Highly flexible binder-free core–shell nanofibrous electrode for lightweight electrochemical energy storage using recycled water bottles

The PET resin for the construction of the PET nanofiber substrate was obtained from the recycled water or other beverage bottles with identification of PETE1 recyclable logos. The solvents, namely trifluoroacetic acid (TFA, ReagentPlus, 99%) and dichloromethane (DCM, anhydrous, ≥99.8%) utilized in the electrospinning process were purchased from Sigma Aldrich and used without further modification. The aniline monomer (ACS reagent, ≥99.5%) was obtained from Sigma Aldrich, and underwent distillation to obtain a golden transparent color before use. Hydrochloric acid (HCl) and ammonium persulfate used in the chemical polymerization of PAni were used without further modifications.

As shown in figures 1 and S1 pictures, the raw PET resin was obtained from various recycled beverage bottles by cutting them into small, thin strips. The PET strips thus obtained were dissolved in a solvent mixture of 1:1 molar ratio TFA:DCM, as commonly used for electrospinning PET. The amount of PET resin dissolved was controlled at 10 wt. %. The solution was then transferred to a sealed glass bottle and magnetically stirred before the electrospinning process. The electrospinning process was carried out using a conventional electrospinning chamber with controlled temperature and relative humidity (R.H.). The chamber was first conditioned to a temperature around 30 °C, with R.H. below 10%. A syringe was filled with the electrospinning solution, then secured onto the syringe pump (New Era Pump Systems, NE 300). The collector was a grade 304 stainless steel sheet. A high voltage DC source was connected to the metallic syringe tip, whilst the fiber collector was grounded. The source DC voltage used for the process was 15 kV, and the distance between the high voltage needle tip and the fiber collector was set to 10 cm. The pumping rate was set to 0.04 ml h⁻¹ and the entire process continued for 30 min to obtain a thin flexible white film. The film was then detached from the collector and placed within the vacuum oven at 60 °C for 12 h.

Zoom In Zoom Out Reset image size

As shown in figure 2, the formation of PAni was carried out via an in situ chemical polymerization process. The oxidant used was ammonium persulfate (APS), and the acidic environment required for protonation was created with 1 M HCl. Upon doping with an acid, PAni emeraldine base can be converted to emeraldine salt, which is electrically conductive and suitable for charge storage applications [30]. A solution with 4 mmol of APS and 20 mL HCl (1 M) was made and the previously obtained PET core nanofiber was placed in the solution. The PET nanofiber core was hydrophobic at first when placed in the solution, and therefore needs to be stirred vigorously for 4 h and ultrasonicated for 10 min before the entire film is submerged under the solution surface. The ultrasonication process was able to increase the surface energy of the PET nanofibers, and is therefore seen to be very important in obtaining uniform core–shell structures throughout the film surface. Another solution with 4 mmol of aniline monomer and 20 mL of distilled water was made and stirred thoroughly to form a white opaque solution. The solution containing aniline was then added dropwise into the oxidant/acid solution. The mixture was then placed in an ice bath at 0 °C for 4 h to allow for polymerization, which can be characterized by a significant color change [31]. After the polymerization process, the PAni@PET film was washed thoroughly with distilled water and ethanol to remove excess monomers and oligomers. The washed PAni@PET fiber was then dried overnight in an oven at 50 °C.

Zoom In Zoom Out Reset image size

Electrochemical characterization was performed using a Solartron 1470E multi-channel potentiostat, with a 2-electrode symmetrical cell setup with 1 cm² pieces of PAni@PET nanofibers serving as the electrodes, an aqueous 1 M sulfuric acid (H₂SO₄) as the electrolyte and thin (0.05 mm) flexible stainless steel sheets as the current collector. CV measurements were performed at varying scan rates, ranging from 5 mV s⁻¹ to 2000 mV s⁻¹, with a voltage window from −0.2 V to 1.0 V, which was selected with respect to the optimum operating condition for PAni-based electrodes. Electrochemical impedance spectroscopy (EIS) was performed from a high frequency of 10⁵ Hz to a low frequency of 0.02 Hz. A galvanostatic charge discharge (GCD) test was performed at a current density of 12 A g⁻¹ and 1.2 A g⁻¹, with the voltage cycling between 0 V and 1.0 V. The specific capacitance was calculated based on the weight of the entire PAni@PET fiber electrode using equation (1). The electrochemical measurements were averaged over five sets of samples and the repeatability has been verified.

$$\begin{eqnarray}&&{C}_{s}=\displaystyle \frac{1}{Sw({\rm{\Delta }}V)}\displaystyle {\int }_{Vi}^{Vf}iV{\rm{d}}V\end{eqnarray} \tag{ 1 }$$

Where $S$ is the scan rate in (V s^–1); $w$ is the mass of the active material in (g); ${\rm{\Delta }}V$ is the potential window in (V) (V_i to V_f), $i$ is the current measured in (A), and $V$ is the voltage applied in (V).

The morphology of the electrode surface structures was analyzed using a Hitachi S-5200 field emission scanning electron microscope (SEM). The core–shell structure was further verified with Hitachi H-7000 transmission electron microscope (TEM). The water contact angle test was carried out with standard sessile drop technique with drops of 20 uL of deionized water on the surface, measured at 25 °C. Brunauer–Emmett–Teller (BET) specific surface area analysis was performed by Quantachrome Instruments Nova 1200e, with CO₂ as the adsorbate gas and a degassing temperature of 200 °C. Fourier transform infrared spectroscopy (FTIR) was performed with a Bruker ALPHA FT-IR Spectrometer, averaged over 25 scans. Thermal gravimetric analysis (TGA) was carried out with TA Instrument Model Q50 at a temperature scan rate of 20 °C min^–1, from 20 °C to 900 °C, in order to verify the composition of the PAni@PET fiber structures. Roughness and surface characteristics were measured via the usage of an atomic force microscope (AFM), namely a Bruker MultiMode 8 with PeakForce tapping mode.

The PET nanofibers were obtained via conventional electrospinning technique, without the aid of rotating collectors for alignment. Alignment of the fibers would disrupt the proper size distribution of the pores that allows the electrolytic ions to travel freely for interfacial charge storage interactions with the electrode. Previous studies noted that mesoporous structures with pore diameters in the range of 2 nm to 50 nm are the most appropriate for supercapacitor electrode applications [10]. However, it was found in our study that a blend of macro-, micro-, and meso-porous structures allowed for more efficient charge transfer at the electrode/electrolyte interface, hence better performance at fast scanning rates, even up to 500 mV s⁻¹. This distribution of various types of pores gave rise to more accessible surface area, and therefore, it was deduced that randomly distributed fiber structures can sufficiently meet the objectives of maximizing accessible surface areas for electrolytic ions. Via the conventional electrospinning technique, PET core fibers with a diameter of 121 nm ± 39 nm were produced. The overlapping or crisscross nature of the PET fibers in the mat depth allowed for formation of high-surface-area 3D structures that effectively increased the available surface for ion access, without hindering the ion transport, as the mix of macro-, micro-, and meso- pores allowed for highly efficient transport routes to form.

As shown in table 1, the PET electrospun core mat thickness is a function of time and therefore, the time allowed for electrospinning process was carefully controlled to obtain film with ideal thickness and flexibility. The observed density was very consistent at around 0.71 g cm⁻³, and the porosity was calculated to be around 50%. The availability of porous structures allows the efficient insertion of oxidant, acid, and the monomer for the in situ polymerization to take place. The tradeoff here is that the formation of thicker films can increase the three-dimensional surface area, allowing higher per-area capacitance values, but the material flexibility can be jeopardized. Also, as was observed, thicker PET substrates can impede the in situ chemical polymerization process, causing uneven PAni attachment, while hardening the film, making it unsuitable for potential flexible applications.

As shown in figure 3(a), by only using a flat, solid, PET film as the substrate, the aniline monomer was unable to reach the surface pores of the PET material and can only form large PAni agglomerates not well-adhered on the PET surface, indicating a weak interaction between PAni and PET. The PET solid flat film was formed via a solvent casting technique with the same recycled PET resin and treatment as for the PET core fibers. Figures 3(b) and (c) showed the morphology of electrospun pure PET nanofiber core. BET surface area analysis with CO₂ adsorbate gas showed a specific surface area of 73.12 m² g⁻¹ for the pure PET electrospun fibers. The increased surface area was created via the interlacing fibrous three-dimensional structure and the interconnected nature throughout the depth fiber membrane. This non-woven textile structure offered ideal pore sizes from the crisscross patterns formed by the randomly distributed fibers, which allowed the oxidant and aniline monomer solutions to polymerize at the liquid−solid interface, which led to the appropriate PAni@PET core–shell formation. The electrospinning parameters were tuned in order to avoid the formation of beads and other non-uniformities. Figures 3(d) and (e) showed the very evenly distributed layer of PAni shell formed via the in situ chemical polymerization method as described.

Zoom In Zoom Out Reset image size

As shown in figure 4(a), the average diameter of these fibers is measured to be 121 nm ± 39 nm, with fiber diameters ranging from 24.2 nm to 301.5 nm. The average thickness of the PAni shell layer was measured to be 69 nm ± 22 nm and gives an overall core–shell structure thickness of around 260 nm, which was also presented in figure 4(a)). Via the thermogravimetric analysis (TGA) measurements, shown in S2, the PAni shell accounted for 22.7% of the overall mass of the core–shell structure. Figures 4(b) and (c) also showed the sample and the SEM image of the pure PET electrospun mat. The low weight and high flexibility were some of the desired features of the substrate material. After the core–shell structure formation, as demonstrated by the bending tests and SEM images shown in figure 4(d) as well as figures S3 and S4, the PET backbone continues to provide flexible structural support and good surface adhesion properties for the active conductive polymer layer to better interact with the electrolytic ions. Figure 4(e) showed the uniformity and the complete coverage of the coating layer on the PET fiber substrate in the PAni@PET core–shell composite. The specific surface area increased slightly to 83.72 m² g⁻¹ from the pure PET fiber, indicating formation of additional micro-porous structures of less than 2 nm in diameters on the PAni shell surfaces.

Zoom In Zoom Out Reset image size

Additionally, TEM was used to further verify the core–shell structure formation of the PAni@PET fiber. As shown in figure 5, core–shell fibrous structures of PAni@PET samples with varying dimensions were scanned and the core–shell formation was clearly shown.

Zoom In Zoom Out Reset image size

Fourier transform infrared spectroscopy (FTIR) was performed on both the pure PET electrospun film, as well as the PAni@PET core–shell fiber structures, as shown in figure S5 and table S1. The results clearly indicated the difference in the IR spectra between the pure PET fibers and the PAni@PET fibers, with the additional characteristic peaks of PAni shown in the PAni@PET fiber composite materials, confirming the composition. The flat surface of the flexible electrode was also ensured in fabricating the electrodes. The PAni@PET composite fiber mat was firmly adhered onto the stainless current collector using carbon paste. As shown in the figure S6, the roughness factor R_a of the surface was measured using AFM technique to be 37.7 nm, similar to that for pure PET fibers at 30.5 nm.

One of the key features of the PAni@PET core–shell fiber structure is its high surface energy. The high surface energy allows better ionic interaction with the typical aqueous electrolyte used for electrochemical capacitors. The pure PET fiber was shown to have a water contact angle of 131.66°, while the PAni@PET core–shell structure measured a water contact angle of 8.36°, as shown in figures 6(a) and (b). It has been reported previously that flat solid pure PET surface has a water contact angle of 72.4°, indicative of partial wetting. With the electrospinning treatment, it was found that the modified PET surface led to a significantly decreased surface energy, indicated by the increase in water contact angle. The decreased surface energy can be attributed to the formation of overlaying nanofibers, which affected the surface interfacial interactions. After the formation of the PAni shell layers on the PET nanofiber surface, the almost complete wetting indicated that the PAni shell was wrapped uniformly throughout the PET electrospun nanofiber surfaces. This can be readily indicated by the hydrophilicity of the PAni@PET fibrous core–shell structure in comparison to the hydrophobic pure PET nanofiber.

Zoom In Zoom Out Reset image size

As demonstrated with the schematics in figures 6(c) and (d), the high surface energy allowed the aqueous electrolyte system to readily penetrate into the available electrode surfaces, which led to better charge storage behaviors and lower charge transfer resistance.

The viability of the PAni@PET core–shell nanofiber structure to be employed in energy storage system has been examined using a two-electrode symmetrical electrochemical cell setup, which was selected to present a better estimation of the actual capacitance offered by the electrode. The specific capacitance calculations from the three-electrode setup quadruples the values shown here and may overestimate the EC performance in some cases. The CV scan rates were varied between 5 mV s⁻¹ to 2000 mV s⁻¹, and the results are shown in figures 7(a) and (b). It was observed that even at very high scanning rates at 2000 mV s⁻¹, the PAni@PET nanofibers still experienced capacitor behaviors indicating that the electrode three-dimensional matrix allowed electrolytic ions to travel freely through the thickness of the electrode material and interact with the PAni shell effectively.

Zoom In Zoom Out Reset image size

When examining the range of CV curves with slower scan rates from 5 mV s⁻¹ to 100 mV s⁻¹, the REDOX peaks were apparent with the current peaks in the charge/discharge processes. The different peaks at specific voltages in the CV diagrams can be attributed to the various oxidation states of PAni. The uniformity of the CV curve shapes indicated that by varying the scan rates between 5 mV s⁻¹ and 100 mV s⁻¹, the charge transfer process taking place at the porous electrode surface was not greatly affected. The specific capacitance at 10 mV s⁻¹ as calculated from equation (1) was 347 F g⁻¹, which presented a dramatic increase from the pristine PAni powder electrodes with carbon binders. The efficient contact between the electrode and the electrolytic ions was also demonstrated by the relatively small decrease in specific capacitance values as the scan rates were increased, as shown in figure 7(c). Even at a very high scan rate of 500 mV s⁻¹ for conductive polymers, the PAni@PET nanofiber electrode still retained a specific capacitance of 244.9 F g⁻¹, which still represented a relatively high capacitance for electrolytic capacitors. This observation can be attributed to the optimally sized porous structures that allowed high surfaces of the active PAni shell layer to be utilized for charge storage without hindering ion transfer.

As show in figure 8(a), the specific capacitance of binder-free PAni@PET nanofibers measured at 10 mV s⁻¹ showed a significant increase in comparison to pristine PAni powder electrodes with carbon ink binders. From the charge-discharge curves, it is also evident that not only the specific capacitance increased with the core–shell PAni@PET structures, the IR drop was also much less. The observed IR drop in PAni@PET electrodes has been significantly reduced, which can be attributed to the internal resistance of the PAni@PET structure and the contact resistance from the electrolyte/electrode interface. The carbon binder clearly introduced higher internal resistance at the electrode/current collector interface and the higher thickness of the pristine PAni electrodes also contributed to the higher IR drops. The discharging time for PAni@PET nanofiber in a typical discharge cycle was 147.4 s, which showed a higher energy density in comparison to previously reported values. With a 1.2 V applied voltage window, it can be calculated that PAni@PET nanofiber electrodes offered an energy density of 69.4 W h kg⁻¹, which is comparable to a typical NiCad battery. Researchers have reported that by utilizing PAni with other substrates, such as reduced graphene oxide and self-supporting graphene sheets, the energy density was reported at 25 W h kg⁻¹ and 46 W h kg⁻¹ respectively [18, 32]. From the galvanostatic 4-cycle test, it was observed that the symmetrical shape of the charge/discharge cycles was also very consistent.

Zoom In Zoom Out Reset image size

Figure 9(a) shows the complex plane Nyquist plot of the impedance of the PAni@PET nanofibers. The curve intercepts the real axis at around a 45° angle indicating an efficient porous electrode behavior. The ESR was estimated at 1 kHz to be around 223 mΩ, which demonstrates a very low contact resistance at the electrode/electrolyte interface. The series resistance of around 1.0 Ω was from the aqueous 1 M H₂SO₄ electrolyte system. The almost vertical slope of the Nyquist plot showed ideal capacitor behaviors as the frequency decreased, which verified the hypothesis that the porous structures created with the electrospinning process aided the charge storage efficiency.

Zoom In Zoom Out Reset image size

One of the key benefits delivered by the PAni@PET nanofibers is the cycling capacity. From figure 10(a), it is clearly shown that the CV measurements taken at the 1st and 1500th cycles are similar in both shape and magnitude. Figure 10(b) shows the galvanostatic charge–discharge behavior comparison between the PAni@PET nanofiber behavior at 1st cycle and 1500th cycle. By the 1500th charging cycle, the capacitance had degraded only from 347 F g⁻¹ to 331.9 F g⁻¹, which represents only a 4.4% decrease in charge storage capacities, a significant improvement over the previously studied PAni-based electrode systems. And the discharging time intervals only decreased from 147.4 s to 142.8 s, which is also an insignificant drop in capacitance. Therefore, it can be concluded that the PAni active surface was well attached on the PET substrate surface, and suffered almost no apparent loss of capacitance after 1500 cycles, which is already a high cycle number for conductive polymer based pseudocapacitor devices.

Zoom In Zoom Out Reset image size

After the cycling performance of PAni@PET fiber electrodes was evaluated, it was noted that the bending performance for this type of flexible electrode is also extremely important. Bending tests were carried out with constructed cells as discussed. Bending was applied to a 90° angle and for 100 cycles, with CV tests carried out before and after the test, as demonstrated in figure 11. From the CV results shown, it was demonstrated that the charge retention capability was not greatly affected after 100 bending cycles. The specific capacitance was 89.6% of the original value measured without bending. The two-electrode cell showed good flexibility and very limited degradation in cell performance after 100 bending cycles. This demonstrated that the performance can be ensured even for very demanding applications where repetitive bending is required.

Zoom In Zoom Out Reset image size