Introduction

The search for sustainable technologies that are able to replace fossil fuels has increased in recent years. Products arising from the burning of these fuels have caused several environmental problems such as greenhouse effect, pollution and acid rain [1]. In this context, hydrogen emerges as a promising alternative, thanks to its high energy density and because its combustion generates only water [2]. Currently, the main industrial process to produce hydrogen is the natural gas steam reforming. However, this process produces carbon monoxide, which is an undesirable product during the use of hydrogen in fuel cells, causing their poisoning. Several purification steps are required to obtain high purity hydrogen by conventional production methods, increasing the cost of the final product [3, 4].

Among several possibilities, the catalytic decomposition of methane reaction (CH4 → C + 2H2) arises as an interesting alternative, since the produced hydrogen is pure, free of carbon oxides. Without catalysts, the breaking of the methane C–H bond requires high temperatures. Nevertheless, in the presence of catalysts, the reaction temperature falls dramatically [5]. The most catalysts used in catalytic methane decomposition are derived from transition metals such as Ni, Co and Fe [6]. The fact that these metals have partially filled 3d orbitals make them the most suitable ones for hydrocarbon dissociation reactions [2, 7, 8].

Besides the produced hydrogen, the deposited carbon in the decomposition of methane has attracted the interest of the scientific community. Depending on the reaction conditions and the type of catalyst used, the carbon produced can be deposited in nanofibers or nanotubes form [9, 10]. These materials have achieved notoriety due to their exceptional mechanical and electrical properties, with potential applications in products such as batteries, capacitors, transistors, special cables, anticorrosive paints, among others [11].

Nickel-based catalysts have been extensively studied due to their high activity at low temperatures, being able to be prepared on a support such as silica and zeolites, or co-precipitated with other metals such as aluminum, for example [1–10, 12–29]. The use of promoters, as Cu, Pd, Pt and Rh, has attracted attention because they act on the stability and performance of the catalysts. Studies indicate that the addition of copper influences the reduction temperature in co-precipitated cobalt-based catalysts [19]. For Ni-based catalysts, the addition of copper seems to act in the metal particles dispersion on the catalyst, contributing to its activity and structural properties [23]. However, the relation between the catalytic activity and the Cu–Ni alloy properties is not completely understood yet [25].

In this respect, the aim of this work was to study the effect of copper on the activity of nickel-based catalysts supported on silica for catalytic decomposition of methane. In order to evaluate the addition of different amounts of copper, we fixed the amounts of Ni (10 wt%) and silica. The influence of copper addition in the catalytic properties and in the characteristics of the carbon deposited after the reaction was analyzed.

Experimental

Catalysts preparation

The catalysts were prepared by wet impregnation method in which aqueous solutions of metal nitrates were mixed with silica Aerosil 200 (Degussa) under continuous stirring for 4 h. After impregnation, the material was dried in an oven at 80 °C for 24 h. Solid samples were then calcined under air flow at 600 °C for 2 h.

Catalysts characterization

The specific surface area of the samples was measured in a multipurpose equipment employing the dynamic nitrogen adsorption method at liquid N2 temperature. Samples were previously pretreated at 250 °C for 1 h to remove moisture and other volatile components. The flow rate used was 30 ml/min of a N2/He mixture containing 30 % of N2. Temperature programmed reduction (TPR) runs were performed at a heating rate of 10 °C/min up to 700 °C under a mixture of 10:90 H2:N2 (30 ml/min).

Temperature programmed oxidation (TPO) analysis were carried out in a TA SDT-Q600 thermobalance. For the TPO analysis, 10 mg of sample was heated at 10 °C/min to the temperature of 900 °C under a flow rate of 100 ml/min of air.

The X-ray diffractograms were obtained using a diffractometer BRUKER D2-Phaser with CuKα radiation at 30 kV and 10 mA. The crystallite sizes were determined by the Scherrer equation. For Raman spectroscopy analysis, a RENISHAW in via spectrometer with 785-nm laser was used.

Activity test

The activity runs were carried out at atmospheric pressure in a quartz tubular reactor with 9 mm inner diameter and 370 mm in length. Quartz wool was used to hold the catalyst bed. The system was heated up using a resistive electric oven and the temperature was measured by a K-type thermocouple. Samples weighing 100 mg were heated up in a 10 °C/min ramp to the desired temperature under a 90 ml/min flow of N2 and 10 ml/min of CH4, adjusted with Sierra Instruments mass flow controllers. The conventional reduction with hydrogen was not carried out. The obtained products were analyzed by a gas chromatography coupled in-line to the reactor using nitrogen as carrier gas and a thermal conductivity detector.

Results and discussion

Catalysts characterization

Table 1 shows the composition of prepared samples and the specific surface area obtained by BET method (SBET) for calcined samples. It was observed that Cu addition to the catalyst promoted an increase in the specific surface area of the sample, with most pronounced effect in the Cu1Ni10 sample. On the other hand, the sample with 3% of Cu did not show important changes in the observed area. This effect indicates that the addition of small amounts of Cu can improve the Ni dispersion on the silica support.

Table 1 Composition of prepared samples and BET specific surface area
Full size table

Fig. 1 shows the XRD patterns of calcined samples in air. All samples showed peaks with 2θ at 37.5°, 43.3° and 62.9°, which correspond to the NiO crystalline phase (ICDS No: 01-1239) [12, 14, 16, 17, 26, 30, 31]. Due to the lower copper content in the samples, it was not possible to observe peaks related to the CuO phase. However, the Cu1Ni10 sample presented much smaller NiO peaks, indicating better dispersion and smaller crystallite size in agreement with the SBET results. This sample also showed a shift of the main reflection from 43.3° to 43.7° indicating that the atoms of Cu entered into the NiO lattice [32]. The samples without Cu (Ni10) and with the higher amount of Cu (Cu3Ni10) exhibited stronger peaks related to the NiO phase, indicating a higher crystallite size.

Fig. 1
figure 1

X-ray diffraction patterns for samples calcined at 600 °C

Full size image

The results of the TPR analysis are presented in Fig. 2. It is possible to observe that the samples showed two main reduction peaks. The main peak, with maxima in the temperature range of 300–360 °C could be attributed to the reduction of larger NiO particles weakly bound to the silica support [18, 24, 33]. The second peak for Ni10 is a shoulder between 400 and 500 °C. This can be related to the reduction of smaller crystals of NiO which have a stronger interaction with the support and are reduced at higher temperatures. For Cu-containing samples, a sharp peak of hydrogen consumption between 190 and 210 °C is also observed in Fig. 2. This peak could be attributed to CuO reduction.

Fig. 2
figure 2

TPR profiles of the samples calcined at 600 °C

Full size image

The addition of Cu significantly changes the reduction temperatures of the samples. The main peak of NiO reduction was shifted to lower temperatures according to the Cu amount added, from 362 °C for Ni10 to 302 °C for Cu3Ni10. It was also verified that Cu addition eliminates the shoulder at high temperatures, suggesting the formation of a Cu–Ni alloy, because the reduction of all species of Ni occurs between 200 and 350 °C.

Catalyst activity

Fig. 3 shows the activity tests between 500 and 750 °C. For samples containing Cu, it was observed that the methane conversion increases until 600 °C and above this temperature there was a strong loss of the catalytic activity. The Ni10 sample showed significant conversion of methane only at 500 °C, indicating the low thermal stability of the sample without copper (Ni10). Taking into account the loss of catalytic activity for the reaction temperature above 600 °C, runs with constant temperature were carried out. These results are presented in Figs. 4 and 5.

Fig. 3
figure 3

Methane conversion at reaction temperatures between 500 and 750 °C (10 ml/min of CH4 and 90 ml/min of N2)

Full size image
Fig. 4
figure 4

Methane conversion with the time-on-stream for the sample Ni10 at different reaction temperatures (10 ml/min of CH4 and 90 ml/min of N2)

Full size image
Fig. 5
figure 5

Methane conversion with the time-on-stream for the samples: a Cu1Ni10, b Cu2Ni10 and c Cu3Ni10 at different reaction temperatures (10 ml/min of CH4 and 90 ml/min of N2)

Full size image

Fig. 4 shows the activity runs for the sample Ni10 at constant temperature. The catalyst activity kept stable only at 500 °C, but with methane conversion lower than 30%. The sample was deactivated after 10 min at 550 °C. These results are in agreement with the test performed at variable temperature, in which this sample exhibited loss of catalytic activity at temperatures above 500 °C. The deactivation for this sample can be related to carbon encapsulation [21, 22] or to the sintering of Ni° crystallites.

The runs at constant temperature for samples containing Cu are showed in Fig. 5. All the samples exhibited nearly constant activity at 500 °C, indicating a good thermal stability at this temperature, but with methane conversions lower than 20%. For the runs carried out at 550 °C, the sample Cu1Ni10 was more stable and had a small decrease in the activity after 5 h of time-on-stream. The samples Cu2Ni10 and Cu3Ni10 presented a higher CH4 conversion at the beginning of the reaction, but with a more pronounced activity decrease with the time-on-stream. At the end of the run, the CH4 conversion of these samples corresponded to approximately 50% of the initial conversion.

When the runs were carried out at 600 and 650 °C, a rapid fall in the catalytic activity was observed for all samples, possibly due to the sintering of the catalysts. The catalyst deactivation was more pronounced for the sample Cu3Ni10 and less pronounced for Cu1Ni10, showing that the increase of Cu amount in the samples does not contribute to their thermal stability. The activity results obtained at constant temperature are in agreement with the surface area presented in Table 1. The higher thermal stability of the Cu1Ni10 sample could be explained by its better dispersion revealed by the highest surface area value and by the smaller crystals evidenced by poorly crystalline diffractogram (Fig. 1). This resulted in a higher resistance to deactivation by sintering.

Carbon and catalyst characterization after the reactions

The TPO characterization results are shown in Fig. 6. Through the DTA curve analysis, it is possible to note oxidation of carbonaceous material at temperatures ranging from 550 to 650 °C, with a maximum between 600 and 625 °C, suggesting the presence of multi-walled carbon nanotubes [1, 25, 34]. It should be possible to note a shift to higher temperatures for samples containing Cu. However, the higher the Cu amounts, the lower the shift observed. The peak relative to the Ni10 sample was smaller and broader than the other samples because the reaction produced less carbonaceous material. In addition, this catalyst exhibited a shoulder with a maximum at 510 °C that could be related to the oxidation of amorphous or encapsulating carbon. The sample Ni10 also exhibits a portion of the TPO profile that oxidizes above 650 °C which can be associated to graphitic carbon. These events could be responsible for the rapid deactivation of the catalyst. These results are consistent with the data of activity shown in Fig. 4, in which the Ni10 sample deactivated in a few minutes at 550 °C.

Fig. 6
figure 6

TPO profiles for the samples after methane decomposition at 550 °C

Full size image

Fig. 7 shows the results of the characterization by Raman spectroscopy. All spectra showed two major bands. The G band is located in the 1580 cm−1 region and is related to the stretching of sp2 C–C bonds in layers of graphitic materials. The D band is located in the 1350 cm−1 region and is related with the breathing modes of the sp3 atoms in graphitic structures, indicating the presence of defects in carbon nanotubes [1, 10, 35]. By comparing the intensities of the D and G bands (ID/IG), it is possible to estimate the type of carbonaceous material produced during the reaction. For samples containing Cu, the ID/IG ratios were close to 1, indicating the presence of multi-walled nanotubes (MWNT). The ID/IG ratio clearly increases with the amount of Cu in the catalyst showing that the defects of MWNT increase proportionally to the addition of Cu. The sample Ni10 showed an ID/IG ratio of 0.93, which suggests a minor presence of defects than samples containing Cu.

Fig. 7
figure 7

Raman Spectroscopy of the samples after methane decomposition at 550 °C

Full size image

The XRD patterns for the samples after reaction at 550 °C are shown in Fig. 8. The samples containing Cu exhibited an intense peak in the region of 26°, which is attributed to the presence of graphitic carbon and carbon nanotubes in the sample. The lower intensity of the peak for the Ni10 sample is related to the smaller amount of carbon deposited during the reaction, which is in agreement with the TPO results shown in Fig. 6. Among the Cu-containing samples, it was noticed a shift of reflection from 26.1° to 26.7° according to increases in the Cu amount. These results indicate the predominance of MWNT in Cu1Ni10 because this sample showed a reflection at 26.1° while the other samples (Cu2Ni10 and Cu3Ni10) exhibited a peak at 26.7° related to graphitic carbon [19].

Fig. 8
figure 8

X-ray diffraction patterns of the samples after methane decomposition at 550 °C

Full size image

Besides this peak, the samples also showed reflections at 44.5° and 51.9°, which are attributed to nickel in the reduced form [25, 26]. The Ni10 sample exhibited sharp peaks indicating a large Ni° crystallite size, differently from the Cu-containing samples. The crystallite size obtained for the samples after reaction at 550 °C is presented in Table 2. The Ni10 sample exhibited the largest nickel crystallites, which demonstrates the sintering of this sample during the reaction. This fact and carbon encapsulation explains the rapid deactivation of this sample, as shown in Fig. 4. The Cu-containing samples exhibited smaller crystallite sizes in comparison to the Ni10 sample, ranging between 8 and 12 nm. The smaller crystallite size was obtained for the sample with the lower amount of Cu (Cu1Ni10) and increases with the copper addition. In addition, the Cu presence could also affect the diffusion of elementary carbon through Ni favoring the formation of carbon nanotubes.

Table 2 Crystallite size for samples after reaction at 550 °C
Full size table

Discussion

The textural characteristics of fresh samples, the catalytic behavior of the calcined samples as well as the characteristics of the spent catalysts and of the carbon produced by the Cu-containing samples were quite different from the sample without Cu (Ni10). These differences are clearly related to the Cu added to the catalysts. It was verified that the addition of Cu in small amounts (1 wt%) promotes the dispersion of Ni particles due to the specific surface area increase and favors the diffusion of carbon. In addition, the size of NiO crystallites in Cu1Ni10 was smaller than in the sample without Cu, as evidenced by the XRD patterns, indicating that that the presence of Cu inhibits the Ni sintering. However, the promoter effect of Cu addition declines with the amount of Cu, as shown by the SBET and XRD results. The main influence of Cu addition was observed on the reducibility of NiO as it is possible to observe in TPR profiles (Fig. 2). The Cu addition shifted the reduction of NiO to low temperatures. Consequently, the thermal stability of samples with higher amounts of Cu decreases. These observations are in agreement with the results of runs at 550 °C or higher temperatures (Fig. 5). It is possible to observe that the samples Cu2Ni10 and Cu3Ni10 exhibit a higher conversion than Cu1Ni10 at the beginning of the reaction, but the conversion of methane decreases more quickly for samples with high amount of Cu. The crystallite size obtained for these samples after reaction at 550 °C (Table 2) corroborates with this hypothesis. On the other hand, the results of Cu-free sample runs (Ni10) clearly demonstrate that their lower thermal stability is related to the high crystallinity of NiO phase (Fig. 1). Therefore, this sample showed strong deactivation in the runs with variable temperature (Fig. 3) and was active only at 500 °C for the runs at constant temperature. The high crystallinity of Ni10 was also responsible for the high initial conversion at 550 °C and for the fast deactivation after 10 min of time-on-stream. These results were confirmed by the XRD pattern obtained after the reaction, resulting in the largest crystallite size (33 nm). In addition to sintering, the sample Ni10 could be deactivated by carbon encapsulation increasing thus the rate of deactivation on this sample.

From these results, it is evident that the key aspect for a high activity and stability of the catalyst is to achieve the lowest Ni° crystallite size during the activation step. The smaller the crystallite size, the more stable the catalyst. On the other hand, the larger the Ni° crystallite size the more susceptible to deactivation by particle aggregation (sintering) or by carbon encapsulation. The addition of Cu inhibited the Ni sintering but also decreased the reduction temperature of Ni; however, the Ni reduction temperature decreases too much for higher amounts of Cu, resulting in metal particles with low thermal stability. One explanation for this behavior is that at low Cu content, the formation of a homogeneous Ni–Cu alloy happens, while the phase segregation occurs at higher amounts of Cu [32].

The characterization of carbon also showed pronounced differences between samples with or without Cu. For Cu-free sample (Ni10), the TPO profile is in agreement with the lower amount of carbon produced and showed that the type of carbon formed was mainly graphitic and some amorphous or encapsulating carbon. The Raman spectra confirm the presence of graphitic carbon due to the lower ID/IG ratio bands for Ni10. The addition of Cu showed a significant influence on the TPO profiles in comparison to the Cu-free sample. It was observed that the peaks of samples with Cu were very sharp and were shifted to high temperatures, at 628 °C for Cu1Ni10, which is typical for MWNT. However, the increase in the Cu amount decreased the carbon oxidation temperature, which suggests an increase of the amount of amorphous carbon and the defects in MWNT, evidenced by the asymmetry of the TPO profiles.

Conclusions

The presence of copper in the samples significantly affects the reducibility and the catalytic properties of Ni/SiO2 for methane decomposition. The sample with 1 wt% of Cu exhibited higher stability than samples with Cu amounts exceeding 1 wt%, due to the formation of a homogeneous Ni–Cu alloy. The addition of small amounts of Cu contributes to the Ni dispersion on the support, facilitates the reduction of NiO with methane and allows obtaining small crystallites of Ni°, while avoiding sintering of the metal particles at high temperatures. The characterization of deposited carbon on Cu-containing samples indicates the presence of multi-walled carbon nanotubes, which present increased defects with the Cu addition. The sample containing 1 wt% of Cu has the best performance for the production of hydrogen from methane decomposition.