Prahlada Thippeswamya,
Vinay Arya
b,
Indranil Dey
c,
Chirodeep Bakli
*b and
Debasis Ghosh
*a
aCentre for Nano and Material Sciences, JAIN (Deemed to be University), Kanakapura Road, Bangalore, Karnataka 562112, India. E-mail: g.debasis@jainuniversity.ac.in; debasisghosh88@gmail.com
bThermofluidics and Nanotechnology for Sustainable Energy Systems Laboratory, School of Energy Science and Engineering, Indian Institute of Technology Kharagpur, Kharagpur, West Bengal 721302, India. E-mail: chirodeep@iitkgp.ac.in
cDepartment of Materials Engineering, Indian Institute of Science, Bengaluru, 560012, India
First published on 14th May 2026
Herein we demonstrate the fabrication of a laser induced graphene (LIG) supported MnO2 symmetric pseudocapacitor with unprecedented electrochemical performance enabled by 17 m NaClO4 water-in-salt hybrid electrolyte with ethylene glycol (EG) as the anti-freezing molecularly crowding agent (MC-WiSE). MnO2 was electrodeposited onto an interdigital patterned conductive LIG template by a fast and facile binder free approach that eliminates the requirement for a separator or metal current collector, effectively minimizing inactive dead mass in the device. The depressed freezing point and suppressed electrochemical activity of water in the MC-WiSE enabled stable cell operation even at −40 °C within a broad voltage window of 0–2.5 V. Under these conditions, the device delivered a high areal capacitance of ∼25.1 mF cm−2 (∼0.32 mA cm−2), corresponding to an energy density of 21.8 µWh cm−2 at a power density of 396.8 µW cm−2, along with excellent cycling stability. The cell also showed exceptional thermal response with an increased energy density of ∼75 µWh cm−2 at a power density of ∼1984 µW cm−2 via activating bulk pseudocapacitance at 60 °C. Furthermore, the ion–ion and ion–solvent interactions in the electrolyte and electrode/electrolyte interfacial interplay were studied via MD simulation which validated the experimental observations of excellent thermal and electrochemical stability.
On the other hand, conventional supercapacitors with stacked electrodes and sandwiched separators either in pouch or cylindrical cell geometry suffer from limited flexibility and bulky size factors, which makes them unsuitable for flexible and miniaturized electronics.9 Thus, developing low-cost, high-performance electrodes in flexible and miniaturized geometry, coupled with high voltage, safer and all climate operational electrolyte can significantly boost the flexible electronics market.
Interdigital patterned laser induced graphene (LIG) on a flexible substrate has been the centre of interest for fabrication of flexible and miniaturized energy storage devices. In contrast to biomass-derived amorphous carbons or carbonaceous composites—which are widely used as supercapacitor electrodes but typically require binders or solution-based processing during electrode fabrication—LIG provides several distinct advantages.10–12 This is due to the unique properties of LIG including its high surface area and excellent electrical conductivity, which enable binder-free growth of active electrode materials, precise micron-scale electrode patterning, and the geometrical benefits of interdigital configurations (e.g. separator-free design).13–15
The potential of LIG in flexible energy storage devices has been demonstrated by several authors with promising electrochemical performance. For example, Awasthi et al. reported flexible LIG-based interdigital microsupercapacitors (MSCs) with optimized operational temperature offering the maximum capacitance of 83.33 mF cm−2 and energy density of 4.62 µWh cm−2.13 Xiong et al. demonstrated the superiority of interdigital geometries of LIG supercapacitors, achieving an energy density of 4.85 µWh cm−2 with a high operational voltage window of 0–1.8 V.16 Coelho et al. reported a chromatography paper derived LIG based MSC with a capacitance of 4.6 mF cm−2 and stability over 10
000 cycles.17 Yuan et al. introduced hydrophilicity to LIG via O/S co-doping, which enhanced wettability in aqueous electrolyte leading to a high capacitance of 53.2 mF cm−2.18
To further improve the performance, LIG has been modified with metal oxides or conducting polymers and the combined double layer capacitive and pseudocapacitive storage mechanism reportedly improved device performance.19–28 Among different pseudocapacitive metal oxides, MnO2 represents a unique combination of high performance and cost efficiency and is environmentally benign, which allowed it to be one of the most explored electrodes for energy storage. For example, Klem et al. reported electrochemically deposited MnO2 onto wax coated paper derived LIG with a high capacitance of 86.9 mF cm−2 (at 0.1 mA cm−2) and energy density of 7.3 µWh cm−2.29 Xi et al. reported the electrochemical growth of MnO2 onto KOH activated LIG with a high capacitance of 18.82 mF cm−2 (at 0.2 mA cm−2) and energy density of 2.61 µWh cm−2 at a power density of 260.28 µW cm−2.30 Yuan et al. reported MnO2 grown on SWCNT bridged LIG with an improved capacitance of 156.94 mF cm−2 and energy density of 21.8 µWh cm−2.31 However, none of these studies addressed sub-zero temperature operability, which is critical for real-world deployment in wearable and flexible electronics, and autonomous devices exposed to extreme conditions.27,28,31–42 The employed aqueous electrolyte and gel polymer electrolyte suffer from poor electrochemical performance at low temperature and poor compatibility with metal oxide interfaces. Also, the limited voltage window restricts the overall energy density.
Suppressing the water electrochemical activity via ‘water-in-salt’ electrolyte formulation has been a widely adopted strategy to suppress the water electrochemical activity and improve the ESW.43 However, the application is limited to ambient temperature mostly, as low temperature often causes salt precipitation due to the reduced solubility factors. Deng et al. demonstrated that via forming polyelectrolyte hydrogel with 24 M CH3COOK, both the operational temperature (−20 to 70 °C) and voltage window (0–2.1 V) can be improved.44 Alternatively, a molecularly crowding additive with hydrogen bonding sites can moderate the electrochemical activity of water leading to an improved ESW. For example, Sekar et al. introduced a sterically engineered hydrogen-bonded electrolyte (D-fructose + 5 m NaClO4) to suppress water electrolysis, achieving >2.3 V ESW.45 The nature of molecularly crowding agents (e.g. acetonitrile or ACN, ethylene glycol, glycerol, dimethyl sulfoxide, etc.) also influences the subzero temperature operation while maintaining a high voltage window. Recently we demonstrated that a water ethylene glycol mixture effectively suppresses salt precipitation at an extremely low temperature of −40 °C even with a high salt concentration of 17 m NaClO4.15 Wu et al. developed a dilute aqueous–aprotic mixture (0.5 m Zn(CF3SO3)2 + 1 m LiTFSI in H2O/acetonitrile), enabling high voltage (2.2 V) and subzero temperature (−30 °C) cell operation for zinc ion capacitors.46 Qin et al. demonstrated a LiTFSI (2 m)–glycerol–PVA gel electrolyte achieving a high operational voltage of 2.4 V and wide temperature window (−20 °C to +60 °C).47 Sun et al. demonstrated the effect of ACN on altering the solvation structure of NaClO4 based WiSE, suppressing salt precipitation, leading to an ESW of 2.3 V and extreme low temperature operation at −50 °C.48 Liu et al. reported DMSO as a co-solvent in an aqueous based system. However, the capacitance in the presence of the organic additives at subzero temperature remains low.
Here, we address key limitations of conventional aqueous energy storage systems, namely limited operational temperatures and restricted voltage windows, as well as the lack of systematic investigation of MnO2 (λ-MnO2) electrochemical behavior across a wide temperature range. By combining the advantages of a miniaturized in-plane interdigitated electrode architecture with LIG-supported MnO2 nanostructures as electrodes, we demonstrate an energy-dense pseudocapacitor that employs a ‘water-lean’ electrolyte (17 m NaClO4) with ethylene glycol as the anti-freezing ‘molecularly crowding’ agent to push the operational voltage window to 2.5 V and achieve reliable operation at extreme temperatures from −40 °C to +60 °C. While conventional ‘water-in-salt’ electrolytes may experience phase instability and salt crystallization due to reduced salt solubility and increased viscosity, the presence of EG alternates the electrolyte salt-solvent microstructure, resulting in a depressed freezing point, and dispersibility of the electrolyte ions even at extremely low temperature (−40 °C). Furthermore, molecular dynamics simulation study was conducted to understand the hybrid electrolyte microstructure, spatial arrangement and dynamic behavior of ions within the electrolyte matrix, for different salt concentrations and EG additive in the bulk and the interfacial interplay with MnO2 as the active electrode. Electrochemically the 2.5 V operational symmetric device with miniaturized electrode architecture showed a high energy density of ∼50.2 µWh cm−2 at RT, retaining an energy density of 17.1 µWh cm−2 at −40 °C and a power density of 793.7 µW cm−2 accompanied by excellent low temperature, room temperature and high temperature stability when tested over 4000, 200
000 and 10
000 cycles, respectively. Furthermore, the interfacial composition and electrode/electrolyte degradation behaviour under long-term operation at an elevated temperature was studied, revealing no detectable electrolyte or salt degradation making it suitable for real-world deployment in wearables, outdoor electronics, and autonomous systems exposed to fluctuating environments including space applications.
To understand the phase purity and crystallinity of the electrodeposited material, XRD (Fig. 1f) was performed, which matched with the JCPDF file no. 00-014-0644, revealing the formation of λ-MnO2 with an orthorhombic crystal system. To further investigate the surface chemical composition and oxidation states, XPS analysis was performed on the electrodeposited material. The XPS survey spectra reveal the presence of key elements carbon (C 1s), oxygen (O 1s), and manganese (Mn 2p) as shown in Fig. S1, validating the successful synthesis of manganese oxide on LIG. Fig. 1g shows high resolution XPS spectra of C 1s, with a main peak centered at 284.8 eV. It can be further deconvoluted into 3 distinct peaks corresponding to C–C/C
C (284.8 eV), C–O (286.1 eV), and C
O (288.4 eV), indicating the presence of oxygen containing surface functionalities in LIG, which can enhance interfacial wettability.23,51,52 The Mn 2p spectrum (Fig. 1h) exhibited two prominent peaks centered at 653.94 eV and 642.24 eV corresponding to Mn 2p1/2 and Mn 2p3/2 spin orbit coupling with a spin-energy separation of 11.7 eV, which is comparable to the previously reported Mn 2p spectra of MnO2.53,54 Furthermore, the deconvolution of the Mn 2p peaks showed the presence of both Mn3+ and Mn4+, with relative contributions of 54.75% and 45.26%, respectively, revealing oxygen defects in the electrodeposited MnO2, which is stabilized by the Mn–OH group.55,56
The high resolution O 1s XPS spectra (Fig. 1i) could be deconvoluted into four distinct BE peaks, located at 529.8 eV (52.02%), 531.05 eV (34.65%), 532.73 eV (11.52%), and 534.48 eV (1.8%), corresponding to Mn–O, Mn–O–H, O–C/O
C (from LIG) and absorbed water/hydroxyl species, respectively.54,57 The presence of Mn–O–H further indicates the oxygen defects at the analogous sites of MnO2.
:
2 V/V), where EG acts both as a ‘molecularly crowding’ and an antifreezing agent to push the performance limitation of aqueous electrolyte by extending the operational voltage as well as the operational temperature window. The key physical properties, such as ionic conductivity, pH, viscosity, freezing point and Na+ transference number across different temperatures of the MC-WiSE are included in the SI (Fig. S2–S5).
The linear sweep voltammetry (10 mV s−1) (Fig. 2a) showed an extended ESW on moving from pure water to 1 m, 5 m, 10 m, and 17 m NaClO4 and MC-WiSE. Increasing the salt concentration alters the ion-solvation structure and the intermolecular hydrogen-bonding network. As the concentration increases, the water molecules become more constrained, leading to increasingly restricted water electrochemical activity and thus requiring a higher overpotential for the water redox reactions. The MC-WiSE showed the highest ESW of ∼2.75 V, and room temperature ionic conductivity of ∼23 mS cm−1, which increased to ∼32 mS cm−1 at 60 °C. Furthermore, the ionic conductivity at different temperatures was also determined employing EIS analysis (Fig. S2) and has been discussed in the SI.
Furthermore, we performed spectroscopic analysis via ATR-FTIR (Fig. 2b) and Raman spectroscopy (Fig. 2c) to illustrate the effect of salt concentration and EG as a molecularly crowding agent on the ion solvation and interionic interactions. A supportive MD simulation study was also conducted to understand the microstructure of the electrolyte ions and solvent molecules (water and ethylene glycol) for different concentrations and at different temperatures.
The ATR-FTIR spectra of 1, 5, 10, and 17 m NaClO4 and the MC-WiSE systems are shown in Fig. 2b. The broad band between ∼3075 and 3700 cm−1 is associated with the water vibration, with three shoulder regions, corresponding to the ice like liquid component (a fully interconnected tetrahedral hydrogen-bond network comprising five adjacent water molecules) at ∼3235 cm−1, isolate or monomeric water molecules at ∼3560 cm−1 and a water molecule with intermediate (1–3) hydrogen bonds at ∼3375–3400 cm−1.15,58
The ability of the perchlorate anion to break the intermolecular hydrogen bonding in water is evident form the increase in the intensity of the 3560 cm−1 band and decrease in the intensity of 3235 cm−1 band with increasing salt concentration from 1 m to 17 m. For the intermediate salt concentration of 5–10 m, the band at ∼3375–3400 cm−1 dominates, revealing an asymmetric environment, and represents the water molecules in the vicinity of cations and water molecules forming hydrogen bonds with the perchlorate anion.58
In MC-WiSE, the water vibrational features closely resembled those of 17 m NaClO4, with the ∼3560 cm−1 band becoming even more dominant and the ∼3400–3410 cm−1 band correspondingly diminished. This indicates that the introduced EG further disrupts the hydrogen-bond network by forming hydrogen bonds with water molecules, thereby increasing the proportion of isolated water molecules that are bonded to EG rather than to other water molecules. This could significantly decrease the water electrochemical activity, as evident in the LSV analysis (Fig. 2a).
Furthermore, the effect of salt concentration on the ion solvation and interionic interaction was analyzed. The low absorption band at ∼931 cm−1 in 1 m NaClO4 can be attributed to the symmetric stretching vibration (v1) of ClO4− anions, which is IR inactive. However, with increasing salt concentration the band showed a blue shift with a slight increase in relative intensity to 934 and 936 cm−1 in 5 m and 10 m NaClO4 and to 938 cm−1 in 17 m NaClO4 or in the MC-WiSE, indicating progressive formation of contact ion pairs with the cations.59 The ion pair formation was also evident from the red shifting of the v3 band of perchlorate anions from 1081 cm−1 in 1 m NaClO4 to ∼1070 cm−1 in either 17 m NaClO4 or in the MC-WiSE.58,59
The lower wavenumber shoulder at 1038 cm−1 of 1 m NaClO4, typically associated with weakly interacting or free ClO4− anions, shifts to 1033 cm−1, with a reduced intensity as the concentration increases to 17 m NaClO4. These shifts in the ClO4− bending region are indicative of progressive solvation shell reorganization and enhanced cation–anion interactions as free solvent molecules become less available in the highly concentrated electrolyte system.15
The Raman analysis (Fig. 2c) further supported the ATR-FTIR analysis of isolated water and contact ion-pair (CIP) formation upon increasing the salt concentration in either water or water + EG solvent. The high-resolution water stretching (2800 to 4000 cm−1) and chlorate stretching region (900 to 1000 cm−1) with deconvolution to understand the nature of hydrogen bonding and the local co-ordination environment are shown in Fig. 2d and f, respectively. With increasing NaClO4 concentration, notable variations were observed in the O–H stretching region, consistent with ATR-FTIR results.
Deconvolution of the Raman O–H stretching region reveals a redistribution of bonding populations with increasing NaClO4 concentration as presented in Fig. 2d. Water molecules can be classified into three bonding states: (i) strongly hydrogen-bonded water (SHW), forming a tetrahedral network with four neighbouring water molecules; (ii) weakly hydrogen-bonded water (WHW) involving 1–3 neighbouring water molecules, and (iii) isolated water (ISW), which is not hydrogen bonded with any neighbouring water molecules. In 1 m NaClO4, the O–H stretching bands associated with SHW, WHW, and ISW appear at approximately 3232, 3428, and 3580 cm−1, respectively, with the corresponding relative concentration of 31.3, 56.7 and 12%.60 With increasing salt concentration (≥5 m), the % of SHW decreases significantly with the corresponding –OH stretching vibration peak shifting to a higher wave number, while the % of ISW increases, reflecting weakened intermolecular hydrogen bonds and cation-dominated restructuring. This clearly suggests that increasing salt concentration causes an increased hydrogen bonding interaction between the chlorate anion and water molecules. The % of SHW, WHW and ISW as a function of salt concentration for 1–17 m NaClO4 and the MC-WiSE is shown in Fig. 2e. Notably, in the presence of EG, the extent of SHW further decreased from ∼9.6% in 17 m NaClO4 to ∼7.5% in MC-WiSE, highlighting enhanced disruption of the hydrogen-bonding network facilitated by EG coordination. Nevertheless, the % of ISW slightly decreased from 58.2% in 17 m NaClO4 to ∼52.3% in MC-WiSE.
Furthermore, deconvolution of the ClO4− stretching band (900–1000 cm−1) revealed a progressive formation of contact ion pairs (CIPs) from solvent separated ion pairs (SSIPs), upon increasing the salt concentration. In 1 m NaClO4, the band at ∼931.6 cm−1 corresponds to free anions (FA, ∼66.2%) and the band at ∼945.7 cm−1 represents the solvent separated ion pair (SSIP). Both the peaks showed a shifting to higher wave numbers for high concentration. At higher concentrations (5 m NaClO4), new peaks at ∼937.6 and ∼950.7 cm−1 emerge, attributed to a SSIP, and contact ion pair (CIP), respectively.61,62 The SSIP fraction dominates at 5 m (75%) but decreases with further increasing salt concentration, while CIP content increases from 24.9% in 5 m NaClO4 to 35.4% in 17 m NaClO4.60 In the MC-WiSE system, the CIP fraction further increased to 45.7%, highlighting the role of EG in promoting ion association via hydrogen-bond-mediated water capture, which indeed brings the counter ions closer (Fig. 2e).
To further understand the solvation environment as a function of salt concentration, EG additive and temperature, an in-detail molecular dynamic simulation study was performed. The baseline aqueous NaClO4 simulations (Fig. 3a) establish the fundamental solvation environment around sodium ions. The system remains homogeneously dispersed under NPT (isobaric-isothermal ensemble) conditions, and the total energy decreases monotonically with concentration (Fig. 3b), indicating enhanced electrostatic stabilization in high-molality environments. The radial distribution functions of sodium-water (Fig. 3c) for all the cases feature a pronounced first peak at ∼0.23 nm, consistent with well-defined hydration shells around sodium. As the salt concentration increases, water availability decreases, leading to a reduction in peak height and coordination number. Simultaneously, the sodium-chlorate (Fig. 3d) radial distribution shows a complementary increase with concentration, reflecting the progressive intrusion of perchlorate into the primary solvation shell. The coordination number of water and chlorate in the first hydration shell of sodium ions is shown in Fig. 3e. The coordination number quantitatively captures the water-anion exchange; the water coordination number decreases from ∼5 to ∼3 across the concentration range, while that of chlorate increases. These changes confirm that sodium transitions from a water-dominated to mixed anion-water solvation environment at high molality, an effect typical of chaotropic salts. The baseline behavior sets the stage for interpreting how ethylene glycol and electrode surfaces further modulate the solvation structure.
Fig. 3f–k depicts the effect of introducing ethylene glycol (EG) into the electrolyte phase. Increasing the EG concentration results in substantial changes in both microscopic interactions and macroscopic transport behavior of ions. The diffusion coefficient of sodium ions in the MC-WiSE (Fig. 3g) is calculated and shows a non-monotonic temperature dependence, with reduced diffusion values at elevated temperatures in the presence of ethylene glycol which would otherwise have higher values due to thermal activation. This reduction arises from the higher viscosity of ethylene glycol and its capacity to form stronger hydrogen bond networks. The sodium-water (Fig. 3h) and sodium-chlorate (Fig. 3i) radial distribution functions depict significant structural rearrangements with increasing concentrations of ethylene glycol. The interactions between sodium and water weaken, as the peak height reduces in the RDF with subtle radial shifts in the first hydration shell. In contrast, sodium-chlorate interactions slightly strengthen, indicating that ethylene glycol indirectly promotes closer cation–anion proximity by competing with water for hydrogen bonding. This behavior is further supported by interaction analysis of various constituents. The interactions between various constituents with an increasing ethylene glycol mass ratio and at different temperatures are shown in Fig. 3j and k, respectively. Interactions between water and ethylene glycol, and ethylene glycol and chlorate, increase markedly with increasing EG ratio. This also results in reduced water–NaClO4 interactions. This reorganization suggests that ethylene glycol disrupts the extended hydrogen bond network, thereby weakening water–ion interactions. Furthermore, the number of water–water hydrogen bonds in 17 m NaClO4 in the water–EG system as a function of EG concentration was quantified employing MD simulation and is included in Fig. S6. Temperature dependent results, at a 17 m concentration, show similar results wherein elevated temperatures weaken the interactions across all species, while preserving the affinity of EG towards both water and chlorate. These modifications suppress the hydration of sodium ions, promote partial ion-pairing, and reduce free-water mobility, thus resulting in reduced ionic mobility and altered electrochemical behavior.
Furthermore, MD simulation was performed to understand the electrode (MnO2) electrolyte interplay for NaClO4 electrolytes of different concentrations (1–17 m in water) and 17 m NaClO4 in a water–EG mixture with different EG concentrations.
In Fig. S7, the RDF of manganese with oxygen atoms of water shows a strong first peak followed by long-range oscillations, characteristic of layered water structuring near solid interfaces. As we increase the sodium chlorate concentration, the structuration diminishes, thus suggesting that high ionic strength disrupts the water layering and weakens hydration near the electrode. Higher concentrations of dissolved ions compress the hydration environment, screen electrostatic interactions, and weaken the layering effect close to the electrode. A similar decline is observed in a manganese-chlorate system which is counterintuitive. This trend suggests that at high ionic strengths, perchlorate ions move towards solvent-shared or contact ion pairs with sodium in the bulk rather than absorbing directly on the electrode. This results in a reduced number of free chlorate ions thus reducing the structured layers near the electrode surface. Furthermore, the overall high ionic crowding in the system suppresses long-range ordering near the interface, resulting in flattened RDF profiles for both water and anions.
Fig. 4 illustrates the role of ethylene glycol inclusion in the electrolyte and its interactions with the electrode. All these studies are performed at 17 m sodium perchlorate concentrations. The inclusion of EG in the electrolyte significantly alters the interfacial structuring, especially around the sodium ions. The RDF for manganese-water (Fig. 4c) shows very little change with addition of EG; however, chlorate ions increase near the electrode with increasing EG (Fig. 4d). This indicates slight competitive displacement of water by chlorate ions. This is indeed helpful in maintaining a water-lean condition at the interface to push the water redox potential window. Interestingly, the manganese-sodium RDF (Fig. 4e) shows that with temperature, peak intensity slightly increases suggesting thermally enhanced ion rearrangement near the electrode. The sodium-water (Fig. 4f) and sodium-chlorate (Fig. 4g) RDF suggest increasing sodium–water and sodium–chlorate interactions with increasing EG concentration. This reflects enhanced anion pairing under EG-rich and high temperature conditions. A particularly insightful descriptor is the ratio of water molecules to sodium ions in the electrode's first hydration layer (Fig. 4h). The ratio decreases dramatically with increasing temperature, indicating that high thermal energy and the presence of EG synergistically dehydrate the interfacial region. This dehydration weakens classical electric double-layer structuring and could have direct implications for interfacial charge transfer kinetics in practical systems. Experimentally we observed a significantly enhanced capacitance, arising mostly from a pseudocapacitive faradaic reaction at an elevated operational temperature of 60 °C. The accompanying sodium diffusion coefficients (Fig. 4i) further support this interpretation. Higher temperatures yield significantly higher mobility, although EG-rich systems remain slower than pure aqueous electrolytes. This interplay between thermal activation, viscosity, solvation disruption, and interfacial confinement explains the intricate transport behavior observed.
Fig. 5a presents the CV profiles of MnO2-SSC across various scan rates of 2–500 mV s−1 at room temperature (RT), whereas the profiles covering only the slow scan rates (2–10 mV s−1) are included in Fig. S8. The voltammogram showed mixed characteristics of double layer capacitive and pseudocapacitive behavior with a broad red-ox peak pair at 1.62 V–1.65 V (at 2 mV S−1). The redox peaks became boarder with increasing scan rates. The redox peaks can be ascribed to the Mn4+/Mn3+ redox couple corresponding to the following redox reaction:
| MnO2 + Na+ + e− ↔ NaMnO2 | (1) |
However, the redox peaks showed increased polarization with increasing scan rate, a behaviour commonly observed in metal oxide-based supercapacitors.63 At high scan rates, the CV profiles became spindle shaped, a behaviour commonly observed in metal oxides with a faradaic redox-reaction as the charge storage mechanism, where restricted ion diffusion and charge-transfer resistance dominate at higher scan rates.
The galvanostatic charge–discharge (GCD) profiles obtained at different currents of 0.63–4.76 mA cm−2 (considering the entire exposed electrode area) under ambient conditions are shown in Fig. 5b. The GCD profiles deviate from perfect linearity and have the typical characteristics of pseudocapacitive charge storage.15 At 0.63 mA cm−2, the MnO2-SSC delivered a maximum capacitance of 57.8 mF cm−2, corresponding to a high energy density of 50.2 µWh cm−2 at a power density of 793.7 µW cm−2. The device also showed excellent rate performance with a high capacitance retention of 37.7, 30.1, 25.2, 14, 11.4 and 8.9 mF cm−2 upon increasing the discharge current to 0.95, 1.27, 1.6, 2.54, 3.18 and 4.76 mA cm−2, respectively. The Ragone plot showing the areal energy and power densities of the MnO2-SSC is compared with many other recent reports in Fig. 5f. Notably, the cell was able to deliver a very high-power density of 5952.4 µW cm−2 while delivering a high energy density of 7.72 µWh cm−2.
Furthermore, the MnO2-SSC was electrochemically tested at an extremely low temperature of ∼−40 °C. Fig. 5c (CV at 2–10 mV s−1 scan rates in Fig. S9) presents the CV profiles of the MnO2-SSC recorded at scan rates ranging from 2 to 500 mV s−1 at an ultra-low temperature of −40 °C. Unlike the CV profiles with well resolved redox peaks as obtained at RT, the CV profile obtained at −40 °C showed somewhat suppressed redox peaks. This must be associated with the increased viscosity and reduced ionic mobility of the electrolyte at extremely low temperatures, along with reduced surface redox kinetics. As a result of this slow ion diffusion, which limits charge compensation speed and leads to an obviously increased internal resistance at extremely low temperature, the CV profiles at higher scan rates showed spindle-like geometry.
This was also reflected in the GCD analysis (Fig. 5d), where the cell showed a high capacitance of 25.2 mF cm−2 at 0.32 mA cm−2, corresponding to a high energy density of 21.8 µWh cm−2 at a power density of 396.8 µW cm−2. However, increasing current at an operational temperature of −40 °C showed a comparatively faster capacitance decay, as the cell retained capacitances of 19.7, 17.3, 12.9, 9.8, 4.7, 2.9 and 0.9 mF cm−2 at the different currents of 0.63, 0.95, 1.27, 1.59, 2.54, 3.2 and 4.76 mA cm−2, respectively.
The areal capacitances obtained at RT and at −40 °C are compared in Fig. 5e. Nevertheless, in spite of this diffusion-controlled limitation at extremely low sub-zero temperature, the large operational voltage enabled a high energy density of 8.48 µWh cm−2 and a high-power density of 1984 µW cm−2. Furthermore, at a very high-power density of 3968.3 µW cm−2, the cell exhibited a high energy density of 2.5 µWh cm−2. The Ragone plots obtained with the MnO2-SSC at ambient temperature and at −40 °C are shown in Fig. 5f, and they have been compared with several other published articles in this domain, affirming its adaptability in extreme environments.64–71
The cycling stability (Fig. 5g) of the MnO2 symmetric pseudocapacitor (MnO2-SSC) was evaluated at a current density of ∼1.6 mA cm−2. With a starting capacitance of ∼12.6 mF cm−2, the cell showed an initial capacitance drop to 7.8 mF cm−2 after 10
000 cycles and then became stable, retaining high capacitances of 6.17 and 5.82 mF cm−2 over 100
000 and 200
000 cycles, respectively.
After 200
000 cycles the cell exhibited a high energy density of 5.05 µWh cm−2 at a power density of 1984 µW cm−2. These values were higher than those in several other reports even when considering their performance before cycling.64,65,68,69,71 To further understand the phase stability and morphological stability of MnO2 post cycling, XRD and FESEM were performed. The post cycling FESEM analysis (Fig. 5j–l) showed that the bur-flower morphology of the electrodeposited MnO2 changed into irregular shaped interconnected nanoparticles, forming a porous architecture. The XRD pattern (Fig. 5i) obtained on a cycled cell after several cycles was very comparable to that of the freshly electrodeposited MnO2. Noteworthily, the redox reaction of MnO2 tends to undergo partial dissolution in aqueous based electrolyte, since the Mn3+ formed during the redox reaction may induce a disproportionation reaction to form Mn2+, which is soluble in aqueous electrolyte. This may be a probable reason behind the initial loss in capacitance.50,72 The MnO2 dissolution increases at acidic pH in a water rich system. However, the ‘water-lean’ electrolyte and a near neutral pH of the MC-WISE (pH = ∼5.8) should stabilize the MnO2 against further dissolution, leading to the observed high cycle stability after the initial capacitance loss.
The applicability of the MnO2-SSC under extreme conditions was further evaluated through a long-term cycling test of a similar cell at −40 °C for 4000 cycles at a current density of 1.6 mA cm−2, as shown in Fig. 5h. While the cell showed an initial drop in capacitance, from 4.8 to 2.75 mF cm−2 after 1000 cycles, it showed quite a stable behaviour afterwards, retaining an end capacitance of 1.71 mF cm−2 after 4000 cycles, leading to a high energy density retention of 1.49 µWh cm−2 at a power density of ∼1984 µW cm−2. Importantly, the MnO2-SSC maintained a high coulombic efficiency of ∼97–98% throughout the test. The practical application of the MnO2-SSC cell under low temperature operation is digitally demonstrated in Fig. 5m, where the cell was able to power blue/red and yellow LEDs or digital clocks, in a subzero (∼−40.5 °C) temperature environment.
To further evaluate the all-temperature applicability of the MnO2-SSC coupled with MC-WiSE, a temperature-dependent rate study was conducted using GCD measurements at a constant current density of 1.6 mA cm−2. The device was tested across a wide temperature range from −40 °C to 60 °C, within a potential window of 0–2.5 V and the GCD profiles obtained at different currents are shown in Fig. 6a, whereas the variation of capacitance and energy density as a function of operational temperature is shown in Fig. 6b. The MnO2-SSC exhibited excellent thermal adaptability, with a regular increment in capacitance with increasing temperature, delivering a high capacitance of 42.4, 61.9 and 86.3 mF cm−2 at 40, 50 and 60 °C, respectively. Notably, at 60 °C, the device showed a high energy density of 74.9 µWh cm−2 at a power density of ∼1984 µW cm−2, which is ∼3.1 and 7.3 times higher than that at RT and −40 °C, respectively.
This progressive enhancement in capacitance with increasing temperature can be attributed to the thermal energy induced increment in ionic mobility/conductivity, which facilitates faster ion transport and redox kinetics. The influence of operational temperature on the charge transfer resistance and diffusion kinetics was also evident from the electrochemical impedance spectroscopy (EIS) study (Fig. 6c). Notably, the solution resistance (Rs) values were found to be 81 Ω, 54 Ω, and 4.5 Ω at −40 °C, RT, and +60 °C, respectively. The charge transfer resistance (Rct) exhibited a decreasing trend with the increase in temperature, and was found to be 103 Ω, 45.5 Ω, and 6.22 Ω, at −40 °C, RT, and +60 °C, respectively. This significant reduction in Rct at elevated temperatures can be attributed to several factors, including the enhanced ionic mobility of the MC-WiSE, along with fastened charge transfer kinetics at elevated temperature leading to enhanced pseudocapacitance.
The enhanced surface redox activity at elevated temperature was further confirmed from the derivative capacitance vs. voltage profile (Fig. 6d). The dQ/dV vs. V profile (during discharge) for the MnO2-SSC in MC-WiSE at RT showed a well-defined peak centred at ∼1.18 V, the intensity of which gradually increased with a peak shifting towards higher voltage (1.36 V at 60 °C) upon increasing the operational temperature, indicating an enhanced surface reaction with increased reaction kinetics upon increasing temperature.
| (MnO2)surface + Na+ + e− ↔ (MnOONa)surface | (2) |
Interestingly, when the operational temperature was increased to 50 °C, an additional peak appears at ∼0.5 V, which further shifted to ∼0.56 V at 60 °C. The low voltage peak indicates the bulk intercalation, which is usually a slow process, and occurs only at elevated temperatures with increased ion diffusion kinetics.73
| (MnO2)bulk + Na+ + e− ↔ (MnOONa)bulk | (3) |
Thus, the enhanced surface pseudocapacitance combining bulk pseudocapacitance at high temperature results in a significantly high capacitance at 60 °C.
To further support the pseudocapacitive charge storage mechanism involving surface or near surface capacitive redox reactions, without involving any bulk phase change, ex situ XRD was performed while operating the cells at a high temperature of ∼60 °C. Ex situ XRD measurements were performed at the fully charged and fully discharged states after cycling (1000 cycles) at 60 °C (Fig. S10a). The diffraction patterns exhibit identical peak positions and profiles in both states, with no detectable phase transition, or emergence of secondary phases. The characteristic pattern of pristine MnO2 remains mostly unchanged, except for a slight shifting of the (131) plane towards a lower two theta value, confirming that no bulk crystallographic transformation occurs during charge–discharge even at elevated temperature. However, the highly amorphous nature of the active materials makes it difficult to accurately monitor any minor regional phase evolution. As has been shown later, when the high temperature cycling was continued for many cycles (10
000 cycles), ex situ XPS study showed the coexistence of reduced Mn species, indicating a likely mixed surface composition, after long-term operation at elevated temperature.
Furthermore, based on the peak potential in the differential capacity profile (Fig. 6d) the cell was interrupted at 1.35 V, 0.55 V and at 0 V during discharge while operating the cell at a high temperature of ∼60 °C. The XRD pattern at different states of discharge (Fig. S10b) of the electrode remained near identical revealing mostly surface or near surface redox reactions without any bulk phase change as the charge storage mechanism.
Furthermore, to validate the operational resilience of the MnO2-SSC in MC-WiSE, a temperature-dependent cycling study was conducted (Fig. 6e), where the cell was sequentially tested for 8 cycles each at −40 °C, RT, and +60 °C, with intermediate reversibility checks at RT between each temperature transition. Initially, the cell delivered an average capacitance of 27.3 mF cm−2 at RT, which decreased to 10.14 mF cm−2 and recovered to 27.7 mF cm−2 upon changing the cell operational temperature to −40 °C and reverting to RT, respectively. While continuing the sequential operational temperature in the order of RT → −40 °C → RT → +60 °C → RT → −40 °C → RT, the cell showed average capacitances of 27.3, 10.14, 27.7, 85, 28, 10, 28.3, and 86.4 mF cm−2, respectively, demonstrating an excellent thermal response with consistent and repeatable capacitance values in spite of rigorous operational temperature alternation. Furthermore, a more aggressive and long-term cycling over −40 °C → RT → +60 °C temperature variation is shown in Fig. S11.
Furthermore, a fresh cell was subjected to cycling at high temperature (∼63 ± 2 °C) for 10
000 cycles (Fig. 6f) where it showed a highly stable performance with a capacitance retention of 102 mF cm−2. Thus, a high-temperature environment promotes faster ion diffusion and more efficient redox kinetics, leading to enhanced charge storage capability.
Additionally, comprehensive postmortem characterization of the cell components after high-temperature cycling (10
000 cycles), including the MnO2@LIG electrode (FESEM, HRTEM, XPS) and the MC-WiSE (NMR, ICP-OES), was carried out to elucidate the structural/phase changes in the active material, Mn dissolution behaviour, possible EG adsorption on the electrode surface, and the electrochemical stability of the MC-WiSE.
The FESEM image (Fig. 7a) showed spherical morphology of the cycled MnO2, which is different from that of the pristine sample, where a burflower-like morphology was observed. No long-range periodic fringes are clearly visible from the HRTEM image (Fig. 7b), indicating the amorphous nature of the cycled material. Also, the SAED pattern (inset of Fig. 7b) exhibits concentric rings composed of discrete spots, indicating a polycrystalline nanostructure with a random orientation. The diffuse background suggests the presence of amorphous or poorly crystalline manganese oxide formed during electrochemical cycling.
EDAX elemental analysis (Fig. 7c) showed the presence of Mn, O, and C along with trace Na and few impurities mostly arising from the adhesion of the PI substrate and Cu grid. The trace Na peak can be from trapping of the Na+ ion, which could not be de-intercalated fully during the reverse process. However, this intercalation is only near the surface and in the short range, which does not cause any bulk phase change. This is consistent with the ex situ XRD analysis in Fig. S10. The absence of any strong Cl- peak indicates that the electrolyte salt does not degrade at the interface giving any insoluble deposit. Furthermore, the STEM-EDX mapping (Fig. 7d–g) showed the coexistence of carbon along with Mn and O, indicating possible chemical adsorption of ethylene glycol (EG) on the MnO2 electrode. This adsorption may partially block the active sites; however, this effect was not further considered in this study.
Furthermore, to probe the effect of long-term high temperature cycling on the degradation behaviour of the active materials XPS analysis of the cycled electrode was performed. The high-resolution Mn 2p XPS spectra (Fig. 7h) showed the characteristic Mn 2p3/2 and 2p1/2 peaks. The deconvoluted Mn 2p3/2 XPS spectra revealed the contribution from Mn3+ (640.9 eV) and Mn4+ (∼642.5 eV), and a subtle contribution from Mn2+ (639.1 eV), indicating partial reduction of MnO2 and formation of mixed valent Mn–Ox after prolonged cycling at elevated temperature.74 The reduction could be associated with the local/short range Na+ insertion effect, formation of hydroxylated-Mn or defected MnOx. Eventually the Mn3+ may undergo a self-disproportionation reaction to produce Mn2+ species, which is soluble in the aqueous based electrolyte. In fact, ICP-OES analysis of the recovered cycled electrolyte confirmed Mn dissolution with ∼53.4 ppm manganese in the electrolyte. This clearly indicates a possible dissolution–redeposition process. Nevertheless, the dissolution is expected to be slower in the water-lean electrolyte, compared to that in the water reach case.
Furthermore, the deconvoluted O 1s XPS analysis (Fig. 7i) supported the contribution from lattice oxygen (Mn–O–Mn, BE ∼529.8 eV), defect oxygen (Mn–Ox, BE ∼530.6 eV), and surface hydroxyl groups (Mn–OH, BE ∼531.3 eV) from the cycled manganese oxide electrode.75 The BE peak at ∼532.2 could be assigned to the O–C/O
C functionality appearing from surface oxidation of LIG.76 Furthermore, the high BE peak at ∼533 eV could be assigned to surface adsorbed water molecules and organic species (EG).77 Notably, this peak concentration in the cycled electrode was much higher compared to that of the pristine material (MnO2@LIG).
The high-resolution C 1s XPS spectra (Fig. 7j) showed distinct deconvoluted BE peaks at ∼284.7 eV, ∼285.9 eV, and ∼288.5 eV corresponding to the C–C/C
C, C–O and C
O bonding environment, revealing a partially oxidized LIG surface.78 The high BE deconvoluted peak at ∼291 eV could be assigned to the π–π* shake-up satellite, characteristic of the graphitic domain.79 The absence of any BE peak corresponding to the O–C
O species indicates that long-term cycling at 60 °C does not induce significant carbonate-type electrolyte decomposition on LIG. However, a slight increase in C–O (32.6% vs. 22.2%) and C
O (13% vs. 9.2%) components in the cycled MnO2@LIG vs. pristine MnO2@LIG suggests mild surface oxidation of the graphene framework, likely originating from electrochemical edge functionalization during prolonged cycling.
Notably, the MC-WiSE recovered from the high temperature cycled cell showed excellent electrochemical stability with no oxidation of ethylene glycol, as observed in the 1H and 13C NMR study (Fig. S12). A comparison of the NMR spectra (in D2O) of the freshly synthesized MC-WiSE and MC-WiSE recovered from the high temperature cycled cell (10
000 cycles operated at ∼63 ± 2 °C) or room temperature cycled cell (1000 cycles) showed no new carbon or proton related signal, indicating excellent electrochemical stability of the MC-WiSE.
Furthermore, the mechanical flexibility of the device was systematically evaluated by performing galvanostatic charge–discharge (GCD) measurements at different bending angles (0°, 60°, 90°, 120°, and 180°), as shown in Fig. S13. The areal capacitance shows only a slight decrease from ∼30.7 to ∼29.1 mF cm−2 at 180° as shown in Fig. S13a, corresponding to a capacitance retention of ∼95%.
The trivial loss in capacitance upon extreme bending can be from an increased internal resistance and minor interfacial strain between the active material (MnO2) and the LIG substrate, affecting the interfacial contact and ion accessibility. Importantly, upon returning to the flat state, the capacitance nearly recovers to its initial value (30.7 mF cm−2), indicating that these effects are primarily elastic and reversible, confirming the robust mechanical integrity and strong adhesion of the electrode materials.
The GCD profiles exhibited nearly identical shapes under different bending conditions, indicating stable charge storage behavior as shown in Fig. S13b.
Furthermore, the device flexibility was digitally demonstrated (Fig. S13c), where the cell successfully powered an LED with no observable change in brightness under both flat and bent conditions.
Supplementary information (SI): detailed experimental procedures for laser induced graphene synthesis and MnO2 electrodeposition, electrolyte formulation, and electrochemical testing; molecular dynamics simulation details, supporting figures (XPS, viscosity, DSC, CV, ex situ XRD, NMR), quantitative analyses of ion pairing and hydrogen bonding, and comparative performance tables of supercapacitors across different temperatures. See DOI: https://doi.org/10.1039/d5ta10251k.
| This journal is © The Royal Society of Chemistry 2026 |