A bi-functional configuration for a metal-oxide film supercapacitor

Abstract Transition metal oxides can introduce high pseudocapacitance to an electric double layer capacitor for storing more electrical charges. Their role can be more than that since they possess high dielectric constant. Here, we propose a self-sustainable bi-functional configuration by eliminating traditional separator of a metal oxide film supercapacitor that is capable of providing good energy storage performance. We take advantage of super-aligned electrospun carbon nanofibers (SA-ECNFs) as an interconnected scaffold, coupling with electrochemical deposition of α-MnO2 layers at different currents to introduce pseudocapacitance while providing dielectric layer functioning as a separator to assemble a state-of-the-art supercapacitor. The good electrochemical performance of galvanic charge-discharge specific capacitance at 141 F g−1 and energy density at 12.5 Wh kg−1 offers the promising applications in the energy storage field.

A separator between two electrodes is a critical component in both lithium-ion batteries and supercapacitors. Its function is to provide an insulator between two electrodes while allowing free ions to flow through. They usually are polymeric membrane or non-woven fabric mat with thicknesses ranging from 25 μm to 200 μm [ [26], [27], [28], [29]]. However the thicknesses of a typical supercapacitor electrode range from 100 nm to 120 μm [30,31], namely significant space (9.6%-50%) inside the cells was taken by the separator (Fig. S1), which may adversely affect their energy storage performance and product manufacturing. In principle, a separator provides low electric conductivity as insulator meanwhile offers channels for ions to transport. Manganese dioxide possesses internal pore structures which allow the permeation and rapid transport of ions to the nanoscopic MnO2 phase [25]. The low electric conductivity [32] of MnO2 meets the separator requirement in batteries and supercapacitors.
In our recent work, we reported a low-current controlled electrodeposition of metal oxides (e.g. MnO2) on super-aligned electrospun carbon nanofibers (SA-ECNFs) (Fig. S2) [33]. The thickness of metal oxides can be tailored by the magnitude of current deposition and time while the aligned ECNFs provide advantages for the formation of uniform packed films [33]. Therefore, we propose to utilize the dielectric property of compact massive MnO2 film loading to function as a separator, meanwhile to introduce pseudocapacitance energy storage to the SA-ECNF scaffold electrode. This is different from a previous report using graphene oxide film as separator in a supercapacitor in which the graphene oxide film demonstrates proton conductivity [34].
In this work, a new configuration of supercapacitor where the MnO2 layer functions as both the pseudo-material and separator is presented. This novel design may save the space of a traditional separator and, more importantly, simplify the assembly procedure in the supercapacitor manufacturing. The SA-ECNFs are synthesized by a facile electrospinning setup with alignment ( Fig. S3). A variety of uniform crystalline α-MnO2 layers of thickness ranging 200 nm-2000 nm onto SA-ECNFs as one of the electrodes in an asymmetrical supercapacitor cell are tested to examine the energy storage performance without using traditional separators.

Results and discussion
Fig. 1a-c schematically illustrate a new supercapacitor configuration of a non-separator design consisting of MnO2/SA-ECNFs hybrid film as one electrode and a CNT paper as the other electrode for current collection. An insulator (3M vinyl electrical tape) was added to prevent the shortage between the end of the CNT paper electrode and the exposed SA-ECNFs end. The potion of SA-ENCFs coated with MnO2 film and the CNT film form an asymmetrical cell without using traditional membrane separator for electrochemical and energy storage measurements. Note that the orange arrows display potential electron flows from current collector on the MnO2 electrode side to the other side (Fig. 1b). The component preparation and testing device assembly are briefly described as follows. The SA-ECNF film was cut into 2 × 1 cm 2 pieces and half-immersed in the electrolyte for electrodeposition. The as-prepared MnO2/SA-ECNFs hybrid film, CNT paper, and insulator were transferred into a glovebox then were assembled in a standard testing platform (Fig. 1c) [35]. The thicknesses of these films were carefully measured to ensure the good contact between the stainless steel plate as current collector and the supercapacitor electrodes (Fig. S4).  Our recent study indicates that the alignment of ECNFs can lead to a more uniform electrodeposition process for MnO2 film formation. It is concluded that the growth of MnO2 on SA-ECNFs follows a three-step mechanism including the final self-cessation based on the electrodeposition time under a small electrodeposition current at μA level. Therefore, a variety of low currents (from 20 μA to 80 μA) were applied by an electrochemical workstation for electrodeposition at different time ranging from 0.5 h to 12 h to reach the self-cessation growth thus reaching the maximum film thickness (resultant details in Table S1) at each applied current. The composites' structure and morphology of the deposition were characterized by SEM ( Fig.  2c-f, Fig. S5). The SEM images show a compact uniform MnO2 films which wrap around the carbon fiber core (except the 20 μA deposition). As discovered in our recent study [33], the MnO2 film keeps growing as electrodeposition time increases until growth stops at a certain time point which depends on the magnitude of the deposition current. In this study, the SEM image shows the 20 μA electrodeposition could not achieve a uniform coating (Fig. 2c). When the electrodeposition current was higher than 20 μA, the current flow could keep depositing MnO2 onto SA-ECNFs after all surface activation sites of the AS-ECNFs were occupied. Clusters merge with one another to form small balls around the SA-ECNFs then keep growing until the fibers are fully covered. Fig. 2d-f shows the fully covered structure of MnO2/SA-ECNFs with single layer thicknesses ranging from 897 nm to 1971 nm under a deposition current ranging from 40 μA to 80 μA. Fig S6 shows the thickness distribution of different deposition currents with the saturation of film growth. As an outstanding type of carbon fibers, SA-ECNFs have a horizontally aligned structure, which highly facilitates the electron flow through the fiber arrays for a uniform electrodeposition. The SA-ECNFs with diameters of 305 ± 53 nm (Fig. S7) were synthesized with a standard electrospinning method followed with stabilization and carbonization [36]. Then MnO2 can be deposited onto the SA-ECNFs to form a thick layer, and its thickness (Fig. S6) and mass (Fig. S8) were measured. In general, the thickness of the MnO2 film increases with the applied constant current until the formation of a compact film for a current of 40 μA or larger.
The as-prepared MnO2/SA-ECNFs films were also characterized to study their chemical composition. The energy dispersive X-ray (EDX) analysis (Fig. S9) shows that the surface composition of the electrochemically deposited electrodes is composed of the elements C, O, and Mn. The atomic ratio of O and Mn is close to 2:1, which implies the formation of MnO2. The chemical composition of the MnO2/SA-ECNFs sample was also investigated by X-ray Photoelectron Spectroscopy (XPS). The high-resolution Mn 2p spectra for MnO2/SA-ECNFs is presented in Fig. S10a. Two strong peaks at 642.5 and 654.0 eV can be clearly seen, corresponding to the Mn 2p3/2 and Mn 2p1/2 spin-orbit peaks of MnO2, respectively [37,38]. The pronounced peak in the O 1s XPS spectrum (Fig. S10b) at 530.3 eV is assigned to Mn−O−Mn in the composite, and the peak at 531.2 eV can be assigned to Mn-OH formation [39,40]. To investigate the MnO2 crystal structure, the as-prepared MnO2/SA-ECNFs under different electrodeposition conditions were examined by X-ray diffraction (XRD). Compared to the results of SA-ECNFs only (Fig. S11a), the XRD patterns could be fully indexed to α-MnO2 (JCPDS No. 44-0141) [41] as showed in Fig. S11b  The resistance of MnO2/SA-ECNFs electrodes with 40 μA (charge transfer resistance of about 330 Ohms), 60 μA (charge transfer resistance of about 511 Ohms) and 80 μA (charge transfer resistance of about 709 Ohms) electrodeposition was firstly deduced from the electrochemical impedance spectroscopy (EIS) Nyquist plots from 100 kHz to 0.01 Hz (Fig. 3a) and a Randles equivalent circuit model fitting (Fig. S12). These plots consist of a semicircle in the highfrequency region, which display an increase in charge transfer resistance of the MnO2/SA-ECNFs electrodes with an increase in electrodeposition current [42,43].
To investigate the energy storage performance of the MnO2/SA-ECNFs full cell, an asymmetric two-electrode supercapacitor system was assembled (Figs. 1 and S4) without the use of cellulose separator and a 6 M KOH aqueous is used as supporting electrolyte. Galvanic charge-discharge tests of different electrodes using the metal oxide-based bi-functional configuration without membrane separators were performed at current densities of 2.5 A g −1 , 5 A g −1 and 10 A g −1 in the asymmetric supercapacitor prototype. The charge-discharge curves are shown in Fig. 3b-d. Electrical shortcut happens during the charge-discharge testing of the 20 μA deposited electrode system due to the insufficient thickness and/or nonuniformity of MnO2 film. Fig. 3e demonstrates the capacitance of the MnO2/SA-ECNFs electrodes with 40 μA, 60 μA and 80 μA deposition is 141 F g −1 , 91 F g −1 , 56 F g −1 , and Fig. 3f presents the energy density of 12.5 Wh kg −1 , 8.1 Wh kg −1 , and 5 Wh kg −1 with respect to the whole electrode mass, respectively. To examine the electrochemical behavior of the electrodes, cyclic voltammetry tests were conducted in a three-electrode setup where a gold electrode taped with MnO2/SA-ECNFs served as the working electrode, platinum wire worked as the counter electrode, and Ag/AgCl as the reference electrode). Fig. S13 compares the cyclic voltammograms of the MnO2/SA-ECNFs electrodes of scan rates at 5 mV s −1 to 50 mV s −1 under different deposition current and time. As it can be seen, deposition under higher current exhibits larger integrated curve area, indicating the potentially higher absolute areal capacitance (80 μA > 60 μA > 40 μA > 20 μA). While considering the different mass loading and thickness of the MnO2 layers, the MnO2/SA-ECNFs under 20 μA deposition current has the highest specific capacitance from CV with respect to mass of the whole electrode. The capacitance calculated from CVs with 20 μA, 40 μA, 60 μA and 80 μA deposition is 652 F g −1 , 396 F g −1 , 142 F g −1 and 89 F g −1 , respectively. Note that the potential window was set to be 0.8 V based on the deviated CV shapes (comparing to the ideal rectangular shape) particularly at higher scan rates. Similar results can be found in our previous reports [33] and can be attributed to the obstructed tunneling of the electrolyte ions in the thick MnO2 layers. The electric double layer capacitance of pure SA-ECNFs electrode has been examined under same conditions [16,33]. The capacitance is less than 2 F/g, about 1% of the total specific capacitance of MnO2/SA-ECNFs electrode. Hence, the contribution of the SA-ECNFs in energy storage is insignificant and not considered in the following discussion.
The contribution of pseudo-capacitance depends on the reversible redox reactions between Mn(IV)/Mn(III) species and K + intercalation/de-intercalation at the MnO2/electrolyte interfaces [ [44], [45], [46]]. After a competing retention test at a current density of 2.5 A g −1 with and without the use of cellulose separator, the non-separator configuration shows an acceptable loss of energy density (Fig. S14) after 200 cycles. Fig. 3b presents a galvanic charge-discharge curve in a bulged triangle with 40 μA deposition current, suggesting pseudo-redox reaction in capacitance contribution. In contrast, Fig. 3c and d indicate concaved triangle curves with deposition current larger than 40 μA. The shape change of the galvanic charge-discharge curve may correlate to the significant increases of the resistance based on other reports of similar MnO2 hybrid electrodes [16,[47], [48], [49]]. The 40 μA deposited electrode displays a good energy storage without using any separator in the testing setup, suggesting the feasibility of a self-sustainable bi-functional MnO2 supercapacitor configuration based on uniformly coated metal oxide film-ECNFs electrodes.
A hypothesis of the bi-function of MnO2 film at the electrode is proposed to describe energy storage mechanism. Assuming that the MnO2 film contributes both the pseudocapacitance and the "separator" function, the assembled MnO2/ENCFs electrode and its functions are depicted in Fig. 4a. The MnO2 layer at the fiber surface is divided into two sublayers. The inter-layer near the carbon nanofiber surface is capable of electron penetration at a specific voltage during the charge/discharge process. The depth (d) is a function of applied voltage based on electron tunneling process in the dielectric material at contact and increases with a bigger applied voltage [50]. Within the inter-layer of MnO2, the redox Mn(IV/III) reaction occurs during the charge/discharge process, thus contributes to the energy storage as the pseudocapacitance portion. The inter-layer thickness is defined as hps and equivalent or correlated to the electron "penetration" depth (d). If the deposited MnO2 layer is thick enough, the outer-layer may act as a resistance barrier to avoid the direct shortage between two electrodes in the cell, thus functions as a "separator" (hsep). In order to realize the self-sustainable bi-functional metal oxide configuration for a supercapacitor, the total thickness of deposited MnO2 should be greater than the electron tunneling depth (d). In this work, 20 μA deposition for the saturated film is not sufficient to reach the tunneling depth. We further analyzed the experimental results based on this proposed energy storage model. In this study, MnO2 saturation thickness depends on the applied magnitude of electrodeposition current. The total thickness (htotal) can be divided into the inner-layer for the pseudocapacitance portion (hps) and the outer-layer for the "separator" function portion (hsep). Note that hps contributes to the reversible redox reactions between Mn(IV)/Mn(III) species at the interlayer MnO2/carbon interfaces while the resistance of hsep is responsible to the separation function between two electrodes allowing free ions to flow and accumulate at the outer-layer MnO2/electrolyte interface to avoid electrical shortage. Since the inner layer MnO2 thickness is limited by the electron tunneling in the metal oxide, the higher htotal would lead to larger resistance (directly) and lower energy density (indirectly) (Fig. 4b). Herein, there is a tradeoff to the energy storage due to the resistance of the outer-layer MnO2 functioning as "separator". To balance the pseudocapacitance and the separator role of MnO2, an optimal electrodeposition current (Iopt) could be modeled. Using the energy density (De) results with respect to the deposition current (I), Fig. 4c shows the best fit of the hsep versus electrodeposition current data (I) by an analytical equation (1): where the (htotal-d)max represents the maximum outer-layer MnO2 thickness allowing free ions to flow and accumulate, which is equivalent to the hsep corresponding to different electrodeposition currents to their maximum thick MnO2 deposition (Fig. 4d). The best fitting gives an Iopt of about 47 μA to reach the maximum energy density storage and the best hsep can be obtained to be 983 nm, which is close to the result obtained from the 40 μA deposition-electrode in this work.
Galvanic charge-discharge tests of the same prototype using a cellulose separator indicate that MnO2/SA-ECNFs electrodes obtained a higher specific capacitance and energy density (Fig.  S15) from the 20 μA deposition-electrode with respect to the whole electrode mass. For the 40 μA or higher current deposition-electrodes, the bi-functional metal oxide configuration offers reasonable specific capacitance and energy density in a comparison to that of configuration with a membrane separator with respect to the electrode mass. Nevertheless, this new metal oxidebased bi-functional configuration can reduce the process of separator installation for supercapacitor production.

Conclusion
This work, for the first time, demonstrates a novel configuration of supercapacitor free of a traditional separator membrane by uniformly electrodepositing MnO2 on aligned SA-ECNFs. The feasibility of MnO2 layer functioning as both the pseudocapacitive material and separator in different thickness by changing the electrodeposition current was explored and confirmed. From the electrochemical performance, MnO2/SA-ECNFs with 40 μA electrodeposition display a capacitance of 141 F g −1 and a specific energy density of 12.5 Wh kg −1 . This research presents a potential revolution of the metal oxide/carbonaceous nanomaterial-based pseudocapacitance energy storage by simply increasing the thickness of the metal oxide as replacement of a membrane separator, thus simplifying the manufacturing process.