Introduction

Nowadays we can observe increasing demand for high power, capacitance, light weight and long-life energy storage materials. For this purpose, supercapacitors are most promising candidates (Poonam et al. 2019; González et al. 2016) because of their high power density, long cycle life, fast charge/discharge rates and low maintenance cost. Carbon-based materials are often the choice for the capacitor electrodes (Najib et al. 2019; Dubey et al. 2019). One of the carbon nanomaterials with a 2D structure, high and accessible surface area, exceptional intrinsic electrical conductivity, high temperature stability is currently widely studied graphene (Ke et al. 2016; Lemine et al. 2018; Li et al. 2019a, b). Despite advantageous properties of its pristine form, there some drawbacks of unmodified graphene. This include, among other its zero band-gap and difficulty in formation of stable dispersions in organic solvents. However, tailored functionalization of graphene can lead to improvement of its properties, along with gaining new features coming from introduced functionalities.

When considering the functionalization of graphene there are two main approaches to achieve this goal described in the literature. First strategy is based on various non-covalent methods that allow to modify graphene surface (Georgakilas et al. 2016; Zhao et al. 2021). For this purpose, molecules with planar conjugated π-systems are usually chosen due to their van der Waals, electrostatic or π–π stacking interactions with graphene surface. One of the main advantages of this type of graphene derivatives is their non-disrupted, extended π-conjugation of carbon atoms on the graphene surface. Resulting graphene retains its remarkable properties, showing less defective structure and higher conductivity when comparing to less conjugated graphene (Su et al. 2013; Jin et al. 2017). These nanomaterials have found applications in the fields of drug delivery (Liu et al. 2013), energy materials (Cai et al. 2019), solar cells (Mahmoudi et al. 2018), water splitting (Li et al. 2019a, b), biosensing (Pumera 2011; Peña-Bahamonde et al. 2018), environmental (Perreault et al. 2015), catalytic (Haag et al. 2014) and biomedical technologies (Shareena et al. 2018).

On the other hand, numerous synthetic procedures for covalent modification of graphene, which provide more stable arrays comparing to those based on secondary interactions, were reported (Bottari et al. 2017). Those reactions can be divided into two main groups: formation of covalent bond by reactions with the oxygen groups present on the graphene oxide surface or formation of covalent bond by addition to the graphene C=C double bond. The second type of reactions involves mainly organic radicals and dienophiles as reagents, and include cycloaddition reactions, which have been successfully employed for functionalization of graphene surface. One of the examples of such approach is the so-called Bingel reaction, well known from the chemistry of fullerenes, but also already employed in functionalization of carbon nanotubes and graphene (Naebe et al. 2015; Jin et al. 2016).

Similarly to graphene, fullerenes due to their unique physical and chemical properties, remarkable electron-accepting features and solubility in organic solvents, along with various methods of their functionalization are also interesting platform for construction of energy storage materials (Wang et al. 2017; Khan et al. 2019; Bairi et al. 2019). It would be very interesting to create novel hybrid nanomaterial composed of graphene and fullerene, as it is expected that binding of fullerenes to graphene surface should not only increase the capacity of resulting hybrid nanomaterial (Ma et al. 2015; Cerón et al. 2019) but also provide adequate stability and prevent restacking of graphene sheets (Wang et al. 2009). Reports on not only performance comparison for capacitors based on C60 or C70 fullerenes (Tran et al. 2009) but also on energy storage application of C70 and its derivatives are scarce (Bairi et al. 2016).

In this contribution, we report herein, synthesis of graphene functionalized with C60 and C70 fullerenes by novel approach based on Diels–Alder reaction of fullerene core with anthracene derivative. Necessary anthracene moieties were covalently bound onto graphene surface via Bingel reaction using corresponding dianthracene malonate, ensuring high stability of obtained nanomaterial. Functionalized graphene showed fullerene binding capability due to Diels–Alder reaction of anthracene moieties with fullerene cores, both C60 and C70. Obtained graphene-fullerene hybrids were studied using electrochemical methods, and impact of the fullerene core type on energy storage properties was determined. Results obtained using cyclic voltammetry (CV) and galvanostatic charge/discharge (GCD) measurements showed that presence fullerene core influences the energy storage ability and introduction of appropriate fullerene molecules can improve the final capacitance of synthesized nanomaterials.

Experimental

Chemicals

9-Anthracenemethanol, ethyl acetate, n-hexane, 1,2-dichlorobenzene (ODCB), malonyl chloride, potassium hydroxide, toluene and Nafion™ (5wt.%) were bought from Sigma-Aldrich. Carbon tetrabromide, 1,8-diazabicyclo(5.4.0)undec-7-ene (DBU), N-methylpyrrolidone and silica gel (0.060–0.2 mm, 70–230 mesh) were purchased from Alfa Aesar. Acetone, dichloromethane, ethanol, magnesium sulfate, methylene blue, trimethylamine and potassium hydroxide were bought from POCh (Poland). C60 and C70 fullerene were purchased from Nano-C. Toluene was dried and purified before use by refluxing with sodium and benzophenone under argon atmosphere, other solvents were analytical grade reagents and were used as received.

Characterization methods and instrumentation

Mass spectra were recorded using Micromass LCT ESI-TOF mass spectrometer equipped with an orthogonal electrospray ionization source. 1H and 13C nuclear magnetic resonance (NMR) spectra were recorded on Varian Unity Plus 500 MHz spectrometer using CDCl3 as a solvent. The infrared experiments were carried out using the Shimadzu FTIR-8400S. X-ray Photoelectron Spectroscopy (XPS) measurements were carried out using a VG ESCALAB 210 electron spectrometer equipped with an Al Kα source (1486.6 eV). XPS data were calibrated using the binding energy of C1s = 284.6 eV as the internal standard. UV–Vis spectra were recorded using Varian Cary 50 UV–Vis spectrophotometer. Scanning electron microscopy analysis was performed using Merlin SEM (Zeiss). Electrochemical experiments were carried out using the potentiostat PGSTAT 30 N (Metrohm, Autolab, The Netherlands). All experiments were performed at room temperature in 3 M KOH as supporting electrolyte, in a three-electrode cell with a silver/silver chloride (Ag/AgCl) as the reference electrode and platinum foil counter electrode. The working electrode was a gold disk of the surface area of 0.2 cm2.

Synthesis of bis(anthracen-9-ylmethyl) malonate (Fig. 1)

Fig. 1
figure 1

Synthesis of bis(anthracen-9-ylmethyl) malonate (BAM)

Full size image

9-Anthracenemethanol (3 g, 14.4 mmol) was added to a stirred solution of triethylamine (1.46 g, 14.4 mmol) in anhydrous dichloromethane (60 ml) at room temperature and under argon atmosphere. The resulting solution was cooled to 0 °C using an ice bath. Afterward, malonyl chloride (1.02 g, 7.2 mmol) diluted with 5 ml of anhydrous methylene chloride was added dropwise in the period of 45 min. Then mixture was allowed to warm to room temperature and was stirred for additional 3 h. Then it was diluted with ethyl acetate, washed twice with water and dried using anhydrous magnesium sulfate. Then, the solvent was removed at reduced pressure. Purification of the product was accomplished by the means of flash column chromatography (1:2.5 ethyl acetate: n-hexane) to give desired product as yellow solid.

Yield: 89%, Mp 191 °C, lit. 178–182 °C (Herranz et al. 2004), the mass spectrum (ESI–MS) showed a [M + Na]+ peak at 507.7 (Figure S1, Supplementary Information, SI); IR (KBr disk) νmax(cm−1) 3085.27, 3036.83, 3001.65, 2960.30, 2916.11, 2848.30, 2359.83, 1747.57, 1722.21, 1383.12, 1352.70, 1330.53, 1171.30, 982.33, 883.91, 735.78, 728.82 (Figure S2, SI); NMR (literature data can be found in Herranz et al. 2004) δ1H 500 MHz; CDCl3; TMS) 8.47 (s, 2H), 8.18–8.20 (q, 4H), 7.97–7.99 (q, 4H), 7.43–7.46 (m, 8H), 6.12 (s, 2H), 3.40 (s, 2H) ppm (Figure S3, SI); δ13C (125 MHz; CDCl3) 166.76, 131.31, 131.07, 129.44, 129.08, 126.77, 125.13, 123.76, 59.97, 41.47 ppm (Figure S4, SI); UV–Vis λmax = 334.4, 350.5, 368.5, 387.6 nm (Figure S5, SI).

Preparation of graphene

Graphene was prepared by exfoliation of graphite in liquid phase using ultrasonic treatment (Quintana et al. 2010; Georgakilas et al. 2010). Briefly, 40 mg of graphite was ultrasonically dispersed in 300 ml of N-methylpyrrolidone (NMP) and sonicated for 30 min. Obtained dispersion was centrifuged carefully for 20 min at 1000 rpm, resulting solid was washed with methanol.

Functionalization of graphene by the Bingel reaction

To a well stirred dispersion of graphene (20 mg) in o-dichlorbenzene (50 ml), bis(anthracen-9-ylmethyl) malonate (49 mg, 0.1 mmol) and tetrabromomethane (33 mg, 01.mmol) were added. Afterward, solution of DBU (75 µl, 0.5 mmol) in o-dichlorbenzene (5 ml) was added slowly under argon atmosphere. Resulting reaction mixture was stirred for 24 h at room temperature. Then it was centrifuged at 1000 rpm for 30 min, and washed twice with toluene and n-pentane. Obtained anthracene functionalized graphene (GA) was dried under vacuum and examined by SEM, FT-IR, XPS and CV, and used in further Diels–Alder reaction with selected fullerenes.

Diels–Alder reaction of fullerenes and anthracene functionalized graphene

Fullerene-graphene hybrid nanomaterials (GAF6/GAF7) were prepared via Diels–Alder reaction using modified method reported by our group previously for anthracene functionalized gold electrodes (Piotrowski et al. 2016). For this purpose, 10 mg of anthracene functionalized graphene was dispersed in 1·10−3 M toluene solution of corresponding fullerene (12 ml). After stirring for 72 h, obtained mixture was centrifuged at 1000 rpm for 30 min. Resulting solids were washed twice with toluene to remove unbound fullerenes and dried under vacuum at 40 °C. Obtained samples of graphene/fullerene composites were subsequently analyzed by SEM, FT-IR, XPS and CV techniques (Figs. 2, 3).

Fig. 2
figure 2

Synthesis scheme for graphene functionalization with bis-anthracene malonate

Full size image
Fig. 3
figure 3

Scheme for functionalization of anthracene covered graphene GA with C60 and C70 fullerenes via Diels–Alder reaction

Full size image

Results and discussion

Scanning electron microscopy (SEM)

Morphology of obtained nanostructures, i.e., pristine graphene and its GA and GAF derivatives was analyzed using scanning electron microscopy (Fig. 4). Graphene obtained by exfoliation of graphite in NMP (Fig. 4A) showed 3D structure build from bent single graphene sheets and flakes. After functionalization of graphene with bis-anthracene malonate, we don’t observe any visible change in the structure of this nanomaterial (Fig. 4B). No significant restacking of single graphene sheets can be seen in SEM image, along with any other visible changes when comparing to unmodified graphene. After Diels–Alder reaction of GA with C60 (Fig. 4C) or C70 (Fig. 4D), we can observe presence of granular aggregates on the graphene surface in both samples. When comparing fullerene modified graphenes: GAF6 and GAF7, we can observe that graphene layers modified with C70 fullerene (GAF7) show higher tendency to restacking. There is also a noticeable difference in the size of observed fullerene aggregates, as C70 adducts with GA show visibly bigger particle size when comparing to C60 analog. To confirm that observation we have performed particle size calculations using ImageJ software (Abramoff et al. 2004) and estimated average diameter of fullerene aggregates on surface of GAF6 and GAF7, which was found to be approximately 24 nm and 34 nm, respectively.

Fig. 4
figure 4

SEM images of pristine graphene (a), graphene covered with anthracene GA (b), graphene functionalized with C60 (GAF6) (c) and C70 (GAF7) (d) via Diels–Alder reaction

Full size image

X-ray photoelectron spectroscopy (XPS)

For the investigation of the graphene-based nanomaterials, C1s region of the X-ray photoelectron spectra can give the biggest insight into characterization of obtained composite materials. Deconvoluted C1s core level spectrum of pristine graphene (Fig. 5A) showed presence of two signals centered at 284.6 and 285.5 eV. This peaks are attributed to graphene carbon atoms with sp2 (at 87% abundance) and sp3 hybridization (13%at), respectively (Kabir et al. 2019; Lesiak et al. 2018), which indicates low content of defects on graphene surface.

Fig. 5
figure 5

XPS C1s spectra of pristine graphene (A), graphene functionalized with anthracene malonate (B), graphene covered with anthracene after Diels–Alder reaction with C60 (C) and C70 (D)

Full size image

Functionalization of the graphene with anthracene moieties ensue visible changes in the XPS results (Fig. 5B). The spectrum of GA in the C1s region was fitted with good correlation into four signals. Main peak observed at 284.6 eV is attributed to sp2 carbons coming from both graphene and introduced anthracene malonate (70%at). Raise of the higher binding energy peaks intensity is caused by increased amount of carbon atoms with sp3 hybridization and was expected after successful addition of anthracene malonate to double bonds of graphene surface. The second biggest peak with 22% abundance is centered at 285.4 eV and is attributed to C–C carbon atoms (Yulaev et al. 2016). The other two peaks are observed due to presence of C–O ester (3%at) and C=O carbonyl (5%at) carbon atoms and where observed at 286.3 and 288.0 eV, respectively (Zuo et al. 2013).

Self–assembly of C60 fullerene on the surface of anthracene functionalized graphene via Diels–Alder reaction had also impact on the C1s region of XPS spectrum (Fig. 5C). Deconvolution of registered signal revealed six components centered at 284.6, 285.6, 286.5, 288.9, 290.1 and 291.3 eV, arising from of all expected forms of carbon: C=C, C–C, C–O and C=O, along with two C60 fullerene π–π satellite peaks (Leiro et al. 2003; Nakao et al. 1997). Intensity of the lowest energy component (89%at), attributed to sp2 carbon, was higher when compared to the anthracene modified graphene. This was observed due to deposition of high amount of C60 fullerene by Diels–Alder reaction and then secondary growth of aggregates on covalently bound seeds. XPS atomic percent values of other carbon atoms were calculated to be: 5%at (C–C), 2%at (C–O) and 2%at (C=O), and remaining π–π* shake-up features were associated with 2%at. Those results confirm successful formation of fullerene-graphene hybrid nanomaterial (GAF6), obtained using Diels–Alder reaction, indicating also that most of fullerenes were physisorbed on the graphene substrate due to formation of aggregates on previously covalently bound C60 molecules.

Similarly, XPS spectrum registered for GAF7 showed noticeable changes in C1s region (Fig. 5D), when comparing to GA. C1s region of registered XPS spectrum was fitted with three signals, which can be attributed to C=C/C–C, C–O, C=O, and were observed at correspondingly 284.6, 286.0 and 288.8 eV (Sohmen et al. 1992). The most important difference when comparing foregoing data to the spectra obtained for GAF6 sample, is the higher content of the sp2 carbon atoms. This result is in good agreement with the observations made for SEM images, where the GAF7 sample showed presence of visibly larger fullerene aggregates.

FT-IR spectroscopy

Formation of graphene-fullerene nanocomposites was also confirmed by means of FT-IR spectroscopy (Fig. 6). When comparing the spectra of pristine graphene (G) and its anthracene derivative (GA), both share the two signals located at around 820 and 1110 cm−1 (Leiro et al. 2003; Güler et al. 2013), frequently reported for grapehenous materials. However, IR spectrum of anhtracene modified graphene indicates serious changes in the composition showing numerous additional bands. Strong signal observed at 1704 cm−1, which is attributed to C=O stretching vibrations indicates introduction of carbonyl groups, coming from malonate adduct. Presence of ester functional groups is also confirmed by C–O stretching bands at 1073 and 1263 cm−1. Aromatic C=C stretching observed at 1563 and 1656 cm−1, along with aromatic C–H stretch at 2966 cm−1, can be assigned to anthracene moieties from introduced malonate, revealing successful functionalization of graphene surface.

Fig. 6
figure 6

FT-IR spectra of pristine graphene (G), anthracene functionalized graphene (GA), Diels–Alder adduct of antracene with fullerene C60 (GAF6) and C70 (GAF7) registered in KBr disk

Full size image

FT-IR spectrum after addition of C60 fullerene via Diels–Alder reaction retains most characteristic signals observed for graphene and its newly introduced anthracene malonate functional groups. At the same time, signals characteristic for C60 molecule were registered at 1422, 564 and 525 cm−1 (Bethune et al. 1991), allowing to conclude that anthracene modified graphene reacts with fullerene yielding GAF6.

Similarly, addition of C70 to anthracene modified graphene also results in presence of additional signals in registered spectrum. Typical C70 related bands were registered in GAF7 sample at 1437, 1130 and 537 cm−1 (Bethune et al. 1991), indicating that C70 also form Diels–Alder adducts with graphene covered with antrhracene malonate.

Electrochemical measurements

In order to investigate the capacitive properties of the synthesized nanomaterials, 3 mg of each sample was mixed with 300 µl of ethyl alcohol and 25 µl of Nafion™ solution and left overnight on a magnetic stirrer in order to obtain stable suspension. Prepared samples were applied on the working electrode by dropping a known volume of suspension until the gold surface was completely covered. Such modified electrodes were ready to use in electrochemical experiment after complete evaporation of the solvent. All measurement were conducted in 3 M KOH. Cyclic voltammograms of the synthesized materials in potential range from 0.2 to  1.1 V with a scan rate of 20 mV s−1 are shown in Fig. 7. The CV curves for all samples show a quasi-rectangular shape, indicative of the electrochemical double-layer capacitive characteristics.

Fig. 7
figure 7

Comparison of cyclic voltammograms of pristine graphene (G), graphene functionalized with anthracene malonate (GA), graphene covered with anthracene after Diels–Alder reaction with C60 (GAF6) and C70 (GAF7) in 3 M KOH with a scan rate of 20 mV s−1

Full size image

The specific capacitance (Cs) of electrode materials can be determined from the slope of the linear charge–discharge graph (Fig. 8) according to the equation Cs = I/(m(dV/dt)), where I is the applied constant current, dV/dt is the slope of the linear portion of the charge curve, and m is the mass of the materials deposited on the working electrode. The calculated specific capacitances of the graphene/fullerene composites were shown in Table 1.

Fig. 8
figure 8

Galvanostatic charge/discharge curves at the current density of 1 A g−1 in 3 M KOH conducted in potential range from 0.2 to  1.2 V for pristine graphene (G), graphene functionalized with anthracene malonate (GA), graphene covered with anthracene after Diels–Alder reaction with C60 (GAF6) and C70 (GAF7)

Full size image
Table 1 Specific capacitance of obtained electrode materials
Full size table

Results indicate that functionalization of graphene with anthracene malonate does not affect the capacitive properties of this material; the values of specific capacitances G and GA samples are almost the same (55.0 F g−1 and 54.6 F g−1, respectively). Among all samples, C60 covered graphene (GAF6) shows the highest value of specific capacitance (65.5 F g−1). Since the capacitive properties of electrode material are closely related to this part of material which is in direct contact with electrolyte solution, the decreased capacitance for GAF7 composite (45,4 F g−1) is in quite good agreement with the SEM images where the GAF7 sample showed presence of visibly larger fullerene aggregates compared to GAF6, resulting in lower active surface area. To support this observation, we have determined specific surface area of both nanomaterials using methylene blue method (Chen et al. 2020; Rubino et al. 1999). Results obtained for GAF6 and GAF7 are in good agreement with SEM images, as specific surface areas were calculated to be approximately 810 m2/g and 540 m2/g, respectively.

The performance and stability of our best energy storage material GAF6 were studied using GCD measurements for the time of 12 h of continuous charging and discharging (Fig. 9). The first GCD curves reveal highly linear and symmetrical shape, but slow loss of this symmetry and tendency of capacitance drop during cycling is observed. As seen in the capacitance versus number of cycle plots, the electrode retained around 60% of its capacitance after 200 cycles.

Fig. 9
figure 9

Potential versus time plot for the first seventeen cycles at the current density of 1 A g−1 in 3 M KOH. Inset figures show capacitance retention (left) and GCD curves for 12 h experiment (right)

Full size image

Conclusions

Graphene functionalized covalently with bis-anthracene malonate using modified Bingel reaction was synthesized and characterized. Facile functionalization method allows to obtain desired graphene/fullerene composites. Reported novel nanomaterial is capable for binding fullerenes via Diels–Alder reaction of introduced anthracene molecules with both C60 and C70 cores.

Obtained graphene/fullerene composites were successfully employed as energy storage materials. Electrochemical measurements revealed enhanced properties of C60 covered graphene (GAF6) when comparing to pristine graphene, its anthracene derivative and corresponding C70 analog, showing nearly 20% increase in specific capacitance value in comparison to starting material. This can be attributed to high specific surface area of C60 fullerene functionalized graphene and fact that C60 molecules surface serve as spacers that prevent graphene sheets from restacking.

Presented approach can be further developed, leading to formation advanced hybrid materials consisting of graphene and fullerenes or their derivatives. It can be applied in materials engineering not only for construction of energy storage devices but many other fullerene-graphene composites, where simple synthesis method and resulting covalent binding of fullerene are of great importance.