1 Introduction

Carbon aerogels (CAs) are novel porous carbon materials that feature high specific surface area, large mesopore volumes, low density, good electrical conductivity and high chemical stability [1]. CAs were first synthesized by Pekala in 1989 [2]. This synthesis involved sol–gel polycondensation of resorcinol with formaldehyde, a solvent exchange, CO2 supercritical drying, and carbonization under a nitrogen atmosphere. CAs are expected various applications such as adsorbents, catalyst supports, electrode material for double layer capacitors, energy storage devices, column packing materials for chromatography, etc. [3–6]. Given the high costs of both resorcinol and the supercritical drying process involved in carbon aerogel synthesis, the use of carbon aerogel becomes limited. Efforts have recently been devoted to reducing the aforementioned costs by either using cheaper precursors or replacing the supercritical drying step.

Various hydroxybenzols with similar structure such as m-cresol, cresol mixtures or phenol, were used to partially or completely replace resorcinol [7–16]. Among the candidates available, phenol is the most abundant and inexpensive precursor, but it was thought that pure phenol–formaldehyde mixtures would not be suitable for the synthesis of carbon gel precursors [17]. One reason for this is that the reactivity of phenol is about 10–15 times lower than that of resorcinol. Another reason is that, although phenol has a hydrophilic hydroxyl group in its structure, it is difficult to obtain hydrophilic colloidal particles from it. If such particles can not be obtained during gel synthesis, phase separation will occur, resulting in the formation of precipitate rather than a carbon gel. Despite these limitations, research shows that phenol–formaldehyde (PF) carbon aerogel can be synthesized. Mukai [12] obtained PF carbon cryogels by tert-butanol freeze-drying with Na2CO3 as a catalyst. On the other hand, Wu [13] prepared PF carbon aerogel by ethanol supercritical drying with NaOH as a catalyst. Jirglová [14] synthesized CAs and xerogels from phloroglucinol and PF mixtures, using Na2CO3 as a catalyst.

Another approach is to avoid supercritical drying and use ambient pressure drying. These researches of the organic gels dried at ambient pressure focus on resorcinol–formaldehyde (RF) gels with or without prior solvent exchange [18–23]. PF organic gels during ambient pressure drying were reported only by Scherdel [15, 16]. PF carbon xerogels can be obtained by ambient pressure drying with Na2CO3 or NaOH as catalysts, however, many samples cracked during drying.

In the present study, graphene-cross-linked PF carbon xerogel (GCPFCX) was prepared. Graphene oxide (GO) was added as a cross-linking agent in the PF sol–gel process. After gelation, the wet gel was converted into an organic xerogel by ambient pressure drying. Organic xerogel carbonization and GO reduction were performed by high-temperature pyrolysis, and GCPFCX was eventually obtained. The microstructure and morphology of the resultant organic xerogel and carbon xerogels were analyzed by Fourier transform infrared (FTIR) spectroscopy, ultraviolet–visible (UV–vis) spectroscopy, X-ray photoelectron spectroscopy (XPS), scanning electron microscopy (SEM), X-ray diffraction (XRD), and surface area analyzer. The electrochemical behavior of the samples as electrode material in electric double-layer capacitors (EDLCs) was investigated by galvanostatic charge/discharge measurement.

2 Experimental

2.1 GO preparation

Graphene oxide was synthesized by a modified Hummers method. Graphite powder (6 g) was mixed with H2SO4 (25 mL), K2S2O8 (5 g), and P2O5 (5 g). The mixture was stored at 80 °C for 4.5 h and then distilled water was added for dilution followed by filtration, washing, and drying. H2SO4 and KMnO4 were slowly added in the pretreated graphite in an ice bath. The mixture reacted at 35 °C for 2 h. Distilled water was subsequently added for hydrolysis. After a 2 h reaction, 30 % H2O2 was added to the mixture. The mixture was allowed to stand for at least 12 h and then the clear supernate was decanted. The remaining precipitate was washed with a 5 % HCl solution followed by washing with distilled water. The final solution was centrifuged and then ultrasonicated for 4 h. GO concentration was calculated as 2.4 %.

2.2 Preparation of organic xerogel

Phenol (P), formaldehyde (F), and catalyst KOH (C) in specific ratios were dissolved in water at room temperature. P/F and P/C had molar ratios of 1/3 and 5/1, respectively. GO was then added to the PF solution; the mass ratios of GO to GO and P were 5, 7.5, 10, and 15 %. The mixture was stirred and placed in a water bath (65 °C) for 2 h. The solution was poured into a glass vial (8 cm length × 2 cm internal diameter) and cured in a water bath at 85 °C for 3 days. After gelation, the gel was removed from the vial and dried in an oven at 80 °C to obtain a GO cross-linked PF organic xerogel (GOCPFOX). The as-produced GOCPFOX were identified as GOCPFOX-x, where x represented the mass ratio of GO.

2.3 Carbon xerogel preparation

The GOCPFOX was carbonized at 800 °C for 2 h under a nitrogen atmosphere to obtain GCPFCX. Similarly, these GCPFCX were denoted as GCPFCX-x. For comparison, the PF organic xerogel (PFOX) and the carbon xerogel (PFCX) were prepared under the same conditions without GO.

2.4 Characterization

Fourier transform infrared measurement was made on a PerkinElmer Spectrum RX1 Spectrometer. The UV–vis absorption spectra were recorded using Pgeneral TU-1810 UV–vis spectrophotometer in the wavelength range of 200–1,100 nm at room temperature. XPS analysis was carried out on a ThermoFisher, ESCALab 250 spectrometer using a monochromatized Al Kα X-ray source. The surface morphology was characterized using a Hitachi S-4700 SEM with an acceleration voltage of 20 kV. XRD patterns were obtained using a D/Max 2500 diffractometer (40 kV and 200 mA) using Cu Kα radiation, performed in the 2θ range of 5°–90° at a scan rate of 10° min−1. Surface area and pore volume were determined by a Sorptomatic 1990 with samples heated to 200 °C under vacuum (10−5 Torr) for 2 h to remove all adsorbed species. The nitrogen adsorption data were taken at 77 K to calculate the surface area using the Brunauer–Emmett–Teller (BET) model. The carbon electrodes were prepared by mixing the samples with acetylene carbon black and polytetrafluoroethylene with a weight ratio of 75:20:5. The samples were pressed onto the nickel foam with a radius of approximately 5.6 mm. Galvanostatic charge/discharge experiments were conducted in 30 % KOH aqueous solution at room temperature using an Arbin battery tester (BT-4, USA) in a two-electrode system which can measure their characterizations more accurately [24].

3 Results and discussion

Figure 1 shows the comparison between the morphology of the PFOX and that of all other GOCPFOX samples after ambient pressure drying. The PFOX sample was cracked. The surface tension of the organic gel formed a liquid–vapor meniscus during ambient pressure drying, which increased the capillary pressure on the pore walls of the gel. With a weaker network strength, the PFOX sample collapsed under the capillary pressure, whereas the GOCPFOX-5 sample was only partially cracked. The samples retained their original shapes when the GO mass ratio was increased to 7.5 %. The GOCPFOX-10 sample exhibited 15 % drying shrinkage, whereas the GOCPFOX-15 sample exhibited only 3 % shrinkage during ambient pressure drying. The bulk densities of GOCPFOX-7.5, GOCPFOX-10 and GOCPFOX-15 were 0.239, 0.187 and 0.170 g/cm3, respectively. The mechanically strong GO in the organic gel increased the gel skeleton strength. Thus, the samples withstood the capillary pressure caused by ambient pressure drying.

Fig. 1
figure 1

Comparison of the samples in appearance. a PFOX, b GOCPFOX-5, c GOCPFOX-7.5, d GOCPFOX-10, e GOCPFOX-15

Full size image

Figure 2 shows the FTIR spectra of the samples. The FTIR spectrum of GO shows the broad O–H stretching peak at 3,382 cm−1, strong C=O peak at 1,726 cm−1, O–H deformation peaks at 1,620 and 1,412 cm−1, C–OH stretching peak at 1,225 cm−1, and C–O stretching peak at 1,051 cm−1 [25]. The peaks indicate the presence of hydroxyl, carboxylic acid, and epoxy groups on the surface of the GO sheets. For the FTIR spectrum of the sample PFOX, the broad band at 3,450 cm−1 was assigned to the OH stretching vibration. The 2,915 cm−1 absorption band belongs to the methylene (CH2) and methyl (CH3) stretching vibrations. The band at 1,607 cm−1 was assigned to the stretching vibration of aromatic rings. Finally, the 1,475 cm−1 band was assigned to the CH2 scissor vibration and the CH3 flexural vibration. Methylene–ether (CH2–O–CH2) was observed at 1,053 and 1,210 cm−1 [26]. The CH2–O–CH2 bridge is a cross-linking bond formed between aromatic rings by polycondensation [9]. All PFOX characteristic bands were observed in the GCPFOX samples. However, the absorption bands at 1,726 and 1,412 cm−1 corresponded to the C=O stretching vibration of the COOH groups, and the O–H bending vibrations were absent in all GCPFOX samples. The carboxylic acid group on the GO surface probably became carboxylate, and hydroxyl groups reacted with formaldehyde when GO was added into the KOH-catalyzed PF solution. Moreover, the absorption band of PFOX at 1,607 cm−1 that corresponded to the stretching vibration of the aromatic rings shifted to 1,600 cm−1 in GOCPFOX probably because of the π–π interaction between the sp2 structure of GO and the aromatic structure of PF. These results suggest that the surface functional groups of GO interacted with PF during sol–gel polymerization.

Fig. 2
figure 2

The FTIR spectra of the samples

Full size image

Figure 3 reveals the UV–vis spectra of the GO, GOCPFOX-7.5, and GOCPFOX-10 samples. In the typical GO spectrum, the π–π transitions of aromatic C–C bonds and n–π transitions of C=O bonds were obtained at a maximum of 240 nm and a tiny shoulder at 300 nm, respectively [27]. For the GOCPFOX-7.5 and GOCPFOX-10 samples, the two peaks moved to 220 and 290 nm because of the π–π interaction between the sp2 structure of GO and the aromatic structure of PF. The result is consistent with that of the FTIR analysis.

Fig. 3
figure 3

The UV–vis spectra of the samples

Full size image

Figure 4 shows the XPS peaks of the PFOX, GO, and GOCPFOX-10 samples. The XPS peak of the C 1s of PFOX consisted of two peaks with binding energies of 284.8 eV (C–C) and 286.5 eV (C–O). The C 1s peak region of GO was further decomposed into four peaks with binding energies of 284.5 eV (C–C), 286.8 eV (C–O), 287.9 eV (C=O), and 288.1 eV (O–C=O) [28]. When combined with the PF gel in the GOCPFOX-10 sample, the peak that corresponded to the C=O of graphite oxide disappeared, and the C 1s of the O–C=O peak increased to 289.06 eV, which indicated that the carboxylic acid groups of GO reacted to form a carboxylate.

Fig. 4
figure 4

The XPS peaks of the PFOX, GO and GOCPFOX-10 samples. a PFOX, b GO, c GOCPFOX-10

Full size image

Figure 5 shows the XRD patterns of the samples after pyrolysis. The PFCX and all GCPFCX samples show two peaks at approximately 22° and 44°, corresponding to (002) and (101) graphite diffraction peaks, respectively. These two peaks suggest that PFCX and GCPFCX are partially graphitized carbon particles [29]. The intensity of the (002) diffraction peaks of the PFCX and all GCPFCX samples are similar. This peak becomes somewhat sharper; moreover, two small diffraction peaks at approximately 30° and 31° appear with the increase in GO mass ratio. The result indicates that the addition of GO improves the partial graphitization structure. On the other hand, the PFCX exhibits a very sharp (101) diffraction peak caused by the large crystallite diameter from nanoparticle stack after ambient pressure drying. The intensity of the peak gradually decreased with the addition of GO, which is opposite to the (002) peak variation. The (101) peak intensities are the same as those of the CAs prepared by supercritical drying when the GO mass ratio exceeded 7.5 %, which clearly indicates that the GO sheets fully support the PF gel network.

Fig. 5
figure 5

XRD patterns of the samples

Full size image

Figure 6a shows the curved, layer-like shape and smooth surface exhibited by the obtained GO. PFCX demonstrates a blocky morphology, as shown in Fig. 6b. The PF organic gel network collapses under ambient pressure drying, resulting in formation of PF nanoparticle stacks after pyrolysis. This finding is consistent with the aforementioned XRD analysis. Similarly, the GCPFCX-5 in Fig. 6c exhibits a block structure although such structure is rather loose and not as dense as the PFCX structure. By contrast, the GCPFCX-7.5, GCPFCX-10, and GCPFCX-15 samples formed a layered structure (Fig. 6d–f). However, the three samples did not exhibit a three-dimensional interconnected nanoparticle structure but rather, a layered one. The layers also gradually became thinner and smoother with the increase in the GO mass ratio. This phenomenon indicates that the mechanically strong GO supports the organic gel network and the addition of GO changes the gel structure.

Fig. 6
figure 6

SEM images of the GO, PFCX and GCPFCX samples. a GO, b PFCX, c GCPFCX-5, d GCPFCX-7.5, e GCPFCX-10, f GCPFCX-15

Full size image

A possible reaction mechanism as characterized above is shown in Fig. 7. When GO was added into the PF solution, the PF molecules were anchored onto the surface of GO through chemical and physical interactions. The chemical interaction resulted from the chemical bonding formed by the oxygen atoms of the hydroxyl groups on the surface of the GO being linked with the carbon atoms of formaldehyde when GO was added into the KOH-catalyzed PF solution. Meanwhile, the physical interaction included π–π stacking between the GO sheets and the PF chain. At very low GO ratios, the GO sheets were fully embedded inside the PF molecules during the sol–gel process such that the GCPFCX-5 sample showed a block morphology similar to that of the PFCX sample. The PF molecules did not entirely envelop the GO when its mass ratio reached 7.5 %; thus, the PF molecular chains only grew along the planar structure of the GO and covered its surface. The GO served as a template in the formation of a layered structure. The higher the GO ratio, the more numerous were the reactive sites and the fewer PF molecules were covered on the GO surface. The PF gels only formed very thin interconnected nanoparticles on the GO surface; thus, the GCPFCX-15 sample exhibited a thinner and smoother layered structure.

Fig. 7
figure 7

Schematic of possible reaction mechanism and structure. a possible reaction mechanism, b different structure of GOCPFOX samples

Full size image

Figure 8 shows the nitrogen adsorption–desorption isotherms of the PFCX sample and the GCPFCX sample. The data are listed in Table 1. The PFCX sample contained several pores, indicating the collapse of the whole PF gel network during ambient pressure drying. When the GO mass ratio was increased from 5 to 10 %, the specific surface increased from 106 to 378 m2/g, and the mesopore volume also increased. The GCPFCX-10 had a mesopore volume of 0.56 cm3/g. The GCPFCX-10 sample demonstrated the presence of mesopores and macropores. These macropores possibly formed between the randomly packed sheets that originated from GO. However, for carbon xerogel derived from ambient pressure drying without solvent exchange, the mesopore volume achieved is still noteworthy. GO supported the PF gel, and increasing GO ratio gradually improved the structural strength. The samples withstood the capillary pressure, preventing network collapse during ambient pressure drying. Thus, the specific surface area and the mesopore volume of the samples after pyrolysis increased when the GO mass ratio was increased from 5 to 10 %. However, the specific surface area of GCPFCX-15 was 243 m2/g. Compared with the GCPFCX-10 sample, the specific surface area and the mesopore volume of the GCPFCX-15 sample decreased because the higher GO mass ratio resulted in fewer PF molecules that grow and cover the GO surface. The specific surface area and the mesopore volume decreased because the thin PF gel did not effectively form interconnected nanoparticles.

Fig. 8
figure 8

The nitrogen adsorption–desorption isotherms of the samples

Full size image
Table 1 Surface area and porosity analysis of samples
Full size table

Figure 9 shows the galvanostatic charge/discharge curves of the PFCX and the GCPFCX samples at a current of 5 mA with potentials ranging from 0 to 1.0 V. The capacitors exhibited a capacitive behavior with almost symmetric charge/discharge curves. The specific capacitance Cs of the electrode is calculated from the discharge curve according to the following equation:

$$ {\text{C}}_{\text{s}} = {\text{I}}_{\text{t}} /{\text{V}}_{\text{m}} , $$
(1)

where I is the discharge current (A), t is the discharge time (s), V is the potential change during the discharge process, and m is the active mass of electrodes (g). The calculated specific capacitance of the PFCX and the GCPFCX samples are listed in Table 1. The specific capacitances of the GCPFCX samples remarkably increased with the addition of graphene. The GCPFCX-10 sample exhibited the highest specific capacity at 116 F/g, which is nearly twice that of PFCX. Conversely, the specific capacity of GCPFCX-15 decreased, although it was still higher than that of PFCX. The variation in specific capacity is consistent with that of BET. The improved electrochemical performance of the GCPFCX samples may be ascribed to its unique morphology. A porous structure with a high specific surface area is first formed, which significantly facilitates the rapid transport of the electrolytes in the electrode during the charge/discharge processes. The graphene sheets dispersed in the GCPFCX samples then formed a conducting network and provided a highly conductive path for electron transport during the charge/discharge processes. Combining these two factors, the ionic diffuse resistance and the charge-transfer resistance is reduced, and the specific capacitance is substantially increased. Hence, the highly dispersed graphene sheets in the GCPFCX samples improved the electrochemical performance. GCPFCX is a potential material for EDLCs because of their low cost and ability to avoid supercritical drying despite values lower than those of carbon aerogel electrodes reported in the literature.

Fig. 9
figure 9

The galvanostatic charge/discharge curves of the samples

Full size image

4 Conclusion

The PFOX sample cracked under ambient pressure drying, whereas the GOCPFOX-7.5 sample retained its original shape. When the GO mass ratio was increased, the GOCPFOX-10 and GOCPFOX-15 samples exhibited low drying shrinkage because GO improved the structural strength of the gel.

The π–π interaction between GO and PF in the GOCPFOX samples was observed using their FTIR and UV–vis spectra. The FTIR spectra showed that the hydroxyl groups on the GO surface reacted with formaldehyde. Furthermore, FTIR and XPS proved that the carboxyl acid groups on the GO surface formed a carboxylate. The XRD patterns indicated that the addition of GO improved the partial graphitization structure.

On the basis of the SEM micrographs, PFCX and GCPFCX-5 exhibited a block structure, whereas the GCPFCX-7.5, GCPFCX-10, and GCPFCX-15 samples demonstrated layered structures, indicating changes in the gel structure with the addition of GO. When GO was added into the PF solution, the PF molecular chains were anchored onto the surface of GO by chemical and physical interactions.

The specific surface area and the mesopore volume of the GCPFCX samples increased and then decreased with the increase in the GO mass ratio. The GCPFCX-10 sample reached the highest value when the specific surface area was 378 m2/g and the mesopore volume was 0.56 cm3/g. The variations in specific capacity were consistent with that of BET. Moreover, the specific capacity of the GCPFCX-10 sample, which may be ascribed to its unique morphology, was the highest among the tested materials at nearly twice that of PFCX. The potential of GCPFCX as material for EDLCs is attributed to its low cost and ability to avoid supercritical drying despite lower values compared with those of the carbon aerogel electrodes reported in the literature.