Abstract
Microcrystalline cellulose was pyrolyzed and catalytically graphitized under temperatures ranging from 1000 °C to 1600 °C in the presence of nickel (Ni). Optimal conditions for graphitization were determined, along with the structure and conductivity of the resulting samples. The optimal conditions were identified as heating at 1400 °C for 3 h with 3 mmol Ni loading per gram of carbon. The samples obtained had excellent graphitic crystallinity comparable to that of commercial graphite. However, in the absence of Ni loading, no obvious graphitic structure appeared after heating under the same conditions, indicating that Ni was an efficient catalyst for the graphitization of cellulose-based carbon. High-resolution transparent electron microscopy (HRTEM) images showed well-defined graphitic structures of more than 30 layers with slice gaps of 0.340 nm. The conductivities of the samples treated under different temperatures varied from 27 S·cm-1 to 54 S·cm-1 under 20 MPa of pressure, and higher temperatures led to higher conductivity due to better graphitic crystallinity. This study fills an important area of research on the catalytic graphitization of cellulose and provides a reference for the preparation of other cellulose-based graphitic materials.
Download PDF
Full Article
Catalytic Graphitization of Cellulose Using Nickel as Catalyst
Chao Chen,a,c Kang Sun,a,b,c Ao Wang,a Siqun Wang,c and Jianchun Jiang a,b,*
Microcrystalline cellulose was pyrolyzed and catalytically graphitized under temperatures ranging from 1000 °C to 1600 °C in the presence of nickel (Ni). Optimal conditions for graphitization were determined, along with the structure and conductivity of the resulting samples. The optimal conditions were identified as heating at 1400 °C for 3 h with 3 mmol Ni loading per gram of carbon. The samples obtained had excellent graphitic crystallinity comparable to that of commercial graphite. However, in the absence of Ni loading, no obvious graphitic structure appeared after heating under the same conditions, indicating that Ni was an efficient catalyst for the graphitization of cellulose-based carbon. High-resolution transparent electron microscopy (HRTEM) images showed well-defined graphitic structures of more than 30 layers with slice gaps of 0.340 nm. The conductivities of the samples treated under different temperatures varied from 27 S·cm-1 to 54 S·cm-1 under 20 MPa of pressure, and higher temperatures led to higher conductivity due to better graphitic crystallinity. This study fills an important area of research on the catalytic graphitization of cellulose and provides a reference for the preparation of other cellulose-based graphitic materials.
Keywords: Catalytic graphitization; Cellulose; Nickel; Conductivity
Contact information: a: Research Institute of Forestry New Technology, Chinese Academy of Forestry (CAF), Xiangshan Road, Beijing, l00091, China; b: Institute of Chemical Industry of Forest Products, CAF; National Engineering Laboratory for Biomass Chemical Utilization; Key and Open Lab of Forest Chemical Engineering, SFA; Key Laboratory of Biomass Energy and Material, Jiangsu Province; Nanjing 210042, China; c: Center for Renewable Carbon, University of Tennessee, 2506 Jacob Drive, 37996, Tennessee, USA; *Corresponding author: biomass2016@163.com
INTRODUCTION
Graphite is the most stable form of carbon under standard conditions. It possesses excellent thermal stability and electric conductivity, which have led to its widespread application in such varied areas as electrode materials, refractory materials, and steelmaking (Takeuchi et al. 2014). Besides, graphite is considered to be an ideal precursor for the production of few-layered graphene sheets, which have attracted increasing interest in recent years (Castarlenas et al. 2014; Amiri et al. 2017). However, natural graphite is not a renewable resource. Thus, some kinds of materials, such as petroleum coke or pitch that forms graphitizable carbon precursor after heating, are commonly utilized for the preparation of graphite. Under extremely high temperatures (always above 2800 °C), the disordered microstructures in the precursors can be transformed into highly ordered graphitic structures. This process increases the energy consumption, the cost, and the complexity (Greene et al. 2002; Yoon et al. 2005; Fan and Han 2011; Huang et al. 2013). As a result, the development of new materials and new graphitization methods for the preparation of graphite or graphitic material is desirable, both to replace the use of scarce natural graphite and to find ways to produce commercial graphite at lower temperatures.
As an inexpensive, cheap, biocompatible, abundant, and renewable natural resource, cellulose is an excellent candidate for the production of graphite (Herring et al. 2003; Sevilla and Fuertes 2010; Glatzel et al. 2013; Hoekstra et al. 2015). Thus, further production of few-layered graphene from cellulose is valuable and promising. However, its use presents great challenges because cellulose forms non-graphitizing amorphous carbon after carbonization, which means that the obtained highly disordered and anisotropic carbon cannot be effectively graphitized even at 3000 °C, which is the heat required to graphitize petroleum coke or pitch (Franklin 1951; Dahn et al. 1995). Therefore, the generally used heating method is not suitable for cellulose. In this case, an alternative method catalytic graphitization – may be preferable.
Catalytic graphitization utilizing transition metals, such as Fe, Co, Ni, Mn, or metalloid element B (Kawanoa et al. 1999; Maldonado-Hódar et al. 2000; Sevilla and Fuertes 2006; Wang et al. 2011; Al-Falahi 2014; Wang et al. 2016), has been widely reported as an effective approach for the structural transformation of graphitizable amorphous carbon obtained from pitch (Zhai et al. 2011; Khokhlova et al. 2015), petroleum coke (Gumaste et al. 2012), or polymers such as phenolic resin or polyacrylonitrile (PAN) (Lu et al. 2006; Tzeng 2006; Chen et al. 2008; Yi et al. 2009; Zhao and Song 2011; Wen et al. 2012; Al-Falahi 2014) to obtain clear graphitic structures at temperatures below 1000 °C.
Of the catalysts named above, Ni has been found to be particularly effective and has been generally used to obtain graphitic structures with relatively high crystallinity (Sevilla and Fuertes 2006). Although catalytic graphitization is effective with a number of carbon materials, there has been much less research into the graphitization of cellulose as compared to petroleum coke or pitch, primarily because the structural transformation of cellulose-based non-graphitizable carbon is very difficult even at extremely high temperatures.
Despite the relative lack of research on the subject, there are indications that catalytic graphitization may also be effective for cellulose or high-cellulose-containing materials. For instance, Sevilla and Fuertes prepared graphitic nanostructures with a coil morphology from cellulose-based hydrochar in the presence of Ni at 900 °C (Sevilla and Fuertes 2010). Herring produced well-defined hollow carbon nanospheres with excellent graphitic structure from Ni-doped cellulose that was laser-pyrolyzed at 2250 °C (Herring et al. 2003). Hoekstra pyrolyzed microcrystalline cellulose spheres homogeneously loaded with base metal salts (nitrates of nickel, cobalt, and iron) under N2 atmosphere in the temperature range from 500 to 800 C, and turbostratic graphitic nanoribbons were formed under such conditions (Hoekstra et al. 2015). Liu prepared activated carbon with 3 to 8 ultra-thin graphitic layers using leaves of Zizania latifolia as a starting material at 900 °C with the assistance of Ni; in that study, approximately 30 graphitic layers were observed when the heating temperature was increased to 1000 °C (Liu et al. 2013).
All of these studies involved the graphitization of cellulose at relatively lower temperatures (900 °C) or extremely high temperatures (2250 °C). However, to date, there is still a lack of research on the graphitization of cellulose at temperatures ranging from 1000 °C to 2000 °C. Taking Liu’s report (ibid.) into account, it is possible that better graphitic structures may be more likely to be generated at heating temperatures above 1000 °C and that therefore, an appropriate temperature for obtaining the best graphitic structure may exist between 1000 °C and 2000 °C.
For this investigation, the authors studied the catalytic graphitization of cellulose at temperatures ranging from 1000 °C to 1600 °C, and the optimal conditions for such graphitization were determined in detail. Finally, the conductivity of all the products was recorded to further evaluate the graphitic crystallinity of the cellulose-based carbons that were obtained.
EXPERIMENTAL
Materials
Microcrystalline cellulose (Aladdin Co., Ltd., Shanghai, China) with an average diameter of 50 μm was put in a quartz boat and carbonized under a nitrogen atmosphere at 500 °C for 1 h. The obtained carbon with a yield of about 28%, was impregnated overnight with nickel chloride (NiCl2) solution at 0.5 mmol/g to 5 mmol/g Ni to carbon. The impregnated samples were dehydrated at 120 °C and then calcined in sealed crucibles at temperatures ranging from 1000 °C to 1600 °C for 3 h, during which time the catalytic graphitization reaction occurred. For preventing the oxidation during graphitization, the dried samples were put on the bottom of the crucibles, and then covered with granular coconut shell carbon until the crucibles were full. The coconut shell carbon that was present consumed the majority of oxygen, maintaining a mionectic atmosphere for graphitization and leading to a relatively high final yield (after removing Ni), i.e. 72% , 66%, and 49% for the samples heated at 1000 °C, 1400 °C, and 1600 °C, respectively. After heating, all of the samples were cooled to room temperature, and then they were separated from coconut shell carbon through a sieve. The powder samples were washed with HCl (37%) to remove the metal residue and finally dehydrated at 105 °C overnight. The control sample was prepared in the same process at 1600 °C, but without the addition of Ni. For convenience, the obtained samples were denoted as C-X-Y (C representing cellulose, and X and Y representing temperature and Ni loading amount, respectively).
Methods
The X-ray powder diffraction (XRD) patterns for all samples were obtained on a Bruker D8 instrument (Bruker Corporation, Karlsruher, Germany) operating at 40 kV and 20 mA with Cu Kα radiation (λ = 0.15406 nm) and with measuring angles ranging from 10° to 80°. The determination of the graphitized samples and control sample was carried out on a Thermo DXR532 Raman microscope (Thermo Fisher Scientific Inc., Waltham, USA) in the range of 50 cm-1 to 3500 cm-1. X-ray photoelectron microscopy (XPS) spectra were recorded on a Kratos Axis Ultra DLD spectrometer (Kratos Analytical Co., Ltd., Manchester, UK). For observation of their morphology, the graphitized samples were first well dispersed in alcohol and treated with an ultrasonic processor followed by coating to copper grids. The coated grids were placed into a JEM2100 transparent electron microscope (Japan Electron Optics Laboratory Co., Ltd., Tokyo, Japan) operated at 200 kV accelerating voltage. The conductivity of all graphitized samples was determined on a ST2722SD powder resistivity instrument (Suzhou Jing-Ge Electronic Co., Ltd., Suzhou, China) under pressures varying from 2 MPa to 20 MPa.
RESULTS AND DISCUSSION
XRD Patterns Utilized for Condition Optimization
To ensure the optimal conditions for graphitization, the two most likely factors, heating temperature and nickel loading amount, were evaluated. First, the influence of temperature was determined by XRD, as shown in Fig. 1(a) with a Ni loading amount 3 mmol/g referring to previous reports. The strong, sharp peak at around 26° (C002), which was widely utilized to determine the degree of graphitization (Pudukudy et al. 2016), indicated that obvious graphitization occurred after the heating temperature reached 1000 °C; the higher the heating temperature, the stronger the peaks were at heating temperatures ranging from 1000 °C to 1400 °C. At heating temperatures higher than 1400 °C, the graphitic crystallinity of the samples no longer increased, which can be deduced from the fact that Samples C-1400-3 and C-1600-3 had almost the same XRD patterns. Moreover, as can be seen in Fig. 1(b), C-1400-3’s characteristic peak had almost the same intensity as that of commercial graphite, indicating the existence of graphitic structures with similarly high crystallinity. It is also notable that the characteristic peaks of all prepared samples slightly shifted to lower 2θ angles (26.2°) than graphite (26.5°), which suggested that the obtained graphitic structure had greater interplanar spacing (d002 = 0.340 nm) than graphite (d002 = 0.335 nm), as calculated according to Bragg’s equation. As has been reported, the graphitic structures obtained by catalytic graphitization have the same layer spacing as the products in this study (Sevilla et al. 2007; Anton 2008). Such structures are similar to graphene sheets or may even be regarded as graphene (Anton 2008; Jiang et al. 2010); thus the authors inferred that prepared graphitic structures were more likely to be composed of few layered graphene (Pudukudy et al. 2016). Furthermore, it was clearly seen that the control sample heated at 1600 °C without Ni had no clear characteristic peak for graphitic structure, indicating that the carbon remains an amorphous structure in the absence of Ni as a catalyst. Thus, the authors concluded that the optimal temperature for graphitization was 1400 °C, the temperature at which graphitic structures with high crystallinity similar to graphite were obtained.
Fig. 1. XRD patterns of cellulose treated with Ni at different temperatures (a), compared with XRD patterns of graphite, C-1400-3, and the control sample (b)
Fig. 2. XRD patterns of samples with different Ni loading amounts
It is known that graphitization occurs only in areas where carbon contacts Ni (Sevilla and Fuertes 2010). So the authors surmised that a higher Ni loading amount might also create more contact areas, resulting in greater graphitization. Thus, the optimal loading amount was determined at 1400 °C by XRD, as shown in Fig. 2. From the clearly increased characteristic peak, it was deduced that the degree of graphitization dramatically increased as catalyst amounts were increased from 0.5 to 3 mmol/g. However, when further loading was added up to 5 mmol/g, the crystallinity barely increased, illustrating that 3 mmol/g was an appropriate loading amount.
Raman, XPS, and TEM analysis
Raman curves for the graphitized sample and the control sample treated without Ni are shown in Fig. 3. It is known that for carbon-based materials, there are two distinct peaks in Raman spectra the D band at around 1340 cm-1 and the G band at around 1570 cm-1, which correspond to diamond-like carbon (sp3– bonded) and graphitic carbon (sp2– bonded), respectively. The intensity ratios of the D band and the G band (ID/IG) are widely used to estimate the degree of graphitization (Sevilla et al.2007), and it is known that a smaller ratio of ID/IG produces a better graphitic structure (Liu et al. 2004; Pudukudy and Yaakob 2015). It is calculated that the ID/IG ratio was approximately 0.5 for the sample C-1400-3, indicating that an obvious graphitic structure appeared with the assistance of Ni, while this ratio for the control sample was approximately 1.2, proving that the control sample remained amorphous carbon structure in the absence of Ni (Zhang et al. 2000). Besides, the sharp 2D band lower than G further indicated the existence of few layered graphene sheets (Pudukudy et al.2017).
Fig. 3. Raman spectra for C-1400-3 (a) and the control sample (b)
Fig. 4. XPS spectrum for the sample C-1400-3
Figure 4 shows the XPS spectra of C-1400-3. Remarkably, there were three peaks (dotted areas) attributable to three series of carbons: graphitic carbon, i.e. C=C groups (EB = 284.8 eV), amorphous carbon, i.e. C-C or CHx groups (EB = 285.5 eV), and carbonyl carbon, i.e. C=O groups (EB = 286.9 eV) (Okpalugo et al. 2005). From the ratio of corresponding peaks, the graphitic/amorphous ratio was calculated as 1.80. This ratio was much higher than the reported value of 0.756 (Sevilla and Fuertes 2010), indicating the existence of a higher proportion of graphitic carbon in the sample C-1400-3, which was treated at a higher temperature than reported.
A clear structural difference can be seen by comparing Figs. 5(a) and (b): the graphitic sample C-1400-3 showed clear ribbon-like structures composed of 10 to approximately 30 graphitic layers. In contrast, the control sample had no characteristic structure or shape, indicating that there was no reaction leading to obvious structure transformation in the absence of Ni. High-resolution transparent electron microscopy (HRTEM) micrographs of C-1400-3 and the control sample are shown in Figs. 4(c) and (d), respectively.
