1 Introduction

The catalytic methanation (CO + 3H2 → CH4 + H2O) has very important applications in various industrial processes, including ammonia synthesis, gasification of coal and Fischer–Tropsch synthesis [1–4]. This reaction even attracts ever-growing interest because it can be used to remove carbon monoxide from feed gas for fuel cell [5–7] and to convert biogas into methane [8, 9]. The supported noble metal catalysts and transition metal catalysts have been developed for the reaction of methanation. From the viewpoint of practice, nickel (Ni) based catalysts are more favored because of the low-cost and easy availability of nickel. However, a problem with Ni based catalysts is the deactivation with time on stream due to carbon deposition and sintering of the metal. A continuous effort has been made in order to develop a sustainable Ni based catalyst for the methanation.

Many studies confirmed that dissociation of carbon monoxide (CO) is the rate-limiting step of the methanation. Once carbon monoxide is dissociated, hydrogen then reacts with carbon from the dissociation to form CH and finally methane (CH4). However, if the hydrogenation of carbon is slower than CO dissociation, the carbon will be aggregated into inactive carbon species, which induces a deactivation of the catalyst [10]. On the other hand, the methanation or CO dissociation was considered to be a structure sensitive reaction [11–14]. Obviously, the catalyst preparation or pretreatment methodologies will have a significant effect on the catalytic properties. Regarding this, many studies have been conducted with various catalyst preparation methods [4, 15–22]. For example, Snel [4] prepared three kinds of iron catalysts by impregnation, ion-exchange and coprecipitation, and observed different catalytic behaviors. Besides, the extent of reduction also showed influence on methane production. Dai et al. [19] studied Co/Al2O3 catalyst. They found that reduction temperature and time have significant effect on the activity and selectivity in CO hydrogenation. Aksoylu et al. [20] compared two Ni/Al2O3 catalysts, prepared by impregnation and precipitation, respectively. These catalysts show different performances in hydrogenation of carbon oxides. Vos et al. [21] confirmed the impact of calcination condition on nickel particle supported by alumina. Size dependence of nickel nanoparticles for the methanation has been observed [23–25]. The preparation or treatment conditions plays a very important role in the size control of the Ni catalyst and thereby the catalytic properties.

Recently, catalyst preparation using various plasmas has attracted many attentions. In our previous studies [26–28], room temperature glow discharge plasma was applied for the treatment of nickel precursor supported on alumina (Al2O3) and silica (SiO2) for the improvement in the reforming catalysts. The further calcination thermally of the plasma treated samples leads to a significant improvement in the activity and the anti-carbon deposit property (thereby the stability) for dry reforming of methane [26–28]. However, the mechanism for how the plasma affects the catalyst preparation is still unclear.

In this work, we attempt to use this glow discharge for the treatment of Ni/SiO2 catalyst in order to improve the catalyst properties for the reaction of methanation. The mechanism for the plasma treatment of Ni precursor will be discussed.

2 Experimental Section

2.1 Catalysts Preparation

The samples of 10 wt% Ni/SiO2 was prepared by conventional incipient impregnation. SiO2 powder was commercially obtained (Aldrich; calcined at 500 °C for 3 h before use; S BET = 340 m2/g). To prepare the catalyst, Ni(NO3)2·6H2O was dissolved in distilled water as the precursor of nickel. The SiO2 powder was then added into the solution under stirring. After being placed at room temperature for 12 h, the sample was dried at 110 °C for 12 h. Then it was divided into two parts: one was treated by the glow discharge plasma for 1 h, followed by calcination thermally at 500 °C for 4 h. The plasma treatment setup was previously described in detail [26, 28]. The sample (about 0.4 g), loaded in a quartz boat, was placed in the ‘positive column’ region of the glow discharge. When the pressure in the discharge tube was about 100 Pa, the glow discharge plasma was generated by applying 900 V to the electrode using a high voltage amplifier (Trek, 20/20B). The signal input for the high voltage amplifier was supplied by a function/arbitrary waveform generator (Hewlett Packard, 33120A) with a 100 Hz square wave. Ultra high pure grade argon (>99.999%) was used as the plasma-forming gas. The discharge was initiated at room temperature without external heating or cooling. According to the temperature measurements by using IR imaging, heating effect of the glow discharge could be ignored and the temperature of the catalyst powder was close to room temperature. During the plasma treatment, nickel nitrate was decomposed under the influence of the glow discharge. After that, calcination thermally was conducted with no use of glow discharge plasmas. High dispersion (or small particle size) and enhanced metal-support interaction were achieved this way. Such plasma treated sample was denoted by Ni/SiO2-P. Another part of the sample was calcined at 500 °C without the plasma treatment. This thermal treated sample was denoted by Ni/SiO2-C. Before use for catalytic reactions, these two samples were pressed, crushed and sieved. The particles with a size between 40 and 60 mesh were used for the activity test.

2.2 Activity Test

For the reaction test of methanation, 50 mg of sample was packed into a quartz tube reactor (4 mm in inner diameter) and was fixed with quartz wool. Before the reaction, the sample was reduced in situ by flowing hydrogen (20 mL/min) at 500 °C for 2 h. Then the reactor was cooled down to desired temperature under Ar. A fluid of CO/H2/Ar (CO:H2:Ar = 1:2:3, GHSV = 48000 mL/h gcat) was purged into the reactor. The product was analyzed on line by gas chromatography [GC; Agilent 6890, with TDX01 column and thermal conductivity detector (TCD)]. A cold trap was equipped before the GC to remove the condensable steam in the exhaust gas. For kinetic analyses, lower temperatures were employed to get conversion of CO below 15%. Under the conditions we employed, the limitation of mass transfer for the reactions on the catalysts have been eliminated.

2.3 Catalyst Characterization

X-ray powder diffraction (XRD) patterns were recorded by an X’Pert Pro diffractometer with a Co Kα radiation at a scanning rate of 6°min−1. The phase identification was made by comparison to the Joint Committee on Powder Diffraction Standards (JCPDSs).

The surface morphology of nickel particles of both fresh and used samples were observed by transmission electron microscopy (TEM), using a Philips TECNAI G2F20 system operated at 200 kV.

Thermal-gravimetric analysis (TGA) was carried out to study the effect of plasma treatment using a Netzsch STA 449 F3 system. The sample (approximately 10 mg) was loaded into an alumina crucible and heated up at a rate of 5 °C/min under nitrogen (30 mL/min). The exhaust gas was analyzed by an Omnistar GD301 mass spectrograph (MS).

3 Results

XRD patterns of fresh samples are shown in Fig. 1. Two peaks, assigned to Ni (111) at 52.2° and Ni (200) at 61.1°, respectively, appear on the patterns. Average particle sizes are calculated from Scherrer formula. It is about 14 nm for Ni/SiO2-C and 10 nm for Ni/SiO2-P.

Fig. 1
figure 1

XRD patterns of fresh samples after reduction at 500 °C for 2 h

Full size image

TEM images of both samples after reduction are shown in Fig. 2. The particle distributions of both samples, counted from TEM images, are shown in Fig. 3. The Ni/SiO2-C sample shows a wide size distribution. The plasma treated sample, however, has a narrow size distribution. Average particle sizes are 17.5 and 9.7 nm, respectively, in good agreement with the results from XRD. The XRD and TEM analyses conform that the plasma treated sample has smaller metal particles.

Fig. 2
figure 2

TEM images of samples after reduction at 500 °C for 2 h

Full size image
Fig. 3
figure 3

Size distributions after reduction at 500 °C for 2 h. a Ni/SiO2-C, b Ni/SiO2-P

Full size image

The average conversions of CO and H2 for 4-h tests versus temperature are shown in Fig. 4. As shown in Fig. 4, the activity of Ni/SiO2-C increased slightly when the temperature was raised from 300 to 325 °C while it ascended greatly from 325 to 375 °C. For the plasma treated sample, a substantial increase in conversions of CO and H2 occurred at lower temperatures. Obviously, the plasma treated sample had a higher activity at the same reaction temperature.

Fig. 4
figure 4

Conversions of CO and H2 versus temperature

Full size image

The TEM images of both samples after the 4-h reaction at 350 °C are shown in Figs. 5 and 6. For both samples, no serious coke deposition was observed. However, it could be seen that nickel particle was covered by carbon species on Ni/SiO2-C. While even with careful search, no visible carbon species could be found on Ni/SiO2-P. In addition, aggregation of nickel could be found on Ni/SiO2-C after the 4-h reaction. From Fig. 5d, a much more broad partition distribution of nickel could be observed on Ni/SiO2-C after reaction (from 7.7 to 57.4 nm, with an average of 21.8 nm). On the contrary, Ni/SiO2-P still showed good dispersion after reaction. No obvious aggregation was observed after reaction. A narrow size distribution (from 5.8 to 20.7 nm) with an average of 10.5 nm was shown in Fig. 6e. The results demonstrate that the plasma treated sample possesses better anti-carbon deposition and anti-sintering properties. A size control via the glow discharge plasma treatment was achieved. The smaller nickel particle size or the higher dispersion is the major reason for the better activity of the plasma treated sample.

Fig. 5
figure 5

TEM images of nickel particles of Ni/SiO2-C after reaction at 350 °C for 4 h (a, b, c) and size distribution (d)

Full size image
Fig. 6
figure 6

TEM images of nickel particles of Ni/SiO2-P after reaction at 350 °C for 4 h (a, b, c) with EDX of image c (d) and size distribution (e)

Full size image

Kinetic analyses were conducted in order to further confirm it. Figure 7 shows reaction rates of CO conversion to methane on both samples versus reciprocal temperature. Apparent activation energies E a ’s are listed in Table 1. Both samples had similar apparent E a ’s. However, Ni/SiO2-P had a pre-exponential factor larger by 2-fold than that of Ni/SiO2-C. It is the smaller particle that provides more active sites per mass of metal, which can be proved by the larger pre-exponential factors of the plasma treated catalyst.

Fig. 7
figure 7

CO conversion rates on Ni/SiO2-C (filled square) and Ni/SiO2-P (square) versus the reciprocal temperature

Full size image
Table 1 Pre-exponential factors, activation energies and turn over frequencies (TOF) of both samples
Full size table

Although more and more studies have been performed worldwide to use plasma treatment for catalyst preparation, it is still unclear for how plasma improves the catalyst preparation. With the purpose of understanding the effect of plasma treatment on the precursor, thermal-gravimetric analysis (TGA) was conducted in this work. The results of TGA are shown in Fig. 8. Two steps of weight loss were observed for both samples. From room temperature to c.a. 270 °C, both samples showed continue loss of weight, referring to the desorption of physical adsorbed water and surface hydroxyl on the support, followed by the decomposition of nickel precursor between 270 and 320 °C. During the second step, the weight reduced by 7.41% (Ni/SiO2-C) and 6.68% (Ni/SiO2-P), respectively. It suggests that partial decomposition of nickel nitrate was induced during the plasma treatment.

Fig. 8
figure 8

Weight changes of impregnated Ni/SiO2-C (solid line) and plasma treated sample (dot line) before calcination versus temperature under nitrogen

Full size image

Our presumption is supported by the MS results of the gaseous products during TGA, as described in Fig. 9. NO (m/z = 30) was formed on both samples. The most significant difference is that, for the plasma treated sample, a peak of N2O (m/z = 44) can be observed during the decomposition while no N2O was detected for the conventional one. As discussed by Sietsma et al. [29], this signal could not be indicative for CO2. For both samples, significant amount of water was detected. However, compared with the conventional sample, the plasma treated precursor had lower signal for physically adsorbed water (from room temperature to 270 °C) but more intensive one corresponding to chemically combined water or hydroxyl species (from 270 to 320 °C, simultaneous with decomposition of nickel compound). Besides, from the curves in Fig. 8, we can observe that the decomposition of nickel salt for Ni/SiO2-C starts at a temperature of 273.3 °C. While for Ni/SiO2-P, the temperature is raised by 6.8 °C, up to 280.1 °C. We attribute the up shift of the decomposition temperature to enhanced interaction between the nickel precursor and the support. Hereby, we conclude that during the plasma treatment, nickel nitrate decomposed partially, and a part of hydroxyl species or water got stronger adhesion to nickel species. The exact structure and composition of the nickel species after plasma treatment is being investigated.

Fig. 9
figure 9

MS patterns of N2O, NO and H2O in the exhaust gases of both samples during TGA. Solid line for conventional one and dot line for plasma treated one

Full size image

4 Conclusion

The room temperature glow discharge plasma was applied for the pretreatment of nickel precursor supported by silica. The smaller nickel particles and better dispersion of the plasma treated catalyst induce higher activity, compared to the conventionally calcined Ni/SiO2 catalyst. The anti-carbon deposition and anti-sintering performance are also enhanced from the plasma treatment.