1 Introduction

A key concept in circular economy is to recycle and reuse. The valorization of biomass waste (e.g. agricultural waste) to produce clean energy thus enables a successful transition from the current linear to a circular economy of societies. The anaerobic digestion of biomass produces mainly biomethane and carbon dioxide, with minor proportion of other gaseous products such as H2S and NH3. The solid–liquid residue (digestate) is rich in nutrients and can be used as organic fertilizer. [1, 2]

Biomethane can play an important role in a circular economy, especially when further converted to Synthesis gas (Syngas, H2/CO). Syngas is crucial in industrial catalysis for liquid fuel production via Fischer–Tropsch (FT) synthesis, yielding clean fuels (free from S, N, etc.), which then emit less harmful pollutants. When further processed by water gas shift, syngas also represents a source of hydrogen, a promising energy vector for, e.g., fuel-cells. In general, fossil-free resources, more efficient processes and alternative routes are needed to reduce greenhouse gases, global warming and to meet international environmental agreements till 2030 [Paris Agreement; ONU, UNFCCC].

Currently, Syngas is mainly produced from methane (natural gas) via the steam reforming of methane (SRM), yielding a H2:CO ratio of 3:1. Alternatively, dry reforming of methane (DRM) utilizing CO2 yields H2:CO of 1:1. The main drawback is that both routes are based on fossil fuels and energy-demanding (endothermic), but these obstacles can be overcome.

First, methane can originate from biogas, also known as biomethane (purified methane from biogas), making it ecofriendly and sustainable. Second, the Partial Oxidation of Methane (POM, reaction 1) produces Syngas with less energy demand, because POM is exothermic, fast, and its H2/CO ratio is 2 (suitable for FT synthesis or methanol production).

$${\text{POM}}: \quad {\text{2 CH}}_{{4}} + {\text{ O}}_{{2}} \to {\text{ 4 H}}_{{2}} + {\text{ 2 CO}} \quad \Delta {\text{H}} = \, - {22}.{\text{6 kJ}}\,{\text{mol}}^{{ - {1}}}$$
(1)

Group VIIIB elements are good catalysts for methane reforming, especially the low-cost non-noble metals (for example Ni, Co, Fe). Industry uses Nickel-based catalysts for SRM, but they are suffering from carbon deposition (coking), which can lead to catalyst deactivation, and/or to an increase of reactor pressure (highly dangerous clogging [3,4,5,6,7]). Different ways have been suggested to minimize carbon deposition, including using solid solutions as catalyst supports (mixed oxides [3]), decreasing the crystallite size of the metal particles [8, 9], using metal alloys [10, 11], adding promoters, etc.

Herein, we used a combined strategy: (i) applying the POM reaction (O2 effectively removes carbon) and (ii) employing Nickel nanoparticles supported on MgO–ZrO2. Nickel nanoparticles supported on solid solutions (denoted “MgO–ZrO2”) work very well for POM, since the support is rich in oxygen vacancies important for activating oxygen, while avoiding total oxidation of CH4 to CO2. Solid solutions based on zirconia also have high oxygen storage capacity and excellent redox properties, which is beneficial for oxidation reactions, especially for POM. According to some reports [12, 13], these oxygen vacancies also aid in coke removal.

The effectivity of Ni/MgO–ZrO2 catalysts in POM was initially reported by Barbero et al. [14]. Later, Sun et al. [15] studied Ni/MgO–ZrO2 catalysts (prepared by coprecipitation) in the tri-reforming of coal bed methane to syngas, but the composition of the catalyst was not determined in detail [15]. It was argued that intense metal-support interaction, good thermal stability, and the basic nature of the catalyst were responsible for the good performance. Other work reported by Al-Fatesh et al. [16] studied the effect of MgO in NiO–ZrO2 catalysts (obtained by wet impregnation) in DRM. According to them, the interaction of NiO–MgO solid solution with the ZrO2 support was crucial and the reason for high CH4 and CO2 conversions. However, these authors [16] did not observe a stabilization of tetragonal zirconia (that is related to the formation of MgO–ZrO2), so that the role of oxygen vacancies during reaction was unknown. Favorable properties of Ni/MgO–ZrO2 catalysts were also reported by Titus et al. ([17], prepared by melt impregnation) for DRM and by Farooqi et al. ([18], co-precipitation/impregnation), applied for bi-reforming of methane. In all these studies, the exact origins of the favorable effects of MgO–ZrO2 (generating oxygen vacancies) and of NiO–MgO on catalytic performance and carbon removal remained rather unclear.

The current Ni/MgO–ZrO2 catalysts (prepared by a one-step polymerization method) were previously studied (with Ni content of 20 wt% in total) [12, 13, 19, 20] and it was demonstrated that two solid solutions were present: NiO–MgO and MgO–ZrO2. Both solid solutions influenced the catalytic behavior of the catalysts when used for oxidative reforming of methane (ORM) and POM, but their exact role was not understood. The major objective of the current study was thus to thoroughly characterize Ni/MgO–ZrO2 catalysts (with Ni at total weight content of 30%), including in situ techniques, to test their catalytic POM activity, and to elucidate the influence of the two solid solutions, NiO–MgO and MgO–ZrO2 on the POM reaction and the carbon deposition rates. The higher nickel content was chosen to aim for coke deposition and to examine the coke removal performance of the solid solutions under POM conditions.

2 Experimental

2.1 Preparation of Catalysts

The Ni-based catalysts were prepared by a One-Step Polymerization (OSP) method, previously reported in [12, 19], using Ni(NO3)2.6H2O, Zr(CO3)2.1.5H2O, Mg(NO3)3.6H2O, citric-acid, and ethylene–glycol (all of analytical degree). This straightforward preparation method was chosen because it yields a very high dispersion of the catalysts´ components and a homogeneous texture. The as-synthesized polymers were subjected to calcination under an air stream in two consecutive steps at 500 and 750 °C, for a total of 5 h.

The molar percent of MgO in the support was varied: 0, 4, 20, 40, and 100%, of the total mol of the MgO–ZrO2 support: 0% means that the catalytic support is pure ZrO2, whereas for 100% the catalytic support is pure MgO. The amount of Ni was constant at 30% regarding the total weight of the catalyst (such a high loading is frequent for Ni reforming catalysts). Overall, this yielded 5 different catalysts under study labeled as NiZr, Ni4MgZr, Ni20MgZr, Ni40MgZr, and NiMg.

2.2 Characterization Methods

2.2.1 X-ray Diffraction (XRD)

The crystal phases were identified by X-ray diffraction, in a Rigaku Multiflex X-ray diffractometer (40 kV, 30 mA), each analysis scanned in the range 2θ = 5°–80° (at 2° min−1), using Cu Kα radiation as source (λ = 1.5406 Å). The crystal phases were identified by matching the available data in the International Center of Diffraction Data (ICDD-JCPDS).

2.2.2 In Situ XRD

The in situ XRD analysis was carried out on a PANalytical X'Pert Pro diffractometer in Bragg–Brentano geometry using Cu Kα1,2 radiation filtered with an BBHD mirror and an X’Celerator linear detector, the wavelength used for each analysis was 1.5406 Å. Each analysis recorded a 2θ region of 35°–45° (containing the principal peaks of Ni0 and NiO). In every test, approximately 100 mg of the powder catalyst was placed in a ceramic sample holder in a temperature-programmable oven. For in situ experiments, an Anton Paar 900 high temperature chamber was used with the sample temperature monitored/controlled via a thermocouple and heating. Each catalyst was in contact with the respective gas flow. The first XRD pattern was acquired at room temperature in a flow of Ar (100 mL·min−1). The catalyst was then heated from room temperature to 750 °C in 30 mL·min−1 H2 stream (approx. 5% H2/He), at a rate of 10 °C·min−1. Subsequent XRD patterns were collected in situ at 50, 150, 250, 350, 450, 550, 650, and 750 °C. After this procedure, the cell was purged with Ar. Another XRD pattern was acquired in situ after 1 h of POM reaction, exposing the reduced catalyst to a mixture of CH4 (33 mL·min−1), O2 (16.5 mL·min−1) and Ar (60 mL·min−1) at 750 °C.

The average crystallite sizes (d) of the catalysts were determined from XRD and in situ XRD using the Scherrer Equation (d = k·λ/(βhkl·cosθ)), where k is the shape factor (= 0.89), λ is the wavelength of CuKα radiation, θ is Bragg’s angle and βhkl is the full width at half maximum (FWHM) of the principal peak of the crystal phase.

2.2.3 Temperature Programmed Reduction (TPR)

Analysis by Temperature Programmed Reduction with H2 (H2-TPR) was carried out in a multi-purpose quartz reactor, with the H2 consumption measured (in-line) with a Thermal Conductivity Detector (TCD). The quantification of H2 consumption of each catalyst was calculated by comparing the corresponding peak area to that of a standard CuO powder. For each TPR analysis, 100 mg of catalyst and a gaseous mixture containing hydrogen (1.96% H2/Ar, flowing at 30 mL·min−1) were used. The analyses were carried out in the temperature range of 25–1000 °C (at a heating rate of 5 °C·min−1).

2.2.4 Electron Microscopy

High-resolution transmission electron microscopy (HRTEM), high-angle annular dark-field scanning transmission electron microscopy (HAADF-STEM), electron energy loss spectrometry (EELS) and energy filtered TEM (EFTEM) were performed using a 200 kV FEI Tecnai F20 S-TWIN analytical (scanning) transmission electron microscope [(S)TEM]. The catalysts were directly deposited on carbon-coated copper grids, and contamination and adsorbed water were removed by plasma cleaning.

2.2.5 Energy-Dispersive X-ray (EDX) Spectroscopy

The chemical composition of the catalysts was determined by energy-dispersive X-ray spectroscopy (EDX), in a LEO 440 scanning electron microscope (using a tungsten filament coupled to an EDX detector), with three different regions of each sample analyzed.

2.2.6 X-ray Photoelectron Spectroscopy (XPS)

All X-ray photoelectron spectroscopy (XPS) experiments were carried out at room temperature in a stainless-steel UHV chamber (35 l, base pressure < 5 × 10–10 mbar) for surface analysis, equipped with a Specs XR50© high intensity nonmonochromatic Al/Mg dual anode X-ray source and a Phoibos 100© hemispherical energy analyzer (EA) with multichannel plate detector. The conditions used were: Al anode at 1486.6 eV, with the value of 285.0 eV for C1s at 0° emission angle serving as a binding energy reference. XPS data were treated by analysis via CasaXPS software.

2.2.7 Specific Surface Area (SSA)

The SSA of each catalyst was measured in a Quantachrome Nova 1200 instrument, through N2 adsorption/desorption isotherms at liquid nitrogen temperature. The results were treated according to the Brunauer–Emmett–Teller (BET) method.

2.3 Catalytic Tests

Catalytic tests were carried out in a fixed-bed down-flow quartz reactor (i.d. = 10 mm), and in every test 100 mg of catalyst was used. Before reactions, the catalysts were pre-treated with H2 (30 mL·min−1) at 750 °C for 30 min (to reduce NiO to Ni0). After H2 pre-treatment, the catalyst surface was cleaned with inert gas (N2) at the same temperature, followed by POM reaction for 5 h at 750 °C, with an inlet gas mixture with a molar ratio of 2 CH4:1 O2, composed of 32 mL·min−1 (CH4) and 16 mL·min−1 (O2), with oxygen added as synthetic air (60 mL·min−1 of nitrogen and 16 mL·min−1 of oxygen), thus the total flow in the reactor was 108 mL·min−1. The reactor was connected in-line to a gas chromatograph (Varian, Model 3800). A thermocouple inserted directly in the catalyst bed was used to measure and control the reaction temperature.

The gas chromatograph was equipped with an automated injection valve, two chromatographic columns and two thermal conductivity detectors (TCD): hydrogen and methane were separated on a 13X molecular sieve packed column, with nitrogen as the carrier gas, while N2, CO2, CH4, and CO were separated on a Porapak-N packed column, with He as the carrier. All catalytic tests were carried out twice and the values plotted in the respective figures are average values.

The carbon deposition rates (mmol·h−1) were calculated as the apparent mass gain of the catalysts after a specific POM reaction time.

The CH4 conversion was calculated as:

$${\text{Conversion CH}}_{{4}} \, = \, \left( {{\text{Mol CH}}_{{{4}\,{\text{in}}}} - {\text{ Mol CH}}_{{{4}\,{\text{out}}}} } \right)/\left( {{\text{Mol CH}}_{{{4}\,{\text{in}}}} } \right)$$

The H2, CO, CO2 selectivities were calculated as:

$${\text{Selectivity i }} = {\text{ Mol of i}}_{{{\text{produced}}}} {\text{/Mol of CH}}_{{{4}\,{\text{converted}}}}$$

where i = product (H2, CO or CO2).

2.4 Post-reaction Characterization of the Used Catalysts

2.4.1 Carbon Deposition

The carbon deposition rates (mmol·h−1) were determined from the apparent mass gain of the catalysts after a specific POM reaction time, with respect to the reduced catalysts (H2 pre-treated at 750 °C for 30 min) before reaction.

2.4.2 Thermogravimetric Analysis (TGA)

The coke produced after the catalytic tests was characterized by thermogravimetric analysis (TGA), using a Shimadzu DTG-60H simultaneous TG/DTA analyzer, using a continuous air stream flow and a heating rate of 10 °C∙min−1.

2.4.3 Scanning Electron Microscopy (SEM)

The used catalysts were analyzed by Scanning Electron Microscopy (SEM), LEO model 440, equipped with an Oxford detector, operating at 20 kV. Each sample was coated with a gold layer to avoid charge build-up.

3 Results and Discussion

3.1 Characterization of Calcined Catalysts

3.1.1 TEM

To investigate the internal elemental distribution of catalyst particles, (S)TEM was applied in combination with EELS and EFTEM. Figure 1a–c display images of NiZr, Ni4MgZr, and NiMg, respectively. In Fig. 1a, a NiO particle with a dominant (100) facet is imaged in high resolution. The corresponding EELS spectrum (not shown) clearly revealed stoichiometric NiO, which was confirmed by EELS quantification and by comparing the energy loss near edge structure (ELNES) of the Ni-L edge with literature. The supporting ZrO2 had the shape of nano-spheres.

Fig. 1
figure 1

a HRTEM of NiZr; b overview and c EFTEM of Ni4MgZr; d SAED pattern of NiMg catalyst

Full size image

For Ni4MgZr, Fig. 1b shows an overview of a conglomerate of particles, which are chemically quantified in Fig. 1c by means of EFTEM. As can be seen in the diffraction pattern in Fig. 1d, the Ni(x)Mg(1−x)O catalyst shows the lattice parameters of MgO (0.412 nm).

3.1.2 X-ray Diffraction

Figure 2a collects the XRD patterns of the fresh catalysts (after calcination), with Fig. 2b and c showing enlarged regions of 25°–35º and 41º–45º, respectively. According to Fig. 2a, peaks assigned to the tetragonal phase of ZrO2 (JCPDS 17-923) are present in all zirconium-containing samples, while the monoclinic phase of ZrO2 is additionally observed only in NiZr. After addition of MgO (Ni4MgZr, Ni20MgZr, and Ni40MgZr), the monoclinic phase was no longer present, because the MgO–ZrO2 solid solution stabilized the tetragonal ZrO2 structure, as reported previously [13, 20]. In Fig. 2b, a continuous shift of the ZrO2 principal peak to larger Bragg angles is observed for higher MgO content, indicating that the ZrO2 crystal lattice was successively contracted upon formation of the MgO–ZrO2 solid solution. Lattice parameters calculated from the (111) peak of the tetragonal phase of ZrO2 are summarized in Table 1.

Fig. 2
figure 2

XRD patterns of the calcined catalysts: a complete XRD patterns, b enlargement of the XRD pattern for 2θ = 20°–35°; c enlargement for 2θ = 41°–45º. T: tetragonal phase of ZrO2, M: monoclinic phase of ZrO2, NiO: cubic crystal face (fcc) or cubic solid solution (fcc) NiO–MgO

Full size image
Table 1 Properties of calcined catalysts calculated from XRD
Full size table

The contraction can be explained by the ionic radius of Mg2+(0.57 Å) being smaller than that of Zr4+ (0.59 Å). Therefore, during catalyst preparation at high temperature the Mg2+ ions substituted some Zr4+ ions in the ZrO2 crystal lattice, forming a substitutional solid-solution [21]. The contraction increases as more MgO is added. The formation of the MgO–ZrO2 solid solution also involves the generation of oxygen vacancies to maintain electro-neutrality due of the lower oxidation state of the cations at some sites (Mg+2 vs. Zr+4). Furthermore, as the MgO quantity was increased, the diffraction peaks of tetragonal ZrO2 became broader.

MgO addition also caused a shift of NiO peaks to lower Bragg angles (Fig. 2c), pointing to the formation of a NiO–MgO solid solution, as noted in previous reports for similar conditions [13, 20]. The formation of the NiO–MgO solid solution leads to an enlargement of the lattice parameter of NiO (Table 1), coherent with the ionic radius of Mg2+ (0.57 Å) being larger than that of Ni2+ (0.55 Å [21]), suggesting that Mg2+ cations entered the cubic lattice of NiO forming a substitutional solid-solution. This is favored by the structural similarity of NiO and MgO, having face-centered-cubic structures, identical cation charges (Ni2+ and Mg2+), and similar bond distances (2.10 and 2.11 Å for Ni–O and Mg–O, respectively) [22].

A distinct main peak of cubic MgO located at 43.10° (JCPDS 78-0430), observed for NiMg, is absent in the XRD patterns of Ni4MgZr, Ni20MgZr, and Ni40MgZr. This indicates that MgO is taking part in both solid solutions (NiO–MgO and MgO–ZrO2), once more suggesting that Mg2+ cations replaced Zr4+ ions in the ZrO2 crystal lattice and Ni2+ ions in the NiO crystal lattice.

The crystallite sizes calculated from the Scherrer formula are included in Table 1. These values suggest that the addition of MgO to NiZr reduced the size of both tetragonal ZrO2 and cubic (fcc) NiO crystallites.

3.1.3 Specific Surface Area (BET)

Table 2 shows the specific surface area of each catalyst, obtained by N2 adsorption using the BET method, with higher values upon increasing MgO content. This agrees with the smaller crystallite sizes of ZrO2 (see Table 1) formed at higher MgO content.

Table 2 Specific surface area of the catalysts obtained by N2 adsorption using the BET method
Full size table

3.1.4 EDX

The elemental composition of each catalyst was determined by energy-dispersive X-ray spectroscopy (EDX), with results shown in Table 3. The obtained values are close to the nominal composition aimed for in the synthesis, i.e. 30%wt. Ni and molar % of Mg of 4, 20, and 40.

Table 3 Chemical composition of the catalysts, obtained by EDX
Full size table

3.1.5 XPS

Complementing EDX, X-ray photoelectron spectroscopy (XPS) was applied to obtain information about the surface composition of the catalysts. Based on the used photon energy and the inelastic mean free paths (IMFPs) of Ni, Zr and Mg (13, 23, and 24 Å nm, respectively), the corresponding probing depth is up to 6 nm. Figure 3 displays the XPS Ni 2p, Mg 2 s, and Zr 3d, spectra, respectively, with Table 4 collecting the corresponding binding energies (EB, eV) and surface composition.

Fig. 3
figure 3

XPS Ni 2p, Mg 2s and Zr 3d spectra of the calcined catalysts

Full size image
Table 4 Binding energy values (EB, eV) and atomic % obtained by XPS analysis
Full size table

The line shapes and energies of both the Ni 2p (Fig. 3) and Ni LMM region (not shown) agree with literature data of NiO (BE Ni 2p3/2: 855.4 eV (highest intensity peak); BE Ni LMM ~ 844 eV) [23, 24]. The increase of the Ni 2p3/2 BE from 854.9 to 856.0 eV upon increasing Mg content points to the formation of a NiO·MgO solid solution [25] or to larger NiO particle size [26, 27].

For NiMg (Fig. 3), the Mg 2s BE of 88.0 eV corresponds to bulk MgO (literature value of 87.8 eV ([28]). A BE shift up to − 0.7 eV upon increasing in content may again be connected to varying amounts of NiO·MgO solid solution.

Figure 3 also shows the Zr 3d core level spectra. For, NiZr, the Zr 3d5/2 binding energy of ~ 182 eV matches well both theoretical [29] and experimental ([30, 31]) data of ZrO2. A minor increase in BE for the Mg containing catalysts may arise due to the change in coordination number of Zr atoms: according to XRD, only tetragonal ZrO2 was present in these samples (in NiZr, ZrO2 is both tetragonal and monoclinic). Independent of Mg content, additional weak doublets of Zr were found for all samples at ~ 1 and ~ 2 eV lower binding energies, likely corresponding to ZrOx suboxides or thin ZrO2 layers [31].

Apart from the mere BEs, the XPS signal intensities are informative. A comparison of the sample surface vs. bulk composition measured by XPS (Table 4) and EDX (Table 3), respectively, reveals significant differences. The amount of Ni at the surface was 3–6 times smaller than in the bulk. Apparently, the abundance of Zr and Mg at the surface leads to smaller Ni signals.

Mg surface segregation was reported for air calcination of Ni–Mg alloy films ([32]), as well as for Ni/MgO, prepared by incipient wetness impregnation, finally leading to the formation of a NiO·MgO solid solution [25, 33]. This picture is further confirmed by the fact that the ratio of Ni(XPS)/Ni(EDX) decreases for increasing Mg amounts in the sample. The surprising observation that Ni(XPS)/Ni(EDX) is still ~ 0.3 for NiZr (with Mg absent) may be explained by incorporation of Ni2+ into ZrO2, partly forming a NiO·ZrO2 solid solution sufficient to stabilize the tetragonal ZrO2 phase [34, 35]

3.1.6 H2-TPR

Following the characterization of calcined samples, TPR was applied to examine the reducibility of oxide species in the catalysts. Figure 4 displays TPR profiles of all samples and the extent of reduction (%) of each catalyst. The reduction of ZrO2 and MgO–ZrO2 supports up to 1000 °C is almost insignificant, as described in [12, 13]. Therefore, the TPR profiles are attributed to the reduction of NiO (NiO + H2 → Ni0 + H2O). Three peaks can be distinguished at low (peak I) and high temperatures (peaks IIa and IIb), related to NiO species weakly and strongly interacting with the support, respectively. According to TPR, already the addition of 4% MgO to NiO/ZrO2 favored the formation of NiO species strongly interacting with the support (peak IIa, located at higher temperatures than peak I). Further addition of MgO (Ni20MgZr and Ni40MgZr) led to even more strongly bound NiO species (peak IIb). As with increasing MgO content all peaks shifted to higher temperatures, continuous formation of a NiO–MgO solid solution is suggested. The oxygen vacancies produced in the MgO–ZrO2 solid solution (as previously reported in [13, 20]]), apparently do not favor NiO reduction. In NiMg, a well-estsablished NiO–MgO solid solution nearly fully inhibits NiO reduction, owing to electron transfer from NiO to MgO.

Fig. 4
figure 4

TPR profiles of catalysts (Red. = reduction %)

Full size image

3.1.7 In Situ XRD Upon H2 Reduction

In this type of analysis, the catalysts were characterized by in situ XRD during reduction in H2 (Fig. 5), analogous to TPR. Figure 5a corresponds to the NiZr catalyst, indicating that the reduction of NiO to Ni0 set in below 350 ºC, in agreement with TPR (Fig. 4). The peak shift to lower Bragg angles upon further heating is related to the expansion of the crystal lattice at high temperature. During H2 reduction, the crystallite size of Ni0 in NiZr increased from 18 to 37 nm.

Fig. 5
figure 5

XRD patterns of catalysts during reduction in H2 (from 25 to 750 ºC). a NiZr, b Ni4MgZr, c Ni20MgZr, d Ni40MgZr, e NiMg

Full size image

Figure 5b–d show the in situ XRD analysis during H2 reduction of Ni4MgZr, Ni20MgZr, and Ni40MgZr, indicating an onset of NiO reduction around 350, 450, and 450 ºC, respectively. The mean crystallite size of Ni0 increased from 15 to 29 nm, 11 to 21 nm, and 11 to 33 nm, respectively. Accordingly, the reached Ni particle size was smaller for MgO-containing catalysts, which is beneficial as smaller Ni0 crystallites are known to favor catalysis and limit carbon deposits. It seems that MgO–ZrO2 solid solutions hinder the crystal growth of Ni0 during reduction, especially for 4 and 20% MgO. For Ni40MgZr, even at 750 ºC NiO was still present, as in this catalyst a larger quantity of NiO–MgO solid solution is formed. The latter can be understood by the in situ XRD of reduction of NiMg (Fig. 5e), with mostly NiO and only a tiny amount of Ni0 present even at 750 ºC. As expected, the NiO–MgO solid solution cannot be reduced under these conditions.

3.2 Catalytic Tests

3.2.1 Conversion and In Situ XRD

The different pre-reduced catalysts were then tested for POM with results summarized in Fig. 6a–e. The average conversions of methane over Ni4MgZr, Ni20MgZr, Ni40MgZr, and NiMg, shown in Fig. 6a, are in a range of 93–96%, much higher than for NiZr (around 60%). However, for NiZr, the pronounced growth of Ni0 crystallites during reduction (Fig. 5a) and during POM (cf. Figure 6f) can partly explain its lower conversion.

Fig. 6
figure 6

Catalytic test for POM. a Conversion of methane for various catalysts. b Selectivity to H2. c Selectivity to CO. d Selectivity to CO2. e H2/CO ratio. f in situ XRD collected at 750 °C, after 1 h of POM

Full size image

Ni4MgZr and Ni20MgZr exhibited the highest catalytic conversion (Fig. 6a), and according to in situ XRD (Figs. 5b,c, and 6f), they also maintained smaller Ni0 crystallite size during reduction and during POM. The methane conversion over Ni40MgZr decreased with reaction time, probably because on the higher amount of MgO more carbonates formed that poisoned the active centers [36]. Ni sintering (Fig. 6f) may also contribute to the decreasing activity. For NiMg, mostly NiO is present during POM, but despite its minute Ni0 surface concentration, the conversion was still quite high. As reported by [22, 37,38,39], the NiO–MgO solid solution provides a good Ni distribution in the catalyst.

For all samples, the initial Ni0 crystallite size of the reduced catalysts increased within 1 h of POM reaction (Figs. 5a–d and 6f): NiZr (from 37 to 42 nm), Ni4MgZr (from 29 to 31 nm), Ni20MgZr (from 21 to 29 nm), and Ni40MgZr (from 33 to 43 nm).

3.2.2 Selectivity and Mechanism

Figure 6b and c display the selectivities to H2 and CO, respectively, monitored over reaction time. The selectivities are high and quite similar for Ni4MgZr, Ni20MgZr and NiMg, but lower for Ni40MgZr and much lower for NiZr.

Two general mechanisms have been proposed for Syngas production by POM: (i) a combustion-reforming mechanism, which initially involves the fast total combustion of methane (TCM), followed by Dry Reforming of Methane (DRM) and Steam Reforming of Methane (SRM), as shown in reactions (2–4); (ii) a pyrolysis mechanism, in which Syngas is produced directly according to reaction (1). Both mechanisms lead to H2 and CO at a molar ratio of 2 (in case of (i), the global sum TCM + DRM + 2 × SRM yields POM).

$${\text{TCM}}: {\text{CH}}_{{4}} + {\text{ 2O}}_{{2}} \to {\text{ CO}}_{{2}} + {\text{ 2H}}_{{2}} {\text{O}} \quad \Delta {\text{H}} = - {89}0{\text{ kJ}}\,{\text{mol}}^{{ - {1}}}$$
(2)
$${\text{DRM}}: {\text{CH}}_{{4}} + {\text{ CO}}_{{2}} \to {\text{ 2CO}} + {\text{ 2H}}_{{2}} \quad \Delta {\text{H}} = { 26}0.{\text{5 kJ}}\,{\text{mol}}^{{ - {1}}}$$
(3)
$${\text{SRM}}: {\text{CH}}_{{4}} + {\text{ H}}_{{2}} {\text{O}} \to {\text{CO}} + {\text{ 3H}}_{{2}} \quad \Delta {\text{H}} = { 225}.{\text{4 kJ}}\,{\text{mol}}^{{ - {1}}}$$
(4)

In the catalytic tests, traces of water and CO2 were indeed detected, which is why the combustion-reforming mechanism is favored for the studied catalysts.

Figure 6d shows the CO2 selectivity during POM. Almost all catalysts have very low selectivity to CO2, only for NiZr the values are higher. Apparently, POM over NiZr involves TCM. The inherent oxygen vacancies present in ZrO2 (tetragonal and monoclinic) are active in oxygen activation increasing the number of O(s) species [40,41,42], responsible for TCM and the resulting CO2.

The MgO–ZrO2 solid solution, and smaller amounts of NiO–MgO, formed in Ni4MgZr, Ni20MgZr and Ni40MgZr, thus favored the activation of CO2 and H2O. As reported in [43, 44] their oxygen vacancies dissociate CO2, resulting in a release of CO and chemisorbed O(s) species (CO2 + oxygen vacancy → CO + O(s)). O(s) can then react with CH4. Similarly, the oxygen vacancies interact with H2O forming H2 and chemisorbed O(s) species [43]. This is consistent with previous reports [45,46,47,48] on the beneficial effects of zirconia based solid-solutions in DRM and SRM.

All observations point to an important catalytic role of the interface sites between Ni0 and MgO–ZrO2, but also of NiO–MgO, with the latter increasing for higher MgO content. The oxygen vacancies present in Ni4MgZr, Ni20MgZr, and Ni40MgZr are already beneficial for POM, but it seems that also the basic centers in these catalysts (owing to the basic character of MgO) may enhance CO2 and H2O activation, i.e. DRM (Reaction 3) and SRM (Reaction 4). Consequently, the selectivity to CO is increased (cf. Figure 6c). The effect of basic centers is apparent from the results of NiMg (highest selectivity to H2 and CO). The role of MgO is to form intermediate carbonate species that rapidly decompose, enabling the further conversion of CO2 formed during POM [49, 50]. In summary, it is suggested that NiMgZr catalysts benefit both from active interfaces/oxygen vacancies (owing to MgO–ZrO2 and to a lesser extent NiO–MgO) and basic centers (owing to MgO).

3.2.3 H2/CO Ratio

Based on the results so far, POM likely occurs via a combustion-reforming mechanism. Nevertheless, on NiZr TCM is still significant, whereas the other catalysts favor subsequent DRM and SRM. This trend is also reflected by the H2/CO ratio (Fig. 6e): as NiZr produced more CO2 than CO, a higher H2/CO ratio (> > 2) is obtained. On the other hand, for the other catalysts (Ni4MgZr, Ni20MgZr, Ni40MgZr, and NiMg) the H2/CO ratio was very close to 2 (which is the stoichiometric ratio of Syngas produced by POM 2CH4:1O2). Once more, this suggests that these catalysts promote DRM and SRM with similar contributions, so that the H2/CO ratio is closer to the stoichiometric value. The role of TCM, DRM, and SRM in POM over different catalysts was also described in [51].

3.3 Post Reaction Analysis

3.3.1 Carbon Deposition Rates

Table 5 collects the (absolute) carbon deposition rate of each catalyst, measured after 5 h of POM, yielding an order of: Ni20MgZr≈Ni40MgZr > NiMg >  > Ni4MgZr > NiZr. Please note that during reaction Ni is mostly metallic (Fig. 6f), which is why the reduced catalysts serve as starting point (the partial oxidation to NiO upon reaction is insignificant). Table 5 also shows the average values of conversion of methane, and selectivity for H2 and CO. Table 5 also compares the carbon deposition rate normalized by the average conversion and relative to NiZr. The best catalyst is thus Ni4MgZr, as it had the highest methane conversion (Fig. 6a), good selectivity to H2 and CO, but still a low carbon deposition rate. It was previously suggested [52, 53] that large nickel particles favor carbon deposition, so that the larger increase in Ni0 crystallite size of Ni20MgZr and Ni40MgZr may contribute to their higher carbon deposition rates (Table 4). The lowest carbon deposition rate observed for NiZr is not a good comparison, as it also had much lower conversion.

Table 5 Carbon deposition rates (based on weighthing the reduced catalysts before reaction and the used catalysts after reaction)
Full size table

Oxygen vacancies (e.g. of the MgO–ZrO2 solid solution) favor oxygen dissociation to form various reactive oxygen species adsorbed on the surface of the catalyst: O2−, O22−, O2−, O− [40, 42], which may remove carbon deposits by reacting with the surface carbon species (CHx + O(s) → CO + x H; x = 1,2,3). However, it appears that a higher MgO amount in the NiO–MgO–ZrO2 mixture was not beneficial for lowering carbon deposits during POM. Although NiMg also had high methane conversion, its carbon deposition rate was quite high, which may be due to the absent MgO-ZrO2 solid solution.

It seems that the excess of MgO in the NiO–MgO–ZrO2 mixture was not beneficial for reducing carbon deposits during the POM reaction, probably because the excess of MgO formed more carbonates that affected the oxygen vacancies required for carbon removal. A previous study [36] demonstrated that this excess also hindered methane decomposition in the course of DRM. Among the catalysts studied herein, a composition of Ni4MgZr was best for limiting carbon deposits.

3.3.2 SEM

The coke formed during POM and deposited on each catalyst was analyzed by SEM, with Fig. 7a–e showing the morphology of the used catalysts. On NiZr, amorphous carbon was formed, as no defined shapes were observed (Fig. 7a). Figure 7b shows that the major part of carbon deposited on Ni4MgZr is amorphous like on NiZr, but very short filaments can be observed as well. The presence of carbon filaments increased progressively for Ni20MgZr and Ni40MgZr, with the filaments also becoming longer. It seems that the addition of MgO affected the carbon morphology, leading to more and longer filaments. Finally, in NiMg, only long filaments were observed.

Fig. 7
figure 7

a–e Scanning Electron Microscopy (SEM) images of used catalysts: a NiZr; b Ni4MgZr; c Ni20MgZr; d Ni40MgZr; e NiMg

Full size image

3.3.3 TGA

Figure 8 shows the TGA profiles of catalysts used for 5 h of POM. The initial weight increase of used catalysts is related to the oxidation of Ni0 present after POM, with the exception of NiMg which does not contain Ni0. The following weight loss upon heating in air is then related to the gasification of carbon deposited on the catalyst (Fig. 8). Every used catalysts showed a pronounced weight loss between 500 and 750 °C, indicating that the coke formed was mainly graphitic carbon, typical of carbon with hierarchical structures (such as carbon nanotubes) [54].

Fig. 8
figure 8

Results of thermogravimetric analysis (TGA) in air of used catalysts after 5 h of POM

Full size image

According to the TGA profiles, the carbon-related weight loss of the used catalysts were: NiZr (not determined, shows an increase due to oxidation), Ni4MgZr (6%), Ni20MgZr (28%), Ni40MgZr (28%), and NiMg (5%). For NiMg, the carbon seems quite stable, as gasification was low. Roughly, the carbon-related weight loss follows the same trend as the carbon deposition rates in Table 5 (Ni20MgZr≈Ni40MgZr>>Ni4MgZr), corroborating that Ni4MgZr is the best catalyst for limiting carbon deposition during POM.

4 Conclusions

Pre- and post-reaction analysis of various Ni-based catalysts, including different MgO–ZrO2 supports, and complemented by in situ XRD during H2 activation and POM reaction, yielded the following conclusions.

  • In catalysts of NiO/MgO/ZrO2, synthesized by a one-step polymerization method and calcined at 750 °C, two solid-solutions were formed upon increasing MgO content: MgO–ZrO2 and NiO–MgO. Their presence was confirmed and characterized by XRD and XPS.

  • Catalysts were pretreated by H2 reduction at 750 °C, which was monitored by H2-TPR and in situ XRD. With higher MgO content, the onset temperature of NiO reduction increased, but NiO persisted for NiO–MgO.

  • In the partial oxidation of (bio)methane to syngas, Ni (30 wt%) supported on various MgO–ZrO2 solid solutions showed high methane conversion (~ 95%), good selectivity to H2 and CO (ratio of 2.2) and low carbon deposition rates, overall outperforming Ni/ZrO2.

  • Ni4MgZr (4 mol% MgO) turned out to be the best catalyst. In situ XRD during POM revealed metallic Ni nanoparticles (average crystallite size of 31 nm), supported on MgO-ZrO2 solid solution, with small amounts of NiO–MgO being present as well.

  • The amount of MgO in the Ni/MgO–ZrO2 catalysts also influenced the morphology of the carbon deposits. The more MgO, the more and longer carbon filaments were formed. The best catalyst (Ni4MgZr) predominantly produced amorphous carbon, with few short filaments.

  • Although NiO-MgO also exhibited high conversion, it also suffered from rather high carbon deposition producing termally stable filaments.