Combustion synthesis of N-doped three-dimensional graphene networks using graphene oxide–nitrocellulose composites
- 86 Downloads
We demonstrate a novel method to prepare high-quality and uniform nitrogen-doped, three-dimensional, graphene networks through combustion of graphene oxide (GO)–nitrocellulose composites. The N-doped 3D graphene networks have tunable porous morphology and the pore size can be controlled by adjusting the concentration of GO in the nitrocellulose matrix. This new method is a simple method to obtain nitrogen-doped graphene networks and has potential applications in energy storage and conversion, catalysis, and sensors.
KeywordsNitrogen-doped (N-doped) three-dimensional (3D) graphene networks Graphene oxide (GO)-polymer composites Nitrocellulose (NC)
Graphene is a single-atom thick sheet packed in hexagonally arrayed sp2-bonded carbon atoms. Graphene has been widely applied in energy and environmental science, organic synthesis and catalysis, material fabrication, biomedicine, chemical and biosensors, and nanoelectronics due to its large theoretical specific surface area (~ 2600 m2/g), high thermal conductivity (~ 5000 W/mK), fast charged carrier mobility (~ 200,000 cm2/V s), great mechanical strength (Young’s modulus is around 1 TPa), and unique optical properties (non-linear Kerr’s coefficient is around 10−7 cm2/W) [1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13]. The properties of graphene strongly depend on its morphology and size. For example, graphene shows excellent semiconducting characteristics when its width is less than 10 nm, but larger size graphene does not exhibit semiconducting characteristics ; two-dimensional (2D) graphene sheets commonly aggregate during processing in solution to reduce the specific surface area and even affect the thermal and mechanical properties due to strong π–π interactions and Vander Waals forces, but three-dimensional (3D) graphene networks can avoid aggregation and take full advantage of graphenes’ intrinsic physical and chemical properties [15, 16, 17, 18, 19, 20]. Apart from the morphology and size control, another efficient method to tune the properties of graphene is chemical doping. To date, there are two popular methods to chemically dope graphene. The first method is absorbing other materials such as metal ions, planar organic molecules, and 2D or liner polymers to the surface of graphene via strong π–π interactions, Vander Waals forces, or chemical reactions . The other method is doping heteroatoms such as nitrogen, silicon, and boron atoms into the hexagonally arrayed sp2-bonded carbon lattice [17, 21, 22, 23, 24, 25, 26].
The preparation of nitrogen-doped graphene (NG) has received attention in recent years because introducing nitrogen into the carbon lattice of graphene can modulate the electron donor, or acceptor, properties intrinsically of graphene, and improve the chemical functionality and electrical properties of graphene, such as the specific capacitance and cycle ability [23, 25, 26, 27, 28, 29]. For example, NGs exhibited higher efficient electrocatalytic activity for the oxygen reduction reaction than graphene and has the possibility to replace expensive heavy metal–based catalysts to fabricate metal-freed fuel cells . Currently, there are mainly two methods developed to prepare NGs, including chemical vapor deposition (CVD) [30, 31, 32] and chemical treatment of graphene/GO [23, 33, 34, 35, 36, 37, 38, 39].
CVD is a popular method widely used to produce high-purity and large scale carbon nanofibers, carbon nanotubes, and graphene [40, 41, 42, 43, 44, 45]. CVD has also been used for producing NG using metal catalysts such as Cu and Ni where the carbon source gas needs to be mixed with a nitrogen-containing gas such as ammonia and pyridine . For example, Wei et al. reported that the NG can be synthesized by using CH4 as a C source and NH3 as a N source . The CVD process was performed at 800 °C and used a 25-nm Cu film on a Si substrate as the catalyst. The doping concentration of nitrogen on graphene could be adjusted via controlling the ratio of CH4 and NH3. Apart from the gas mixture, liquid organic materials (such as acetonitrile  and pyridine ) and even solid organic materials (such as 1,3,5-triazine ) have also been used as a C source and a N source to prepare NGs. Jin et al. indicated that NGs with around 2.4% nitrogen doping concentration can be prepared using a CVD process with pyridine as the sole source of both C and N, the copper foil as the catalyst and 1000 °C as reaction temperature . Lu et al. reported that NGs can be synthesized using 1,3,5-triazine as the sole source of both C and N and a Cu foil as a catalyst . They also showed that the doping concentration of nitrogen on graphene could be adjusted via controlling growth temperature. Although the CVD method can be used to synthesize high-purity and large-scale NGs, the manufacturing is relatively high cost and the procedures are complex and time consuming.
Another popular method widely used to prepare NGs is chemical treatment of graphene/GO where NGs can be synthesized by directly modifying graphene [23, 34, 35]. There are two main ways: plasma and thermally treat graphene with some nitrogen-rich compounds, such as nitrogen [23, 35] and ammonia . The graphene can be chemically synthesized [34, 35] or mechanically exfoliated graphene . For example, Wang et al. synthesized N-doped graphene nanoribbons (GNRs) through high-power electrical Joule-heating GNRs in ammonia gas . Wang et al. prepared NGs through nitrogen plasma treatment of graphene-chitosan composite . Moon et al. synthesized N-doped graphene quantum sheets from monolayer graphene by nitrogen plasma . Lin et al. synthesized N-doped graphene by ammonia plasma . NGs can also be synthesized via plasma or by thermally treating GO with some nitrogen-rich compounds, such as nitrogen , ammonia (NH3) [37, 38], hydrazine hydrate (N2H4) , pyrrole , urea , and melamine . For example, Zhao et al. have prepared N-doped 3D graphene networks via a two-step thermal treatment method . First, they hydrothermally treated GO aqueous suspension mixed with 5 vol% pyrrole at 180 °C for 12 h to produce a nitrogen-containing gel. Then they freeze-dried the gel and annealed it at 1050 °C for 3 h under Ar atmosphere to synthesize N-doped 3D graphene networks. They indicated the mechanism that pyrrole can be used to prepare NGs because of the conjugated structure and the electron-rich N atom of pyrrole can make it attach to the surfaces and galleries of GO sheets through hydrogen bonding and π–π interactions. When the GO-pyrrole composite is thermally treated, pyrrole can be used as a nitrogen source to modify graphene sheets . Jeong et al. prepared NGs via a three-step plasma and thermal treatment method . First, the GO was reduced by a hydrogen plasma process; then, the nitrogen plasma was used to introduce nitrogen on graphene sheets; finally, the samples were annealed at 300 °C for 3 h to remove residual functional groups on the NG surface. Although the cost of chemical treatment of graphene/GO to synthesize NGs is lower than CVD, the process still require specific instruments such as plasma and tube oven and the procedures are still complex and time consuming.
Combustion synthesis, also called self-propagating high-temperature synthesis (SHS), is a highly effective, low-cost, and simple method to produce various nanomaterials [49, 50, 51]. In theory, the use of combustion could also provide a high-temperature environment to reduce GO to graphene, but there is no work about the preparation of graphene or NGs using combustion synthesis as far as we know. Herein, we demonstrate a combustion method to prepare high-quality and uniform N-doped 3D graphene networks through burning of GO-nitrocellulose (NC) composites over a large area. NC is a low-cost industrial polymer and is also an important energetic material [52, 53, 54]. The combustion of NC can provide a flame temperature higher than 1000 °C and the burning of NC is fast and complete. Furthermore, GO can be easily dispersed in NC matrix to form a composite material and the rich nitrogen in NC can be used as a nitrogen source to react with GO at a high temperature to prepare NGs . The as-prepared N-doped 3D graphene networks have tunable porous morphology and the pore size can be controlled by adjusting the concentration of GO in the NC matrix. Compared with the CVD and the chemical treatment of graphene/GO methods, this combustion synthesis is a lower cost, faster (the burning process only need several seconds), and simpler method to obtain nitrogen-doped 3D graphene networks, which has potential applications in energy storage and conversion, catalysis, and sensors.
GO powders were purchased from Graphene Laboratories Inc. (Calverton, NY). NC (4 ~ 8% in ethanol/diethyl ether, 11.8–12.2 wt.% nitrogen) was purchased from Sigma-Aldrich Inc. NC solution was made to solid films before using by evaporating organic solvents at ambient and drying at 70 °C in a vacuum oven until a constant weight is achieved in air.
Preparation of GO–NC composites
GO powders were dispersed in de-ionized water by a tip sonicator (Bandelin Sonoplus) for 60 min to make 3 mg/mL GO–water solution. Solid NC was dissolved in acetone to make 5% (mass concentration) NC–acetone solution. Then, the different amount of fresh GO solution was added into the NC–acetone solution and stirred for 6 h at room temperature to make various ratios of GO–NC solutions. The various GO–NC composites were made through adding various ratios of GO–NC solutions to glass petri dishes (VWR International) of 7.5 × 7.5 cm and then evaporating solvents at ambient and drying at 70 °C in a vacuum oven until a constant weight is achieved in air. Four GO–NC composites, including GO–NC-1, GO–NC-2.5, GO–NC-5, and GO–NC-10, which including 1%, 2.5%, 5%, and 10% GO respectively, are prepared.
Transmission electron microscopy analysis
The prepared N-doped 3D graphene networks were dispersed in acetone by a tip sonicator (Bandelin Sonoplus) for 10 min. The N-doped 3D graphene networks for TEM (Hitachi 8100) observation was prepared by placing drops of the N-doped 3D graphene acetone suspension onto the carbon-coated copper grid (Ultrathin Carbon Film on Holey Carbon Support Film, 400 mesh, Copper, Ted Pella, Inc.), which was then dried under ambient conditions prior to being introduced into the TEM chamber. The samples were measured by using an acceleration voltage of 60 kV.
Scanning electron microscopy analysis
The morphologies of N-doped 3D graphene networks were examined by Hitachi S570 SEM. All samples were sputter coated a thin layer of gold prior to them (about 5 nm) to ensure good conductivity and imaging.
Atomic force microscope analysis
The thickness of graphene sheets from N-doped 3D graphene networks were obtained with a Nanoscope IIIa multimode AFM (Veeco, Santa Barbara, CA) operating in tapping mode. The AFM tips (Tetra15/Au/15, Nanotechnology LLC) is a silicon cantilever with Au conductive coating, with the resonant frequency 200 ~ 400 kHz and force constant 20 ~ 75 N/m. All images were collected at a scan rate of 0.5 Hz with a driving frequency of 325 kHz (256 × 256 lines scan). All samples were made by placing drops of the N-doped 3D graphene network acetone suspension onto a silicon substrate (1 cm × 1 cm), and the solvent was then dried using spin-coater with 3000 rpm for 3 min. Polished silicon wafers were purchased from Nova Electronic Materials Ltd. The diced silicon substrates were cleaned by a solution containing 5 parts NH4OH (20%), 1 part H2O2 (30% solution), and 1 part de-ionized water and heated to 50 ~ 70 °C for 3 h, rinsed with de-ionized water, and dried at 70 °C in N2 environment. The substrates were stored in desiccators prior to use.
X-ray photoelectron spectroscopic analysis
The XPS analysis of pure GO, pure NC films, GO-NC-1, and N-doped 3D graphene networks was performed using the PHI 5000 VersaProbe II Scanning Microprobe with the PHI MultiPak Version 9.3 software (Physical Electronics Inc., MN, USA). All spectra were acquired with a monochromatic aluminum X-ray source (hν = 1486.6 eV), a 100-μm spot size in point mode, both electron and ion neutralization, and a hemispherical analyzer pass energy of 29.35 eV. All samples were mounted with double-sided copper tape to the XPS sample holder. All spectra were collected at a 45° take-off angle. Curve fitting of the C1s and O1s spectra was performed using a Gaussian–Lorentzian peak shape after performing a Shirley background correction.
TGA was carried out in a TGA I 1000 (Instrument Specialist Inc.). The temperature was raised from 25 to 1000 °C at the rate of 10 °C/min at air or Helium (He) atmospheres.
3 Results and discussions
Relative atomic percent composition of BMRGOs (not weight percent)
C atomic %
O atomic %
N atomic %
The C1s and O1s XPS spectra of pure GO powders, pure NC film, GO-NC-1, BMRGO-1, and BMRGO-5 are shown in Fig. 5b, c. The C1s peak in pure GO located at 284.6 eV is characteristic of the C–C and C–H and the peaks in 286.5 and 288.1 eV are assigned for the C–OH and C=O groups, respectively [26, 57]. The pure NC film and GO-NC-1 show the C–OH and C–O–NO2 C1s bonding peaks located at 286.9 and 288.3 eV, respectively. However, after burning GO-NC-1, the C1s spectrum in BMRGO-1 is almost pure C–C bonds (284.6 eV) and contains a very small peak at 286.5 eV, which can be assigned to residual C–OH bonds [26, 57]. When RGO is doped with nitrogen, a new small peak at around 287.5 eV will be observed [23, 26, 30, 31, 33, 34, 38, 48]. This small peak reflects the different bonding structure of the C–N bonds . The C1s spectrum in BMRGO-1 has a peak at 287.8 eV, which indicates that the nitrogen is successfully doped in the graphene sheets after burning of the GO-NC-1. Previous studies also reported that N-doped GO has a new small peak at ~ 289 eV compared with RGO [23, 26, 30, 31, 33, 34, 48]. This new peak is ascribed to the physisorbed oxygen on the graphene sheets. The C1s spectrum in BMRGO-1 has a small peak at 289.7 eV. This result also indicates that the nitrogen doping occurs in the graphene sheets after burning of the GO-NC-1. There is no 288.1 eV peak, which indicates all C–O–NO2 groups are degraded during the burning, again consistent with the overall composition change. The C1s peak in BMRGO-5 is similar as in BMRGO-1. Therefore, the nitrogen is also doped in the graphene sheets after burning GO-NC-5 composite.
The O1s spectrum in pure GO contains C=O and C–OH groups located at 531.0 eV and 532.7 eV, respectively . The O1s peak in the pure NC film, GO-NC-1, and GO-NC-5 is located at 533.5 eV and represents the O–NO2 groups. After burning GO-NC-1, the O1s spectrum in BMRGO-1 has two small peaks at 530.6 and 533.9 eV, which are characteristic of some new chemical species that are obtained from the C=O or C–OH groups . These results also indicate some C=O and C–OH groups on original GO sheets are reduced during burning the GO-NC-1 composite. The O1s spectrum in BMRGO-5 has two small peaks at 530.6 and 532.6 eV, which is similar with BMRGO-1 and also support the conclusion that BMRGO-5 is reduced GO. The atomic ratio of carbon to oxygen in BMRGO-1 and BMRGO-5 is 79.0:13.3 and 78.2:14.7, respectively. Therefore, during burning GO–NC composites, most functional groups C=O and C–OH on original GO sheets are transferred to some new chemical species that contain oxygen and some of them were reduced to C=C or C–N bonds. All in all, the XPS results clearly demonstrated that burning of GO–NC composites can prepare nitrogen-doped 3D-graphene networks, but the degree of GO reduction is lower than thermal decomposition of GO–NC composites .
In summary, we have demonstrated a simple method to large area fabricates high-quality and uniform N-doped 3D graphene networks through combustion of the GO–NC composites. The highly ordered N-doped 3D graphene networks assemblies with tunable open porous morphologies and the pore size can be effectively controlled by tuning the concentration of GO in the NC matrix at original GO–NC composites. The BMRGO 3D networks made from 1% GO doping concentration owns the highest BET SSA and the doping concentration of nitrogen on graphene is around 2.8%. The BET SSA of BMRGO 3D networks was reduced significantly when the GO doping concentration was higher than 5%, but the doping concentration of nitrogen on graphene is not changed obviously, which was around 2.8%. In addition, the original GO–NC hybrid materials are soft and flexible and can be easily fabricated to different sizes and shapes, which allow adjustment of the size and shape of the N-doped 3D graphene networks. Furthermore, the N-doped 3D graphene networks can also be prepared on different substrates, such as indium tin oxide (ITO), metals, glass, silicon wafers, and mica papers, through depositing and burning original GO–NC hybrid materials on these substrates. This is a simple method to obtain the N-doped 3D graphene network–modified electrode or substrates. Therefore, the N-doped graphene networks prepared by this facile combustion method has various potential applications, including energy storage and conversion, catalysis, and sensors.
This work was supported by ONR (N00014-11-1-0424) and the U.S. Department of Homeland Security under Award Number 2008-ST-061-ED0001.
Compliance with ethical standards
The authors declare that they have no competing interests.
- 16.Ji X, Zhang X, Zhang X (2015) Three-dimensional graphene-based nanomaterials as electrocatalysts for oxygen reduction reaction. J Nanomater 2015:1–9Google Scholar