Promotional effect of magnesium oxide for a stable nickel-based catalyst in dry reforming of methane

The generation of synthesis gas (hydrogen and carbon monoxide mixture) from two global warming gases of carbon dioxide and methane via dry reforming is environmentally crucial and for the chemical industry as well. Herein, magnesium-promoted NiO supported on mesoporous zirconia, 5Ni/xMg–ZrO2 (x = 0, 3, 5, 7 wt%) were prepared by wet impregnation method and then were tested for syngas production via dry reforming of methane. The reaction temperature at 800 °C was found more catalytically active than that at 700 °C due to the endothermic feature of reaction which promotes efficient CH4 catalytic decomposition over Ni and Ni–Zr interface as confirmed by CH4–TSPR experiment. NiO–MgO solid solution interacted with ZrO2 support was found crucial and the reason for high CH4 and CO2 conversions. The highest catalyst stability of the 5Ni/3Mg–ZrO2 catalyst was explained by the ability of CO2 to partially oxidize the carbon deposit over the surface of the catalyst. A mole ratio of hydrogen to carbon monoxide near unity (H2/CO ~ 1) was obtained over 5Ni/ZrO2 and 5Ni/5Mg–ZrO2, implying the important role of basic sites. Our approach opens doors for designing cheap and stable dry reforming catalysts from two potent greenhouse gases which could be of great interest for many industrial applications, including syngas production and other value-added chemicals.

, and Ce [56][57][58] . A brief literature survey of promoter/ modifiers that were utilized over Ni-doped different supports is given in Table 1.
Use of strong solid base as CaO and MgO showed significant improvement and facilitated the catalytic performance with prompt adsorption of slightly acidic CO 2 during dry reforming reaction over Ni-based catalysts. CaO coprecipitated Ni supported ZrO 2 was well studied for different types of carbon species deposited over the catalyst surface during dry reforming of methane 59 . MgO modified Ni system is known for outstanding coking tolerance 60 . Chunwen Sun et al. showed that MgO modification might help to stabilize the lattice oxygen sites of NiO which effectively decrease the carbon deposition or graphitic layer formation 61 . Garcia et al. prepared the Ni/MgO-ZrO 2 -MgO (MgO loading in the range of 1-5 wt%) catalysts by co-precipitation method and found out that the CO 2 and CH 4 conversions were less than 35% 47 . Asencios and Assaf loaded Ni and Mg with different ratios on zirconia support by wet impregnation method and found out that catalyst with 20 wt% Ni and 20 mol% Mg has the best performance, where the activity was less than 80% in the oxidative reforming of methane 49 . Most of the research outputs in the literature used high loading of Ni or MgO (as high as 35 mol%) for the dry reforming reaction as Montoya et al. via sol-gel method 56 and Titus et al. via melt impregnation 50 .
Herein, we prepared four catalysts via incipient wetness impregnation method, where the support was mesoporous zirconia, nickel as the active catalyst, and magnesium oxide as a promoter. We varied the amount of magnesium oxide to find its optimum loading for the best catalytic performance. Furthermore, we optimised the performance by varying reaction temperature. Catalysts were characterized by TGA, N 2 physisorption, XRD, H 2 -TPR, and CO 2 -TPD. To understand the surface chemistry in optimizing the catalytic activity along with the stability of the modified catalyst, CO 2 -TPD, H 2 -TPD and O 2 -TPO of spent catalyst were also performed. A very fine-tuning, among catalytic activity and characterization results were performed; this will help to better understand the surface behaviour towards syngas production from dry reforming of methane.

Results and discussion
The catalytic activity of 5NixMgZr catalysts (x = 0, 3, 5, 7) in terms of CH 4 conversion, CO 2 conversion, and H 2 / CO mole ratio at 700 °C are shown in Fig. 1(A-C) and at 800 °C are shown in Fig. 1(D-F). The TGA results of spent catalysts are shown in Fig. 1(G,H), respectively. It is worth noting that without magnesium oxide modification, catalyst 5Ni/ZrO 2 shows lower catalytic activity than that of magnesium oxide modified catalyst in all cases. At the reaction temperature of 700 °C, 5Ni/xMg-ZrO 2 catalysts showed approximately 50-60% CH 4 conversion and 65-75% CO 2 conversion which were comparable to those in the recent publications 10,47,49,54,62,63 . The TGA results of these spent catalysts also showed carbon deposition. Interestingly, when the reaction temperature was set at 800 °C, it gave a stable performance with constant high conversion up to 500 min in the time-on-stream test (TOS) and no noticeable carbon deposition. Over 5Ni/3Mg-ZrO 2 catalyst, 85% CH 4 conversion, 92% CO 2 conversion and 0.94 H 2 /CO ratios were achieved constantly up to 500 min in the TOS. On the target of H 2 /CO = 1, the performance of the 5Ni/5Mg-ZrO 2 catalyst was found to be the best as it showed 82% CH 4 conversion, 87% CO 2 conversion. The 5Ni/7Mg-ZrO 2 catalyst performance was a little bit lower than that of 5Ni/5Mg-ZrO 2 (78% To understand the surface behaviour of the DRM reaction, we characterised the catalyst thoroughly and discussed the characterization results herein. The surface area analysis indicated that after the addition of MgO, type IV adsorption-desorption curve with H1 hysteresis loop ( Figure S1) was built up. It indicates the narrow distribution of mesopores. The H 2 -TPR surface reduction profiles of fresh 5Ni/xMg-ZrO 2 catalysts are shown in Fig. 3A. 5Ni/ZrO 2 has one small reduction peak in the temperature range of 140-200 °C that attributed to the free NiO species, a shoulder reduction peak at the temperature range of 200-300 °C for "NiO weakly interacted with ZrO 2 support" and a strong peak at 300-450 °C for "NiO that interacted strongly with ZrO 2 support". After the addition of 3.0 wt% MgO, these three peaks diminished and reduction peaks in the intermediate and high-temperature ranges appeared. The high reduction temperature for MgO modified samples could be correlated to the high inherent stability expected for NiO-MgO-solid solution with respect to pure NiO. From the XRD results, also after MgO modification, NiO-MgO-solid solution was found 50 . The intermediate temperature reduction peak in the range of 450-700 °C could be attributed to "NiO-MgO-solid solution weakly interacted with ZrO 2 support" whereas high-temperature reduction peak in the range of 700-900 °C could be claimed to "NiO-MgO-solid solution strongly interacted with ZrO 2 support". As MgO loading was increased from 3.0 wt% to 5.0 wt%, the TCD signal intensity of the intermediate temperature reduction peak was decreased and high-temperature reduction peak was increased. These observations indicated that a higher amount of "NiO-MgO-solid solution strongly interacted on ZrO 2 " was present in 5Ni/5Mg-ZrO 2 than 5Ni/3Mg-ZrO 2 , thus 5 wt% MgO was the optimum loading. At 7 wt% MgO loading, both types of high-temperature peaks were suppressed in comparison to those for 5Ni/5Mg-ZrO 2 . The H 2 -TPR surface reduction profile of spent 5Ni/3Mg-ZrO 2 is shown in Fig. 3B. It showed that TPR peaks in the intermediate and high-temperature regions had got suppressed. Also, it was noticeable www.nature.com/scientificreports/ that a lower reduction temperature peak (0-400 °C) remained preserved as well as shifted to a lower temperature. The H 2 -TPR surface reduction profile of spent 5Ni/5Mg-ZrO 2 indicated the suppression and shifting of high-temperature region peaks to intermediate temperature regions (Fig. S2). These observations indicated that NiO supported on ZrO 2 was less involved whereas "NiO-MgO-solid solution interacted with ZrO 2 support" are significantly involved in DRM. Apart from that, the elimination of carbon deposit by hydrogen gas during methane gasification reaction (C + 2H 2 → CH 4 ) over spent catalyst system was also possible 64 . The CO 2 -TPD profiles of 5Ni/xMg-ZrO 2 are shown in Fig. 3C. Without magnesium oxide modification, the catalyst showed a sharp peak at lower temperature (weak basic sites) region and in intermediate temperature (medium basic sites) regions, but a broad peak in higher temperature regions (strong basic sites). This profile indicated a wide distribution of basic sites. However, after loading of 3.0 wt% MgO, only weak basic sites remained preserved; the rest disappeared. Surprisingly, basic modifier addition caused the disappearance of basicity. XRD of the same sample showed the appearance of NiO-MgO-solid solution as well as the rise of ZrO 2 crystallinity. This means that after the addition of basic 3.0 wt% MgO, basic MgO was engaged in the nurture of NiO-MgO solid solution and supported the crystallinity, thus it caused the disappearance of basicity. It caused the removal of intermediate strength as well as strong strength basic sites from the surface. Again, at 5 wt% MgO loading, peak reappeared in the intermediate temperature region whereas it broadened in high-temperature regions. As the TGA profile of the spent catalyst did not show markable carbon deposition, it is interesting to observe the basic profile of the spent catalyst.
The CO 2 -TPD profile of fresh as well as spent 5Ni/3Mg-ZrO 2 & 5Ni/5Mg-ZrO 2 catalyst are shown in Fig. 3D and Figure S3 respectively. Figures 3D and S3 also include O 2 -TPO and "CO 2 -TPO followed by O 2 -TPO" of spent 5Ni/3Mg-ZrO 2 and 5Ni/5Mg-ZrO 2 catalysts, respectively. It is obvious from the fresh and spent CO 2 -TPD samples that there was a significant decrease in the intensity of basic sites after the reaction over the spent catalysts. However, unlike the fresh samples, the spent catalysts showed a small peak in CO 2 -TPD. Again, a consumption (negative) peak in O 2 -TPO of spent 5Ni/3Mg-ZrO 2 and spent 5Ni/5Mg-ZrO 2 catalyst samples were also seen at about the same temperature region. Interestingly, O 2 -TPO (carried out after CO 2 -TPD) of spent 5Ni/3Mg-ZrO 2 and spent 5Ni/5Mg-ZrO 2 catalysts had no such O 2 consumption peak. It can be explained that www.nature.com/scientificreports/ O 2 consumption peak in O 2 -TPO was due to oxidation of residual carbon by O 2 into CO 2 . So, the small evolution peak in CO 2 -TPD profile also indicated the oxidation of residual carbon deposit by CO 2 . As the carbon deposit on the surface of the catalyst was already oxidized by CO 2 during CO 2 -TPD profile so when O 2 -TPO was carried out after CO 2 -TPD, no evolution peak was found. This confirmed the oxidation of the carbon deposit by CO 2 over the surface of the catalyst 45,55 . Oxidation of carbon deposit by lattice oxygen of ZrO 2 and thereafter simultaneous compensation of the oxygen vacant sites by CO 2 (through losing one of its oxygen to the vacant site) might be a possible route of oxidation of carbon deposit by CO 2 .
To study the conditions and sites of CH 4 decomposition, CH 4 -temperature programmed surface reaction (CH 4 -TPSR) experiment over ZrO 2 , 5Ni/ZrO 2 and 5Ni/3Mg-ZrO 2 were carried out (Fig. 4). It shows a decrease in the methane concentration with temperature over catalysts due to methane decomposition reaction on the surface. For ZrO 2 , a single prominent consumption peak at 870 °C temperature was noticed due to CH 4 interaction at ZrO 2 surface 53 . After the addition of Ni, apart from the high-temperature peak, a lower temperature CH 4 consumption peak at about 350 °C and an intermediate temperature broad peak in the range of 400-800 °C were observed. Low temperature and intermediate temperature peaks could be claimed to the catalytic decomposition of CH 4 over Ni active sites as well as Ni-Zr interface 53 . MgO containing catalysts (i.e. 5Ni/3Mg-ZrO 2 ) also showed the intense peak at high temperature (about 800 °C), attributed to the effect of the temperature. At higher reaction temperature (about 800 °C), an endothermic feature of DRM reaction promotes more efficient catalytic decomposition of CH 4 over Ni and Ni-Zr interface over 5Ni/3Mg-ZrO 2 catalyst systems. This could explain the excellent CH 4 conversion over the magnesium modified catalyst system. It is worth noting that the hightemperature peak is near to the reaction temperature region according to the CH 4 -TPSR profiles. That means if dry reforming of methane was carried out in the temperature region of 700 °C, an advantage of high temperature favourable endothermic feature (about 800 °C) of DRM reaction would be missing as shown in Fig. 4. It might be an indication of lower catalytic conversion at the lower reaction temperature.
Discussion. Thermal decomposition of CH 4 and thereby oxidation of carbon deposits by CO 2 towards dry reforming of methane is albeit possible with little activity i.e. 1.6% CH 4 conversion, 3.6% CO 2 conversion, H 2 / CO = 0.14. So, the catalytic role is utmost demanded in DRM. The summary of the catalytic activity of different catalysts towards dry reforming of methane is shown in Fig. 5. At 700 °C reaction temperature, comparable CH 4 conversion, and CO 2 conversion, were observed. At high reaction temperature, about 800 °C, an endothermic feature of DRM reaction was ruled over. It efficiently promotes catalytic decomposition of CH 4 over Ni and Ni-Zr interface and thereafter oxidation of deposit by CO 2 . So, at 800 °C, all catalysts showed high CH 4 and CO 2 conversion as well as nearly no carbon deposit over the surface of the catalysts. Yang et al. 60 also claimed MgO modified Ni system as outstanding coking tolerance and Chunwen et al. 61 explained the effective reduction of carbon deposit by MgO modified Ni system by stabilization of lattice oxygen sites of NiO by MgO. 5Ni/ZrO 2 had free NiO species, "NiO species interacted with support" and a wide range of basicity. CO 2 uptake at basic sites, catalytic decomposition of CH 4 at Ni and Ni-Zr and oxidation of deposits by CO 2 pivoted the way of high-performance dry reforming reaction. It showed a constant 76% CH 4 conversion, constant 84% CO 2 conversion and 0.99 H 2 /CO ratios for 130 min, then a ratio of 0.98 for 300 min and finally a ratio of 0.96 for 500 min.
After modifying the catalyst with 3.0 wt% MgO, NiO-MgO-solid solution was built up. With a wide range of NiO-MgO-solid solution interaction (weakly as well as strongly with support ZrO 2 ), 5Ni/3 Mg-ZrO 2 promoted the efficient catalytic decomposition of CH 4 over Ni, Ni-Zr interface and thereafter oxidation of deposit by CO 2 . Thus, 5Ni/3Mg-ZrO 2 showed high 85% CH 4 conversion and 92% CO 2 conversion with H 2 /CO ratio ~ 0.96. The CO 2 -TPD, as well as O 2 -TPO profile of spent catalysts, showed an extra peak in TPD and a negative (consuming) peak in TPO, respectively which both related to the oxidation of residual carbon deposits on the www.nature.com/scientificreports/ surface of the catalyst. The CO 2 -TPD along with the O 2 -TPO results showed that CO 2 is capable of oxidizing carbon deposit over the surface of the catalyst. Removal of carbon deposits by hydrogen gas through methane gasification reaction (C + 2H 2 → CH 4 ) is also possible 64 . It resulted in stable catalytic activity up to 500 min in the TOS test. Furthermore, modifying the catalyst with 5 wt%MgO in 5Ni/5Mg-ZrO 2 , it showed more amount of "NiO-MgO solid solution strongly interacted with ZrO 2 support" as well as a wide variety of basic sites. That catalyst showed a constant conversion (82% CH 4 conversion and 89% CO 2 conversion) as well as H 2 /CO ratio = 1 for 250 min in the TOS then slightly decreased to 0.99 for another 250 min, with overall 500 min TOS. Thus, it could be concluded that 5 wt% MgO loading is optimum loading for an active and stable catalyst for methane dry reforming reaction. Further increase in magnesium oxide loading to 7 wt% MgO caused a decrease in NiO-MgO-solid solution that interacted weakly or strongly with the ZrO 2 support and consequently the loss of strong basic sites. Thus, decreasing the CH 4 conversion to 79% as well as CO 2 conversion to 86% and H 2 /CO ratio to 0.98 were noticed.

Conclusion
Magnesium promoted NiO supported mesoporous zirconia, 5Ni/xMg-ZrO 2 (x = 0, 3, 5, 7) were prepared and tested for the methane dry reforming reaction. Higher activity was found at 800 °C than that at 700 °C due to favourable endothermic feature of DRM reaction which promotes efficient CH 4 decomposition over Ni and Ni-Zr interface and successive oxidation of carbon deposits by CO 2 . By modifying the catalyst (5Ni/ZrO 2 ) with MgO as a promoter, NiO-MgO-solid solution was formed. It was found that for high constant CH 4 and CO 2 conversions, NiO-MgO-solid solution played a significant role during the DRM. The 5Ni/3Mg-ZrO 2 catalyst showed a constant 85% CH 4 conversion and 92% CO 2 conversion up to 500 min on stream at H 2 /CO mole ratio ~ 0.96. The highly constant performance of magnesium oxide modified catalysts was due to the ability of CO 2 to oxidize the carbon deposits during the DRM, thus maintaining the catalytic stability. However, with a further loading (> 5.0 wt% Mg) such as in 5Ni/5Mg-ZrO 2 which showed a higher amount of "NiO-MgO-solid solution strongly interacted with ZrO 2 support" along with a wide variety of basic sites as well. Thus, it showed a constant 82% CH 4 conversion and 89% CO 2 conversion and H 2 /CO mole ratio ~ 1. It is hoped that these findings could inspire finding more stable and less expensive synthesis gas production catalysts, including from two potent greenhouse gases emissions such as methane and carbon dioxide.

Experimental
Materials. Nickel  Catalyst preparation. A two-step procedure, based on incipient wetness impregnation as described elsewhere 21 , was followed for synthesizing the desired catalysts. The first step was to dope the support with a metal oxide promoter, while the second step was to load nickel oxide over the promoted support. The detailed description of each synthesis step is given below.  www.nature.com/scientificreports/ formation of a colourless paste, which was mechanically stirred until complete dryness at room temperature. The addition of water and drying processes were performed three times to ensure homogeneous distribution of Mg (CH 3 CO 2 ) 2 within the matrix of meso-ZrO 2 . The solid mixture was then grounded and calcined in a muffle furnace, at 600 °C for 3 h in the static air atmosphere. The resultant materials were designated as xMg-ZrO 2 catalysts where x is wt% of MgO (x = 0, 3, 5, 7).

Synthesis of mesoporous zirconia supported nickel oxide promoted with magnesia (NiO/
MgO-meso-ZrO 2 ). The required amount of Ni (NO 3 ) 2 .6H 2 O to obtain 5.0 wt/wt% of NiO loading was mixed and was crushed with the required amount of MgO-meso-ZrO 2 of the desired MgO wt/wt% loading, forming a green solid mixture. Drops of ultrapure water were then added to get a paste. By continuous mechanical stirring, the paste was dried at room temperature. The wetting and drying processes were repeated three times. Afterwards, calcination was performed at 600 °C for 3 h in static air atmosphere. Overall, 5 wt% NiO loaded catalyst sample is designated as 5Ni/xMg-ZrO 2 catalysts where x is wt% of MgO (x = 0, 3, 5, 7).
Catalyst characterization. The details of instrument specifications and procedures are described in the supporting information and described elsewhere 21 .
Catalyst test. DRM was carried out in a fixed-bed stainless steel tubular micro-reactor (ID = 9 mm) at atmospheric pressure. A load of 0.10 g catalyst was activated under 20 SCCM H 2 flow at 800 °C for 60 min. Then 20 sccm of N 2 was fed to the reactor for 20 min at 800 °C to remove adsorbed H 2 . Afterwards, CH 4 , CO 2 , and N 2 were dosed at flow rates of 30, 30 and 5 sccm, respectively. A GC (GC-2014 Shimadzu) unit, equipped with a thermal conductivity detector and two columns, Porapak Q and Molecular Sieve 5A, was connected in series/ bypass connections to have a complete analysis of the reaction products. The following equations were used to calculate the conversion of each reactant and the H 2 /CO mole ratio, respectively 21 .