Photo-assisted charging of carbon fiber paper-supported CeO2/MnO2 heterojunction and its long-lasting capacitance enhancement in dark

It is important to develop green and sustainable approaches to enhance electrochemical charge storage efficiencies. Herein, a two-step in-situ growth process was developed to fabricate carbon fiber paper-supported CeO2/MnO2 composite (CeO2/MnO2—CFP) as a binder-free photo-electrode for the photo-assisted electrochemical charge storage. The formation of CeO2/MnO2 type II heterojunction largely enhanced the separation efficiency of photo-generated charge carriers, resulting in a substantially enhanced photo-assisted charging capability of ∼20%. Furthermore, it retained a large part of its photo-enhanced capacitance (∼56%) in dark even after the illumination was off for 12 h, which could be attributed to its slow release of stored photo-generated electrons from its specific band structure to avoid their reaction with O2 in dark. This study proposed the design principles for supercapacitors with both the photo-assisted charging capability and its long-lasting retainment in dark, which may be readily applied to other pseudocapacitive materials to better utilize solar energy.


Introduction 
The rapid consumption of fossil fuels and the ensuing environmental pollution have led to increasing demands in developing sustainable energy supplies throughout the world [1][2][3]. Among various energy resources, solar energy is virtually unlimited and regarded as one of the most promising solutions to address these concerns [4][5][6][7][8][9]. However, solar energy is low-density and intermittent, which has to be stored to compensate the fluctuating availability of the Sun and when these semiconductor-based active materials were illuminated to enhance their energy conversion and charge storage efficiencies. For example, a photorechargeable zinc ion battery was proposed using V 2 O 5 as the photo-active cathodes, which showed a significant capacity increase under illumination, and its photoconversion efficiencies reached ~1.2% [25]. An et al. [26] reported a nanopore Cu@Cu 2 O hybrid array with an increased capacitance of ~37.9% under photo-assisted charging, which was attributed to photo-generated charge carriers. Furthermore, photo-assisted charging of zinc ion-based capacitors, batteries, as well as supercapacitors was also successively proposed and showed application potentials [17][18][19][20][21][22][23][24][25][26][27]. However, they generally lacked of the capability to retain their extra charge storages from the photo-assisted charging process when the illumination was switched off, which largely limited their practical applications.
Although MnO 2 has drawbacks of low actual specific capacitance and poor cycling stability due to its low conductivity and intrinsic redox reactions [28], it has emerged as a promising pseudocapacitive material among various active materials for supercapacitors owing to its high theoretical capacitance, low cost, environmental friendliness, and natural abundance [29,30], and has been attracting extensive research interests [31][32][33][34][35]. However, there is still no report in the literature on the photo-induced capacitance enhancement behavior of supercapacitors with traditional pseudocapacitive MnO 2 -based materials as the active component, which may be related to the easy photo-generated charge carrier recombination in them to offset the photoassisted charging effect. Due to their great application potentials, it is of great interest to develop strategies to endow the photo-assisted charging capability to MnO 2 -based supercapacitors to enhance their capacitance. Furthermore, it would be even better if their capacitance enhancement by the photo-assisted charging could be retained in dark for an extended period of time.
Herein, we developed a two-step in-situ growth process to fabricate carbon fiber paper-supported CeO 2 / MnO 2 composite (CeO 2 /MnO 2 -CFP) as a binder-free photoelectrode, which demonstrated a good photo-assisted charging capability and could retain a large part of its capacitance enhancement from the photo-assisted charging for an extended period of time in dark. CFP served as the support material for the growth of MnO 2 , which could provide long-range electron-transport pathways, avoid the use of polymer binders, and effectively increase the contact area of the electrode. Subsequently, a small amount of CeO 2 nanoparticles were in-situ grown on the surface of MnO 2 -CFP for the formation of CeO 2 /MnO 2 -CFP composite electrode. As a photosensitive semiconductor, CeO 2 has been extensively explored for a wide range of energy-related applications, including photocatalysis, supercapacitor, and lithium batteries [36][37][38][39][40][41][42]. The construction of a type II CeO 2 /MnO 2 heterojunction could not only provide additional photo-induced charge carriers upon proper light irradiation in this material system, but also enhance the charge carrier separation in MnO 2 to make a better use of its photo-generated charge carriers to substantially promote its photo-assisted charging capability. Under visible light illumination, the as-prepared CeO 2 / MnO 2 -CFP electrode demonstrated a specific capacitance of ~303 F·g −1 at 0.25 A·g −1 , which was ~53 F·g −1 higher than that of the MnO 2 -CFP electrode. Furthermore, it could retain over half of its photoinduced capacity enhancement (~56%) even after the visible light irradiation was shut off for 12 h, which could be attributed to its slow release of stored charges from its specific electronic band structure to assure its long-lasting capacity enhancement in dark after the visible light charging was over.

2 Synthesis of the MnO 2 -CFP sample
MnO 2 -CFP composites were prepared via a robust www.springer.com/journal/40145 hydrothermal process with modifications [43][44][45]. Prior to the synthesis, CFP (2 cm × 3 cm) was pretreated successively with acetone, 10% hydrochloric acid, deionized water, and ethanol under ultrasonic cleaning, and finally dried in an oven overnight at 60 ℃. In a typical synthesis, 0.75 mM KMnO 4 was dissolved in deionized water (35 mL) under constant magnetic stirring for 30 min, and then transferred into a Teflon-lined stainless-steel autoclave. In the meantime, CFP was placed into the KMnO 4 solution, and a hydrothermal reaction was undertaken at 150 ℃ for 6 h. After cooling down to room temperature naturally, CFP coated with a brown-black product was taken out from the autoclave, washed several times with deionized water and ethanol, and then dried at 60 ℃ in air for 12 h to obtain the MnO 2 -CFP sample. For comparison purpose, MnO 2 nanoparticles were also synthesized with the same hydrothermal process except for the placing of CFP into the KMnO 4 solution.

3 Synthesis of the CeO 2 /MnO 2 -CFP sample
0.024 g Ce(NO 3 ) 3 ·6H 2 O was dissolved in 30 mL deionized water with moderate stirring magnetically for 30 min. Then, the as-obtained solution and a piece of the as-prepared MnO 2 -CFP substrate were transferred into a Teflon-lined stainless-steel autoclave and heated at 90 ℃ for 6 h in an oven to deposit CeO 2 nanoparticles onto MnO 2 through the redox reaction between Ce 3+ and MnO 2 [46]. After cooling down to room temperature naturally, it was taken out from the autoclave, washed several times with deionized water and ethanol, and then dried at 60 ℃ in air for 12 h to obtain the CeO 2 /MnO 2 -CFP sample. For comparison purpose, CeO 2 nanoparticles were also synthesized by a water bath method, in which 0.6 g Ce(NO 3 ) 3 ·6H 2 O was firstly added into 200 mL of deionized water and stirred for 10 min. Then, 5 g hexamethylenetetramine (C 6 H 12 N 4 ) was added into the solution under magnetic stirring for another 10 min. Finally, the solution was placed in a water bath at 90 ℃ for 1 h under continuous stirring to obtain CeO 2 nanoparticles.

4 Material characterization
The crystal structures of the as-prepared samples were investigated by the X-ray diffractometer (Empyrean, Malvern Panalytical, the Netherlands). Their morphologies were observed by both the field emission scanning electron microscope (FESEM; JSM-7610F, JEOL, Japan) and the transmission electron microscope (TEM; JEM-2100F, JEOL, Japan) equipped with a mapping system. The X-ray photoelectron spectroscopy (XPS) was obtained on an X-ray photoelectron spectrometer (ESCALAB 250Xi, Thermo Scientific, USA) with a monochromatized Al Kα X-ray source, and a spectrophotometer (UV-3600Plus, Shimadzu Corporation, Japan) was used to measure the optical absorbance spectra of these samples. The element ratios were confirmed by the inductively coupled plasma optical emission spectrometer (5110 ICP-OES, Agilent, USA).

5 Electrochemical measurements
The electrochemical performances of the as-prepared samples were measured by an electrochemical workstation (CHI760E, Shanghai Chenhua Instrument Co., Ltd., China) in a 1 M Na 2 SO 4 electrolyte at room temperature with a three-electrode system. CFP, MnO 2 -CFP, and CeO 2 /MnO 2 -CFP samples of 1 cm × 1 cm in area were directly used as working electrodes. The CeO 2 working electrode was prepared by a slurry coating procedure. CeO 2 nanoparticles, acetylene black, and polyvinylidene fluoride (PVDF) were first mixed uniformly in a weight ratio of 8 : 1 : 1, and then dispersed in N-methylpyrrolidone (NMP) to form a uniform slurry. The slurry was coated on a CFP sample of 1 cm × 1 cm in area, and then dried at 60 ℃ for 12 h. A platinum plate was used as the counter electrode, and an Ag/AgCl electrode was used as the reference electrode. The mass loadings of active materials on CFP were all controlled at ~1.4 mg·cm −2 . The light source was a 300 W xenon lamp (PLS-SXE300+, Beijing Perfect Light Technology Co., Ltd., China) with filters to provide visible light illumination (400-780 nm) or visible/ infrared illumination (400-2200 nm). A circulation cooling water system was used to keep the temperature stable during the performance measurement.

1 Creation of the CeO 2 /MnO 2 -CFP sample
The CeO 2 /MnO 2 -CFP sample was prepared by a facile, two-step in-situ growth method with CFP as the substrate. Figure 1(a) compares the X-ray diffraction (XRD) patterns of MnO 2 -CFP and CeO 2 /MnO 2 -CFP samples. Both samples demonstrated the typical peaks of CFP at 2θ of ~26.4° and 54.5° from their CFP substrates, while a tiny peak at 2θ of ~12.5° could be observed in the XRD patterns of both samples, which could be ascribed to the (002) peak of MnO 2 (JCPDS: 18-0802). Thus, the XRD analysis results confirmed the successful formation of MnO 2 on the CFP surface. The inset in Fig. 1(a) shows their magnified XRD patterns from 2θ of ~28° to 80°, which demonstrated that several new diffraction peaks at 2θ of ~28.54°, 33.07°, 47.47°, and 56.33° emerged for the CeO 2 / MnO 2 -CFP sample, corresponding to (111), (200), (220), and (331) peaks of the cubic CeO 2 phase (JCPDS: 81-0792), respectively. Thus, the CeO 2 /MnO 2 -CFP sample was successfully created by the second in-situ growth step of our approach through the decoration of CeO 2 nanoparticles on the MnO 2 -CFP sample. Figure S1(a) in the Electronic Supplementary Material (ESM) shows the SEM image of the MnO 2 -CFP sample, which demonstrated that a dense MnO 2 layer was uniformly coated on the surface of carbon fibers, resulting in a three-dimensional interweaved structure. This structure was highly porous with a large surface area, which was beneficial for the efficient transportation and diffusion of electrons and electrolyte ions. Figure  1(b) shows the SEM image of the CeO 2 /MnO 2 -CFP sample, which clearly demonstrated that the deposition of CeO 2 nanoparticles on the MnO 2 -CFP sample did not change its whole three-dimensional interweaved structure. The inset in Fig. 1(b) shows the magnified image of the dense layer on carbon fibers of the CeO 2 /MnO 2 -CFP sample, which demonstrated that brighter CeO 2 nanoparticle clusters were deposited on the MnO 2 layer. This observation was consistent with the CeO 2 deposition process, in which a redox reaction between Ce 3+ ions and MnO 2 happened to oxidize Ce 3+ to Ce 4+ for the creation of CeO 2 nanoparticles [46].  www.springer.com/journal/40145 that Mn, Ce, and O elements were uniformly distributed in the CeO 2 /MnO 2 composite. The composition ratio between CeO 2 and MnO 2 was evaluated further by the ICP-OES, and the result was shown in Table S1 in the ESM. The atomic ratio between Mn and Ce could be calculated at ~7.78 : 1. Figure 1(d) shows the representative high-resolution TEM (HRTEM) image of the CeO 2 /MnO 2 -CFP sample. In the lower part of the image, two sets of lattice planes with d-spacing values of ~0.244 and ~0.475 nm could be clearly observed with the interfacial angle of ~85.73°, which matched well with the (006) and (101) planes of the MnO 2 phase, respectively. In the upper part of the image, parallel lattice planes with d-spacing values of ~0.191 and ~0.271 nm could be clearly observed, which could be assigned to the (220) and (200) planes of the cubic CeO 2 phase, respectively. The HRTEM observation demonstrated that nanosized CeO 2 crystals were decorated on MnO 2 crystals with good contacts between them, which was consistent with the CeO 2 deposition process through the redox reaction between Ce 3+ and MnO 2 and beneficial for charge carriers to transfer between them [46].

2 Chemical composition of the CeO 2 /MnO 2 -CFP sample
The XPS was further used to investigate the surface chemical compositions and element valence states of obtained samples. Figure 2(a) shows the XPS survey spectrum of the CeO 2 /MnO 2 -CFP sample, and that of the MnO 2 -CFP sample could be found in Fig. S1(b) in the ESM. Both samples had XPS signals of C, Mn, and O elements, while additional XPS peaks belonged to Ce element could be clearly observed on the survey spectrum of the CeO 2 /MnO 2 -CFP sample as expected. Thus, it further demonstrated clearly that CeO 2 nanoparticles were successfully decorated on the MnO 2 -CFP sample surface in the second-step in-situ growth of our approach.  Mn 4+ 2p 3/2 , respectively; and the other two fitted peaks at ~653.2 and ~641.9 eV could be assigned for Mn 3+ 2p 1/2 and Mn 3+ 2p 3/2 , respectively. For the MnO 2 -CFP sample, the two fitted peaks at ~654.1 and ~642.6 eV belonged to Mn 4+ 2p 1/2 and Mn 4+ 2p 3/2 , respectively; and the other two fitted peaks at ~653.2 and ~641.9 eV could be assigned for Mn 3+ 2p 1/2 and Mn 3+ 2p 3/2 , respectively [47][48][49][50]. It could be found that the ratio of Mn 4+ /Mn 3+ decreased from 2.46 for the MnO 2 -CFP sample to 1.38 for the CeO 2 /MnO 2 -CFP sample after the decoration of CeO 2 nanoparticles. This observation was consistent with the fact that CeO 2 nanoparticles were grown on the MnO 2 surface through the redox reaction [46], which could form a good contact interface between them to benefit the electron transfer between them.

3 Band structure and photo-generated charge carrier separation behaviors of the CeO 2 /MnO 2 -CFP sample
The diffuse reflectance spectrum measurement was used to reveal the optical properties of the obtained samples. Figures 3(a) and 3(b) show the light absorbance curves of MnO 2 and CeO 2 nanoparticles, respectively, which were approximated by the Kubelka-Munk function from their diffuse reflectance data [55,56]. As expected, MnO 2 nanoparticles demonstrated a wide range of light absorption from the ultraviolet (UV) range into the near-infrared region [57]. The inset in Fig. 3(a) shows the Tauc plots ((F(R)hν) 2 vs. hν) of MnO 2 nanoparticles (direct band gap semiconductor) constructed from their light absorbance data [58,59], from which its band gap value could be determined at ~1.57 eV. For CeO 2 nanoparticles, they demonstrated a light absorption from the UV range just into the visible light region [60]. The inset in Fig. 3(b) shows the Tauc plots ((F(R)hν) 1/2 vs. hν) of CeO 2 nanoparticles (indirect band gap semiconductor) constructed from their light absorbance data [61], from which its band gap value could be determined at ~3.06 eV.  Figure 3(e) shows the band structure of the CeO 2 / MnO 2 heterojunction from the above analysis results, which demonstrated that it had a type II semiconductor heterojunction. Their aligned band structure could facilitate the separation of photo-generated charge carriers, which could minimize the well-known photo-generated charge carrier recombination problem in MnO 2 [63][64][65]. Thus, more photo-generated charge carriers could participate in the photo-assisted charging process, which should ensure the desired photo-assisted charging capability to supercapacitors with MnO 2 -based active materials.  under visible light illumination. The photocurrent densities of the CeO 2 /MnO 2 -CFP sample were obviously higher than those of the MnO 2 -CFP sample, which was consistent with the theoretical prediction that the CeO 2 /MnO 2 heterojunction could facilitate the photogenerated charge carrier separation and transfer as expected from their matching band structure, as shown in Fig. 3(e). Furthermore, another clear difference existed between their photocurrent behaviors. The photocurrent of the CeO 2 /MnO 2 -CFP sample showed a gradual decrease behavior when the light illumination was shut off, while that of the MnO 2 -CFP sample decreased sharply to zero when the light illumination was just shut off. This observation indicated that part of the photo-generated electrons could be effectively retained by the CeO 2 /MnO 2 -CFP sample under visible light illumination, and then slowly released to generate the current in dark when the illumination was shut off, which was consistent with their band structure analysis result.

4 Photo-assisted charging of the CeO 2 /MnO 2 -CFP sample
The electrochemical performances of pure CFP, the MnO 2 -CFP sample, and the CeO 2 /MnO 2 -CFP sample electrodes were evaluated in the traditional threeelectrode system. Figure 4 MnO 2 -CFP electrode was much higher than that of the MnO 2 -CFP electrode. From their CV curves, the overall specific capacitance of the MnO 2 -CFP electrode at the scan rate of 5 mV·s −1 was calculated at ~182 F·g −1 in dark and ~190 F·g −1 under visible light illumination, representing an increase of only ~4%. For the CeO 2 /MnO 2 -CFP electrode, however, its overall specific capacitance at the scan rate of 5 mV·s −1 increased from ~187 F·g −1 in dark largely to ~229 F·g −1 under visible light illumination, representing an significant increase of ~22%. Figure S2 in the ESM further shows their CV curves in dark and under visible light illumination at scan rates of 10 and 20 mV·s −1 . Their specific capacitance and increase percentage data were summarized in Table S2 in the ESM. It was clear that the photo-induced capacitance increase in the CeO 2 /MnO 2 -CFP electrode was from ~22% to ~27% when the scan rates were from 5 to 20 mV·s −1 , while that of the MnO 2 -CFP electrode was only from ~4% to 10%. Figure 4(d) shows the CV curves of the CeO 2 / MnO 2 -CFP electrodes in dark, under visible light illumination, and under visible/infrared illumination (400-2200 nm) at the scan rate of 5 mV·s −1 . It demonstrated clearly that the photo-induced capacitance enhancement of the CeO 2 /MnO 2 -CFP electrode could be affected by the light source. Under visible/infrared illumination, its overall specific capacitance at the scan rate of 5 mV·s −1 could be calculated at ~240 F·g −1 , which represented an increase of 28% compared with that in dark and a further increase of 5% compared with that under visible light illumination. Due to its narrow gap semiconductor nature (E g = ~1.57 eV), part of the infrared illumination could generate electron-hole pairs in MnO 2 for the photo-assisted charging to enhance its capacitance. Thus, the CeO 2 /MnO 2 -CFP electrode could effectively use the abundant solar energy input to provide a green and economic way to enhance its capacitance.

5 Galvanostatic charging and discharging (GCD) behaviors of the CeO 2 /MnO 2 -CFP electrode
The photo-induced capacitance enhancement of the CeO 2 /MnO 2 -CFP electrode was further confirmed by the GCD profiles. Figures 5(a)

6 Stability and retainment of photo-assisted charging effect in dark of the CeO 2 /MnO 2 -CFP electrode
The capacitance stability is critical for the potential application of a supercapacitor. Figure 6(a) shows the cycling performances of the CeO 2 /MnO 2 -CFP electrode in dark and under visible light illumination at the charging and discharging current density of 1 A·g −1 .
The result demonstrated that its intrinsic capacitance in dark and photo-induced capacitance enhancement were both quite stable during the cycling usage. After 100 cycles, no deterioration was observed for both its intrinsic capacitance in dark and its photo-induced capacitance enhancement. This observation indicated that the CeO 2 /MnO 2 -CFP electrode had a strong structure, and the interfaces of CeO 2 /MnO 2 and MnO 2 /CFP were robust. Actually, its specific capacitance showed a slight increase at the beginning of the cycling experiment, which may be attributed to its required electrochemical activation during the initial stage [66,67]. More interestingly, the CeO 2 /MnO 2 -CFP electrode demonstrated a retainment capability of photo-induced capacitance enhancement in dark for an extended period of time after the illumination was switched off. Figure 6(b) shows the CV curves of the CeO 2 /MnO 2 -CFP electrode at the scan rate of 5 mV·s −1 in dark, under visible light illumination, and in dark after the illumination was switched off for 12 h. It demonstrated clearly that the CeO 2 /MnO 2 -CFP electrode could keep part of its photo-induced capacitance enhancement in dark for a quite long period of time. Even after the illumination was switched off for 12 h, its specific capacitance was still ~222 F·g −1 , which was still ~22 F·g −1 higher than that in dark and represented a retainment of its photo-induced capacitance enhancement of ~56%. For comparison, the CV curves of the MnO 2 -CFP electrode at the scan rate of 5 mV·s −1 in dark, under visible light illumination, and in dark after the illumination was switched off for only 0.5 h were shown in Fig. S4 in the ESM. After the illumination was switched off for only 0.5 h, its CV curve already changed back to overlap with its CV curve in dark, which suggested that the MnO 2 -CFP electrode could not retain part of its photo-induced capacitance enhancement after the illumination was switched off. Thus, the comparison results clearly demonstrated that the decoration of a small amount of CeO 2 nanoparticles on the MnO 2 -CFP electrode could also endow the CeO 2 /MnO 2 -CFP electrode a long-lasting capability to retain a large part of its photo-induced capacitance enhancement in dark.

7 Photo-assisted charging mechanism of the CeO 2 / MnO 2 -CFP sample
It is well known that the electrochemical energy storage is dominated by two mechanisms: One is the surface capacitive process, and the other is the surface redox reaction and/or insertion/intercalation-based process [68]. According to the theory by Ardizzone et al. [69] and Baronetto et al. [70], the charge Q has two components, as described in Eq. (1): Q total = Q surface + Q pseudo (1) where Q total is the total charge that can be stored, Q surface is the surface contribution that mainly stems from physical adsorption, and Q p s e u do is the pseudocapacitive contribution. Figure 7(a) shows the  plots of Q total vs. v −1/2 of the CeO 2 /MnO 2 -CFP electrode in dark and under visible light illumination, in which two distinct regions existed with the scan rate of 20 mV·s −1 (7.07 (V·s −1 ) −1/2 ) as the boundary. When the scan rate was over 20 mV·s −1 (v −1/2 was lower than 7.07 (V·s −1 ) −1/2 ), the Q total value decreased linearly with the scan rate increase, indicating a diffusion-controlled energy storage process. Therefore, the intersection of the extrapolated plot with the y-axis of this region could reveal the charge associated with the most accessible area (Q surface ) [71]. Figure 7(a) demonstrates that the capacitance (Q surface ) values of the CeO 2 /MnO 2 -CFP electrode from the most accessible area in dark and under illumination were almost identical at ~7.83 and ~7.47 C·g −1 , respectively, which indicated that visible light illumination had no obvious impact on the adsorption of protons onto the interface between the CeO 2 /MnO 2 -CFP electrode and the electrolyte. For both of them, two distinct parts could be found in their Nyquist plots, including an imperfect half semicircle in the middle-to high-frequency regions, and a sloped straight line in the low-frequency region. The straight line in the low-frequency region had a finite slope, which could represent the diffusive behaviors of the electrolyte in the electrode pores and ions in active materials. For both of them, their slopes of straight lines in the low-frequency region under visible light illumination were higher than those in dark, which suggested that visible light illumination could enhance their electric conductivities due to photo-generated charge carriers. The imperfect half semicircle in the middle-to high-frequency range could be associated with the surface properties of the electrode and corresponded to the charge-transfer resistance (R ct ). Figure S5 in the ESM shows the illustration of the equivalent circuit with a set of resistors and constant phase elements (CPE, denoted as Q) in series and parallel, where R s is the series resistance, Q dl and Q ps are the double-layer capacitance and pseudo-capacitance, respectively, and W is the Warburg diffusion. Table S3 in the ESM summarizes their parameters of different elements in their equivalent circuits. For both of them, their R ct decreased under visible light illumination, which suggested that they had an easier charge transfer in them with illumination [72]. The R ct of the CeO 2 / MnO 2 -CFP electrode decreased from ~2.77 Ω in dark to ~1.62 Ω under visible light illumination, while that of the MnO 2 -CFP electrode decreased from ~8.50 Ω only to ~5.40 Ω. This observation demonstrated that the charge transfer resistance in the CeO 2 /MnO 2 -CFP electrode was much lower than that in the MnO 2 -CFP electrode, which further verified that the decoration of a small amount of CeO 2 nanoparticles on the MnO 2 -CFP sample surface did facilitate the photogenerated charge carrier separation and transfer.
For both samples, their Q dl values were much smaller than their Q ps values, which was consistent with the Trasatti method analysis result and further verified that their capacitive behaviors were dominated by the pseudocapacitive mechanism. The Q ps of the CeO 2 /MnO 2 -CFP electrode in dark was ~2.40×10 −1 F, and it increased to ~2.78×10 −1 F under visible light illumination, which represented an increase of ~16% from the photo-assisted charging effect. For the MnO 2 -CFP electrode, its Q ps increased from ~2.38×10 −1 F in dark to ~2.55×10 −1 F under visible light illumination, which represented an increase of only ~7% from the photo-assisted charging effect. These results were consistent with their CV curve and GCD behavior analysis results, which further confirmed that the decoration of a small amount of CeO 2 nanoparticles on the MnO 2 -CFP electrode largely improved its photo-assisted charging capability.

8 Mechanism of long-lasting photo-induced capacitance enhancement in dark of the CeO 2 /MnO 2 -CFP sample
As shown in Fig. 3( Figure 8 shows the high-resolution XPS scans over Mn 2p peaks of the CeO 2 /MnO 2 -CFP sample under visible light illumination and in dark after the illumination was switched off for 0.5 h, and they were compared with that in dark, as shown in Fig.  2(b). The Mn 4+ :Mn 3+ atomic ratio of the CeO 2 /MnO 2www.springer.com/journal/40145 CFP sample decreased from ~58.0% : 42.0% in dark to ~49.2% : 50.8% under visible light illumination, which clearly verified that the reduction of Mn 4+ to Mn 3+ did happen in the CeO 2 /MnO 2 -CFP sample upon visible light illumination, as its band structure analysis result predicted. After the illumination was switched off for 0.5 h, its Mn 4+ :Mn 3+ atomic ratio only increased to ~51.7% : 48.3%, which suggested that most stored electrons were not released immediately in dark after the illumination was switched off. Thus, the CeO 2 /MnO 2 -CFP sample could have the desired capability to retain a large part of its capacitance enhancement by the photo-assisted charging in dark for an extended period of time. Because the Mn 4+ /Mn 3+ reduction potential (~0.95 V vs. NHE) was more positive than both the one-electron reduction potential of O 2 (−0.05 V vs. NHE) and the two-electron reduction potential of O 2 (0.68 V vs. NHE), those stored photogenerated electrons from the reduction of Mn 4+ to Mn 3+ could not release in dark by reducing O 2 , as happened from the normal photocatalytic memory effect [73][74][75][76]. Thus, their release in dark could be largely slowed down. Even after 12 h in dark, ~56% of its capacitance enhancement from the photo-charging was still retained, which was far higher than that in Ref. [77] on the h-WO 3 /Bi 2 WO 6 material system (16% retainment after only 5 h in dark) from the photocatalytic memory effect. To create photo-generated electron storage by reducing Mn 4+ to Mn 3+ with the proper Mn 4+ /Mn 3+ reduction potential successfully avoided their reaction with O 2 , which largely slowed down their release and was critical for the observed long-lasting photo-assisted capacitance enhancement of the CeO 2 /MnO 2 -CFP electrode in dark.

Conclusions
In summary, a two-step in-situ growth approach was developed to create a composite system of CeO 2 / MnO 2 -CFP, which could serve as a binder-free electrode for supercapacitors. The CFP support could allow the rapid electrolyte diffusion through the hollow/open framework and the fast electron transfer though the carbon skeleton, while the elimination of binders could further enhance its conductivity and stability. The electrochemical energy storage of the CeO 2 /MnO 2 -CFP electrode was found to be dominated by the pseudocapacitive mechanism. The decoration of a small amount of CeO 2 nanoparticles significantly enhances its photo-assisted charging capability, which could be attributed to the formation of a type II CeO 2 /MnO 2 heterojunction to largely enhance the separation and transfer of photo-generated charge carriers for their participation in the photo-assisted charging process. For example, the photo-assisted charging increased its specific capacitance relatively steadily for ~41-49 F·g −1 when the charging and discharging current densities were from 0.25 to 2.5 A·g −1 , while that of the MnO 2 -CFP electrode without CeO 2 decoration was only ~10-20 F·g −1 . Furthermore, the CeO 2 /MnO 2 -CFP electrode possessed a superior retainment effect on its photo-enhanced capacity in dark for an extended period of time, which could be attributed to its slow release of stored photo-generated charges due to the more positive potential of Mn 4+ /Mn 3+ than the one-and two-electron reduction potentials of O 2 . Even after the visible light illumination was shut off for 12 h, it still retained over half of its photo-enhanced capacity. By optimizing the CeO 2 /MnO 2 mass ratio, sizes and shapes of CeO 2 and MnO 2 nanostructures, and the atomic ratio of Mn 3+ /Mn 4+ , the photo-assisted charging capability and its retainment in dark of the CeO 2 /MnO 2 -CFP electrode could be further enhanced. This study provided the principles on the design of supercapacitors with both the photo-assisted charging capability and its retainment in dark for an extended period of time, which could be readily applied on various pseudocapacitive material systems to advance the development of solar energy utilization devices.