## Introduction

The demand for renewable energy has grown rapidly in recent years due to the rapid decline of fossil fuels and growing concerns about environmental pollution. Meanwhile the demand for sustainable and clean energy is becoming more critical owing to the emergence of various electronic devices1,2,3. Therefore, the search for next generation energy storage materials and devices is very important. Supercapacitors have received a great deal of attention from the research community as energy storage devices due to their low cost, high power density, and high efficiency4,5,6,7. A supercapacitor consists of two electrodes immersed in an electrolyte and separated by an ion conducting but electron insulating membrane. The mechanism of charge storage in supercapacitors can be non-faradaic (electrochemical capacitor) or faradaic (pseudocapacitor)1. Various carbon materials with high surface area, high conductivity, morphology, size, and pore size distribution can be synthesized at large scale. In general, pure carbon materials such as activated carbon, graphene nanosheets, nanotubes, and nanocages exhibit non-faradaic double-layer energy storage mechanism, i.e. there is no electron transfer at the electrode electrolyte interface and energy storage is electrostatic in nature8. Meanwhile, fast reversible redox reactions occur in faradaic pseudocapacitors during the charge–discharge process9. Among pseudocapacitors, transition metal oxides or transition metal hydrides are mostly used due to their high theoretical specific capacitance and fast redox reactions on their surfaces10,11. Noble metal oxides such as RuO2 and IrO2 have been studied as electrode materials in the past12,13. However, the use of noble metal oxides for supercapacitors is limited due to their high cost. Instead, the use of more abundant and cheaper transition metal oxides has been explored, which has made it feasible to design supercapacitor materials with high theoretical capacitance. For example, porous nanostructured NiO and its composites have been studied as electrodes for supercapacitor because of their low cost and high theoretical capacity14,15,16. However, NiO has poor electroconductivity and therefore low charge–discharge rate and reversibility. Taking the benefit of the higher conductivity of carbon materials and high theoretical capacitance of NiO, alternatives have been explored by combining NiO with activated carbon or carbon black16,17,18.

Bitumen is an abundant natural resource in Canada which has been widely used as raw material for petroleum products. Canada alone produced 2.8 million barrels per day of crude bitumen in 201719. Unlike conventional crude oil, bitumen is rich in other elements such as nitrogen, sulphur, and heavy metals. Additionally, asphaltenes, the insoluble component obtained from partial upgrading of bitumen, are also cheap and abundant carbon-rich resource. The molecular complexity of bitumen can be reduced by fractionating it using different solvents using ASTM standards20. Based on this method, bitumen can be fractionated into saturates, aromatics, resins, and asphaltenes. Asphaltenes are a solubility class that is soluble in light aromatics such as benzene and toluene but is insoluble in light paraffins such as the n-pentane or heptane21. Recently, there has been a surge in the synthesis of novel carbon materials such as nanosheet, nanoporous carbon, etc. from fossil fuels including pitch, coal, and asphaltenes22,23,24,25. Bitumen and asphaltenes are rich in polycyclic aromatic hydrocarbons which can be transformed into highly ordered carbon nanostructures including nanotubes and nanosheets. By using a melamine sponge template and asphaltene extracted from crude oil as the precursor26, or asphaltene from coal and an in-situ sheet-structure-directing agent from urea thermal polymerization27, the interconnected porous carbon were derived with an electrochemical capacitance of 200 F g−1 at 5 mV s−126, or the porous carbon nanosheet with a graphitized-like ribbon structure with 282.9 F g−1 at 100 A g−127 in a three-electrode test, respectively. Although bitumen and asphaltenes are very promising raw materials for carbon supercapacitors, there is no detailed study relating the physical and electrochemical properties of nanocarbon obtained from different fractions of bitumen. Furthermore, due to the presence of transition metals in bitumen, it can be directly used to synthesize transition metal oxide–carbon composites (TMO/C) which are known to exhibit superior performance as supercapacitors due to high conductivity of carbonaceous material and high pseudo capacitance of TMOs28.

Various kinds of two-dimensional (2D) materials have been used to assist in the formation of planar carbon nanosheets. Some examples of such materials which provide a guiding surface for the formation of carbon nanostructures are montmorillonite clay, Zn(OH)2 nanosheets, Mg(OH)2 nanoplates, MoS2 nanosheets, amino functionalized graphene oxide, NaCl, Na2SiO3, vermiculite, etc25,29,30,31,32,33,34. We herein report a Mg(OH)2 nanoplate template guided synthesis of porous carbon nanomaterials using bitumen and asphaltene fractionated from the same bitumen and their in-situ KOH activation. Mg(OH)2 nanoplates were chosen as template due to its cost effectiveness, simple preparation, and overall good performance of the carbon nanostructures prepared on Mg(OH)2 substrate. Asphaltenes were fractionated from the bitumen by precipitation using hexane. In addition, the asphaltene obtained was further partitioned into two fractions using N,N-dimethylformamide (DMF) as the solvent. The nanoporous carbon formed presents high surface area and a distribution of micro and mesopores which results in high conductivity, specific capacitance, and retention. The asphaltene fraction obtained from nickel complexes-containing bitumen led to the formation of NiO nanoparticles upon pyrolysis. The NiO/C composite obtained from asphaltenes exhibited the highest capacitance. The specific capacitance of the NiO/C composites obtained from DMF fractionated asphaltene was also measured. Interestingly, the capacitance decreased when the asphaltenes were fractionated using DMF. There was a significant decrease in the capacitance of NiO/C composite obtained from DMF insoluble fraction of asphaltenes, which was ascribed to the lower NiO content after DMF treatment. The NiO/C composite obtained from DMF soluble fraction of asphaltenes was higher than the insoluble fraction but lower than unfractionated asphaltenes. The proposed rationale for lower capacitance than unfractionated asphaltene is the higher Ni content in DMF soluble fraction of asphaltene resulted in NiO/C composites with lower conductivity.

## Experimental

### Chemicals

Toluene (anhydrous, 99.8%), n-hexane (99%), N,N-dimethylformamide (ACS reagent, ≥ 99.8%) and hydrochloric acid (ACS reagent, 37%) were purchased from Sigma Aldrich and used as received. Magnesium chloride hexahydrate (crystalline), sodium hydroxide (pellets, 98%), and potassium hydroxide (pellets, 85%) were purchased from Fisher Scientific. Oil sand sample was provided from an Alberta oil sand company. All solutions were prepared in deionized water (resistivity ≥ 18.2 MΩ cm).

### Synthesis of nanoporous carbon

Bitumen was extracted from the Alberta oil sand sample using toluene as the solvent. Asphaltenes were obtained through precipitating the bitumen in toluene solution with hexane. Asphaltenes were further separated into DMF-insoluble and DMF-soluble fractions by being dissolved in DMF and followed by filtration. Mg(OH)2 was prepared by slow reaction of MgCl2 and NaOH solution as found in the literature35. The precipitated Mg(OH)2 was filtered and washed with DI water and dried. For the preparation of porous nanocarbon, 2 g of bitumen or asphaltenes was mixed with 4 g of Mg(OH)2 and 8 g of KOH. The mixture was transferred to a high temperature crucible and placed inside the tube furnace under N2 atmosphere (flow rate 300 mL min−1). The sample was heated to 300 °C at the rate of 5 °C min−1 in N2 atmosphere and kept for 30 min. Finally, the temperature was raised to 800 °C at the rate of 5 °C min−1 and kept at that condition for another 1 h. After the completion of the reaction, the sample was cooled to room temperature and washed with HCl followed by DI water. The nanoporous carbon samples obtained from bitumen were labeled as BCNS and the nanoporous carbon obtained from hexane precipitated asphaltenes were labeled as ACNS1. The nanoporous carbons from DMF-insoluble and DMF-soluble fractions of asphaltenes were labeled as ACNS2 and ACNS3, respectively.

### Material characterization

Nanoporous carbons were characterized by scanning electron microscope (SEM), Fourier-transform infrared spectroscopy (FTIR), surface area analysis, pore size distribution, and X-ray powder diffraction (XRD) analysis. The characteristic peaks and bands were acquired by FTIR with an ATR sampling accessory (Perkin Elmer 400 FT-IR). 32 scans were performed from 500 to 4000 cm−1 to acquire the FTIR spectra. The XRD spectra were obtained by using a Rigaku multiplex X-ray diffractometer with a Cu X-ray source which was operated at 40 kV voltage and 40 mA current. A Micrometric ASAP 20 surface analyzer was used to measure the surface area of the samples by using the Branauer-Emmett-Teller (BET) method (N2 gas adsorption–desorption). Using the same equipment, pore size distributions were calculated by Barrett-Joyner-Halenda (BJH) formalisms using desorption isotherms. All the samples were degassed for 6 h at 300 °C prior to measurements. SEM images were acquired using Quanta FEG 250 field emission scanning electron microscope. All measurements were carried out under high vacuum at either 2.5 or 5 kV.

### Electrochemical measurement

The working electrode was fabricated by mixing 90% nanoporous carbon or NiO/carbon composite and 10% PTFE in 2-propanol. The mixture was sonicated for 30 min to form a homogeneous mixture which was then loaded on a 1 × 1 cm2 nickel foam current collector. About 1 mg of carbon was loaded on each nickel foam. After evaporating the solvent, the nickel foam with carbon was further dried at 95 °C in an oven for 1 h. For the three-electrode test, 6 M KOH was used as the electrolyte, a Pt plate as the counter electrode, and a Ag/AgCl electrode as the reference electrode. All the electrochemical tests were performed at 25 °C.

The working electrode was tested by cyclic voltammetry (CV) and galvanostatic charge discharge using PARSTAT 4000A electrochem station. Electrochemical impedance spectra were acquired between 105 and 0.01 kHz using the same instrument. From the CV curve, the specific capacitance $${C}_{sp}$$ of the carbon electrode under the three-electrode system was calculated by Eq. (1):

$${C}_{sp}=\frac{\int Idv}{mv\Delta V}$$
(1)

where $$\int Idv$$ is the integrated area of the CV curve, $$m$$ is the mass of the electrode material, $$v$$ is the potential scanning rate (V s−1), and $$\Delta V$$ is the potential window of the CV.

From charge-discharge experiments, the specific capacitance $${C}_{sp}$$ of the carbon electrode under the three-electrode system was determined by Eq. (2):

$${C}_{sp}=\frac{I\Delta t}{m\Delta V}$$
(2)

where $$I$$ is the applied current, $$\Delta t$$ is the discharge time, $$m$$ is the mass of the electrode material and $$\Delta V$$ is the potential.

## Results and discussion

Figure 4A shows the CV curves of ACNS1 at the scan rates of 10, 25, 50, 100, and 200 mV s−1 over the potential range 0 to − 1 V. The large current response and quasi rectangular shape suggests reversible electrochemical double layer capacitance whereas the slight redox peaks at − 0.2 V reveals the pseudocapacitive behavior of the ACNS1 sample which is due to NiO. Galvanostatic charge–discharge of the ACNS1 at current densities 1, 2.5, 5, 10, and 20 A g−1 are shown in Fig. 4B. The galvanostatic charge–discharge curves deviate from symmetry, which implies that the supercapacitive behavior of ACNS1 resulted from both pseudocapacitance and EDLC. Charging–discharging times were longest at 1 A g−1 and decreased as the current was increased. The CV and galvanostatic charge–discharge curves of BCNS, ACNS2, and ACNS3 are shown Fig. 5. The CV curves of these electrodes were similar in shape to the ACNS1 electrode but had lower current values. The galvanostatic charge discharge curves of BCNS electrode was nearly symmetric indicating dominant EDLC phenomenon. On the other hand, the charge-discharge curves of ACNS2 and ACNS 3 were similar in shape to the ACNS1 electrode due to the pseudocapacitance along with EDLC. For comparison, the CV and galvanostatic charge–discharge curves of reduced graphene oxide (rGO) were also measured (Fig. 5). The gravimetric capacitances obtained from CV and galvanostatic charge–discharge curves of BCNS, ACNS1, ACNS2, ACNS3, and reduced GO are shown in Fig. 4C,D, respectively. The data show that ACNS1 had the highest capacitance compared to the other samples. The gravimetric capacitance of ACNS1 is 359 F g−1 at potential scan rate 10 mV s−1. ACNS3 had specific capacitance of 365 F g−1 at the same scan rate but it is lower than ACNS1 at higher scan rates. Similarly, gravimetric capacitance measured from galvanostatic charge–discharge measurements were highest for ACNS1 as shown in Fig. 4D. The calculated gravimetric capacitance from GCD measurement at 1 A g−1 was 380 F g−1. Capacitances of BCNS, ACNS2, and reduce GO were lower than that of ACNS1 and ACNS3. The higher specific capacitance $${C}_{sp}$$ for ACNS1 and ACNS3 is due to the pseudocapacitance of NiO combined with the EDLC of carbon. For ACNS2 and ACNS3, which are prepared from the DMF insoluble and DMF soluble fractions of asphaltene, respectively, the change of the NiO/C ratios in these samples during the fractionation process led to the decrease in capacitance. The Ni content decreased in the DMF insoluble fraction which led to NiO/C composite with lower NiO content. Thus, the pseudocapacitive contribution was decreased leading to overall decrease of the $${C}_{sp}$$ of ACNS2. Meanwhile, the DMF soluble fraction, which was rich in Ni, formed NiO/C composite with higher NiO content and the capacitance increased again in ACNS3. However, the capacitance was lower than ACNS1 which may be due to the increased resistance of the NiO component and was further discussed with the following impedance measurement results (Fig. 6). The effect of NiO content in NiO/C composite on the capacitance has been reported in prior studies. For example, Lota and coworkers prepared NiO-activated carbon composites with three different ratios: 34% NiO and 66% activated carbon, 17% NiO and 83% activated carbon, and 7% NiO and 93% activated carbon. Their results indicated that the low amount of NiO (7%) resulted in the highest capacitance40. Moreover, the smaller NiO crystal size in ACNS1 could be beneficial since it can provide higher specific surface area for charge storage. It should also be noted that most of the literature reported also apply 5–10% conductive carbon black to improve the conductivity of the electrodes23,41. In this work, we prepared electrodes without adding carbon black. Table 2 summarizes the capacitance of various carbon and carbon-composite materials compared to the one reported in this study, indicating very comparable performance achieved in this work.