**1. Introduction**

Nowadays, hydrogen is recognized as a clean fuel and energy carrier since its combustion produces only water as product [1,2]. However, how to produce hydrogen from primary energy sources (such as hydrocarbons) in an efficient and economic way should be further researched and developed [3–6]. In the past few decades, the most effective approach was catalytic reforming of hydrocarbons. Currently, over 50% of the world's hydrogen supply is from steam reforming of hydrocarbons [7].

Nowadays, H2 is mainly produced by steam reforming of CH4 and other high-energy density liquid fuels, including ethanol, gasoline, diesel, or jet fuel [8–10]. An interesting option is hydrogen production from diesel steam reforming. *n*-decane, one of the main components of diesel, is considered as an ideal source of hydrogen since its availability, easy handling and storage and, relatively high H/C ratio (produce 31 mol of H2 per mole of reacted *n*-decane) [11,12]. However, the *n*-decane

steam reforming reaction is different with CH4, CH3OH, C2H5OH etc., and always accompanied by other side effects (cracking, isomerization, hydrogen transfer reaction). More so, the catalysts used in *n*-decane steam reforming reaction are easily to lose activity caused by carbon deposition, especially at higher temperatures [13–15]. Therefore, the catalysts use in *n*-decane steam reforming reaction are put forward higher requirements.

$$\text{C}\_{10}\text{H}\_{22} + 20\text{H}\_2\text{O} \rightarrow 31\text{H}\_2 + 10\text{CO}\_2$$

Hydrocarbons steam reforming reactions have been extensively investigated over noble and transition metals (Pt, Pd, Rh, Ni, Co, etc.) and several oxide supports (Al2O3, CeO2, MgO, ZrO2, zeolite, etc.) [16–23], so as to develop excellent catalysts to obtain hydrogen as high yield as possible together with high resistance of coke deposition. Transition metals (especially Ni-based) catalysts, which have the high C–C and C–H bonds breaking activity, have been proved to be very effective for hydrocarbons steam reforming reactions as noble metal catalysts [16–18]. Moreover, the lower cost improved its applicability. Therefore, more and more researchers focused on studying hydrocarbon steam reforming over Ni-based catalysts [16–21]. However, coking is easily deposited on the surface of the active phase Ni, which can lower the catalytic activity [24–27]. Therefore, various promoters were introduced into Ni-based catalysts to improve catalytic activity and coking resistance. Lanthanide metals (La, Ce), alkali metals (Na, K), and alkali earth metals (Mg, Ca, Sr, Ba) promoters [28–33], have been found to be effective for improving coking-resistant capacity. However, the addition of these additives influenced Ni dispersion, due to a part of the promoter is in an intimate contact with nickel [34–36].

In order to improve the anti-coking ability of Ni-based catalyst, and have a slight influence on catalytic activity, many scholars introduced another active metal into Ni-based catalyst to form bi-metallic catalysts [37–42]. Wang et al. [37] introduced Pd into Ni-alumina catalysts, the catalytic activity and stability was obviously improved. Vizcaino et al. [43] found that Cu modified Ni-based catalyst showed better anti-coking ability. The addition of Cu is helpful for the process of eliminating the deposited carbon. In our previous work [44], we have added M (Fe, Co, Cu, Zn) as a promoter into the Ni/Ce-Al2O3 catalyst in order to improve the anti-coking ability. Clearly, Co doped Ni/Ce-Al2O3 showed an excellent coking-resistant effect. But, the catalytic activity have a slightly reduction at high temperature (650~800 ◦C). In another study [45], we added Co as another active species into Ni/Ce-Al2O3 to form Ni-Co bi-metallic catalyst and investigated the catalytic activity, stability and coking inhibition effect during *n*-decane reforming. The results showed that the introduction of Ni and Co synchronously can effectively suppress carbon deposition and obviously improve catalytic activity. There was obvious synergistic effect between Ni and Co. However, the difference among different Co content on the Co-Ni/Ce-Al2O3 bi-metallic catalyst has not been discussed. Consequently, it would be valuable to investigate the influence of the content of Ni on *n*-decane steam reforming.

In this paper, the steam reforming experiments of *n*-decane over *x*%Co-Ni/Ce-Al2O3 catalysts with different Co loading were carried out. The effect of different Co loading on catalytic activity and the amount of deposited carbon were discussed. The purpose of this work is screening the suitable catalysts for steam reforming process in order to maximize *n*-decane conversion and H2 yield, and minimize the formation of byproducts and carbon deposition. This work provided some positive suggestions for catalysts preparation and optimization by studying the structure-activity correlations.

#### **2. Results and Discussion**

## *2.1. Catalytic Performance*

#### 2.1.1. *n*-Decane Conversion and H2 Selectivity

*n*-decane steam reforming is used as the probe reaction. The initial activity tests over the series catalysts were performed from 650 to 800 ◦C in order to examine the influence of temperature and different promoters on catalytic performance. *n*-decane conversion and H2 selectivity are considered the main parameters to check the advantages and disadvantages of the catalysts, and the results are shown in Figure 1a–d. The catalytic activities over these catalysts gradually increase with the temperature. Obviously, the presence of Ni or/and Co can effectively promote the rate of the steam reforming reaction, as well as the selectivity of H2 and *n*-decane conversion. Moreover, the synchronous introduction of Co and Ni further enhanced the catalytic activity compared with the 6%Ni/Ce-Al2O3(NCA) catalyst. This demonstrates that the addition of Co could provide sufficient Ni active sites for the reactants. In addition, the catalytic activity of *x*%Co-Ni/Ce-Al2O3 (CNCA) bi-metallic catalysts with different Co content increases firstly and then decreases with Co addition. The catalytic activity reaches the best when the Co content is 12%. This indicated that moderate Co is favor for promoting the activity. There is a synergistic effect between Co and Ni.

**Figure 1.** *n*-decane conversions and H2 selectivity over the series catalysts at 650 ◦C (**a**), 700 ◦C (**b**), 750 ◦C (**c**), and 800 ◦C (**d**).

#### 2.1.2. Thermal Stability and Regeneration of C12-NCA

To better understand the effect of the ordered co-modification in *n*-decane steam reforming, the 12%Co-Ni/Ce-Al2O3 (C12-NCA) catalyst was screened out with a 6 h stability test at 750 ◦C and 800 ◦C, and the results are displayed in Figure 2a. It can be seen in Figure 2a that the H2 selectivity and *n*-decane conversion have a slightly change within 6 h. The C12-NCA catalyst has a good thermal stability. The C12-NCA catalyst also screened out for regeneration experiment. Carbon deposition on used C12-NCA catalyst was removed by oxygen enriched calcinations at 650 ◦C. The regenerative C12-NCA catalyst was carried out again in the same reactor as the fresh ones, and the contrast results are presented in Figure 2b. It is found that the *n*-decane conversions and H2 selectivity over the reused-1(-2) C12-NCA catalyst are approximately equal to the results of fresh one. Therefore, the C12-NCA catalyst is renewable.

**Figure 2.** Thermal stability (**a**) and regeneration (**b**) over the C12-NCA catalyst.

#### *2.2. Fresh Catalyst Characterization*

#### 2.2.1. N2 Adsorption-Desorption Measurements

Table 1 shows the results of N2 adsorption-desorption results of the fresh and used catalysts. The surface areas of different samples in this work are in the range of similar CA support, even if the introduction of Ni and Co species by impregnation method. The value has a slightly decrease with the addition of Co and Ni, and gradually drops with the increase of Co. The loss can be attributed to the fact that the internal surface area of the CA pore system is progressively covered by Ni, Co species forming a layer [45–47].


**Table 1.** The textural properties of the fresh and used catalysts.

\* The numbers in the parentheses represent the surface area of used catalysts.

On the other hand, the surface areas of all catalysts used decrease with different levels after *n*-decane reforming reactions. CA and NCA catalysts decreased by 56% and 44% respectively compared with the fresh ones. Fortunately, the falling range gradually reduced with the addition of Co. Co as the active species showed a better carbon-resistant ability. The results are consistent with the results of the catalyst characterization.

#### 2.2.2. X-ray Diffraction (XRD) Analysis

Figure 3 depicts the X-ray diffraction analysis of the fresh CA, NCA and bi-metallic CNCA catalysts. All the samples present similar characteristic features of γ-Al2O3 at 2θ = 45.7◦, 66.8◦; cubic fluorite structural CeO2 at 2θ value of 28.5◦, 56.3◦; and the Ce crystallite at 2θ = 34.7◦, 49.8◦ and 59.2◦ by Bragg's refections [44,48]. For all the catalysts, there was no CoO*<sup>x</sup>*, CoAl2O4, NiO, or NiAl2O4 diffraction peaks detected. This was probably due to the highly dispersion of CoO*x* and NiO particles are not easy to be detected by XRD [49]. Moreover, the synchronous addition of Co and Ni did not form Ni-Co alloy phase. This indicated that active species were strongly interacted with CA support, and all of the as-prepared have a good thermal stability. It also can be seen that the degree of crystallization of all the fresh catalysts are smaller, suggesting that these catalysts are stable at high temperatures, which is coincides with the surface area analysis [50,51]. This is in agreemen<sup>t</sup> with the XRD analysis previously shown and with the observations made in other studies.

**Figure 3.** X-ray diffraction (XRD) diffraction spectrum of the series catalysts.

2.2.3. H2-Temperature-Programmed Reduction (H2-TPR) and NH3-Temperature-Programmed Desorption (NH3-TPD) Analysis

Figure 4 shows the reduction profiles of the CA, and Co, Ni modified CA. It can be seen that there is one or two H2 consumption peaks for CA support at the region of 250~350 ◦C, which could be assigned to the reduction of a small amount of CeO2 to CeO*x* [28,29]. For NCA catalyst, reduction peaks of ~460 ◦C and ~823 ◦C which are attributed to the reduction of NiO and NiAl2O4 [40]. The highest reduction temperature is between 780 and 900 ◦C indicate the existence of species of NiO with strong interaction with Al2O3, resulting from the formation of the NiAl2O4 [42]. For CNCA catalysts with different Co loading, two reduction peaks around ~300 ◦C and ~610 ◦C are ascribed to the reduction of Co2O3 and CoO, respectively [41]. Significantly, the reduction temperatures differences of NiO and Co2O3 between these CNCA catalysts can be attributed to the existence of different interaction between Co and Ni. It is noteworthy that Co addition can obviously promote the reduction of NiO, and reach the optimal effect at 12% Co loading. There was obvious synergistic effect between Ni and Co, which is consistent with the results of the work reported by Jiao et al. [48,52–56].

NH3-TPD technology was used to investigate the acidities of the series catalysts, such as total amount, nature, and strength distribution, in order to look for the possible interpretation for the above experimental results, and the profiles are shown in Figure 5. The area of desorption peak goes hand in hand with the total amount of surface acid sites, while the peak temperature is closely related to the strength of individual acid site. The peak temperature in the range of 80 to 200 ◦C is regarded as weak acid sites, and the desorption temperature between 200 to 400 ◦C is considered the medium acid sites, while the peak temperature locates at 400 to 700 ◦C corresponds to strong acid sites. Figure 5 shows that CA and NCA catalysts have three desorption peaks at the region of 100~500 ◦C, which is regarded as the desorption peak of the weak and medium acid. Moreover, all the Co doped NCA catalysts have one strong desorption peak at 100~400 ◦C. Obviously, the desorption peak temperature moves to a low temperature area by adding Co, suggesting that the amount of acid sites decrease with the introduction of Co. In our previous studies [12,18], we found that the larger acidity and active strong acid centers are easy to give rise to rapid deactivation of the catalyst due to carbon deposition. It is noteworthy that the addition of Co modifier increases the basicity of the NCA catalyst. This process

will result in preventing alkenes further reacting into aromatic or heavier products, which is beneficial to reduce the carbon deposition over catalysts and prolong the work life of catalysts.

**Figure 4.** H2-temperature-programmed reduction (H2-TPR) results of the series catalysts.

**Figure 5.** NH3-temperature-programmed desorption (NH3-TPD) results of the series catalysts.

#### 2.2.4. Transmission Electron Microscope (TEM) Analysis

In this work, the catalytic activity over the C12-NCA catalyst is better than C15-NCA. However, there is somewhat different texture and structural properties between the two. In order to further study the difference between C12-NCA and C15-NCA, the TEM analysis was used to find the influence of microscopic appearance and dispersion on catalytic activity. Figure 6 shows TEM images and Ni or/and Co particle size distributions of C12-NCA and C15-NCA catalysts. It is found that the Ni or/and Co particle size over C12-NCA is mainly focused on 11–20 nm, while the value for C15-NCA is about 16–25 nm. The average particle size of Ni or/and Co of C15-NCA is significantly larger than C12-NCA. Obviously, poor Ni or/and Co distribution over C15-NCA are observed. It may be the reason of the weaker catalytic activity of C15-NCA catalyst. The results indicate that 6% Ni and 15% Co loading is easy to aggregate after 800 ◦C calcination, which may be due to redundant Co enriching on the catalyst surface and exceed the monolayer saturation capacity of the CA support.

**Figure 6.** Transmission electron microscope (TEM) micrographs of the C12-NCA (**a**) and C15-NCA (**b**) catalysts.

In order to further understand the Ni dispersion, the energy dispersive spectrometer (EDS) mappings of C12-NCA and C15-NCA catalysts were examined. As shown in Figure 7, the red sections correspond to Co, the blue sections correspond to Ce, the green sections correspond to Ni, while the yellow sections correspond to Al. Obviously, Ni and Co particles evenly dispersed on the surface of the CA support over C12-NCA. However, some regions overlap each other for C15-NCA. This may be caused by the partial sintering of 6% Ni and 15% Co loading, and form bigger crystallite and lower metal dispersion. Poor nickel dispersion not only notably impact the physiochemical properties of the catalyst, but also impact the catalytic performance in steam reforming of reactions. Therefore, the dispersion state of active species is one of the main reasons of the activity difference among these catalysts.

**Figure 7.** Energy dispersive spectrometer (EDS) mappings of the C12-NCA and C15-NCA catalysts. (**<sup>a</sup>**–**f**) belongs for C12-NCA catalyst; (**a1**–**f1**) belongs for C15-NCA catalyst.
