Saima Nawazab,
Yaqoob Khan*b,
Shaimaa A. M. Abdelmohsenc,
Sadia Khalidb,
Emma M. Björkd,
Muhammad Asim Rasheede and
M. Siddiq*a
aDepartment of Chemistry, Quaid-i-Azam University, Islamabad 45320, Pakistan. E-mail: m_sidiq12@yahoo.com; Tel: +92 5190642147
bNanoscience and Technology Department, National Centre for Physics, QAU Campus, Shahdra Valley Road, Islamabad 45320, Pakistan. E-mail: yaqoob@ncp.edu.pk; Fax: +92 512077389; Tel: +92 3455235423
cDepartment of Physics, College of Science, Princess Nourah Bint Abdulrahman University, P. O. Box 84428, Riyadh 11681, Saudi Arabia
dNanostructured Materials, Department of Physics, Chemistry and Biology (IFM), Linköping University, SE-581 83 Linköping, Sweden
eDepartment of Physics and Applied Mathematics, Pakistan Institute of Engineering and Applied Sciences (PIEAS), Islamabad, 45650, Pakistan
First published on 10th June 2022
Mesoporous silicon (mSi) obtained by the magnesiothermic reduction of mesoporous silica was used to deposit polyaniline (PANI) in its pores, the composite was tested for its charge storage application for high performance supercapacitor electrodes. The mesoporous silica as confirmed by Small Angle X-ray Scattering (SAXS) has a Brunauer–Emmett–Teller (BET) surface area of 724 m2g−1 and mean pore size of 5 nm. After magnesiothermic reduction to mSi, the BET surface area is reduced to 348 m2g−1 but the mesoporousity is retained with a mean pore size of 10 nm. The BET surface area of mesoporous silicon is among the highest for porous silicon prepared/reduced from silica. In situ polymerization of PANI inside the pores of mSi was achieved by controlling the polymerization conditions. As a supercapacitor electrode, the mSi–PANI composite exhibits better charge storage performance as compared to pure PANI and mesoporous silica–PANI composite electrodes. Enhanced electrochemical performance of the mSi–PANI composite is attributed to the high surface mesoporous morphology of mSi with a network structure containing abundant mesopores enwrapped by an electrochemically permeable polyaniline matrix.
Transition metal oxides and hydroxides, transition metal dichalcogenides and carbon based materials have been studied as electrode materials for charge storage applications.12 After the discovery of graphene, rapid development of other new emerging layered structures such as MXenes,13 layered nanoclay, phosphorene, bismuthene and 2D graphene analogues have excellent charge storage capabilities.14 Many structural evolutions, chemical changes and hybridizations of previously used traditional materials in industry are being made to utilize the synergistic effect of constituent materials to ultimately achieve better electrochemical performance. These hybrids/nanocomposites integrate the advantageous attributes and compensate for the disadvantages associated with its single components. New advances include porous 2D and 3D graphene materials,15 transition metal oxide–hydroxide heterostructure (e.g. Co3O4/Co(OH)2 heterostructure via interfacial layer control12), spinel-type Co3O4 and its modification in spinel cobaltites MCo2O4 (M = Co, Mn, Zn).16 Recently Huan Pang et al. work suggests that MOFs (metal–organic frameworks) and its various derivatives such as multimetallic MOFs (i.e. bimetallic and trimetallic MOFs)17 and MIL-96-Al18 with controllable shapes and sizes also possess enhanced charge storage (sulfur storage).
For a material to be an efficient supercapacitor electrode, there should be some spaces inside structure to store charge e.g. MnO2 exist in many structural phases (i.e., α, β, γ, λ and δ) each of which differs in its shape, size and dimensions of tunnels. So the specific capacitance (Cs) of β-MnO2 [narrow (1 × 1) tunnels] is lower while δ-MnO2 (interlayer separation ∼ 7 Å) exhibits high Cs values being layered structure.2,19–21 N. Munichandraiah et al. synthesized nanostructured MnO2 samples with different crystal structures, and investigated as electrode materials for electrochemical capacitors in aqueous 0.1 M Na2SO4 solution. The Cs values are 240 F g−1 for α-MnO2 and 236 F g−1 for δ-MnO2. Alternatively, they are as low as 9 F g−1 for β-MnO2 and 21 F g−1 for λ-MnO2.19
Beside materials containing intrastructural spaces, porous structures are becoming suitable candidates for supercapacitors. Recently Vlad and Balducci showed higher normalized capacitance relative to their BET surface areas of Ni3(hexaiminotriphenylene)2 metal–organic framework (MOF) as compared to activated carbons, carbon nanotubes, zeolite-templated carbon, carbide-derived carbon and graphene.22
Porous materials can be categorized on the basis of pore size i.e. macroporous (>50 nm), mesoporous (2–50 nm) and microporous (≤2 nm).2,23–25
Recently porous silicon has been widely studied for variety of applications, most specifically as anode material in lithium-ion batteries.26–28 Previously established synthetic route of mesoporous silicon from silicon is physical etching which require long duration, various pre-steps and involves many harsh corrosive reagents.29 Three dimensionally structured silicon replicas could be produced from parent silica diatom by magnesiothermic reduction (i.e., vaporized/liquefied magnesium at higher temperatures reduces different metal oxides). Currently, magnesiothermic reduction, being simple and lower temperature method to convert silica to nanostructured silicon is trending which usually results in porous silicon contaminated by unreacted silica.30
We have developed a scheme to synthesize the mesoporous silicon by magnesiothermic reduction of mesoporous silica. The focus of this research was to keep intact the porosity of mesoporous silica in resultant silicon which can be used in the energy areas like supercapacitors and hydrogen generation in the form of composites. Ultimately, we can avail the advantage of this new technique by using it to produce mesoporous silicon having good cycling stability. Our developed synthetic route comprises of just two steps. Firstly, mesoporous silica was synthesized from silica precursor via soft template method. Resultant mesoporous silica was converted into mesoporous silicon by magnesiothermic reduction.
Polyaniline (PANI) due to its easy synthesis, high controllable electrical conductivity (due to its conjugated structure having delocalised p-electrons along its backbone) and environmental stability is considerably attractive choice among several conducting polymers (CPs).31,32 It offers many advantages being pseudo-capacitive electrode material containing high theoretical specific capacitance, as short path lengths for ionic transport allow faster ionic diffusion within the polymer network so energy is delivered relatively at rapid rate.33 Its high surface area in contact with the electrolyte, allows comparatively fast or high charge/discharge rates. However, the shrinkage and swelling occurring due to doping–dedoping phenomena result in poor mechanical characteristics and low cycle life limit (i.e. poor cycling stability) when used as individual electrode material.34,35 The strategies used to overcome the issue with its cyclic life mainly include irradiation, sonication during fabrication or compositing it with fillers which increase volume of polymer, enhances porosity and provide room for its swelling.33 Compositing with other materials such as carbon based materials (carbon nanotubes and graphene), inorganic oxides (such as SnO2, MnO2, TiO2), sulphides and hydroxides and other metal compounds have been proved to enhance cyclic stability as well as maximize the capacitance of resultant composites.6,15,33,36–42
In order to alleviate the limitation associated with PANI-matrix for charge storage application, PANI was added inside synthesized mesoporous silicon (which contains mixed pores within mesoporosity range) to increase its charge storage, cyclic stability and specific capacitance. Besides this, PANI was also added in the mesoporous silica under same conditions and was compared with simple PANI and PANI with mesoporous silicon. This convenient and scalable synthesis procedure developed here indicates a significant potential towards cost-effective electrode materials for practical applications. Moreover the surface area of mesoporous silicon (348 m2g−1) attained by magnesiothermic reaction is amongst the highest published in literature so far.30
SiO2 + 2Mg → Si + 2MgO | (1) |
When SiO2 (silica) and Mg (magnesium) react during magnesiothermic process, Si (mesoporous silicon) forms and MgO (magnesium oxide) is produced as a by-product (Fig. 1).
The powder thus obtained was washed with 1 M hydrochloric acid (HCl) and then vacuum filtered. A greyish brown fine powder of mesoporous silicon (mSi) was attained as end product.
![]() | (2) |
Specific capacitance of three electrode system can be estimated also by slope of discharge curve of GCCD measurement by following formula44 given in eqn (3).
![]() | (3) |
In the XRD patterns of as reduced mSi, peaks from MgO were observed (ICCD#: 01-077-2364, 01-077-2109) (Fig. 3) along with reflections from Si indicating the complete reduction of SBA to mSi. After acid washing, the reflections from MgO disappeared and only the three prominent peaks of Si at 28.4°, 47.3° and 56.1° were present (ICDD#: 00-026-14181). There is no evident peak of mesoporous silica so it shows that magnesiothermic reduction with consequent acid washing left no residues of parent material (Fig. 3).
![]() | ||
Fig. 3 X-ray diffractograms of mesoporous silicon reduced from mesoporous silica before (mSi–Mg) and after (mSi) the acid treatment. |
PANI exhibits X-ray diffraction peaks at 14.4°, 19.6°, and 25.4° for (121), (113) and (322) reflections respectively.45,46 Intensity of these peaks decreases slightly in XRD diffractograms of composites of mSi–P & SBA–P, which is due to the interaction between PANI and filler (ESI Fig. S1†).
![]() | ||
Fig. 6 BJH pore size & volume analysis – Brunauer–Emmett–Teller (BET) of (a) SBA and (b) mSi by Halsey: Faas correction. |
For mSi, the N2 adsorption–desorption isotherm still shows a type H1 hysteresis loop followed by a type H3 hysteresis. This indicates that the metallic Si retains part of the mesopores, even though some pore shrinkage is observed in the pore size distribution in Fig. 6. The strong H3 loop is most probably due to the slit shaped pores formed between Si crystals, see Fig. 4.47–52 The specific surface area of mSi is 348 m2 g−1. This significant decrease in specific area is due to a decreased mesoporosity upon the formation of crystalline Si. The narrow pore size distribution of mSi indicates a homogeneous material reduction.
![]() | ||
Fig. 7 X-ray photoelectron spectroscopy (XPS) of mSi–P showing peaks of (a) Si 2p, (b) N 1s, (c) C 1s and (d) O 1s. |
In the XPS survey spectrum of mSi–P, peaks from Si 2p were due to mSi while peaks from C 1s and N 1s were observed due to the presence of polyaniline in the composite. N 1s was fitted into three components arising from the –NH2+, NH+, and
N– moieties in PANI. C 1s was fitted into four components consisting of C–C, C
C, C–N/C
N, and COOH groups. The three components under the O 1s spectra corresponds to adsorbed O, Si–O and, C–O groups.
Cyclic voltammograms of polyaniline (P) (Fig. S4(a)†) at various scan rates exhibit a redox peak which is attributed to the transition of PANI from leucoemeraldine (semiconducting state) to emeraldine (conductive form). This redox process caused pseudo capacitance of PANI. Apparently, the mSiP composite (Fig. 8(a)) has a similar electrochemical response as that of the P, but peak current of mSi–P increases greatly which implies a larger electrode capacitance. High electrochemical utilization of mSi–P is due to its larger specific surface area as compared to P and hence more electroactive sites.
![]() | ||
Fig. 8 Cyclic voltammogram (CV) of (a) mSi–P and (b) mSi at different scan rates on GCE in 1 M H2SO4 electrolyte vs. SCE. |
The PANI matrix is electrochemically permeable i.e. it does not block the counter-ions to reach the mSi because of the close contact established between the two during in situ polymerization inside the pores of mSi.
Comparison of the shapes of the cyclic voltammograms (Fig. 8(a), S4(b) and (c)†) depicts that the contribution of the mSi and SBA to the total capacitance of mSi–P and SBA–P composite electrodes is of the double-layer type which is also apparent in the form of CV curves for pure mSi (Fig. 8(b)) and SBA (Fig. S4(b)†). The electrochemical behaviour of mSi–P (Fig. 8(a)) is still better than SBA–P (Fig. S4(c)†) which is because of enlarged pores of mSi than SBA which facilitate the charge transfer. P (Fig. S4(a)†) is showing good capacitance response but its efficiency decreases after multiple cyclic charge–discharge phenomena due to changes in its volume and physical properties which is inevitable in its pure polymer matrix.
The rectangular shape and mirror images are observed in the CV curves for all samples (Fig. 8 and S4†), indicating high electrochemical reversibility.36–38,56,57
At 5 mV s−1 scan rate and 1.2 V potential, the specific capacitance of mSi is 19.85 F g−1 while for P its value is 178 F g−1. Specific capacitance of mSi–P is 214.45 F g−1.
Fig. 9 and S5† show the cyclic voltammograms recorded at 20 mV s−1 scan rate for 50 cycles in 1 M H2SO4 aqueous electrolyte solution. The observed CV scans are quite stable for 50 cycles and no significant decrease in current is observed for any material which depicts that materials have high cycling stability. Fig. 9(a) is showing the good cyclic stability of the electrode which is due to the cohesive inter-molecular contact between P and mSi which inhibits the dissolution of filler. Fig. 10 and S6† show specific capacitance decreases by increasing scan rate.
![]() | ||
Fig. 9 Cyclic voltammetry (CV): cyclic stability of (a) mSi–P and (b) mSi at scan rate of 20 mV s−1 for 50 cycles on GCE in 1 M H2SO4 electrolyte vs. SCE. |
Further, relative contributions of the capacitance from the bulk or surface mechanism is derived by analyzing the CV data at various scan rates (5–250 mV s−1) shown in Fig. S10† and using the power law58,59 given in eqn (4).
i = avb | (4) |
According to power law, slope of the plot log(i) i.e. redox peak current versus log(v) i.e. various scan rates provides the b-value which can be used to estimate the controlling mechanism of the active material in electrode. For ideal capacitor type materials, b-value is close to 1 where pseudocapacitive behavior is dominant. If the value of b is close to 0.5, it depicts the material is battery type and contains diffusion controlled phenomenon as dominant one. Fig. S10† shows a plot between log(i) and log(v) showing the cathodic and anodic peak currents of mSi–P as indicated in Fig. 8(a). The b-values from the anodic and cathodic peaks are 1.009 and 1.022 respectively. It suggests that total contribution to capacitance is mainly due to peudocapacitance phenomenon and mSi–P is showing mainly capacitive behavior. Using the Dunn method of differentiation,60 it is estimated that 99.78% of the total capacitance is from pseudocapacitive contribution and 0.22% is diffusion controlled behavior.
EIS is performed in the frequency range of 106 to 10−1 Hz (at AC voltage = 10 mV rms, open circuit potential amplitude = 0.34 and stabilization time = 10 s) to evaluate the resistance of electrochemical phenomenon.
Fig. 11 and S7† show the Nyquist plots of samples. Fig. 11(a) shows the frequency response of mSi–P/1 M H2SO4 (aq.) system in the plotted form of two impedance components against each other, one of which is real component i.e. Z′ and other one is imaginary component i.e. Z′′. The Nyquist plots are fitted using an equivalent circuit model shown in the inset of Fig. 11, where Rs is solution ohmic resistance, Rct is charge transfer resistance, CPE is constant phase element and W is Warburg impedance. In Fig. 11(a), absence of semicircle in the domain of high frequency shows less Rct. Rct calculated by fitting equivalent circuit model is 232.6 × 10−6 Ω. Rct existing between mSi–P electrode and 1 M H2SO4 (aq.) electrolyte is considerably less due to the highly conductive cross-linked mSi and PANI. It may relate to a high material electrochemical activity (pseudo capacitance), and indicates the surface properties of mSi–P electrode is favouring facile fast charge transfer kinetics within electrodes.61
Large impedance values observed for mSi electrodes in Fig. 11(b) are indicative of high Rct (1.868 × 103 Ω) which is greater than that of mSi–P composite (Fig. 11(a)).
In Fig. 11(a) at low frequency region a steeper curve is evident of decreased Warburg impedance which depicts the accelerated diffusion and adsorption rate of counter-ions of electrolyte in the system (in/on the electrode material). The reason behind it is that the network of PANI enwrapping mSi (having mesoporous configuration) with high surface area provide more space with short and equal diffusion path length for transportation of redox species (i.e. counter-ions) of 1 M H2SO4 electrolyte consequences in excellent capacitive behavior.
![]() | ||
Fig. 12 EIS (electrochemical impedance spectroscopy): Bode plots of (a) mSi–P and (b) mSi vs. OC (open circuit) at 10 mV rms AC perturbation (in 1 M H2SO4 electrolyte). |
Nyquist plot (Fig. 11(b)) for mSi shows larger Rct and Warburg impedance as compared to mSi–P; Fig. 11(a) which is due to high obstruction in the movement of counter-ions and higher variations in ion diffusion path lengths respectively.
The slope of each electrode (Fig. 11 and S7†) decreases with increasing AC frequency (potential) which demonstrates an enhanced Warburg resistance which is consequently causing reduction in effective charge storage at the electrode. Fig. 12 and S8† show the Bode plots of samples.
The electrochemical stability was examined by galvanostatic charge/discharge (GCCD) measurements in 1 M H2SO4 aqueous solution by consecutive charge–discharge cycles within the potential range of −0.2 to 1 V at discharge current of 1 μA and capacity of 0.0025 A h. GCCD measurements are executed to attain the quantitative information of the electrochemical capacitance for mSi–P, mSi, P, SBA and SBA–P electrodes (Fig. 13 and S9†).
The higher capacitance observed for mSi–P composite is due to low Rct (also discussed in impedance analysis) which is the consequence of its unique morphology (with more electroactive sites due to mesoporosity of mSi for reversible redox reaction) while having continuous conductive network of PANI.
The specific capacitance calculated by discharge curve of GCCD curves for mSi–P is equal to 68.21 F g−1 at 0.187 A g−1 current density.
Capacitance retention and coulombic efficiency of mSi–P for 1000 cycles in 1 M H2SO4 electrolyte is shown in Fig. 14.
![]() | ||
Fig. 14 GCCD (galvanostatic cyclic charge–discharge): capacitance retention and coulombic efficiency of mSi–P vs. cycle number for 1000 cycles (in 1 M H2SO4 electrolyte). |
Footnote |
† Electronic supplementary information (ESI) available: Polyaniline inside the pores of high surface area mesoporous silicon as composite electrode material for supercapacitors. Fig. S1: X-ray diffraction (XRD) pattern of polyaniline (P), SBA–P and mSi–P composites. Fig. S2(a): scanning electron microscopy (SEM) images of mSi, P and mSi–P composite. Fig. S2(b): energy dispersive X-ray (EDX) pattern of mSi–P composite. Fig. S3: Fourier transform infrared spectroscopy (FT-IR) of SBA (mesoporous silica), polyaniline (P) and composites (mSi–P and SBA–P). Fig. S4: cyclic voltammogram (CV) of (a) P, (b) SBA and (c) SBA–P at different scan rates on GCE in 1 M H2SO4 electrolyte vs. SCE. Fig. S5: cyclic voltammetry (CV): cyclic stability of (a) P, (b) SBA and (c) SBA–P at scan rate of 20 mV s−1 for 50 cycles on GCE in 1 M H2SO4 electrolyte vs. SCE. Fig. S6: scan rate vs. specific capacitance. Fig. S7: EIS (electrochemical impedance spectroscopy): Nyquist plots of (a) P, (b) SBA and (c) SBA–P vs. OC (open circuit) at 10 mV rms AC perturbation (in 1 M H2SO4 electrolyte). Fig. S8: EIS (electrochemical impedance spectroscopy): Bode plots of (a) P, (b) SBA and (c) SBA–P vs. OC (open circuit) at 10 mV rms AC perturbation (in 1 M H2SO4 electrolyte). Fig. S9: galvanostatic cyclic charge–discharge (GCCD): discharge curve; potential vs. real time (a) P (at charge density = 0.143 A g−1), (b) SBA (at charge density = 0.250 A g−1) and (c) SBA–P (at charge density = 0.388 A g−1) and in 1 M H2SO4 electrolyte. Fig. S10: plot of log(i) versus log(v) for the anodic and cathodic current peaks of mSi–P. Fig. S11: proposed mechanism of electrochemical reaction during charging and discharging in 1 M H2SO4 electrolyte. See https://doi.org/10.1039/d2ra01829b |
This journal is © The Royal Society of Chemistry 2022 |