Graphene and carbon nanotubes by CH4 decomposition over CoAl catalysts

Materials Chemistry and Physics 226 (2019) 6–19

Contents lists available at ScienceDirect

Materials Chemistry and Physics journal homepage: www.elsevier.com/locate/matchemphys

Graphene and carbon nanotubes by CH4 decomposition over CoeAl catalysts

T

Camila O. Calgaroa,b, Oscar W. Perez-Lopeza,∗ a b

Department of Chemical Engineering, Federal University of Rio Grande do Sul–UFRGS, Engenheiro Luiz Englert Street s/n, 90040-040, Porto Alegre, RS, Brazil Department of Chemical Engineering, Instituto Federal Sul-Rio-Grandense (IFSul), Campus Pelotas, RS, Brazil

HIGHLIGHTS

GRAPHICAL ABSTRACT

were formed on samples con• CNTs taining CoeAl mixed oxide phase with low Co/Al ratio.

I /I of CNTs increases as the Co/ • The Al ratio and reaction temperature inG

D

creases.

formed on Co100 presented • Carbon Raman spectrum of few-layer graphene.

cobalt carbide phase is related to • The the formation of graphene. O phase and high temperature • Co favored the graphene formation. 3

4

ARTICLE INFO

ABSTRACT

Keywords: Graphene Carbon nanotubes CoeAl oxides Methane decomposition

Catalysts containing Co and Al were prepared to obtain graphene and carbon nanotubes (CNTs) via the catalytic decomposition of CH4. The catalysts were prepared by co-precipitation with molar percentages of Co between 50 and 100%. The effects on the formed carbon of the catalyst composition, reaction temperatures of 500–900 °C, and activation with hydrogen were evaluated. The carbon that formed during the reactions was analyzed using Raman spectroscopy, scanning electron microscopy, transmission electron microscopy, and temperature-programmed oxidation (TPO). The Raman spectra of the samples containing Al were characteristic of CNTs, with the IG/ID ratio increasing as the Co/Al ratio and reaction temperature increased. The Co100 sample presented a Raman spectrum indicative of few-layer graphene. The TPO analysis revealed cobalt carbide oxidation for the Co100 sample, and this phase was related to the formation of graphene. CNTs were favored by CoeAl mixed oxides and reaction temperatures of 500–700 °C. Graphene was favored by the Co3O4 phase and higher reaction temperatures.

1. Introduction C + 2H2) is an Catalytic methane decomposition reaction (CH4 alternative route to producing COX-free hydrogen and nanostructured carbon [1–4]. Nanostructured carbon deposited after the catalytic decomposition of methane is mainly in a filamentous form such as fibers or nanotubes. This nanostructured carbon is of great importance in

∗

nanoscience due to its unique electrical, chemical, and mechanical properties [5]. Graphene is another interesting form of carbon that can be obtained with this process [6]. Graphene has attracted attention due to its theoretically extraordinary electrical and thermal conductivity, structural strength, enormous surface area, and excellent chemical stability [7–9]. There have been many studies on graphene-based flexible and wearable electronics

Corresponding author. E-mail addresses: [email protected] (C.O. Calgaro), [email protected] (O.W. Perez-Lopez).

https://doi.org/10.1016/j.matchemphys.2018.12.094 Received 18 October 2018; Received in revised form 28 December 2018; Accepted 29 December 2018 Available online 02 January 2019 0254-0584/ © 2019 Elsevier B.V. All rights reserved.

Materials Chemistry and Physics 226 (2019) 6–19

C.O. Calgaro, O.W. Perez-Lopez

due to the properties of graphene [10]. The production of graphene with minimal crystalline defects and a controllable number of layers is highly desirable for use in advanced technologies [11]. There are three primary methods to produce graphene nano-sheets: micromechanical exfoliation, epitaxial growth of graphene films, and chemical processing (including oxidation, exfoliation, and reduction). However, these methods are still limited by their high costs [6]. In addition, most commercial applications demand the efficient production of large quantities of nanostructured carbon arrays, as well as high standards of purity, length, and structural quality [12,13]. Alternatively, catalytic methods at atmospheric pressure have recently appeared for graphene formation [6]. Chemical vapor deposition (CVD) is a powerful process to produce graphene [14]. Tu et al. [8] and by Mu et al. [15] achieved the formation of few-layer graphene using the CVD method from mixtures of CH4 and H2 with Cu sheets as a substrate. Graphene has been obtained via the catalytic decomposition of CH4 over Co catalysts in reactions at 1000 °C through direct reduction of the catalyst with CH4 and its production was related to the formation of an intermediary phase (cobalt carbide) [6,16]. Weibel et al. [17] produced 2–8 layers of graphene on MgO powder grains via the CVD method in CH4/Ar atmosphere. The graphene layers were located along the MgO grain boundaries, as opposed to dispersed discrete particles or flakes [17]. Awadallah et al. [18] obtained fewlayered graphene nano-platelets on the surface of reduced metallic sheets (Fe, Co, and Ni) via methane decomposition. Few-layered graphene nano-platelets were obtained over a metallic Co catalyst, and a multilayer material was obtained over Fe and Ni [18]. For graphene/ carbon nanotube hybrids on bifunctional catalysts, CNTs preferred to grow on metal catalyst nanoparticles and graphene was favored to deposit on the surface of Cu or metal oxide [19]. However, further studies are needed due to the limited understanding of the growth mechanism of graphene, the non-linear kinetics, and practical design for the production of graphene or carbon nanotubes (CNTs) [19]. The catalysts generally used for methane decomposition include transition metals from group VIII [16,20,21]. Ni, Fe, and Co catalysts stand out because of their advantages like low cost, better activity and stability [22]. Co nanoparticles have been employed as catalytically active sites for the growth of CNTs [23]. Cobalt catalysts with different supports (alumina, silica, and magnesia) were studied in methane conversion. The carbon capacity was higher for Co-alumina catalyst [24]. Co/SiO2 catalysts have been used to develop a kinetic model of CH4 decomposition that describes the stable catalyst activity associated with filamentous carbon formation and the catalyst deactivation associated with the formation of encapsulating carbon [25]. The catalyst composition may influence the characteristics of the CNTs formed. Some studies have focused on theoretically describing the carbon-metal interaction [26]. Zhang et al. [27] found that NiMo alloy with very small size was responsible for the growth of thin-walled CNTs. The Mo phase was related to the formation of large-diameter thick-walled CNTs via the decomposition of CH4. In addition, changes in the gas composition during CVD were studied during the growth of CNTs, and the degree of methane decomposition (CeH bond dissociation) was closely related to the nanotube yield [28]. The operating conditions of the reactional system are linked to the purity of produced CNTs [29]. The reaction temperature may also influence the type of carbon formed. Simon et al. [30] observed that when they increased the reaction temperature, a structural transition from carbon nanofibers to CNTs occurred on ceramic Al2O3 substrates with palladium as a catalyst in the CVD technique. Temperature rather than the reaction time or flow rate is considered to be a key parameter that controls the wall thickness of CNTs. Low temperatures lead to the formation of multiwalled CNTs (MWCNTs), and high temperatures favor the formation of nanotubes with fewer walls [31]. Gong et al. [32] observed that annealing CNTs at high temperatures removed some of the impurities in the formed material. Pudukudy et al. [33] observed that increasing the reaction temperature from 700 to 900 °C during CH4

decomposition increases the crystallinity and the degree of graphitization of the nanocarbon over unsupported mesoporous nickel ferrites. However, methane decomposition requires further study because the carbon produced is the primary cause of catalyst deactivation in most cases [21]. Deactivation mainly occurs via the deposition of amorphous carbon on the surface of the catalyst [21,34] and when carbon encapsulates the metal particles [21,34,35]. CoeAl catalysts derived from hydrotalcites exhibit great catalytic performance for methane decomposition to hydrogen production [21,36,37]. Hydrotalcites are commonly prepared by co-precipitation [21,38]. The thermal treatment of these materials typically yields mixed oxides with high thermal stability and large surface areas [21,39,40]. Moreover, well-dispersed metallic particles can be obtained after reduction [21,41,42]. These characteristics are also important for the synthesis of CNTs because their parameters depend on the crystallite size of the active species of the catalyst [5]. In the present study, co-precipitated CoeAl catalysts with molar percentages of Co between 50 and 100% were prepared to obtain CNTs and graphene via the catalytic decomposition of CH4. While most studies for the production of graphene through the CVD method use catalyst sheets such as Cu, this study used catalyst particles that can be used for co-production of hydrogen and structured carbon. The effect of the catalyst composition, reaction temperature, and catalyst activation were evaluated on the characteristics of carbon produced. This study builds on previous studies by expanding the investigation of CoeAl composition and reaction conditions on the carbon characteristics [37]. 2. Material and methods 2.1. Catalyst preparation The catalysts were prepared via the coprecipitation method [34,43] with molar percentages of Co between 50 and 100% and Al between 0 and 50%, and was previously described by Calgaro and Perez-Lopez [21]. A solution containing 1 M of the metal precursors Co(NO3)26H2O and Al(NO3)39H2O was continuously added to a stirred reactor with an alkaline solution of Na2CO3 (2 M). The temperature and pH were kept constant at 50 °C and 8.0 ± 0.1, respectively. The precipitate was maintained under agitation for 1 h and then vacuum filtered, washed with deionized water, and dried over night at 80 °C. The obtained material was calcined in a tubular quartz reactor with a synthetic air flow (50 mL/min) at 600 °C for a period of 6 h using a heating rate of 10 °C/min. The particle size of the samples used in the catalytic tests was between 0.355 and 0.500 mm [21]. 2.2. Catalyst characterization Thermogravimetric (TG) analysis and differential thermal analysis (DTA) were carried out in a thermobalance (SDT-Q600, TA Instruments). First, 10 mg of fresh catalyst was heated to 800 °C at 10 °C/min under air flow at 100 mL/min. Temperature-programmed reduction (TPR) analyses were performed with calcined samples (0.1 g) at a heating rate of 10 °C/min up to 850 °C under a mixture of H2:N2 at a ratio of 10:90 with a total flow rate of 30 mL/min [21]. Raman spectroscopy (Raman spectrometer: Horiba Scientific - IHR550) was used to identify the phases in the catalysts after calcination. An excitation laser with a wavelength of 531.1 nm was supplied through an optical fiber. The average crystallite size was determined in previous work [21] through X-ray diffraction using the Scherer equation at 2θ = 36.9° for fresh samples (Co3O4), and at 2θ = 44.5° for reduced samples (Co°). 2.3. Carbon production The equipment and operational conditions were described in previous work by Calgaro and Perez-Lopez [21]. Methane decomposition reactions for carbon production were carried out at atmospheric 7

Materials Chemistry and Physics 226 (2019) 6–19

C.O. Calgaro, O.W. Perez-Lopez

pressure in a tubular quartz reactor, which was heated in a resistive electric oven. Quartz wool was used to support the catalyst bed. The temperature was measured using a K-type thermocouple, and the gas flow rates were established by digital mass-flow controllers (Sierra Instruments). The flow rate used in the tests was 100 mL/min at a ratio of 1:9 of CH4:N2. The pressure was monitored using a manometer. The reaction runs were carried out with approximately 0.1 g of catalyst. The reactor was heated to the desired temperature at a heating rate of 10 °C/min under a flow of a mixture of CH4 and N2 [21]. The runs were carried out at a constant temperature (500, 600, and 700 °C) as a function of time on stream. Runs with the Co100 sample were carried out at constant temperatures of 500, 700, 800, and 900 °C. The effect of H2 reduction on samples containing Al was evaluated. The samples were activated in situ at 700 °C under H2 flow. The heating was carried out at 10 °C/min under a mixture of H2 and N2 at a ratio of 1:9 at 100 mL/min. The samples were maintained at the reduction temperature for 1 h, after which the H2 flow was replaced with CH4 flow, and the reactions were started [21]. 2.4. Carbon characterization The carbon formed in the catalytic decomposition of CH4 was characterized using Raman spectroscopy, transmission electron microscopy (TEM), and temperature-programmed oxidation (TPO). The Raman spectroscopy was performed with a Raman spectrometer (Horiba Scientific - IHR550) and an excitation laser with a wavelength of 531.1 nm supplied through an optical fiber. The TEM analyses were performed using a JEOL JEM 1200 EX II electron microscope at 120 kV. Some samples and the carbon produced on the Co100 sample were investigated using high-resolution TEM (HRTEM) (TEM, FEI Tecnai G2 T20). The samples were prepared by dispersing carbon in isopropanol via ultrasonication and placing a droplet onto the surface of a copper grid covered with amorphous carbon. The carbon obtained at 500 and 600 °C was examined using fieldemission scanning electron microscopy (SEM) analysis (FEI Inspect F50, in a secondary electron beam). The samples were coated with a thin gold film prior to the SEM. The TPO analyses were carried out in a thermobalance (TA Instruments, SDT-Q600) from room temperature up to 800 °C at 10 °C/min under air flow at 100 mL/min [21]. X-ray photoelectron spectroscopy (XPS) was used to investigate some samples before and after reaction. XPS was performed in an Omicron-SPHERA station using Al Kα radiation (1486.6 eV). The anode was operated at 225 W (15 kV, 15 mA). The Co2p and C1s region was recorded with a higher resolution. X-ray diffraction (XRD) patterns were obtained in previous work [21] for fresh and spent samples using a BRUKER D2-Phaser diffractometer with Cu-Ka radiation at 30 kV and 10 mA.

Fig. 1. DTG of the uncalcined samples.

that between 100 and 300 °C, the Co100 and Co90 samples exhibit 3 and 4 wt loss peaks, respectively. The other samples exhibited only 2 peaks in this temperature range. A shift of the second peak to higher temperatures occurred when the Co/Al ratio decreased from 4 to 1 (i.e., as the Al fraction increased from 20 to 50%). This finding is possibly due to the higher amount of Al3+ ions present in these samples, which influenced the formation of the Co spinel structure. This trend was observed in peaks above 200 °C, where the formation of spinel occurred [38]. Weight loss occurred up to 200 °C (the first peak in Fig. 1) due to the removal of water molecules. The layered structure collapsed after the removal of physisorbed water, which was completed at about 100 °C. The collapse is related to the elimination of interlayer water with maximum rates occurring around 125 °C and 190 °C, depending on the sample. The weight loss between 200 and 300 °C (second peak in Fig. 1) is linked to the dehydroxylation of layers and the decomposition of interlayer carbonate anions [38,44]. No other event was observed above 400 °C, which indicates that stable CoeAl oxide was obtained at this temperature [40,45]. Temperature of 600 °C was adopted for the thermal treatment of hydrotalcites based on previous studies [34,39,43]. According the previous work of Calgaro and Perez-Lopez [21] the TPR results showed that pure cobalt (Co100) presented only a single reduction peak at temperatures lower than 500 °C, while the samples containing Al presented a second reduction peak at high temperatures. This indicates that the presence of Al+3 ions in the samples made it difficult to reduce the Co ions. The first peak corresponds to the reduction of Co3O4, and the second peak is related to the reduction of CoeAl mixed oxides. These results were also observed in another previous studies [34,37,39]. Table 2 lists the reduction areas from the TPR analysis. An increase in Co resulted in an increase in the total area of reduction and the amount of Co3O4 [21]. In contrast, the reduction in the area of CoeAl mixed oxides increased from Co90 to Co70 (from 9.0 to 2.3 in the CoeAl molar ratio). The Co70, Co60, and Co50 samples presented similar area% of CoeAl mixed oxides, which demonstrates that mixed oxides were obtained preferentially in this composition range. Fig. 2 shows the Raman spectra of the calcined catalysts. The Raman spectrum of the Co100 sample shows 5 bands at 192, 479, 519, 617,

3. Results and discussion 3.1. Catalyst characterization The sample names and respective Co/Al ratios are listed in Table 1. The samples are identified by the molar percentage of Co. The thermogravimetric analyses of the uncalcined samples are presented in Fig. 1. The derivative thermogravimetry (DTG) curves in Fig. 1 show Table 1 Sample names and Co/Al nominal ratios. Sample Co100 Co90 Co80 Co70 Co60 Co50

Al amount (mol%) 0 10 20 30 40 50

Molar ratio of Co/Al ∞ 9.0 4.0 2.3 1.5 1.0

8

Materials Chemistry and Physics 226 (2019) 6–19

C.O. Calgaro, O.W. Perez-Lopez

Table 2 Phases obtained from TPR profiles and respective reduction areas. Sample

Area of Co3O4 (a. u.)

Area of CoeAl mixed oxides (a. u.)

Total area (a. u.)

Co100 Co90 Co80 Co70 Co60 Co50

7843 (100%) 3464 (55%) 2303 (40%) 972 (21%) 805 (22%) 485 (17%)

– 2857 3417 3603 2818 2418

7843 6321 5720 4575 3623 2903

(45%) (60%) (79%) (78%) (83%)

Fig. 3. Average crystallite size of fresh and reduced samples.

makes it difficult to identify and differentiate these mixed oxide phases using Raman spectra. Fig. 2 shows that the Co90 and Co80 samples presented similar Raman spectra. Furthermore, in the Raman spectra of the Co70, Co60, and Co50 samples, the band with the largest Raman shift was similar in shape and intensity. These results are in agreement with the TPR analysis of these samples (Table 2). Fig. 3 shows the average crystallite size determined through XRD for fresh samples and after H2 reduction at 700 °C [21]. From Fig. 3, it was observed that the average crystallite size increases as the amount of Co in the sample increases (Co50 < Co60 < Co70 < Co80 < Co90 < Co100) for fresh samples and for after H2 reduction. The largest increase in crystallite size occurred from Co80 to Co100 sample and is due to the increase of percentage of Co3O4 phase in the fresh catalyst (Table 2).

Fig. 2. Raman spectra of the calcined catalysts before the reactions.

and 688 cm−1, which are similar to those reported for Co3O4 [46,47]. The bands at 192, 519, and 617 cm−1 can be attributed to the vibrational mode F2g, and the bands at 479 and 688 cm−1 can be attributed to the vibrational modes Eg and A1g, respectively [47]. The Raman spectra of the other samples also presented these five bands at similar shifts. As the molar percentage of Al increased, the intensity of the last peak decreased between 650 and 750 cm−1, and broadening of all peaks occurred. The peaks corresponding to the Co3O4 phase agree with the TPR analysis (Table 2) in that the phase was noted in all of the samples. These characteristic bands result from the crystallinity of Co3O4 spinel with Co2+ and Co3+ located in the tetrahedral and octahedral sites, respectively [48]. For the Co50, Co60, Co70, and Co80 samples, the Raman spectra show an intermediate peak between 571 and 593 cm−1. The bands around 580 cm−1 are due to the vibration of the AleO bond and are attributed to the first symmetrical species F2g [49]. According to Mwenesongole [49], Co2+Co3+AlO4 spinel exhibits Raman displacement peaks at 197, 481, 521, 619, and 688 cm−1. The highest intensity peaks occurred at 197 and 688 cm-1 Álvarez-Docio et al. [50] obtained Raman shift peaks at 204, 417, 513, 646, and 775 cm−1 for the CoAl2O4 spinel, and the highest intensity peaks occurred at 204 at 513 cm−1. The samples containing Al in this study did not exhibit peaks between 400 and 420 cm−1 or 750 and 790 cm−1 corresponding to the CoAl2O4 phase. However, the Raman spectra of the samples were more similar to those of the Co2AlO4 phase. In a previous study [21], more than one phase of CoeAl mixed oxide was identified in CoeAl samples by deconvolution of the TPR profiles. This

3.2. Characterization of the obtained carbon 3.2.1. Carbon characterization by Raman spectroscopy The Raman spectra of the carbon that formed in the reactions at 500, 600, and 700 °C are shown in Figs. 4–6, respectively. These figures also present the intensity ratios of the G and D bands and of the 2D and G bands. The first ratio (G/D) provides information on the structural quality of nanotubes [51], and the second ratio (2D/G) is related to the number of graphene layers [8]. The Raman spectra show that all the samples presented the three characteristic bands: D, G, and 2D [6]. According to Soldano et al. [52], the Raman spectra of carbon materials show similar characteristics in the region of 800–2000 cm−1. The G band occurs around 1560 cm−1 and corresponds to the E2g stretching mode of graphite. The D band appears at 1360 cm−1 and corresponds to structural defects or imperfections in the graphite layers [6,52,53]. The 2D band appears around 2700 cm−1 and is the most characteristic band of graphene. The shape, position, and relative intensity of the band depend on the number of graphene layers [52]. For single-walled CNTs (SWNTs), the shape of the 2D peak is similar to that measured for graphene, and they have a similar position and full width at half maximum [54]. Figs. 4, Figure 5, and Fig. 6 show that increasing the temperature from 500 to 700 °C decreases the intensity of peak D and consequently increases the IG/ID ratio. For most samples, lower IG/ID ratios persist with lower reaction temperatures. In addition, lower IG/ID ratios occur when the amount of Co present is decreased. Therefore, high reaction temperatures and catalysts with high levels of Co yield higher IG/ID 9

Materials Chemistry and Physics 226 (2019) 6–19

C.O. Calgaro, O.W. Perez-Lopez

Fig. 6. Raman spectra of the carbon obtained in the reactions carried out at 700 °C.

Fig. 4. Raman spectra of the carbon obtained in the reactions carried out at 500 °C.

Fig. 7. Raman spectra of the carbon obtained in the reactions with the Co100 sample.

defects in the carbon produced on the Co100 sample. In addition, the sample exhibited I2D/IG ratios of 0.54 and 0.41 after reactions at 500 and 700 °C, respectively, with a 2D band at approximately 2689 cm−1. These results suggest the presence of only a few layers of graphene because the Raman spectrum is very similar to that identified by Tu et al. [8] for trilayer graphene and by Liu et al. [55] for a graphene with less than five layers. According to Tu et al. [8], when the number of graphene layers increases from monolayer to multilayer, the 2D peak widens and the G peak becomes more pronounced. In addition, there is a slight blue shift in the 2D band (2699–2686 cm−1). Moreover, Tu et al. [8] found that trilayer graphene had a Raman spectrum with 0.4 ≤ I2D/IG < 1. The Co90 sample in the reaction at 700 °C (Fig. 6) exhibited similar Raman

Fig. 5. Raman spectra of the carbon obtained in the reactions carried out at 600 °C.

ratios. Larger differences in IG/ID ratios between the samples were observed for the reactions at 700 °C. In addition to the D, G, and 2D bands, Figs. 4 and 5 show the presence of initial peaks between 192 and 688 cm−1 corresponding to the phases containing Co. Fig. 7 shows the Raman spectra for the Co100 sample after reactions occurring between 500 and 900 °C. The Raman spectra at 500 and 700 °C for these samples were very different from those obtained for the samples containing Al in that the D band had a much lower intensity than the G and 2D bands. This observation implied a lower level of 10

Materials Chemistry and Physics 226 (2019) 6–19

C.O. Calgaro, O.W. Perez-Lopez

Fig. 8. Scanning electron microscopy images of carbon obtained in the reactions at 600 °C: (a) Co90, (b) Co80, (c) Co70, (d) Co60, and (e) Co50.

spectra to those of the Co100 sample after reactions at 500 and 700 °C. However, the Co90 sample had a larger D band and thus a lower IG/ID ratio. This finding may be related to the formation of graphitic material with more structural defects. Due to the possible formation of graphene on Co100, the sample was used for reactions at higher temperatures (800 and 900 °C). The intensity of the D band was negligible for the Co100 sample at 500–900 °C. Fig. 7 shows that the I2D/IG ratio increased with increasing reaction temperature between 700 and 900 °C. The highest I2D/IG ratio of 1.35 was obtained at a reaction temperature of 900 °C. I2D/IG ratios of 1.23 and 1.35, were obtained at 800 and 900 °C, respectively, which indicated the formation of graphene with fewer layers than at lower temperatures. Based on work by Tu et al. [8], an I2D/IG ratio in the range of 1 < I2D/IG < 2 corresponds to bilayer graphene. Wimalananda et al. [56] also identified bilayer graphene using Raman analysis and noted I2D/IG ratios of 1.199, 1.395, and 1.510, which are similar to those obtained at 800 and 900 °C in this work. These I2D/IG results demonstrate the influence of the reaction temperature on the production of

graphene in the Co100 sample. In addition to the influence of the reaction temperature on the formation of graphene layers, the catalyst phase is also essential because the characteristic Raman spectrum of graphene was obtained for only the Co100 sample, which contains only the Co3O4 phase. The Raman spectrum of the Co90 sample was more similar to the Raman spectrum of Co100. The Co90 sample had a larger proportion of the Co3O4 phase relative to CoeAl mixed oxides than the other samples containing Al (Table 2). Higher Co/Al ratios in the catalyst are associated with higher probabilities of obtaining graphene. Jana et al. [6] obtained graphene nanosheets via CH4 decomposition on bulk Co3O4 catalysts when the catalyst activation was carried out with CH4 at a reaction temperature of 1000 °C. They did not detect the carbide phase, which consists of a thin CoeC interface. However, the presence of this phase was not completely ruled out. Jana et al. [6] believe that the carbide phase may play a role in graphene formation. Some studies have demonstrated the formation of cobalt carbides such as Co2C during activation for Fischer-Tropsch synthesis [57,58]. Cobalt carbides are unstable and easily decompose to metallic cobalt 11

Materials Chemistry and Physics 226 (2019) 6–19

C.O. Calgaro, O.W. Perez-Lopez

Fig. 9. Transmission electron microscopy images of carbon obtained in the reactions at 700 °C: (a, b) Co90, (c, d) Co80, (e, f) Co70, (g, h) Co60, and (i, j) Co50.

12

Materials Chemistry and Physics 226 (2019) 6–19

C.O. Calgaro, O.W. Perez-Lopez

Fig. 10. Transmission electron microscopy images of carbon obtained with the Co100 sample at: (a) 500 °C, (b) 700 °C, (c) 800 °C, and (d) 900 °C.

and carbon. Therefore, they have rarely been observed using ex situ techniques [59], which makes it difficult to identify them using the characterization techniques employed in this study. Therefore, cobalt carbides may represent an intermediate path for graphene formation.

MWCNT with a diameter of around 30 nm and an inner diameter of 10 nm (Fig. 9 (b)). The Co60, Co70, and Co80 samples exhibited CNTs with different diameters (Fig. 9 (c), (e) and (g)). This indicates sintering of the catalyst during the reaction at 700 °C, which has been noted previously [21]. The high-resolution TEM images of these samples show that the range of nanotube diameters was approximately 8–15 nm and that the nanotubes are mostly multi-walled (Fig. 9 (d), (f), and (h)). However, nanotubes with few walls also formed, as evidenced by the images of the Co50 and Co60 samples (Fig. 9 (j) and (h)). The Co50 sample exhibited CNTs that were more uniform in diameter (Fig. 9 (i)). The diameter of the CNTs in the Co50 sample was approximately 12 nm in the high-resolution TEM image (Fig. 9 (j)). Fig. 10 shows TEM images of the carbon obtained at different reaction temperatures with the Co100 sample. The image taken after the reaction at 500 °C shows the edge of the graphene. However, the layers of the graphene are not aligned (Fig. 10 (a)), and the image from the reaction at 700 °C (Fig. 10 (b)) shows layers that are aligned but difficult to identify due to the excess of sample. The Co100 sample produced some layers of graphene at temperatures of 800 and 900 °C, which were already identified in the Raman spectrum. But TEM analysis showed more than three layers because of the sample prepare before TEM analysis presenting excess of sample. The distance between the carbon layers identified in Fig. 10(c) and (d) is about 0.34 nm, and the layers have an orientation of roughly 110°. This estimated interlayer spacing is similar to that obtained by Reina et al. [60] for graphene prepared on polycrystalline Ni films (about 0.35 nm) and by Gubernat et al. [61] for graphene ribbons through SiC (varied from 0.350 to 0.390 nm). Fig. 10(c) and (d) indicate that there are sheets of graphene on top of one another.

3.2.2. Carbon characterization by SEM and TEM Fig. 8 shows SEM images of the samples after reaction at 600 °C, and Fig. S1 (please check supplementary material) after reaction at 500 °C. The Co50, Co60, Co70, and Co80 samples exhibited large numbers of carbon filaments at both reaction temperatures, but Co90 presented different SEM images. The Co90 sample exhibited very few carbon filaments and wider carbon structures with a straight surface. The differences in the type of carbon formed are due to the composition of the catalysts and differences in the Co phases of the samples. The Co90 sample mainly consisted of the Co3O4 phase and yielded carbon structures similar to those formed by the pure Co sample (Co100). However, the Co50, Co60, Co70, and Co80 samples were mostly formed of CoeAl mixed oxides phases, as shown in Table 2, which favor the formation of carbon filaments. The CNTs had diameters between 10 and 40 nm for the Co50, Co60, Co70, and Co80 (SEM images; Fig. 8). In addition, the longest nanotubes were produced at 600 °C for the Co50 and Co60 samples (Fig. 8 (d) and (e)). This may be related to the higher stability of these samples at this reaction temperature. The carbon produced on samples at 700 °C was evaluated using TEM analysis (Fig. 9). The highest IG/ID ratios and the largest differences in Raman spectra between the samples were observed at this temperature. The Co90 sample exhibited few CNTs compared with the other samples (highlighted with circles). In addition, TEM analysis of this sample (Fig. 9 (a)) revealed greater sintering of the catalyst during the reaction (dark regions). A high-resolution image of the Co90 sample reveals a 13

Materials Chemistry and Physics 226 (2019) 6–19

C.O. Calgaro, O.W. Perez-Lopez

Fig. 12. TPOs of the Co100 sample after reactions: (a) weight variation and (b) DTA.

The other samples containing larger amounts of Al exhibited significant weight loss, which was more pronounced between 400 and 550 °C and mainly due to the oxidation of the CNTs produced during the reactions. This situation was confirmed in the SEM and TEM images, which showed a significant amount of CNTs for these samples (Fig. 8, S1, and Fig. 9). The largest weight loss occurred for the reactions carried out at 500 °C and 600 °C due to the increased formation of CNTs (Fig. S2 and Fig. S3), which was related to the high activity and stability of the catalysts. Based on the DTA results after the reactions at 500 °C, 600 °C, and 700 °C (Fig. 11(b), S2(b), and S3(b)), two carbon oxidation peaks around 436 °C and 490 °C were observed for the Co50, Co60, Co70, and Co80 samples. These peaks were lacking in the Co50 sample processed at 700 °C. These differences as a function of oxidation temperature are likely related to the number of walls of nanotubes, and the TEM images of these samples revealed the formation of nanotubes with different numbers of walls (Fig. 9). The Co50 sample showed only a peak of carbon oxidation and was characterized by more homogeneous nanotube formation, which was also observed in the TEM images for this sample (Fig. 9 (i) and (j)). The TPO analysis of Co100 revealed a very different TPO profile than that of the other samples (Fig. 12). A small degree of weight loss occurred up to 425 °C (Fig. 12 (a)) due to the oxidation of amorphous carbon in the reactions at 500 and 700 °C. The reaction at 500 °C was

Fig. 11. TPOs of samples after reactions at 700 °C: (a) weight variation and (b) DTA.

3.2.3. Carbon characterization using TPO analysis Figs. 11 and 12, and Figs. S2 and S3 (please check supplementary material) show the TPO results for all of the samples after the reactions. The TPO profiles can be used to identify the different types of carbon in the catalysts because they oxidize at different temperatures [62–64]. The different types of carbon produced over the catalysts containing Co can be classified into three temperature ranges: 150–400 °C for amorphous carbon, 400–550 °C for CNTs, and 550 °C or higher for graphitic carbon [21,34,37,39]. The Co90 sample exhibited limited weight loss at temperature below 300 °C and a more significant weight gain at temperatures higher than 300 °C. This weight gain corresponds to the oxidation of metallic Co particles, which remained after the reaction [34,43]. We observed limited weight loss at 400–550 °C for this sample because only a little carbon formed during the reactions (Fig. 11 (a), S2 (a), and S3 (a)). This finding is in agreement with the SEM and TEM images, which show very limited formation of CNTs (Fig. 8, S1, and Fig. 9). The first DTA peak in Fig. 11 (b), S2 (b), and S3 (b) for the Co90 sample corresponds to the oxidation of metallic Co, and the second peak corresponds to carbon oxidation. 14

Materials Chemistry and Physics 226 (2019) 6–19

C.O. Calgaro, O.W. Perez-Lopez

with the Co100 sample played a fundamental role in the formation of graphene. A peak between 450 and 650 °C was observed in the DTA results. This peak increased in intensity as the reaction temperature increased (Fig. 12 (b)). In addition, the oxidation temperature also increased with the reaction temperature, indicating that more structured carbon was obtained at higher reaction temperatures. However, for the reactions at 500, 700, and 800 °C, the DTA peak was mainly related to the oxidation of Co°. For the reaction at 900 °C, the DTA peak was associated with carbon oxidation. Therefore, the reaction at 900 °C resulted in a higher peak intensity and higher oxidation temperature associated with graphene formation. It is worth noting that the weight gain above 600 °C occurs with little thermal effect, which indicates that the oxidation of the cobalt carbide occurred with a smaller release of energy. 3.2.4. Influence of reduction with hydrogen Fig. 13 shows the Raman spectrum of the CoeAl samples after reduction and reaction at 700 °C. All of the samples exhibit the characteristic carbon bands D, G, and 2D. All of the H2-reduced samples had IG/ID ratios greater than one. The Co90 sample exhibited the highest IG/ ID ratio among the samples without reduction (Fig. 6). The effect of the catalyst reduction with hydrogen before the reaction was compared at 700 °C for the Co60 sample. The IG/ID ratio was higher for the carbon obtained in the reaction with the H2-reduced sample than that without reduction. This indicates a smaller amount of defects in the formed carbon after H2 reduction (Fig. 14). The TEM images of these two samples (Fig. 15) revealed greater uniformity of the CNTs that formed on the H2-reduced sample, which is in agreement with the Raman analysis. These differences may be related to the activation of the catalyst. During the reaction without reduction, heating with CH4 produced carbon at different temperatures, which may have influenced the uniformity of the nanotubes. Fig. 15 shows that the two samples formed nanotubes with diameters of approximately 10 nm. The TPO analysis of the carbon obtained with the Co60 sample previously H2-reduced and without reduction showed a similar amount of produced carbon (Fig. 16). Both samples presented roughly 40% weight loss. Only the DTA results showed a small difference in the temperature of the main peak corresponding to carbon oxidation. The temperature was 433 °C for the H2-reduced Co60 sample and 449 °C for the Co60 sample without reduction. This small difference in the carbon oxidation temperature agrees with the differences observed in the TEM images (Fig. 15). The presence of nanotubes with larger diameters and possibly more layers can lead to a slightly higher oxidation temperature for the Co60 sample without reduction. The Raman, TEM, and TPO analyses of the Co60 sample with and without reduction with hydrogen (Figs. 14–16) revealed that the reduction step with hydrogen influenced the quality and homogeneity of the CNTs.

Fig. 13. Raman spectra of the carbon obtained in the reactions carried out at 700 °C after H2 reduction.

3.2.5. Comparative analysis of some samples before and after the reaction XPS spectra of Co90 and Co60 fresh samples and after reaction at 700 °C are shown in Fig. 17. All the samples presented peaks at around 285 eV and 536.0 eV ascribed to C1s and O1s, respectively [65]. Peaks related to carbon are expected after the reactions due to the formation of graphene and/or carbon nanotubes. Carbon peaks exhibited for the fresh samples were possibly due to CO2 adsorption of the environment. Both fresh samples and Co90 after reaction presented peaks between 770 and 810 eV ascribed to Co2p [65–67], and peaks between 65 and 90 eV ascribed to Al2p [68]. However, the Co60 sample after the reaction showed no peaks related to Co2p and Al2p, possibly due to the large deposition of carbon nanotubes on the surface of the catalyst as already seen by SEM images (Fig. 8). The Co2p XPS spectra (Fig. 18) presented two major peaks at around 788 and 804 eV, attributed to the typical Co2p3/2 and Co2p1/2 orbitals, respectively. The major peaks showed binding energy slightly higher than those reported in the literature for Co3O4, CoAl2O4, and Co2AlO4 phases [66,68]. These phases were identified in the previous

Fig. 14. Comparative Raman spectra of the Co60 sample with and without reduction with H2 after reaction at 700 °C.

characterized by a small degree of weight loss from 500 °C to 800 °C, which is related to the oxidation of structured carbon. The reaction at 700 °C was associated with a small degree of weight loss between 600 and 700 °C, and the sample after the reaction at 900 °C exhibited noticeable weight loss between 500 and 600 °C. These weight losses can be attributed to the graphene-like carbon oxidation identified by the Raman and TEM analysis (Figs. 7 and 10). In addition, the Co100 sample exhibited weight gain after all of the reactions. The largest portion of the weight gain was due to metallic Co oxidation between 400 and 600 °C, and another part was due to cobalt carbide oxidation above 600 °C. The weight gain for the reaction at 900 °C occurred only above 600 °C, indicating that under these conditions, the Co100 catalyst was predominantly in the form of cobalt carbide. The cobalt carbide that formed during the reaction at 900 °C 15

Materials Chemistry and Physics 226 (2019) 6–19

C.O. Calgaro, O.W. Perez-Lopez

Fig. 15. Transmission electron microscopy images of the Co60 sample after reaction at 700 °C: (a) sample without reduction and (b) the H2-reduced sample.

work for Co90 and Co60 samples [21]. Through the Gaussian deconvolution, the 2p3/2 peak can be separated into Co2+ and Co3+, proving the existence of the 3 phases (Co3O4, CoAl2O4, and Co2AlO4) identified by TPR analysis previously [21]. The Co90 sample after reaction also presented a deconvolution peak ascribed to Co0 [69], what was expected since this analysis was made after the reaction. The other peaks that are not identified in Fig. 20 correspond to satellite peaks [66,68]. The C1s XPS spectra (Fig. 19) showed a main peak with different bending energy for all samples. Co90 and Co60 fresh samples showed a peak at 292 eV which can be attributed to OeC=O [65] and related to CO2 contamination. Co90 after reaction presented a main peak at 290 eV. From Gaussian deconvolution for Co90 (Fig. 20 (a)) it can be see a small peak at 286 eV, which although displaced could be related to CeC, because that sample produced low carbon and a peak at 290.6 related to OeC=O [65]. However, for Co60 sample (Fig. 20 (b)), which produced very much carbon, the main peak was related to CeC at 285 eV and a small peak ascribed to CeO=C at 288 eV. For Co60 sample previously reduced and after reaction (Fig. 20 (c)), which also produced great amount of carbon, the deconvolution of XPS spectra presented two peaks at 284 eV and 285 eV related to CeC, corresponding to sp2 and sp3 hybridized states, respectively [70]. The small peak at 288 eV was ascribed to CeO=C. Fig. 21 shows the XRD pattern of Co60 in three steps: fresh sample, after H2 reduction, and after reaction at 700 °C. Co60 fresh sample exhibited peaks at 31.3, 36.9, 44.8, 55.8, 59.6, and 65.3°, which can be ascribed to the Co3O4 or CoeAl mixed oxides phases [21]. Since Co60 fresh sample presented more than one phase and due to these phases presenting similar crystallographic parameters, it is difficult to distinguish them by XRD [21]. After H2 reduction at 700 °C, the Co60 sample presented two main peaks ascribed to Co° at 44.5 and 51.8° [21]. These same peaks related to Co° were exhibited for Co60 after reaction, for both unreduced and H2-reduced sample. In addition, the peak at 26.5° corresponding to carbon [21] was exhibited for both Co60 samples after reaction.

Fig. 16. TPO analysis of the Co60 sample after reaction at 700 °C without reduction and with H2 reduction before the reaction.

4. Conclusions This study investigated carbon produced by methane decomposition over CoeAl catalysts. The amount of the Co3O4 bulk phase strongly decreased for samples with Co/Al ratios less than 4. The formation of CoeAl mixed oxides occurred preferentially in these samples. The Co3O4 phase reduces at temperatures below 500 °C, and CoeAl mixed oxides reduce above 650 °C. The Raman spectra of the samples containing Co and Al after the

Fig. 17. XPS wide scan spectra of Co90 and Co60 fresh samples and after reactions at 700 °C.

16

Materials Chemistry and Physics 226 (2019) 6–19

C.O. Calgaro, O.W. Perez-Lopez

Fig. 18. Co2p XPS spectra of catalyst before and after reaction: a) Co90 fresh, b) Co60 fresh, and c) Co90 after reaction at 700 °C.

Fig. 21. XRD analysis of Co60 sample – fresh, H2 reduced and after reaction at 700 °C. Fig. 19. C1s XPS spectra of Co90 and Co60 fresh samples and after reactions at 700 °C.

ID ratio, which correlated with an improvement in the quality of the carbon produced. The TPO analysis for the Co100 sample demonstrated a weight gain at temperatures higher than 600 °C, which was attributed to the oxidation of cobalt carbide. The carbide phase was related to the formation of graphene. In summary, CNTs were favored in samples with low Co/Al ratios in which the CoeAl mixed oxide phase was predominant and reaction temperatures were between 500 and 700 °C. The IG/ID ratio and the quality of CNTs increased with the reaction temperature and with reductions in the catalyst with hydrogen. However, graphene was favored by high Co/Al ratios, pure Co3O4 phase, and higher reaction temperatures.

reaction revealed that the IG/ID ratio increased as the Co/Al ratio and reaction temperature increased. The Raman spectra of samples containing Al were characteristic of CNTs. The Co100 sample presented a Raman spectrum characteristic of graphene. This sample presented an I2D/IG ratio characteristic of double-layer graphene after reactions at 800 and 900 °C. The SEM and TEM images revealed CNTs that became more numerous as the Co/Al ratio decreased. The TEM images showed MWCNTs. The effect of the activation with hydrogen increased the IG/

Fig. 20. C1s XPS spectra deconvolution after reaction: a) Co90, b) Co60, and c) Co60 after H2 reduction and reaction at 700 °C.

17

Materials Chemistry and Physics 226 (2019) 6–19

C.O. Calgaro, O.W. Perez-Lopez

Acknowledgment

(2009) 405–409. [24] L.B. Avdeeva, D.I. Kochubey, S.K. Shaikhutdinov, Cobalt catalysts of methane decomposition : accumulation of the filamentous carbon, Appl. Catal. A-Gen 177 (1999) 43–51. [25] Y. Zhang, K.J. Smith, A kinetic model of CH4 decomposition and filamentous carbon formation on supported Co catalysts, J. Catal. 231 (2005) 354–364. [26] V. Jourdain, C. Bichara, Current understanding of the growth of carbon nanotubes in catalytic chemical vapour deposition, Carbon 58 (2013) 2–39. [27] Q. Zhang, Y. Liu, L. Hu, W. Qian, G. Luo, F. Wei, Synthesis of thin-walled carbon nanotubes from methane by changing the Ni/Mo ratio in a Ni/Mo/MgO catalyst, N. Carbon Mater. 23 (4) (2008) 319–325. [28] H. Ago, N. Uehara, N. Yoshihara, M. Tsuji, M. Yumura, N. Tomonaga, et al., Gas analysis of the CVD process for high yield growth of carbon nanotubes over metalsupported catalysts, Carbon 44 (2006) 2912–2918. [29] S. Douven, S.L. Pirard, F.Y. Chan, R. Pirard, G. Heyen, J.P. Pirard, Large-scale synthesis of multi-walled carbon nanotubes in a continuous inclined mobile-bed rotating reactor by the catalytic chemical vapour deposition process using methane as carbon source, Chem. Eng. J. 188 (2012) 113–125. [30] A. Simon, M. Seyring, S. Kämnitz, H. Richter, I. Voigt, M. Rettenmayr M, et al., Carbon nanotubes and carbon nanofibers fabricated on tubular porous Al2O3 substrates, Carbon 90 (2015) 25–33. [31] J. Song, L. Wang, S. Feng, J. Zhao, Z. Zhu, Growth of carbon nanotubes by the catalytic decomposi- tion of methane over Fe-Mo/Al2O3 catalyst : effect of temperature on tube structure, N. Carbon Mater. 24 (4) (2009) 307–313. [32] Q.M. Gong, Z. Li, Y. Wang, B. Wu, Z. Zhang, J. Liang, The effect of high temperature annealing on the structure and electrical properties of well-aligned carbon nanotubes, Mater. Res. Bull. 42 (2007) 474–481. [33] M. Pudukudy, Z. Yaakob, M.S. Takriff, Methane decomposition over unsupported mesoporous nickel ferrites : effect of reaction temperature on the catalytic activity and properties of the produced nanocarbon, RSC Adv. 6 (2016) 68081–68091. [34] N.A. Hermes, M.A. Lansarin, O.W. Perez-Lopez, Catalytic decomposition of methane over M – Co – Al catalysts ( M = Mg , Ni , Zn , Cu ), Catal. Lett. 141 (2011) 1018–1025. [35] H.Y. Wang, A.C. Lua, Methane decomposition using Ni – Cu alloy nano-particle catalysts and catalyst deactivation studies, Chem. Eng. J. 262 (2015) 1077–1089. [36] U. Sikander, S. Sufian, M.A. Salam, A review of hydrotalcite based catalysts for hydrogen production systems, Int. J. Hydrogen Energy 42 (2017) 19851–19868. [37] L. Zardin, O.W. Perez-Lopez, Hydrogen production by methane decomposition over Co-Al mixed oxides derived from hydrotalcites : effect of the catalyst activation with H2 or CH4, Int. J. Hydrogen Energy 42 (2017) 7895–7907. [38] A. Białas, M. Mazur, P. Natka, B. Dudek, M. Kozak, A. Wach, et al., Hydrotalcitederived cobalt – aluminum mixed oxide catalysts for toluene combustion, Appl. Surf. Sci. 362 (2016) 297–303. [39] C. Escobar, O.W. Perez-Lopez, Hydrogen production by methane decomposition over Cu – Co – Al mixed oxides activated under reaction conditions, Catal. Lett. 144 (2014) 796–804. [40] F. Cavani, F. Trifiro, A. Vaccari, Hydrotalcite-type anionic clays: preparation, properties and application, Catal. Today 11 (1991) 173–301. [41] D. Li, Y. Ding, X. Wei, Y. Xiao, L. Jiang, Cobalt-aluminum mixed oxides prepared from layered double hydroxides for the total oxidation of benzene, Appl. Catal. Gen. 507 (2015) 130–138. [42] Z. Wang, Z. Jiang, W. Shangguan, Simultaneous catalytic removal of NOx and soot particulate over Co – Al mixed oxide catalysts derived from hydrotalcites, Catal. Commun. 8 (11) (2007) 1659–1664. [43] O.W. Perez-Lopez, A. Senger, N.R. Marcilio, M.A. Lansarin, Effect of composition and thermal pretreatment on properties of Ni – Mg – Al catalysts for CO2 reforming of methane, Appl. Catal. Gen. 303 (2006) 234–244. [44] P. Jana, V.A. de La Peña O'Shea, J.M. Coronado, S.P. Serrano, Mild temperature hydrogen production by methane decomposition over cobalt catalysts prepared with different precipitating agents, Int. J. Hydrogen Energy 37 (2012) 7034–7041. [45] A. Vaccari, Preparation and catalytic properties of cationic and anionic clays, Catal. Today 41 (1998) 53–71. [46] B. Jongsomjit, J. Panpranot, J.G. Goodwin, Co-support compound formation in alumina-supported cobalt catalysts, J. Catal. 109 (2001) 98–109. [47] M.P. Woods, P. Gawade, B. Tan, U.S. Ozkan, Preferential oxidation of carbon monoxide on Co/CeO2 nanoparticles, Appl. Catal. B Environ. 97 (2010) 28–35. [48] I. Zacharaki, C.G. Kontoyannis, S. Boghosian, A. Lycourghiotis, C. Kordulis, Cobalt oxide supported on alumina catalysts prepared by various methods for use in catalytic afterburner of PEM fuel cell, Catal. Today 143 (2009) 38–44. [49] E. Mwenesongole, A Raman- and XRD Study of the Crystal Chemistry of Cobalt Blue, University of Pretoria, Pretoria, 2008 Master of Science (Chemistry). [50] C.M. Álvarez-Docio, J.J. Reinosa, A. Del Campo, J.F. Fernandéz, 2D particles forming a nanostructured shell: a step forward cool NIR reflectivity for CoAl2O2 pigments, Dyes Pigments 137 (2017) 1–11. [51] A.R. Biris, A.S. Biris, D. Lupu, S. Trigwell, E. Dervishi, Z. Rahman, et al., Catalyst excitation by radio frequency for improved carbon nanotubes synthesis, Chem. Phys. Lett. 429 (2006) 204–208. [52] C. Soldano, A. Mahmood, E. Dujardin, Production, properties and potential of graphene, Carbon 48 (2010) 2127–2150. [53] T. Belin, F. Epron, Characterization methods of carbon nanotubes : a review, Mater. Sci. Eng. B 119 (2005) 105–118. [54] A.C. Ferrari, J.C. Meyer, V. Scardaci, C. Casiraghi, M. Lazzeri, F. Mauri, et al., Raman spectrum of graphene and graphene layers, Phys. Rev. Lett. 187401 (2006) 1–4. [55] M. Liu, X. Zhang, W. Wu, T. Liu, Y. Liu, B. Guo, et al., One-step chemical exfoliation of graphite to ∼ 100 % few-layer graphene with high quality and large size at

The authors are grateful for the financial support provided by CAPES and CNPq. Thanks for analytical support by LOR (Laboratory of Organometallic and Resins) and LabCEMM (Central Laboratory of Microscopy and Microanalysis) from Pontifical Catholic University of Rio Grande do Sul/PUCRS, Porto Alegre, Brazil. Thanks to Prof Claudio Radtke of the Institute of Chemistry/UFRGS by XPS measurements. Appendix A. Supplementary data Supplementary data to this article can be found online at https:// doi.org/10.1016/j.matchemphys.2018.12.094. References [1] L. Zhou, L.R. Enakonda, M. Harb, Y. Saih, A. Aguilar-Tapia, S. Ould-Chikh, et al., Fe catalysts for methane decomposition to produce hydrogen and carbon nano materials, Appl. Catal. B Environ. 208 (2017) 44–59. [2] H.F. Abbas, W.M.A. Wan Daud, Hydrogen production by methane decomposition: a review, Int. J. Hydrogen Energy 35 (2010) 1160–1190. [3] J.L. Pinilla, R. Utrilla, M.J. Lázaro, R. Moliner, I. Suelves, A.B. García, Ni- and Febased catalysts for hydrogen and carbon nano fi lament production by catalytic decomposition of methane in a rotary bed reactor, Fuel Process. Technol. 92 (2011) 1480–1488. [4] K. Salipira, N.J. Coville, M.S. Scurrell, Carbon produced by the catalytic decomposition of methane on nickel: carbon yields and carbon structure as a function of catalyst properties, J. Nat. Gas Sci. Eng. 32 (2016) 501–511. [5] M. Pudukudy, Z. Yaakob, M.S. Takriff, Methane decomposition into COx free hydrogen and multiwalled carbon nanotubes over ceria, zirconia and lanthana supported nickel catalysts prepared via a facile solid state citrate fusion method, Energy Convers. Manag. 126 (2016) 302–315. [6] P. Jana, V.A. de la Peña O'Shea, J.M. Coronado, D.P. Serrano, Co-production of graphene sheets and hydrogen by decomposition of methane using cobalt based catalysts, Energy Environ. Sci. 4 (2011) 778–783. [7] G. Tian, Q. Zhang, M. Zhao, H. Wang, C. Chen, Fluidized-Bed CVD of unstacked double-layer templated graphene and its application in supercapacitors, Inorganic Matrials: Syntesis and Processing 61 (2015) 747–755. [8] Z. Tu, Z. Liu, Y. Li, F. Yang, L. Zhang, Z. Zhao, et al., Controllable growth of 1 – 7 layers of graphene by chemical vapour deposition, Carbon 73 (2014) 252–258. [9] K.S. Novoselov, A.K. Geim, S.V. Morozov, D. Jiang, Y. Zhang, S.V. Dubonos, et al., Electric field effect in atomically thin carbon films, Science 306 (2004) 666–669. [10] H. Kim, J. Ahn, Graphene for flexible and wearable device applications, Carbon 120 (2017) 244–257. [11] B. Huet, J. Raskin, Role of Cu foil in-situ annealing in controlling the size and thickness of CVD graphene domains, Carbon 129 (2018) 270–280. [12] L.M. Cao, Y.S. Chen, C.L. Yang, Y.Q. Song, J. Yang, J.P. Jia, Selective fabrication of carbon nanowires, carbon nanotubes, and graphene by catalytic chemical liquid deposition, Mater. Res. Bull. 55 (2014) 229–236. [13] S. Ren, P. Rong, Q. Yu, Preparations, properties and applications of graphene in functional devices: a concise review, Ceram. Int. 44 (2018) 11940–11955. [14] S. Naghdi, K.Y. Rhee, S.J. Park, Review article A catalytic , catalyst-free , and rollto-roll production of graphene via chemical vapor deposition : low temperature growth, Carbon 127 (2018) 1–12. [15] W. Mu, Y. Fu, S. Sun, M. Edwards, L. Ye, K. Jeppson K, et al., Controllable and fast synthesis of bilayer graphene by chemical vapor deposition on copper foil using a cold wall reactor, Chem. Eng. J. 304 (2016) 106–114. [16] P. Jana, V.A. de La Peña O'Shea, J.M. Coronado, D.P. Serrano, H2 production by CH4 decomposition over metallic cobalt nanoparticles: effect of the catalyst activation, Appl. Catal. Gen. 467 (2013) 371–379. [17] A. Weibel, D. Mesguich, G. Chevallier, E. Flahaut, C. Laurent, Fast and easy preparation of few-layered-graphene/magnesia powders for strong, hard and electrically conducting composites, Carbon 136 (2018) 270–279. [18] A.E. Awadallah, A.A. Aboul-Enein, U.F. Kandil, M.R. Taha, Facile and large-scale synthesis of high quality few-layered graphene nano-platelets via methane decomposition over unsupported iron family catalysts, Mater. Chem. Phys. 191 (2017) 75–85. [19] T. Chen, Q. Zhang, M. Zhao, J. Huang, C. Tang, F. Wei, Rational recipe for bulk growth of graphene/carbon nanotube hybrids: new insights from in-situ characterization on working catalysts, Carbon 95 (2015) 292–301. [20] P. Jana, V.A. de La Peña O'Shea, J.M. Coronado, D.P. Serrano, Cobalt based catalysts prepared by Pechini method for CO2-free hydrogen production by methane decomposition, Int. J. Hydrogen Energy 35 (2010) 10285–10294. [21] C.O. Calgaro, O.W. Perez-Lopez, Decomposition of methane over Co3-xAlxO4 (x =0-2 ) coprecipitated catalysts : the role of Co phases in the activity and stability, Int. J. Hydrogen Energy 42 (2017) 29756–29772. [22] U.P.M. Ashik, W.M.A.W. Daud, H.F. Abbas, Production of greenhouse gas free hydrogen by thermocatalytic decomposition of methane – a review, Renew. Sustain. Energy Rev. 44 (2015) 221–256. [23] X. Xiang, L. Zhang, H.I. Hima, F. Li, D.G. Evans, Co-based catalysts from Co/Fe/Al layered double hydroxides for preparation of carbon nanotubes, Appl. Clay Sci. 42

18

Materials Chemistry and Physics 226 (2019) 6–19

C.O. Calgaro, O.W. Perez-Lopez ambient temperature, Chem. Eng. J. 355 (2019) 181–185. [56] M.D.S.L. Wimalananda, J. Kim, J. Lee, Rapid synthesis of a continuous graphene film by chemical vapor deposition on Cu foil with the various morphological conditions modified by Ar plasma, Carbon 120 (2017) 176–184. [57] J. Xiong, Y. Ding, T. Wang, L. Yan, W. Chen, H. Zhu, et al., The formation of Co2C species in activated carbon supported cobalt-based catalysts and its impact on Fischer – tropsch reaction, Catal. Lett. 102 (2005) 265–269. [58] E. Patanou, N.E. Tsakoumis, R. Myrstad, E.A. Blekkan, The impact of sequential H2 -CO-H2 activation treatment on the structure and performance of cobalt based catalysts for the Fischer-Tropsch synthesis, Appl. Catal. A-Gen. 549 (2018) 280–288. [59] H. Su Zhao, K. Sun, J. Liu, W. Li, Structural and electronic properties of cobalt carbide Co2C and its surface stability: density functional theory study, Surf. Sci. 606 (2012) 598–604. [60] A. Reina, X. Jia, J. Ho, D. Nezich, H. Son, V. Bulovic, et al., Large area, few-layer graphene films on arbitrary substrates by chemical vapor deposition, Nano Lett. 9 (2009) 30–35. [61] M. Gubernat, A. Fraczek-Szczypta, J. Tomala, S. Blazewicz, Catalytic graphene formation in coal tar pitch - derived carbon structure in the presence of SiO2 nanoparticles, Ceram. Int. 44 (2018) 3085–3091. [62] H.A. Oliveira, D.F. Franceschini, F.B. Passos, Support effect on carbon nanotube growth by methane chemical vapor deposition on cobalt catalysts, J. Braz. Chem. Soc. 23 (2012) 868–879.

[63] J.E. Herrera, D.E. Resasco, In situ TPO/Raman to characterize single-walled carbon nanotubes, Chem. Phys. Lett. 376 (2003) 302–309. [64] B. Kitiyanan, W.E. Alvarez, J.H. Harwell, D.E. Resasco, Controlled production of single-wall carbon nanotubes by catalytic decomposition of CO on bimetallic Co – Mo catalysts, Chem. Phys. Lett. 317 (2000) 497–503. [65] J. Lee, J.W. Kim, J.S. Lim, T.J. Kim, S.D. Kim, S. Park, Y. Lee, X-ray photoelectron spectroscopy study of cobalt supported multi-walled carbon nanotubes prepared by different precursors, Carbon Science 8 (2007) 120–126. [66] S. Mo, Q. Zhang, Q. Ren, J. Xiong, M. Zhang, Z. Feng, et al., Leaf-like Co-ZIF-L derivatives embedded on Co2AlO4/Ni foam from hydrotalcites as monolithic catalysts for toluene abatement, J. Hazard Mater. 364 (2019) 571–580. [67] Q. Ren, Z. Feng, S. Mo, C. Huang, S. Li, W. Zhang, et al., 1D-Co3O4, 2D-Co3O4, 3DCo3O4 for catalytic oxidation of toluene, Catal. Today (2018), https://doi.org/10. 1016/j.cattod.2018.06.053. [68] E. Garbowski, M. Guenin, M. Marion, M. Primet, Catalytic properties and surface states of cobalt containing oxidation catalysts, Appl. Catal. 64 (1990) 209–224. [69] T. Zeng, H. Zhang, Z. He, J. Chen, S. Song, Mussel-inspired approach to constructing robust cobalt-embedded N-doped carbon nanosheet toward enhanced sulphate radical-based oxidation, Nat. Publ. Gr. 6 (2016) 33348 2–10. [70] S. Al Khabouri, S. Al Harthi, T. Maekawa, Y. Nagaoka, M.E. Elzain, A. Al Hinai, et al., Composition, electronic and magnetic investigation of the encapsulated ZnFe2O4 nanoparticles in multiwall carbon nanotubes containing Ni residuals, Nanoscale Res. Lett. 10 (2015) 262 1-10.

19