Interface-Confined High Crystalline Growth of Semiconducting Polymers at Graphene Fibers for High-Performance Wearable Supercapacitors Suchithra Padmajan Sasikala, Kyung Eun Lee, Joonwon Lim, Ho Jin Lee, Sung Hwan Koo, In Ho Kim, Hong Ju Jung, and Sang Ouk Kim* National Creative Research Initiative Centre for Multi-Dimensional Directed Nanoscale Assembly, Department of Materials Science & Engineering, KAIST, Daejeon 34141, Republic of Korea S Supporting Information *
ABSTRACT: We report graphene@polymer core−shell fibers (G@PFs) composed of N and Cu codoped porous graphene fiber cores uniformly coated with semiconducting polymer shell layers with superb electrochemical characteristics. Aqueous/organic interface-confined polymerization method produced robust highly crystalline uniform semiconducting polymer shells with high electrical conductivity and redox activity. When the resultant core−shell fibers are utilized for fiber supercapacitor application, high areal/volume capacitance and energy densities are attained along with long-term cycle stability. Desirable combination of mechanical flexibility, electrochemical properties, and facile process scalability makes our G@PFs particularly promising for portable and wearable electronics. KEYWORDS: graphene, fiber, supercapacitor, conducting polymer, interfacial polymerization solution-phase mixing of GOLC with polymers.8−11,19 Either direct wet-spinning from this solution mixture and subsequent chemical reduction or hydrothermal treatment of this solution mixture in closed tube vessel produced fibers composed of reduced GO sheets (rGO) and organic semiconductors. In these approaches, nonetheless, inhomogeneous physical mixing of GO and polymer, inefficient reduction of GO, and low crystallinity of polymers encapsulated by dielectric surfactants (e.g., PEDOT−PSS) commonly led to the final fiber properties far lesser than the expected performance by such a combination. Moreover, how to avoid the chemical damage of organic semiconductors during the harsh chemical/thermal reduction conditions and how to ensure the overall process scalability remain formidable challenges.8−11,19 The limited porosity and low surface area of these fibers also lead to the inherently limited electrochemical performance. In this work, we report idealized graphene/organic semiconducting polymer hybrid core/shell fibers (G@PFs) with high electrical conductivity and good electrochemical perform-
F
iber supercapacitors hold great promise for wearable energy devices owing to their inherent one-dimensional form factor that can effectively adapt to complex deformation modes where the rigid bulk or planar counterparts may fail.1 Graphene fiber (GF) is a promising component for such purposes along with the synergistic ideal combination of exceptional graphene properties such as large surface area, high electrical and thermal conductivity, low density, and high mechanical stability.2−5 While pristine GFs are electrically conductive without high electrocatalytic activity, redox capability can be imparted at GF surfaces by incorporating pseudocapacitive inorganic materials or semiconducting organic polymers.6−11 In particular, organic semiconducting polymers with inherent low density, mechanical flexibility, and high electronic and optoelectronic properties are anticipated to offer versatile benefits for flexible electronics when combined with GFs.12−14 GFs can be obtained by facile wet-spinning from graphene oxide liquid crystal (GOLC) followed by chemical/thermal reduction.15−18 Notably, incorporation of organic semiconductors into GFs is very difficult due to the unavoidable damage of organic materials during the GO reduction process. Accordingly, there have been only a few attempts for graphene− polymer composite fibers thus far, which principally rely on the © 2017 American Chemical Society
Received: July 17, 2017 Accepted: August 7, 2017 Published: August 7, 2017 9424
DOI: 10.1021/acsnano.7b05029 ACS Nano 2017, 11, 9424−9434
Article
www.acsnano.org
Article
ACS Nano
Figure 1. Schematic illustration displaying the fabrication steps for G@PFs.
GFs obtained from Ca(II) coagulating ion and HI and hydrazine reduction (both followed by thermal treatment) show the conductivity of 98.7 and 15.7 S/cm, respectively. Apart from the improved electrical conductivity, the emerging 2D peak and increased ID/IG value compared to that of the GOF in the Raman spectrum further validated the reduction process (Figure S4a−d). The increase of ID/IG is usually explained as an increase in the number of sp2 domains upon reduction.24 Whereas NH3 in the reduction atmosphere induced 3.8 atom % of N-doping, the Cu(II) ions integrated into the GO fiber during the wet-spinning process concurrently get doped into GF in H2 atmosphere, which is evident from Xray photoelectron spectroscopy (XPS) that detects 2.1 atom % of Cu at 932.62 eV (Figure S3a−c). Notably, presence of Cu(0) can also be detected in X-ray diffraction (XRD) analysis of GFs originally spun from a higher concentration of Cu(II) coagulating ions (>10 wt %) after the thermal reduction step under H2 (Figure S3d). Nevertheless, a possibility of concurrent doping of adsorbed Cu particles at the graphene crystal lattice cannot be ignored. The Cu and N codoping may be the reason for the higher ID/IG value in the Raman analysis of GFs after thermal reduction (Figure S4c,d).25 The noticeable conductivity difference between fibers made from Cu(II) and Ca(II) as coagulating ions proposes that Cudoping effectively enhances the electrical conductivity. Cu is known to offer an n-type doping and causes a shift in the Fermi level of graphene from the Dirac point to the conduction band.26 Importantly, the Cu- and N-doping increase the electrolyte wettability of graphene sheets. As a consequence, the electrochemical surface area and electric double-layer capacitance are greatly enhanced. In addition, these dopant sites provide pseudocapacitance, and thus the electrochemical performance is effectively improved for supercapacitor application. Note that the electrical conductivity of our GFs is comparable to that of the previously reported GFs (Table S2) but inferior to that of the high temperature (3000 °C)treated GFs and fibers consisting of compactly stacked giant graphene sheets. Such a compact stack of GFs may be detrimental to capacitive performance due to limited electrolyte accessibility inside the fibers. The macroporous structure of our GF is beneficial in this respect. In our further experiments, we have used Cu and N codoped macroporous GFs, obtained after the HI and thermal reduction, as a standard sample.
ance based on a straightforward and reliable synthetic route. Our protocol exploits highly confined aqueous−organic interface restricted polymerization of semiconducting polymers at Cu and N codoped GF surfaces. The heteroatom dopant sites at GFs not only allow the manipulation of physical properties but also facilitate good wettability toward electrolytes.20,21 The GF core has high surface area and contains macropores that act as an electrolyte reservoir for supercapacitor electrode applications. The restricted polymerization of organic semiconductors at the aqueous/organic interface confines a slow polymer chain growth at GF surface and thereby yields crystalline semiconducting polymer shells. These features enable our G@PFs to accomplish high electrical conductivity (387.1 S/cm), specific capacitance (417.9 F/cm3), and energy density (7.0 mWh/cm3) among carbon-based fiber supercapacitor devices ever reported thus far (Table S1).
RESULTS AND DISCUSSION In our synthetic strategy (Figure 1), GO fibers (GOFs) are wetspun from GOLC into a coagulation bath containing 5 wt % Cu(II) ions in an ethanol/water mixture (Figure S1a). The airdried GOF shows a solid ribbon-like morphology with highly aligned wrinkles, which is typical of GOFs (Figure S1b). In order to generate porous heteroatom-doped GFs, a chemothermal reduction strategy is employed. First, we compared the chemical reduction effects from hydroiodic acid (HI) and hydrazine vapor on the GOFs, both of which are then followed by a rapid thermal reduction at 700 °C under H2/NH3 mixture gas (6:4 sccm). The resulting GFs exhibited a rapid increase of fiber diameter. Porous morphology is generated while reductive byproduct gas is released (Figure S1c−g).22,23 The macroporous morphology will be beneficial in fibers intended for supercapacitor electrodes because the pores provide open channels for the infiltration of electrolytes, and thus a whole volume of fibers can be utilized for electrochemical reaction. The surface area is also found to increase from 39.1 to 1174.4 m2/g during this reduction step (more details in S1: Surface area measurement and Figure S2a,b). Notably, the HI-treated fiber shows a conductivity of 148.6 S/cm after thermal treatment, considerably higher than that of the hydrazine-treated counterpart (39.4 S/cm), which indicates more efficient deoxygenation and π-conjugation restoration by HI treatment (Figure S3a−d). Comparative experiments with 9425
DOI: 10.1021/acsnano.7b05029 ACS Nano 2017, 11, 9424−9434
Article
ACS Nano
Figure 2. SEM images showing the cross sections of (a−d) G@PEDOT fiber with increasing magnification displaying graphene−polymer core−shell fiber morphology. Inset of (d) is a photograph of G@PEDOT fiber being threaded into a button.
Figure 3. (a) TEM image of G@PEDOT. (b) TEM image of localized PEDOT platelets as identified by EDS spectra (inset). (c) HRTEM of PEDOT platelet displaying its lattice structure. Inset is the electron diffraction pattern of the same area; (d,f) TEM and (e,g) HRTEM images displaying lattice structure of PPy and PAni, respectively. SAED patterns of PPy and PAni are on the insets of (e) and (g), respectively. The samples are prepared by cryo-cutting the fractured surface of the G@P fiber and depositing it in the holey carbon grid. (h) XPS survey of G@ PEDOT, G@PPy, and G@PAni fibers.
9426
DOI: 10.1021/acsnano.7b05029 ACS Nano 2017, 11, 9424−9434
Article
ACS Nano
Figure 4. Electrical and electrochemical performances of GFs and G@PFs. (a) Photograph displaying 2 cm long G@PEDOT fiber as part of a conductive circuit with 3 V battery to light a green LED. (b) CV curves for G@PEDOT, G@PPy, and G@PAni fiber supercapacitors at 10 mV/s. The volumetric current density is calculated for a single fiber electrode covered in electrolyte. (c) Galvanostatic charge−discharge profiles at ∼0.2 A/cm3 and (d) volumetric capacitances at increasing current densities.
dominantly hydrophobic, we first immerse GFs in iron perchlorate dissolved in dichloromethane (/n-butanol), which is followed by immersing in iron perchlorate dissolved in water. The iron-perchlorate-treated GF surface absorbs a thin layer of water after this step. Subsequent immersing in the organic monomer solution (EDOT, pyrrole, or aniline dissolved in dichloromethane/n-butanol) triggers polymerization. Significantly, this sequential aqueous and oil phase treatment enables highly restricted slow growth of conductive polymers at the GF surface and thus enables crystalline polymer structures. Notably, this simple synthetic route based on the sequential solvent emersion at room temperature ensures the scalability for large-scale production (more details are given in the Experimental Section). For the minimal deterioration of the original macroporous fiber morphology, we carefully optimized our design strategy to attain a thin layer (typical thickness = 2 ± 1 μm), as analyzed from high-resolution scanning electron microscopy (SEM) of polymer shells (Figure 2 and Figure S6). Note that the crystalline polymer shell growth (>62 wt % in the hybrid fiber) leads to a minor decrease of the surface area from 1174.4 m2/g (of GF) down to 1111.8, 1096.1, and 1088.3 m2/g for G@PEDOT, G@PPy, and G@PAni, respectively (Figure S2). The slight decrease in surface area can be attributed to polymer growth on the surface pores of graphene fiber. High-resolution transmission electron microscopy (HRTEM) coupled with energy-dispersive X-ray spectroscopy (EDS) is used to characterize the surface polymer layer (Figure 3). A typical (100) lattice plane of PEDOT (d-spacing = 0.584)
As mentioned above, previous solution mixing methods for polymer−graphene hybrid fiber systems had suffered from inhomogeneous mixing and insulative surfactant coverage (e.g., insulating PSS template for PEDOT). Electrochemical polymerization at the GF surface could be an alternative route but also reveals limitations in terms of process scalability and compositional uniformity along the long fiber structure.27,28 At the surface of our Cu and N codoped GFs, conventional in situ polymerization of semiconducting polymers (routinely practiced for fiber/film synthesis), such as poly[3,4-ethylenedioxythiophene] (PEDOT), polypyrole (PPy), and polyaniline (PAni), results in inhomogeneous and rapid growth of low crystalline polymers (Figure S5a−f).29 Accordingly, the electrical conductivity of the resultant fiber is decreased to G@PPy > G@PAni. The specific capacitance of GF and G@ PFs is calculated from CV and galvanostatic charge−discharge (CCD) profiles (Figure 4b−d) and is listed in Table 1. Note that the capacitance calculation involving a low mass loading as in the case of low density graphene fibers leads to an exceptionally high gravimetric capacitance which is not representative of actual capacity of the material in real practice. Instead, volumetric/areal capacitance values are preferred to mass specific capacitance in order to assess the actual feasibility of fiber supercapacitors.37 The G@PEDOT fiber-based supercapacitor displays an ultrahigh volumetric capacitance (considering the entire device volume, Cs,volume) of 78.8 F/cm3 at a current density of 0.5 A/cm3. The volumetric capacitance of a single fiber electrode (Cf,volume) is measured to be either 263.1 F/cm3 (based on the entire volume of single fiber including the surrounding electrolyte) or 417.9 F/cm3 (based on the volume of single fiber only). The volumetric capacitance of our G@ PEDOT fiber is either comparable to or better than previous literature values for fiber supercapacitors (Table S1). Similarly, G@PPy and G@PAni fiber electrodes also reveal reasonably high Cf,volume (Table 1 and Figure 4d). The Cf,volume of our polymer-free GFs (21.2 F/cm3 with electrolyte and 29.1 F/cm3 without electrolyte) is also comparable to that of the previously reported graphene-only fibers (Table S1). The Cu- and Ndopant sites at GF may incorporate polar character and thereby enhance the wettability with electrolytes. Additionally, pseudocapacitance can be also incorporated by such a
has been resolved for the G@PEDOT shell in a HRTEM and selected area electron diffraction (SAED), verifying a crystalline nature (Figure 3a−c). SAED patterns of G@PPy and G@PAni also exhibit welldefined crystalline patterns of PPy and PAni, respectively (Figure 3d−g). The XPS analysis of the G@PEDOT fiber displayed Cl2P and S2P spectra (more details in Figure S7). The deconvoluted high-resolution S2P spectra displayed a S+2P peak, which represents partially oxidized S, which is expected to be charge balanced by ClO4− ions. The doping level was calculated as 31.7% from the areal ratio of Cl2P to S2P peaks. Such a high doping level is in agreement with the crystallinity and high conductivity behavior of PEDOT.33,34 However, similar calculation of doping level for G@PPy or G@PAni fibers was not possible in this study because the N contribution from N-doped graphene sheets overlaps with the N contribution of PPy and PAni. Nevertheless, EDS (inset in Figure 3b) and XPS survey (Figure 3h) detected an insignificant level of Fe contamination and significant Cl/ ClO4-doping for all samples, which reflects positive charge accommodation in the polymer layers. The doping of the semiconducting polymer layer may additionally improve the charge mobility, which may contribute to the exceptionally high electrical conductivities (387.1, 282.7, and 259.4 S/cm for G@ PEDOT, G@PPy, and G@PAni fibers, respectively).33,34Figure 4a presents the utilization of a G@PEDOT fiber as a circuit wire for light-emitting diodes (LEDs), illustrating the high electrical conductivity, which also makes it possible to use G@ PFs as electrochemical capacitor (EC) electrodes without additional conductive additives. For the electrochemical characterization of G@PF electrodes, microfiber devices are assembled by aligning two identical G@PFs (length ∼ 1.5 cm, diameter ∼ 70−80 μm, mass ∼ 0.144 mg, density ∼ 1.5 g/cm3) in parallel and coating with PVA/H2SO4 gel electrolyte to achieve a symmetrical solid-state supercapacitor structure (Scheme S1). The devices are superthin with a maximum thickness of
