH method.e PVFmeso indicates the pore volume fraction of mesopores.f Average pore width was estimated by the BET method (4V/A by BET).
Pure CNF	131	43	0.7	0.1	14	2.81
CNFs-Co-2	416	307	0.30	0.11	33	2.91
CNFs-Co-3	457	340	0.29	0.12	34	2.94
CNFs-Co-4	483	388	0.3	0.13	36	2.95

---

On the other hand, the pore size distribution and pore volume in active materials are very significant characteristics for supercapacitors and CO₂ adsorption. The enhanced mesopore volume provides a low resistance pathway and short diffusion route for ions because of the admission of a larger amount of electrolyte into the pores. Therefore, the adjustment of salt content directly effects the performance of supercapacitors. In order to comprehensively investigate the hierarchical porous structures of CNFs/CNFs-Co, pore size distribution analysis was performed using a two dimensional nonlocal density functional theory (2D-NLDFT) method.³⁹ Fig. 3b depicts the 2D-NLDFT pore size distribution curves. As expected, CNFs-Co exhibited both microporous (pore width < 2 nm) and mesoporous (pore volume 2–50 nm) structures, more polydisperse peaks appeared in the range of 2–15 nm, and the volume of pores increased greatly indicating an efficient enhancement of the pore size. The detailed parameters of surface area and porous structures are listed in Table 1.

3.2 Fractal dimension analysis

Fractal analysis has proven to be useful to describe the geometric and structural properties of fractal surfaces and porous structures. The fractal dimension is the number used to quantify those properties which satisfy the requirement of self-similarity.¹⁰ The surface fractal dimension (D) determined in this work characterizes the micro/mesopores surface irregularity of CNFs-Co, investigated by using N₂ adsorption analysis. The N₂ adsorption analysis method was based on the modified Frenkel–Halsey–Hill (FHH) theory of multilayer gas adsorption.³⁷

ln(V/V_mono) = A[ln(ln(p₀/p))] + constant (5)

where V is the amount of N₂ adsorbed at each equilibrium pressure (p), V_mono is the amount adsorbed of monolayer coverage, and p₀ is the saturation pressure. Thus, a plot of ln(V/V_mono) versus ln(ln(p₀/p)) shows a linear trend, and the slope A could be used to calculate D utilizing the expression: A = D − 3, according to the dominant forces of liquid–gas surface tension at high coverage.

The FHH plots shown in Fig. 4 determine a diverse linear trend with different slopes in the high coverage region. The resultant D values of CNFs, CNFs-Co-2, CNFs-Co-3, and, CNFs-Co-4, were about 2.81, 2.91, 2.94, and 2.95 respectively, indicating a typical surface fractal feature. It can be recognized that the D values exhibit a change with the increment of salt content, confirming the enhancement of mesoporous structures. This phenomenon may be contributed to by three aspects: (i) the decomposition of the salts with gas release increased the porosity of the carbon matrix, (ii) a catalytic reaction between the salts and the carbon matrix occurred, resulting in the consumption of carbon,³⁹ and (iii) the extrusion of Co nanoparticles in the carbon matrix created defects in the carbon matrix.

Fig. 4 Plots of ln(V/V_mono) against ln(ln(p₀/p)) reconstructed from the N₂ adsorption isotherms of the relevant CNFs-Co materials.

To further confirm of the existence of Co₃O₄ in CNFs-Co and their distribution in the carbon matrix, TEM analysis with SAED was carried out. Fig. 5 shows the TEM images of CNFs-Co-2 after calcination at 850 °C, demonstrating the intact structure of individual CNFs and also a uniform distribution of Co₃O₄ NPs in the carbon matrix (Fig. 5a). The naturally turbostratic stacking structure revealed graphitic carbon layers parallel to each other along the axis of the fiber as well as apparent bends and branches. The ambiguous lattice fringes of the nano shell and clear diffraction rings in Fig. 5b support the presence of ordered carbon layers and Co₃O₄. The region below the turbostratic stacking structure displays amorphous carbon phase. CNFs-Co nanofibers are undoubtedly recognizable from the HR-TEM images, showing a lattice distance of 0.2 nm (Fig. 5c). Fig. 5d shows the carbon matrix entailed regions with different levels of carbon layers. The FE-SEM and HR-TEM images show no significant change to the morphology and internal structure after measuring the properties of optimized samples after supercapacitance and CO₂ adsorption performance (see S2 in ESI†).

Fig. 5 (a) HR-TEM images of CNFs with a salt content of 2%. (b) SAED pattern, (c) showing lattice distance of CNFs-Co-2, (d) HR-TEM image showing amorphous carbon layers.

The microstructures of the CNFs derived from various concentrations of salts were investigated using Raman spectroscopy. For all samples, pure CNFs, CNFs-Co-2, CNFs-Co-3 and CNFs-Co-4, main peaks centered at 1344, 1339, 1332, 1327 cm⁻¹ (D bands) and 1582, 1598, 1599, 1604 cm⁻¹ (G bands) can be observed respectively, indicating that organic substances underwent fruitful decomposition following carbonization (Fig. 6). The G band originates from the in-plane stretching motion of sp² bonded carbon atoms, and the highly ordered graphitic structure, while the D band corresponds to the amorphous sp³ carbon amorphous carbon. The high intensity ratio means a high degree of graphitization and the relative intensities of the G and D bands (I_G/I_D) are used to evaluate the degree of graphitization. The values of I_G/I_D for pure CNFs, CNFs-Co-2, CNFs-Co-3 and CNFs-Co-4 are ∼0.86, 0.86, 0.92 and 0.97 respectively, based on the fitting curves, indicating that the carbon exists mainly in an amorphous form in the samples. The degree of crystallinity of carbon (R) calculated from the ratios of I_G/I_D for CNFs confirms the enhancement of the graphitization degree with the addition of salts.

Fig. 6 (a) Raman spectra of CNF fibers with various salt concentrations of (a) 0, (b) 2, (c) 3, and (d) 4 wt% and (b) representative XRD graphs of CNFs with salt contents of (a) 0, and (b) 4 wt%.

Besides the graphitic nanofibers, embedded Co₃O₄ nanoparticles were also obtained from thermal decomposition, which could explain the burst-nucleation and crystal growth mechanism. Normally, this conversion would contain three related processes: firstly, the pyrolysis of Co(acac)₂ generates a large amount of cobalt via a series of complicated decomposition reaction and redox reaction. Secondly, the salt concentration exceeds the critical concentration and aggregates to form clusters. Thirdly, in the burst nucleation process, the monomer concentration decreases, and nuclei gradually grow into Co₃O₄ nanoparticles by the consumption of monomers.²⁷

XPS analysis also validated the growth of Co₃O₄ (Fig. S4†). High resolution XPS spectra of Co 2p and O 1s and C 1s for Co₃O₄ are shown in Fig. S4a.† Peaks at 782.4 eV for 2p_3/2 and 797.2 eV for 2p_1/2 show the contribution of cobalt in the doped CNFs (Fig. S4b). The O 1s peak at 540 eV is attributed to the typical metal oxygen bonds and (Fig. S4c†) the carbon 1s spectrum is used for calibration (295.5 eV), shown in Fig. S4d.†The presence of Co₃O₄ is further confirmed with XRD examination. Fig. 6b shows the XRD patterns of the pure CNFs and CNFs-Co-4 (see S3 in ESI† for CNFs-Co-2 and CNFs-Co-3).

As expected, it can be observed that the locations and relative intensities of the diffraction peaks matched well with carbon and standard Co₃O₄ patterns, indicating a successful synthesis of Co₃O₄ and carbon composites. The broad diffraction peak at a 2θ value of 43° can be allocated to the specific reflection (100) diffraction planes of ordered graphitic carbon. The additional broad diffraction peak at 2θ = 26° is indexed to the characteristic (002) reflection of amorphous carbon, which has further been verified by Raman characterization. This proves that the Co NPs are decomposed and carbothermally reduced to CNFs-Co. After annealing in a N₂ atmosphere, three diffraction peaks at 2θ angles of about 36.2°, 51.5° and 75.8° are observed, which can be assigned to the (111), (200), and (220) reflections of the Co₃O₄ NPs. In addition, the size of Co₃O₄ NPs is calculated to be about ∼100 nm.

3.3 Electrochemical measurement

The higher electrochemical performance might result from the following structural features. Firstly, the ultrathin structure of the CNFs-Co facilitates the electrolyte ion and electron transport. Secondly, the porous nature increases the capacity of the electroactive sites. Thirdly, the 3D ordered nanostructure composed of these ultrathin porous nano sheets on the CNFs backbone facilitates the electron transfer and the electrolyte penetration. CV analysis was carried out, in which two fragments of CNFs-Co membranes were used respectively as the positive and negative electrode to construct a coin cell, with H₂SO₄/PVA gel as the electrolyte/separator (see S5 in ESI†).

The electrochemical properties of the CNFs-Co flexible composite membranes are characterized by CV analysis. The CV curves of CNFs-Co-4 in the voltage range of 0–0.5 V are shown in Fig. 7. The typical CV response at various scan rates (5–15 m s⁻¹) was measured in 1 M H₂SO₄ electrolyte (Fig. 7a). The rectangular shape of the CV curves suggested potential capacitive behavior. Increasing the scan rate from 5 to 15 mV s⁻¹ resulted in decreased specific capacitances (SC) of 911, 698 and 510 F g⁻¹, and 380, 298, and 236 F g⁻¹ for CNFs-Co-4 (Fig. 7a) and CNFs-Co-2 respectively (see S6 in ESI†). A higher scanning rate inhibited the electrolyte ions in accessing the full surface of CNFs-Co by decreasing the diffusion time, leading to reduced capacitance. The gradual activation process of the electroactive Co₃O₄ NPs linearly enhanced the SC of the hybrid porous electrode during the first 200 cycles, which was even maintained after 1000 cycles. Comparing to CNFs, the improved hierarchical structure of CNFs-Co-4 facilitated electrolyte migration to the electrode, resulting in considerable rate capability, whereas the higher surface area and total pore volume of the mesoporous structure could be credited for the better capacitance (see S7 in ESI† for comparison). Fig. 7b shows the specific capacitance and scan rates of pure CNFs and CNFs-Co-4 that give the capacitance behavior at different scan rates. By increasing the scan rate, the specific capacitance decreases due to the diffusion of electrolyte.

Fig. 7 (a) Cyclic voltammetry of CNFs-Co-4, (b) charge/discharge curves at different current densities for CNFs-Co-4, (c) Nyquist plots of supercapacitors, and (d) cycle number vs. specific capacitance of CNFs-Co-4. The inset in (c) shows galvanostatic charge/discharge curves for supercapacitors.

The frequency range of EIS analysis was carried out at 0.1–10 000 Hz. In the Nyquist plots, the intercept at the real part is a combination of the ionic resistance of the electrolyte, the intrinsic resistance of the substrate and active material, and the contact resistance of the active material/current collector interface. An ideal polarizable capacitance gives rise to a straight line along the imaginary axis. In our system, polarizable capacitance gives rise to a straight line along the imaginary axis that shows less resistance (Fig. 7c). The inset in Fig. 7c shows the high values of the Nyquist plot. It is intriguing to note that the equivalent series resistance (ESR) is not altered much, confirming ideal capacitive behavior with the change in salt content in CNFs-Co-4. The near 90° phase angles at high frequency show inductive behavior, while the near ∼90° phase angles at low frequency indicate capacitive behavior. It is commonly recognized that the electrochemical capacitance is dependent upon both the surface area and the electrical conductivity of the electrode material. Fig. 7d shows SC vs. cycle number for the CNFs-Co-4 membrane electrode. It is shown that the SC was measured at 1000 cycles, demonstrating cyclic stability and high specific capacitance. To further demonstrate the advantage of the synergistic effect in this electrode design, the cycling performance of the Co-CNFs at different current densities was recorded. Even though suffering from the sudden change of the current delivery, the electrode exhibits stable capacitance at each current density value. After 1000 continuous cycles at varying current densities, the capacitance was stable in Fig. 7d. This result highlights the capability of the pseudo-capacitive material-based CNFs-Co electrode to meet the requirements of both long cycle lifetime and good rate capability, which are important merits for practical energy storage devices.

The normalized capacitance retention tested at a current density of 0.5 A g⁻¹ over 1000 cycles, is demonstrated in Fig. 8. Comparing to pure CNFs, the loading of Co₃O₄ resulted in higher SC retention for both CNFs-Co-4 (76%) and CNFs-Co-2 (78%). The inset is the corresponding charge–discharge curves of 30 cycles. Consequently, the good stability of the doped CNFs mainly derived from the three dimensional network structures, leads to good ion mobility. The contact between the PVA/H₂SO₄ electrolyte and electrode increases due to the high specific surface area.

Fig. 8 Cyclic performance of CNFs-Co-4 at the current density of 0.5 A g⁻¹. The inset is the corresponding charge–discharge curves of 30 cycles.

The charge curves are closely symmetric to their corresponding discharge curves in the potential range, which showed a high reversibility between charge and discharge processes. It can be seen that all the charge and discharge curves have a symmetric nature indicating a rapid current–voltage response and good coulombic efficiency. It is significant to point out that the electrospun doped CNFs show electrochemical activity (galvanic charge/discharge) in 1 M H₂SO₄ solution in the voltage range at 0.5 V. The superior compatibility and robust adhesion between these well-dispersed and stabilized fine Co₃O₄ nanoparticles and the conductive CNFs may be responsible for the SC and excellent electrochemical stability. It is challenging to significantly increase the specific surface area, electrical conductivity, stability and SC, which are achieved in this study by the incorporation of different concentrations of Co(acac)₂ compared to pure CNFs, making CNFs-Co promising candidates for supercapacitors and lithium ion batteries.

3.4 CO₂ adsorption studies

The capture of CO₂ has attracted considerable attention in recent years due to its main anthropogenic contribution to climate change. It was explained previously that porous CNFs are potential materials for CO₂ adsorption.³⁹ Thus, we have investigated CO₂ adsorption capacity of these CNFs-Co porous materials at 25 °C under atmospheric pressure (1 bar). The CO₂ adsorption isotherms are presented in Fig. 9a which show the capacity for CNFs-Co-4, CNs-Co-3 and CNFs-Co-2 are 5.4, 5.1 and 4.9 mmol g⁻¹ at 25 °C (1 bar) respectively. Fig. 9b shows CO₂ uptake at 1 bar and 0 °C for CNFs-Co-4 reaches up to 6.9 mmol g⁻¹. The more curved isotherm indicates higher CO₂ adsorption at a low temperature on the surface of the CNFs containing Co₃O₄ NPs on the adsorbent surface and micropores. The increase in the CO₂ adsorption capacity is mainly due to the existence of a large fraction of fine micropores (pore size < 1 nm). It was formerly publicized that the physical adsorption of CO₂ on porous CNFs under ambient conditions is governed by the volume of fine micropores.

Fig. 9 (a) CO₂ adsorption isotherms for CNFs-Co measured at 25 °C, and (b) CO₂ adsorption isotherms for CNFs-Co-4 measured at 25 °C and 0 °C.

The CNFs doped with Co₃O₄ NPs possess similar volumes of micropores, and one can reason that these three samples might have similar CO₂ capacities. It is clearly seen that there was a small CO₂ uptake rise for CNFs-Co-4 as compared to CNFs-Co-3 and CNFs-Co-2, indicating the effect of Co₃O₄ NPs in the CNFs on CO₂ adsorption. These values are advanced than those of many reported carbons and are comparable with the salt containing carbons. Therefore, it can be concluded that the overall CO₂ adsorption of the CNFs-Co studied is a combination of two contributions: a major contribution related to adsorption in the fine micropores and a minor contribution due to the adsorption of Co₃O₄ NPs.

4. Conclusion

In this study, a novel and effective strategy has been developed to produce a flexible hierarchical hybrid nanostructured membrane for multifunctional applications. CNFs-Co was successfully synthesized using electrospinning and subsequent carbonization technique. Owing to the high content of salt and the ordered microporous structure, CNFs-Co possess a tunable porous structure with a high surface area of 483 m² g⁻¹. The resultant supercapacitors exhibit excellent electrochemical performance and high CO₂ adsorption. Such a high specific capacitance of 911 F g⁻¹ at a scan rate of 5 mV s⁻¹, in a 1 M H₂SO₄ solution presents great potential and good cycling stability for practical application as electrode material for high performance supercapacitors. Moreover, substantial CO₂ adsorption capacities 6.9 and 5.4 mmol g⁻¹, corresponding to an equilibrium pressure of 1 bar were measured at 25 °C and 0 °C respectively. The structural stability was confirmed by FE-SEM and HR-TEM; the images show that electrolyte and gaseous flow at 1 bar have no adverse effect on the membrane morphology. Furthermore, the material used is low-cost and approachable. Therefore, this design could be applied to various other electroactive and mixed metal oxides for fabricating multifunctional materials in the future.

Acknowledgements

This work is supported by the National Plan for Science, Technology and Innovation (MAARIFAH), King Abdul-Aziz City for Science and Technology, Kingdom of Saudi Arabia, Award Number (11-NAN1915-02).

References

1. N. P. Wickramaratne, J. Xu, M. Wang, L. Zhu, L. Dai and M. Jaroniec, Chem. Mater., 2014, 26, 2820–2828.
2. G. Wang, L. Zhang and J. Zhang, Chem. Soc. Rev., 2012, 41, 797–828.
3. L. Ji, A. J. Medford and X. Zhang, J. Mater. Chem., 2009, 19, 5593.
4. G. Wang, C. Pan, L. Wang, Q. Dong, C. Yu, Z. Zhao and J. Qiu, Electrochim. Acta, 2012, 69, 65–70.
5. A. G. Pandolfo and A. F. Hollenkamp, J. Power Sources, 2006, 157, 11–27.
6. L. Kandic, M. Mitric and N. Ignjatovic, Adv. Mater. Processes, 2006, 518, 507–512.
7. Y. Li, Z. Y. Fu and B. L. Su, Adv. Funct. Mater., 2012, 22, 4634–4667.
8. S. Cavaliere, S. Subianto, I. Savych, D. J. Jones and J. Rozière, Energy Environ. Sci., 2011, 4, 4761–4785.
9. R. Tummala, R. K. Guduru and P. S. Mohanty, J. Power Sources, 2012, 209, 44–51.
10. H. Xu, L. Gao, Q. Zhang, J. Li, J. Diwu, X. Chou, J. Tang and C. Xue, J. Nanomater., 2014, 1–9.
11. A. Y. Lo, Y. Jheng, T. C. Huang and C. M. Tseng, Appl. Energy, 2015, 153, 15–21.
12. D. G. Gromadskyi, J. H. Chae, S. A. Norman and G. Z. Chen, Appl. Energy, 2015, 159, 39–50.
13. Y. F. Yuan, X. H. Xia, J. B. Wu, X. H. Huang, Y. B. Pei, J. L. Yang and S. Y. Guo, Electrochem. Commun., 2011, 13, 1123–1126.
14. S. Goriparti, E. Miele, F. De Angelis, E. Di Fabrizio, R. Proietti Zaccaria and C. Capiglia, J. Power Sources, 2014, 257, 421–443.
15. J. Liu, J. Jiang, C. Cheng, H. Li, J. Zhang, H. Gong and H. J. Fan, Adv. Mater., 2011, 23, 2076–2081.
16. F. Zhang, C. Yuan, J. Zhu, J. Wang, X. Zhang and X. W. D. Lou, Adv. Funct. Mater., 2013, 23, 3909–3915.
17. Y. H. Hsu, C. C. Lai, C. L. Ho and C. T. Lo, Electrochim. Acta, 2014, 127, 369–376.
18. J. Li, X. Cheng, A. Shashurin and M. Keidar, Graphene, 2012, 01, 1–13.
19. G. Abellán, J. A. Carrasco, E. Coronado, J. Romero and M. Varela, J. Mater. Chem. C, 2014, 2, 3723–3731.
20. G. Ferey, C. Serre, T. Devic, G. Maurin, H. Jobic, P. L. Llewellyn, G. De Weireld, A. Vimont, M. Daturi and J. S. Chang, Chem. Soc. Rev., 2011, 40, 550–562.
21. M. Zhi, C. Xiang, J. Li, M. Li and N. Wu, Nanoscale, 2013, 5, 72–88.
22. S. D. Kenarsari, D. Yang, G. Jiang, S. Zhang, J. Wang, A. G. Russell, Q. Wei and M. Fan, RSC Adv., 2013, 3, 22739–22773.
23. J. Wei, D. Zhou, Z. Sun, Y. Deng, Y. Xia and D. Zhao, Adv. Funct. Mater., 2013, 23, 2322–2328.
24. T. Zhang, C. H. Kim, Y. Cheng, Y. Ma, H. Zhang and J. Liu, Nanoscale, 2015, 7, 3285–3291.
25. N. Iqbal, X. Wang, J. Yu, N. Jabeen, H. Ullah and B. Ding, RSC Adv., 2016, 6, 4382–4386.
26. Z. Yu, L. Tetard, L. Zhai and J. Thomas, Energy Environ. Sci., 2015, 8, 702–730.
27. J. Zhou, W. Li, Z. Zhang, W. Xing and S. Zhuo, RSC Adv., 2012, 2, 161–167.
28. Y. Yang, L. Ge, V. Rudolph and Z. Zhu, Dalton Trans., 2014, 43, 7028–7036.
29. F. Su, C. Lu, A. J. Chung and C. H. Liao, Appl. Energy, 2014, 113, 706–712.
30. X. Wang, N. G. Akhmedov, Y. Duan, D. Luebke, D. Hopkinson and B. Li, ACS Appl. Mater. Interfaces, 2013, 5, 8670–8677.
31. Z. Chen, S. Deng, H. Wei, B. Wang, J. Huang and G. Yu, ACS Appl. Mater. Interfaces, 2013, 5, 6937–6945.
32. Y. G. Zhu, Y. Wang, Y. Shi, J. I. Wong and H. Y. Yang, Nano Energy, 2014, 3, 46–54.
33. Y. Miwa, A. R. Drews and S. Schlick, Macromolecules, 2006, 39, 3304–3311.
34. J. Gu and F. Sansoz, Carbon, 2014, 66, 523–529.
35. A. P. Perissinotto, C. M. Awano, D. A. Donatti, F. S. de Vicente and D. R. Vollet, Langmuir, 2015, 31, 562–568.
36. Z. Zhang, J. Deng, X. Li, Z. Yang, S. He, X. Chen, G. Guan, J. Ren and H. Peng, Adv. Mater., 2015, 27, 356–362.
37. G. Orsini and V. Tricoli, J. Mater. Chem., 2012, 22, 23861–23870.
38. G. Zhang and X. W. David Lou, Sci. Rep., 2013, 3, 1470.
39. Y. Zhao, L. Zhao, K. X. Yao, Y. Yang, Q. Zhang and Y. Han, J. Mater. Chem., 2012, 22, 19726–19731.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: 10.1039/c6ra06077c

---