Investigation of carbon dioxide photoreduction process in a laboratory-scale photoreactor by computational fluid dynamic and reaction kinetic modeling

Abstract

The production of solar fuels via the photoreduction of carbon dioxide to methane by titanium oxide is a promising process to control greenhouse gas emissions and provide alternative renewable fuels. Although several reaction mechanisms have been proposed, the detailed steps are still ambiguous, and the limiting factors are not well defined. To improve our understanding of the mechanisms of carbon dioxide photoreduction, a multiphysics model was developed using COMSOL. The novelty of this work is the computational fluid dynamic model combined with the novel carbon dioxide photoreduction intrinsic reaction kinetic model, which was built based on three-steps, namely gas adsorption, surface reactions and desorption, while the ultraviolet light intensity distribution was simulated by the Gaussian distribution model and Beer-Lambert model. The carbon dioxide photoreduction process conducted in a laboratory-scale reactor under different carbon dioxide and water moisture partial pressures was then modeled based on the intrinsic kinetic model. It was found that the simulation results for methane, carbon monoxide and hydrogen yield match the experiments in the concentration range of 10⁻⁴ mol·m^–3 at the low carbon dioxide and water moisture partial pressure. Finally, the factors of adsorption site concentration, adsorption equilibrium constant, ultraviolet light intensity and temperature were evaluated.

Graphical abstract

Keywords

carbon dioxide photoreduction / computational fluid dynamic simulation / kinetic model / Langmuir adsorption

1 Introduction

Supported metal catalysts are widely used in the coal chemical technology, petrochemical industry, fine chemicals and other fields [1−3]. To obtain highly efficient supported metal catalysts, increasing the number of accessible active sites is a direct and effective way to boost the specific activity of a catalyst by increasing the metal loading [4−6]. For example, Ni loadings up to ~40–60 wt % for Ni-based catalysts are practiced for the commercial methanation process, thereby enhancing the catalytic activity with more accessible Ni active sites [7]. However, excessively high metal loadings can weaken the metal-support interactions, giving rise to the serious sintering of the metal with an increased particle size, which is adverse for the catalytic activity [8−10]. It is well known that small-size Ni particles with a high dispersion are continuously sought for a higher activity [11]. Therefore, increasing the metal loading but simultaneously maintaining a high metal dispersion is an effective strategy to prepare supported metal catalysts with higher performance.

As a common fact, smaller metal particles have a greater tendency to aggregate and sinter at high temperatures, resulting in a decreased metal dispersion. Moreover, the dispersion-reducibility dependence of supported metal catalysts is closely related to the metal-support interactions, i.e., stronger metal-support interactions associated with a higher metal dispersion but a lower metal reducibility and vice versa. To date, breaking the reducibility-dispersion dependence of the catalyst with higher metal loadings is still a great challenge for the development of high-performance supported metal catalysts. To this end, substantial efforts have been pursued including the regulation of metal-support interactions [12,13], and the confinement of metal nanoparticles in a well-defined structure such as core-shell, yolk-shell, and ordered mesoporous structures [14−17]. In this respect, an important contribution for breaking the reducibility-dispersion dependence of Co-based catalysts is achieved by loading the pre-synthesized Co nanoparticle precursor on the surface-protected delaminated ITQ-2 zeolite [18,19]. However, the Co loading is only 10 wt %, which is still far from around 20 wt % required for the industrially relevant Co loadings. Meanwhile, the metallic Co nanoparticles are significantly sintered after a time-on-stream (TOS) of 7–8 h under realistic Fischer-Tropsch synthesis conditions of 220 °C, 2.0 MPa, and a H₂/CO molar ratio of 2. Recently, carbon supported metal catalysts with sub-3 nm transition metal nanoparticles of Ni, Co, and Fe at high loadings up to ~41.0% are prepared by the millisecond pyrolysis of the metal-organic frameworks (MOFs) materials, and the aggregation of metal nanoparticles is kinetically inhibited as a result of the encapsulation of the carbon matrix derived from MOF ligands [20,21]. Unfortunately, the complicated route together with the expensive MOFs may limit its scaled preparation of a large amount of catalysts.

In fact, solid-state grinding has been developed as a simple but effective method to directly introduce guest precursors into the channels of ordered mesoporous substrates filled with templates. Moreover, the subsequent removal of the template via the extraction or calcination leads to mesoporous composites with highly dispersed oxides such as ZnO, Co₃O₄, CuO, or NiO with a high loading up to around 36.0 wt % [22−25]. Without the competitive adsorption of solvent and unnecessary solvent evaporation, solid-state grinding is not only economical of time and energy, but also is beneficial to an increased dispersion of metal oxides induced from the enhanced interactions between metal oxide precursors and the residual template in the confined space of ordered mesoporous structures [25]. This is clearly reflected by simply changing the pretreating temperature of template-containing ordered mesoporous alumina (OMA), in which the interactions between Ni precursor and residual P123 templates can be facilely tuned. As a result, highly dispersed metallic Ni particles embedded in the mesochannels show a good activity and stability in the CO methanation reaction, which is due to the absence of the dispersion-reducibility dependence and the limitation of high temperature reduction [26]. Thus, this strategy developed in our previous work [26] is expected to break the loading-reducibility-dispersion dependence of supported metal catalyst for achieving a high-performance catalyst.

Herein, the solid-state co-grinding is demonstrated to be an efficient method for breaking the loading-reducibility-dispersion dependence of supported Ni catalysts. Results indicate that the ordered OMA mesostructure was still preserved even after loading up to 40 wt % NiO. Importantly, the similarly high reduction extent, high dispersion, and small Ni nanoparticles were obtained for the Ni/OMA catalysts when the NiO loading was increased from 20 to 40 wt %. The catalysts were probed for the CO methanation as a model reaction, and the similarly high turnover frequency (TOF) was achieved for all of the Ni/OMA catalysts. Over the catalyst with the highest NiO loading of 40 wt %, both a high low-temperature activity and a good long-term stability under harsh conditions were achieved.

2 Experimental

2.1 Chemicals

The commercial reagents with an analytical grade were used without any further purifications. Nickel nitrate hexahydrate (Ni(NO₃)₂·6H₂O), nitric acid (67 wt %), hydrochloric acid (37 wt %) and anhydrous ethanol were obtained from Sinopharm Chemical Reagent Co., Ltd. Aluminum isopropoxide (Al(OPrⁱ)₃) and triblock copolymer ((EO)₂₀(PO)₇₀(EO)₂₀, Pluronic P123, M_n = 5800) were purchased from Sigma-Aldrich.

2.2 Procedure for synthesizing catalysts

The template-containing OMA precursor was synthesized via the modified evaporation induced self-assembly (EISA) method as reported in our previous work [27]. Typically, 2.1 g Pluronic P123 was dissolved in 40 mL anhydrous ethanol at room temperature, and then 3.2 mL nitric acid and 4.08 g Al(OPrⁱ)₃ were successively added under a magnetic stirring. After further stirring for 6 h, the obtained mixture was transferred to a petri dish and the EISA process was carried out by evaporating the solvent at 60 °C for 48 h. Finally, the resultant yellow xerogel was further treated at 120 °C for 12 h to obtain the template-containing OMA precursor, which is named as OMA-P. For comparison, OMA support was prepared by calcining the yellow xerogel at 600 °C for 4 h with a temperature ramp of 1 °C·min^–1, and named as OMA-600.

The solid-state grinding procedure as reported in our previous work [26] was applied to prepare Ni/OMA catalysts. After thoroughly mixing desired amounts of Ni(NO₃)₂·6H₂O and OMA-P (Table S1, cf. Electronic Supplementary Material, ESM), the hybrid in a mortar was ground manually for 1 h at room temperature. Then, the solid was transferred to a tubular furnace, and the calcination in an air flow was performed at 600 °C for 4 h with a temperature ramp of 2 °C·min^–1. Finally, xNi/OMA catalysts were obtained, where x represents the weight content of NiO, i.e., 20, 25, 30, and 40 wt %, respectively. In the same way, OMA-600 was used to prepare 40Ni/OMA-600 for a comparison purpose. In addition, a catalyst counterpart with 40 wt % NiO was synthesized via the one-pot EISA method, and referred to as 40Ni-OMA.

2.3 Catalyst characterizations

The N₂ adsorption-desorption isotherms were measured on a Micromeritics ASAP 2020 instrument at –196 °C. The samples were degassed at 300 °C for 8 h under vacuum condition prior to each test. Small-angle and wide-angle X-ray diffraction (XRD) patterns were measured on a Bruker D8 advance X-ray diffractometer with Cu/K_α radiation at 40 kV and 40 mA. The samples were scanned from 0.5° to 6° at the speed of 1 °·min⁻¹ and 10° to 90° at the speed of 6 °·min^–1, respectively.

Temperature-programmed reduction of hydrogen (H₂-TPR), H₂ pulse chemisorption, and O₂ titration experiments were conducted on a Micromeritics AutoChem 2920 analyzer. For H₂-TPR, all samples were pretreated at 450 °C under a pure Ar flow for 1 h and then cooled to 50 °C. After changing the gas to 10 vol % H₂/Ar mixed gas, the temperature was raised from 50 to 1000 °C with a temperature ramp of 10 °C·min^–1. For the H₂ pulse chemisorption experiment, the sample was pre-reduced with purified H₂ at 700 °C for 1 h, and then cooled to 35 °C under pure Ar flow. Subsequently, the 10 vol % H₂/Ar mixed gas was pulse-dosed consecutively until a constant area of the thermal conductivity detector (TCD) peak. For the O₂ titration experiment, followed by pre-reducing at 700 °C for 1 h in pure H₂ flow, the pre-reduced sample was purged at 600 °C under pure Ar flow for 0.5 h, and then was re-oxidized by pulsing 3 vol % O₂/Ar mixed gas consecutively at 600 °C until a constant area of TCD peak. The reduction degree was calculated by the total amount of oxygen consumption that is determined by a pre-calibrated TCD.

Transmission electron microscopy (TEM) and high-resolution TEM (HRTEM) were carried out on a FEI Tecnai G2-F20 transmission electron microscope with an accelerating voltage of 200 kV. Thermogravimetric and differential scanning calorimetry (TG-DSC) analyses were performed on a NETZSCH STA 449 F3 instrument. The TG-DSC experiment was conducted from 35 to 1000 °C with a heating rate of 10 °C·min^–1 in an air flow.

2.4 Catalytic reaction

The Ni-based catalysts were evaluated for the CO methanation in a quartz-lined stainless-steel tubular reactor (i.d. = 8 mm). Prior to the reaction, the loaded catalyst (50.0 mg, 40–60 mesh) diluted with 4.0 g of quartz sands (40–60 mesh) was reduced at 700 °C for 1 h in a pure H₂ flow. After reduction, the reaction was performed under the conditions of H₂/CO/N₂ = 3/1/1, 0.1 MPa, 300–450 °C, and the gas hourly space velocity (GHSV) of 240000 mL·g^–1·h^–1. The tail gas passing through an ice-water trap was analyzed by an online gas chromatograph, and the detailed columns and analysis conditions were reported in our previous work [27]. The CO conversion (X_CO), reaction rate indexed by converted CO (r_CO), CH₄ selectivity ($S_{{\mathrm{C}\mathrm{H}}_4}$), CH₄ yield ($Y_{{\mathrm{C}\mathrm{H}}_4}$), space-time yield of CH₄ (${\mathrm{S}\mathrm{T}\mathrm{Y}}_{{\mathrm{C}\mathrm{H}}_4}$) and TOF were calculated as follows.

$X_{\mathrm{C}\mathrm{O}}=(F_{\mathrm{C}\mathrm{O},\mathrm{i}\mathrm{n}}-F_{\mathrm{C}\mathrm{O},\mathrm{o}\mathrm{u}\mathrm{t}})/F_{\mathrm{C}\mathrm{O},\mathrm{i}\mathrm{n}}\times 100\mathrm{\%},$

$S_{\mathrm{C}{\mathrm{H}}_4}=F_{\mathrm{C}{\mathrm{H}}_4,\mathrm{o}\mathrm{u}\mathrm{t}}/(F_{\mathrm{C}\mathrm{O},\mathrm{i}\mathrm{n}}-F_{\mathrm{C}\mathrm{O},\mathrm{o}\mathrm{u}\mathrm{t}})\times 100\mathrm{\%},$

$Y_{\mathrm{C}{\mathrm{H}}_4}=F_{\mathrm{C}{\mathrm{H}}_4,\mathrm{o}\mathrm{u}\mathrm{t}}/F_{\mathrm{C}\mathrm{O},\mathrm{i}\mathrm{n}}\times 100\mathrm{\%},$

${\mathrm{S}\mathrm{T}\mathrm{Y}}_{\mathrm{C}{\mathrm{H}}_4}=\mathrm{G}\mathrm{H}\mathrm{S}\mathrm{V}\times Y_{\mathrm{C}{\mathrm{H}}_4},$

$r_{\mathrm{C}\mathrm{O}}=(F_{\mathrm{C}\mathrm{O},\mathrm{i}\mathrm{n}}-F_{\mathrm{C}\mathrm{O},\mathrm{o}\mathrm{u}\mathrm{t}})/(m_{\mathrm{c}\mathrm{a}\mathrm{t}}\times 22.4{\mathrm{L}\cdot \mathrm{m}\mathrm{o}\mathrm{l}}^{-1}),$

$\mathrm{T}\mathrm{O}\mathrm{F}=(F_{\mathrm{C}\mathrm{O},\mathrm{i}\mathrm{n}}-F_{\mathrm{C}\mathrm{O},\mathrm{o}\mathrm{u}\mathrm{t}})/(N_{\mathrm{N}\mathrm{i}}\times 22.4{\mathrm{L}\cdot \mathrm{m}\mathrm{o}\mathrm{l}}^{-1}),$

where F_{i, in} and F_{i, out} are the flow rates of species i (CO or CH₄) at the inlet and outlet of the reactor, respectively. m_cat is the amount of the catalyst loaded into the reactor (50.0 mg). N_Ni is the number of the metallic Ni atoms (mol) over the loaded catalyst.

3 Results and discussion

To investigate the reducibility and dispersion, the weight content of NiO (x) over xNi/OMA catalysts was designed to be 20, 25, 30, and 40 wt %, respectively. To determine the exact NiO content, the calcined catalysts were fully digested in a mixed solution of concentrated nitric acid (67 wt %) and hydrochloric acid (37 wt %). After this, the solution was subjected to the analysis with inductively coupled plasma mass spectroscopy (M90, Bruker). As shown in Table S1, the error between the designed and measured NiO contents over all of the catalysts was within 2.5%. Thus, the catalysts are safely named with the nominal loading of NiO, and the results are discussed as follows.

3.1 Crystal structures and textural properties

The XRD patterns of fresh and reduced catalysts are shown in Fig.1. The small-angle patterns (Fig.1(a)) indicated that all of the catalysts synthesized via the solid-state grinding had a two-dimensional ordered mesoporous structure even after loading up to 40 wt % NiO. It was evidenced by the appearance of a strong diffraction at 2θ = 0.7° – 0.9° and a weak peak at 2θ = 1.6°, which can be indexed as the (100) and (110) lattice plane, respectively [26]. However, there were no XRD diffractions for 40Ni-OMA synthesized via the one-pot EISA method in the small-angle region, indicating the absence of ordered mesoporous structure. In the cases of the wide-angle patterns of the calcined catalysts (Fig.1(b)), very sharp and strong diffractions at 2θ of 37.4°, 43.7°, and 63.2° were observed for 40Ni/OMA-600 and 40Ni-OMA, which can be easily assigned to the NiO phases. These results indicate the formation of larger NiO particles. Moreover, the peaks at 2θ of 45° and 65.6° assigned to the NiAl₂O₄ spinel were clearly displayed for 40Ni/OMA-600 and 40Ni-OMA. In contrast, very broad and weak diffractions at 2θ of 37.4°, 43.7°, and 63.2° were found for all of the Ni/OMA catalysts, indicating the presence of highly dispersed NiO particles. In addition, the peak intensity of the NiO phase was slightly increased with increase of the NiO loading from 20 to 40 wt %.

Fig.1 (a) Small-angle and (b) wide-angle XRD patterns of the fresh catalysts and (c) wide-angle XRD patterns of the reduced catalysts ((i) 20Ni/OMA, (ii) 25Ni/OMA, (iii) 30Ni/OMA, (iv) 40Ni/OMA, (v) 40Ni-OMA, and (vi) 40Ni/OMA-600).

Full size|PPT slide

As indicated from the XRD patterns of the reduced catalysts (Fig.1(c)), the metallic Ni phase was prevalent over all the catalysts by the detected peaks at 2θ of about 44.6°, 51.8°, and 76.6°, which are assigned to the (111), (200) and (220) lattice planes of metallic Ni (JCPDS 01-070-1849), respectively. Moreover, the intensity of these diffractions was very similar for the Ni/OMA catalysts, and the corresponding mean crystal size of metallic Ni was determined to be about 4.4–4.9 nm based on the (111) diffraction and the Scherrer’s equation, as listed in Tab.1. In contrast, 40Ni/OMA-600 showed the largest Ni crystal size of 25.8 nm, which is much larger than those of 40Ni/OMA (4.9 nm) and 40Ni-OMA (11.2 nm). In addition, the main diffractions of γ-Al₂O₃ at 37.6° and 67.0° were clearly observed over all of the catalysts, which is attributed to the accelerated crystallization of the alumina during the high-temperature reduction [27].

Tab.1 Summary of the crystal and reductive properties of different catalysts

Samples	Ni particle size/nm			Degree of reduction/%	Dispersion/D%c)
	XRDa)	TEM	H2-chemb)
40Ni/OMA-600	25.8	30.6 ± 8.8	38.8	86.3	2.5
40Ni-OMA	11.2	12.8 ± 3.8	18.7	83.5	5.2
40Ni/OMA	4.9(6.6)	5.2 ± 1.3	7.2	95.4	13.5
30Ni/OMA	4.5	4.8 ± 1.3	7.0	91.0	13.8
25Ni/OMA	4.4	4.7 ± 1.2	7.2	93.0	13.4
20Ni/OMA	4.4	4.6 ± 1.1	7.1	91.9	13.7

(a) The number in parenthesis was for the spent catalyst after a TOS of 120 h; (b) it was calculated from the equation of dNi = 97/(D%) [27]; (c) it was determined from the chemisorption results by H2 pulses.

TEM and HRTEM images and particle size distributions of reduced catalysts are presented in Fig.2. Apparently, the ordered cylindrical mesopores oriented along the [110] direction were observable in the cases of Ni/OMA and 40Ni/OMA-600, demonstrating that the mesoporous structure is successfully retained even after the addition of as high as 40 wt % NiO and high-temperature calcination/reduction. As expected, no ordered cylindrical pore was found for 40Ni-OMA. As shown in Fig.2(e), a clear lattice spacing of 0.20 nm assigned to the Ni (111) crystal plane was observed for 40Ni-OMA, indicating the presence of metallic Ni [28]. As for the particle size distribution of Ni, Ni/OMA catalysts showed a very similar particle size distribution of metallic Ni (4.6–5.2 ± 1.3 nm) as the NiO loading increased from 20 to 40 wt %. However, the Ni particles with a larger size and non-uniform size distribution were observed in the cases of 40Ni-OMA (12.8 ± 3.8 nm) and 40Ni/OMA-600 (30.6 ± 8.8 nm). This changing pattern of the particle size distribution of metallic Ni is in good accordance with that determined by XRD.

Fig.2 TEM/HRTEM images and particle size distributions for the reduced catalysts of (a) 20Ni/OMA, (b) 25Ni/OMA, (c) 30Ni/OMA, (d, e) 40Ni/OMA, (f) 40Ni-OMA, (g) 40Ni/OMA-600, and (h) the spent 40Ni/OMA catalyst after a TOS of 120 h.

Full size|PPT slide

The textural properties of the calcined catalysts and OMA-600 were assessed and summarized in Fig.3 and Tab.2. Clearly, all the samples displayed a typical type IV isotherm with a sharp H1 hysteresis loop, demonstrating the existence of mesoporous structure. In comparison with OMA-600, a clear drop for the total volume of absorbed N₂ and the amount of absorbed N₂ in the hysteresis-loop region occurred over Ni/OMA and 40Ni/OMA-600 catalysts, indicating the decreased pore volume and pore size. This is related to the inclusion of NiO species into meso-channels. As shown in Tab.2, regardless of the NiO loadings, Ni/OMA catalysts showed very similar textural data, i.e., the Brunauer-Emmett-Teller (BET) surface area of 220.1–237.4 m²·g^–1, the total pore volume of 0.29–0.31 cm³·g^–1, and the average pore size of 4.7–5.1 nm. In contrast, 40Ni/OMA-600 catalyst showed the lowest BET surface area (104.0 m²·g^–1) and total pore volume (0.17 cm³·g^–1).

Tab.2 Calculated textural properties of the fresh catalysts

Samples	BET surface area/(m2·g–1)	Total pore volume/(cm3·g–1)	Mean pore size/nm
40Ni/OMA-600	104.0	0.17	6.3
40Ni-OMA	185.7	0.58	12.0
40Ni/OMA	223.9	0.31	5.1
30Ni/OMA	220.1	0.29	4.7
25Ni/OMA	225.3	0.29	4.8
20Ni/OMA	237.4	0.30	4.8
OMA-600	282.7	0.58	8.2

Fig.3 (a) N2 adsorption-desorption isotherms and (b) pore-size distributions of the fresh catalysts.

Full size|PPT slide

3.2 Reducibility and dispersion

To evaluate the metal-support interactions, the H₂-TPR analyses were performed and the profiles are presented in Fig.4. With increase of the NiO loadings from 20 to 40 wt %, the intensity of H₂-consumption peaks was generally enhanced, which is indicative of the increased amount of reducible NiO species over the catalyst. In the case of Ni/OMA catalysts, a broad hydrogen-consumption peak showed a good normal distribution at the temperature region of 480–850 °C, suggesting the presence of very similar metal-support interactions. The temperature at the peak maximum was gradually increased from 658 °C for 20Ni/OMA to 671 °C for 40Ni/OMA. It is mainly attributed to the reduction of the confined Ni species with strong interactions within the OMA framework and/or non-stoichiometric amorphous NiAl₂O₄ on the surface [26]. In contrast, 40Ni-OMA presented two distinct shoulder peaks at the maximum peak centered at 701 °C, and the shoulder peak was more pronounced in the low-temperature region of 330–600 °C. In addition, two completely separate reduction peaks for 40Ni/OMA-600 were exhibited in the low-temperature region of 280–530 °C and the high-temperature region of 530–860 °C, respectively. Moreover, the two reduction peaks are essentially the same in terms of peak areas, indicating the presence of a considerable amount of large-size NiO particles weakly interacted with support. These results are consistent with those of XRD and TEM.

Fig.4 H2-TPR profiles for the fresh catalysts of (i) 20Ni/OMA, (ii) 25Ni/OMA, (iii) 30Ni/OMA, (iv) 40Ni/OMA, (v) 40Ni-OMA, and (vi) 40Ni/OMA-600.

Full size|PPT slide

To investigate the effect of the metal-support interaction on the reducibility of NiO, the O₂ titration was used to measure the reducibility and the results are shown in Tab.1. The reducibility of Ni/OMA catalysts could be kept at a very similar level of more than 91.0% as increasing the loading of NiO from 20 to 40 wt %, and the highest reduction degree of 95.4% was achieved over 40Ni/OMA. In contrast, the reduction degree of 40Ni/OMA-600 was determined to be 86.3%, which is much lower than that of 40Ni/OMA and slightly higher than that of 40Ni-OMA (83.5%). As indicated in Tab.1, regardless of the NiO loadings, Ni/OMA catalysts showed very similar dispersion of Ni (13.4%–13.8%), which is much higher than those of 40Ni-OMA (5.2%) and 40Ni/OMA-600 (2.5%) counterparts. The mean particle size of Ni determined by the H₂-chemisorption was slightly larger than those calculated by XRD and TEM characterizations, but the changing patterns of the Ni particle size measured by different techniques remained the same. Moreover, irrespective of the NiO loadings, a similar Ni particle size of about 7.0 nm was obtained for the Ni/OMA catalysts, which is well consistent with the observations of XRD and TEM.

3.3 Catalytic performance

The performance of different catalysts for the CO methanation was measured at a temperature region of 300–450 °C, and the results are presented in Fig.5. In the cases of 30Ni/OMA and 40Ni/OMA, the CO conversion continuously increased with increase of the reaction temperature from 300 to 400 °C, and then remained at a constant level (about 99%) with a further increase of the temperature up to 450 °C. In contrast, in the cases of 20Ni/OMA, 25Ni/OMA, 40Ni-OMA, and 40Ni/OMA-600, the CO conversion continuously increased with increase of the reaction temperature from 300 to 450 °C. As for products, no hydrocarbons besides methane were detected, and the CH₄ selectivity showed two maximums in the temperature region of 300–450 °C (Fig.5(b)). In addition, CO₂ as the main byproduct is probably formed via the water-gas shift and/or Boudouard reactions [29]. Thus, the CH₄ yield of 30Ni/OMA and 40Ni/OMA showed a volcanic trend with increase of the reaction temperature from 300 to 450 °C, and reached the highest at around 420 °C (Fig.5(c)). Except for 30Ni/OMA and 40Ni/OMA catalysts, the CH₄ yield gradually increased with increase of the reaction temperature in cases of the other catalysts. As for the productivity of the catalyst (Fig.5(d)), it was found that the STY of CH₄ over all the tested catalysts showed the same trend as the CH₄ yield. Among them, 40Ni/OMA showed the highest STY of 709.6 ${\mathrm{m}\mathrm{o}\mathrm{l}}_{{\mathrm{C}\mathrm{H}}_4}\cdot {{\mathrm{k}\mathrm{g}}_{\mathrm{c}\mathrm{a}\mathrm{t}}}^{-1}\cdot {\mathrm{h}}^{-1}$ at 320 °C and 1835.3 ${\mathrm{m}\mathrm{o}\mathrm{l}}_{{\mathrm{C}\mathrm{H}}_4}\cdot {{\mathrm{k}\mathrm{g}}_{\mathrm{c}\mathrm{a}\mathrm{t}}}^{-1}\cdot {\mathrm{h}}^{-1}$ at 420 °C, respectively.

Fig.5 Catalytic activity and selectivity indexed by (a) CO conversion, (b) CH4 selectivity, (c) CH4 yield, and (d) STY of CH4 for the low-temperature CO methanation under the conditions of 0.1 MPa and 240000 mL·g–1·h–1.

Full size|PPT slide

Through the correlation analysis in the temperature region of 300–340 °C given in Fig.6(a), there is a linear relationship between the r_CO and the NiO loading over the Ni/OMA catalysts. Similarly, there is a linear correlation between the STY of CH₄ and the NiO loading over Ni/OMA catalysts in the same temperature region (Fig.6(b)). Moreover, as indicated in Fig.6(c), TOFs of ~24.0 h^–1 at 300 °C were kept the same over the Ni/OMA catalysts irrespective of the NiO loading. These results can be attributed to the fact that the series of Ni/OMA catalysts have very similar Ni particle size, dispersion and reduction degree (Tab.1). At the same NiO loading of 40 wt %, the TOF of 40Ni/OMA was much higher than those of 40Ni-OMA (14.5 h^–1) and 40Ni/OMA-600 (6.6 h^–1), attributable to the smaller Ni particles, higher Ni dispersion and reduction degree (Tab.1). Thus, it can be concluded that the low-temperature activity could be correlated with the Ni particle size, dispersion and reducibility [30]. It is generally revealed that a higher catalytic activity is achieved over the catalyst with smaller Ni particle size, higher dispersion and reduction degree of metallic Ni.

Fig.6 The correlation between the NiO loading over Ni/OMA and (a) the rCO at 300–340 °C, (b) STY of CH4 at 300–340 °C, and (c) TOF at 300 °C under the conditions of 0.1 MPa and 240000 mL·g–1·h–1.

Full size|PPT slide

Since the good low-temperature catalytic activity (709.6 ${\mathrm{m}\mathrm{o}\mathrm{l}}_{{\mathrm{C}\mathrm{H}}_4}\cdot {{\mathrm{k}\mathrm{g}}_{\mathrm{c}\mathrm{a}\mathrm{t}}}^{-1}\cdot {\mathrm{h}}^{-1}$ at 320 °C) of 40Ni/OMA for the CO methanation, the long-term durability was tested at high temperature (600 °C), and the results are given in Fig.7. Clearly, no observable decrease indexed either by the CO conversion or the selectivity, yield and STY of CH₄ was found over 40Ni/OMA catalyst for a TOS of 120 h. Meanwhile, the spent 40Ni/OMA catalyst was characterized by XRD, TEM and TG. As shown in Tab.1 and Fig. S1 (cf. ESM), the mean particle size was slightly increased from 4.9 to 6.6 nm for the spent 40Ni/OMA, demonstrating that the small particle size is well retained. In addition, the diffraction peak at 2θ = 26° assigned to graphic carbon was not detected by XRD analysis (Fig. S1), indicating the absence of graphited coke deposition. TG results (Fig. S1) revealed that the amount of the coke deposited over 40Ni/OMA was 0.1%. Thus, 40Ni/OMA shows a high activity as well as good anti-sintering and anti-coking properties leading to a long-term stability at high temperature, which could be attributed to the small Ni particles with high dispersion and reduction degree at high NiO loading, as well as the spatial confinement of the OMA channels.

Fig.7 Long-term durability of 40Ni/OMA catalyst for the CO methanation reaction under the conditions of 0.1 MPa, 600 °C, and 240000 mL·g–1·h–1 ((a) CO conversion and CH4 selectivity, (b) CH4 yield, and (c) STY of CH4).

Full size|PPT slide

3.4 Specific activity and mechanism for the CO methanation

Indeed, the activity for the CO methanation reaction over Ni-based catalysts is closely affected by many factors, such as Ni loading, Ni dispersion and reducibility, Ni particle size, Ni-support interaction, and the properties of support [31−34]. Generally, the Ni-based catalysts with smaller Ni particles, higher dispersion and reducibility tend to have higher CO methanation activity. For this reason, when NiO loadings were kept to be around 40 wt %, the low-temperature activity indexed by the CO conversion at 300 °C (Fig.5(a)) was decreased in the order of 40Ni/OMA (36.6%) > 40Ni-OMA (15.6%) > 40Ni/OMA-600 (8.8%). Although the specific activity is more reasonable for a comparison purpose, TOF at 300 °C (Fig.6(c)) over each catalyst still remains the same order as that based on the CO conversion, i.e., 24.4 h^–1 for 40Ni/OMA > 14.5 h^–1 for 40Ni-OMA > 6.6 h^–1 for 40Ni/OMA-600. In contrast, whether r_CO or STY of CH₄ is used as an indicator of the low-temperature activity, it showed a linear increase with increase of the NiO loading over the catalyst in the temperature range from 300 to 340 °C (Fig.6(a) and 6(b)), indicating the same intrinsic activity of the catalysts. Moreover, the TOFs of the Ni/OMA catalysts at 300 °C (Fig.6(c)) could be kept at about 24.0 h^–1 regardless of the NiO loadings. These results can be reasonably attributed to the fact that the Ni/OMA catalysts have very similar textural properties (Fig.3 and Tab.2), Ni particle size, dispersion and reducibility (Tab.1), as well as very similar metal-support interactions (Fig.4).

Generally, two types of catalytic mechanisms for the CO methanation reaction are proposed, i.e., associative and dissociative pathways. In the case of the associative pathway, adsorbed H_ad is associated with adsorbed carbonyl (CO_ad) to form COH_ad, CHO_ad or CHOH_ad, which are the intermediates for the formation of methane via the breaking of C–O bonds. In contrast, the dissociative pathway characterizes the direct dissociation or disproportion of adsorbed CO_ad to surface-carbon (C_ad) intermediate, which is the key difference between the two mechanisms. In our case, the two pathways cannot be differentiated without the spectroscopic results of the reaction. However, the reaction pathway of the CO methanation reaction over Ni-based catalysts is closely related to the Ni reducibility and dispersion, Ni particle size, metal-support interaction, and the properties of support [34]. Considering the similar properties of Ni/OMA catalysts such as the same support and comparable dispersion, reducibility and size distribution of Ni, it is reasonably assumed that the same exposited active sites are accessible for the CO methanation reaction. This is confirmed from the linear correlation of the NiO loading over Ni/OMA catalysts with r_CO and STY of CH₄ together with the same TOF. Thus, the same mechanism for CO methanation can be concluded for Ni/OMA catalysts.

3.5 Nature on breaking the reducibility-dispersion dependence

As revealed from the XRD and TEM results, all of the Ni/OMA catalysts with a NiO loading up to 40 wt % have a highly ordered mesoporous structure (Fig.1 and Fig.2) and almost the same mean Ni crystal sizes of about 5.0 nm (Tab.1). According to the H₂ chemisorption and O₂ titration results, the Ni dispersion and reducibility are also maintained almost the same, i.e., 13.5% and more than 91.0% (Tab.1). Thus, the small Ni particles with high Ni dispersion and reducibility are achieved for all of the Ni/OMA catalysts independent of the NiO loadings. To explore the nature, the two reference catalysts were comparatively analyzed. As reported in our recent work [26], OMA-P with an ordered mesoporous structure as the precursor for solid-state grinding contains almost all of the P123 template. This indicates that the P123 template in OMA-P plays a key role in dispersing nickel species, which is supported by the very low Ni dispersion of 40Ni/OMA-600 (2.5%) counterpart prepared by using OMA-600 without the P123 template as the support. However, this is contradictory with the very low Ni dispersion of 5.2% for the 40Ni-OMA counterpart derived from the P123 template-containing precursor. When the NiO loading exceeds 22 wt %, it was reported that the ordered mesoporous structure of Ni-OMA catalysts prepared by the one-pot EISA method completely collapsed [35]. Following this fact, the encapsulation effect is weakened and the Ni dispersion is severely deteriorated due to the susceptible agglomeration. In our case, the absence of the ordered mesoporous structure and the agglomeration of the Ni particles are really observed for 40Ni-OMA as reflected by the XRD and TEM results. Thus, to achieve a high Ni dispersion, the aluminum precursor with both a residual P123 template and an ordered mesoporous structure is required for solid-state grinding.

Additionally, the Al species over all of the Ni/OMA catalysts are predominantly amorphous Al₂O₃ without detectable crystallized NiAl₂O₄ (Fig.1(b)). However, the Al species in the counterparts of 40Ni/OMA-600 and 40Ni-OMA are dominated by amorphous alumina and well-crystallized NiAl₂O₄ species (Fig.1(b)). These understandings are also reflected from the H₂-TPR results for all of Ni/OMA catalysts (Fig.4), i.e., a highly symmetric hydrogen-consumption peak with similar peak temperatures of 658–671 °C originated from the reduction of the confined Ni species with strong interactions with the OMA framework. Thus, all of Ni/OMA catalysts have very similar interactions between Ni and the OMA framework, leading to similar dispersion and reducibility of Ni. In contrast, 40Ni-OMA presented two distinct shoulder peaks at the maximum peak centered at 701 °C, while two completely separate reduction peaks for 40Ni/OMA were exhibited in the low-temperature region of 280 to 530 °C and the high-temperature region of 530 to 860 °C, respectively. Those results indicate the presence of different interactions between Ni and Al₂O₃.

Based on the above analyses, it is understandable to speculate the reason why Ni/OMA has very similar metal-support interactions. Because of its very low melting point (~57 °C) and high hydrophilicity, nickel nitrate tends to melt during the solid-phase co-grinding process and the subsequent calcination at low temperatures. As a result, nickel nitrate is easily introduced into the confined accessible space of the OMA-P with P123, resulting in essentially the same degree of interaction between Ni and Al₂O₃. Moreover, very similar textural properties are also achieved for Ni/OMA catalysts with NiO loading from 20 to 40 wt %, which are quite different from those of 40Ni/OMA-600 and 40Ni-OMA.

Owing to the simultaneous presence of P123 template and an ordered mesoporous structure in OMA-P, the Ni/OMA catalysts prepared by co-grinding the hybrid of OMA-P and nickel nitrate show very similar Ni-support interactions up to 40 wt % NiO, leading to small Ni particles (~5.0 nm) with high dispersion (~13.5%) and reducibility (≥ 91.0%) for all of the Ni/OMA catalysts. Thus, co-grinding the hybrid of OMA-P and nickel nitrate is effective for the preparation of Ni/OMA catalysts without the limitation of the loading-dispersion-reducibility dependence at a NiO loading up to 40 wt %.

4 Conclusions

In summary, the solid-state co-grinding is proved to be an efficient method for breaking the loading-reducibility-dispersion dependence of supported Ni catalysts. Specifically, the OMA mesostructure was still retained even after loading up to 40 wt % NiO. More importantly, the uniform Ni nanoparticles (~4.0–5.0 nm) with very similar Ni dispersion (~13.5%) and reduction extent (≥ 91.0%) were achieved for all of the Ni/OMA catalysts as a result of very similar interactions between Ni and OMA and textural properties. Thus, similar TOFs of ~24.0 h^–1 at 300 °C were obtained for all of the Ni/OMA catalysts independent of NiO loadings. For the catalyst with the highest NiO loading of 40 wt %, 40Ni/OMA demonstrated a high low-temperature catalytic activity with STY of methane over 325.8 ${\mathrm{m}\mathrm{o}\mathrm{l}}_{{\mathrm{C}\mathrm{H}}_4}\cdot {{\mathrm{k}\mathrm{g}}_{\mathrm{c}\mathrm{a}\mathrm{t}}}^{-1}\cdot {\mathrm{h}}^{-1}$ at 300 °C, but also a high long-term stability of 120 h with no observable deactivation under the conditions of 600 °C and an extremely high GHSV of 240000 mL·g^–1·h^–1. This work demonstrates an efficient way to break the loading-dispersion-reducibility dependence of Ni-based catalysts, which can be reasonably extended to the development of industrially relevant supported metal catalysts.

References

[1] Lu X, Xie P, Ingham D B, Ma L, Pourkashanian M. Modelling of CO₂ absorption in a rotating packed bed using an Eulerian porous media approach. Chemical Engineering Science, 2019, 199: 302–318

[2] Izumi Y. Recent advances in the photocatalytic conversion of carbon dioxide to fuels with water and/or hydrogen using solar energy and beyond. Coordination Chemistry Reviews, 2013, 257(1): 171–186

[3] Tan L, Peter K, Ren J, Du B, Hao X, Zhao Y, Song Y. Photocatalytic syngas synthesis from CO₂ and H₂O using ultrafine CeO₂-decorated layered double hydroxide nanosheets under visible-light up to 600 nm. Frontiers of Chemical Science and Engineering, 2021, 15(1): 99–108

[4] Song G, Wu X, Xin F, Yin X. ZnFe₂O₄ deposited on BiOCl with exposed (001) and (010) facets for photocatalytic reduction of CO₂ in cyclohexanol. Frontiers of Chemical Science and Engineering, 2017, 11(2): 197–204

[5] Hafez A M, Zedan A F, AlQaradawi S Y, Salem N M, Allam N K. Computational study on oxynitride perovskites for CO₂ photoreduction. Energy Conversion and Management, 2016, 122: 207–214

[6] Schneider J, Matsuoka M, Takeuchi M, Zhang J, Horiuchi Y, Anpo M, Bahnemann D W. Understanding TiO₂ photocatalysis: mechanisms and materials. Chemical Reviews, 2014, 114(19): 9919–9986

[7] Qian R, Zong H, Schneider J, Zhou G, Zhao T, Li Y, Yang J, Bahnemann D W, Pan J H. Charge carrier trapping, recombination and transfer during TiO₂ photocatalysis: an overview. Catalysis Today, 2019, 335: 78–90

[8] Vorontsov A V, Valdes H, Smirniotis P G, Paz Y. Recent advancements in the understanding of the surface chemistry in TiO₂ photocatalysis. Surfaces, 2020, 3(1): 72–92

[9] Thompson W A, Fernandez E S, Maroto-Valer M M. Review and analysis of CO₂ photoreduction kinetics. ACS Sustainable Chemistry & Engineering, 2020, 8(12): 4677–4692

[10] Tahir M, Amin N S. Photocatalytic CO₂ reduction and kinetic study over In/TiO₂ nanoparticles supported microchannel monolith photoreactor. Applied Catalysis A, General, 2013, 467: 483–496

[11] Thompson W A, Fernandez E S, Maroto-Valer M M. Probability Langmuir-Hinshelwood based CO₂ photoreduction kinetic models. Chemical Engineering Journal, 2020, 384: 123356

[12] Marczewski A W. Analysis of kinetic Langmuir model. Part I: integrated Langmuir equation (IKL): a new complete analytical solution of the Langmuir rate equation. Langmuir, 2010, 26(19): 15229–15238

[13] Bloh J Z. A holistic approach to model the kinetics of photocatalytic reactions. Frontiers in Chemistry, 2019, 7: 128

[14] Li H, Yi F, Li X, Gao X. Numerical modelling of mass transfer processes coupling with reaction for the design of the ozone oxidation treatment of wastewater. Frontiers of Chemical Science and Engineering, 2021, 15(3): 602–614

[15] Chu F, Chen S, Li H, Yang L, Ola O, Maroto-Valer M M, Du X, Yang Y. Modeling photocatalytic conversion of carbon dioxide in bubbling twin reactor. Energy Conversion and Management, 2017, 149: 514–525

[16] Chen H, Chu F, Yang L, Ola O, Du X, Yang Y. Enhanced photocatalytic reduction of carbon dioxide in optical fibre monolith reactor with transparent glass balls. Applied Energy, 2018, 230: 1403–1413

[17] Verbruggen S W, Lenaerts S, Denys S. Analytic versus CFD approach for kinetic modelling of gas phase photocatalysis. Chemical Engineering Journal, 2015, 262: 1–8

[18] Olivo A, Thompson W A, Bay E R B, Ghedini E, Menegazzo F, Maroto-Valer M, Signoretto M. Investigation of process parameters assessment via design of experiments for CO₂ photoreduction in two photoreactors. Journal of CO₂ Utilization, 2020, 36: 25–32

[19] Wang X, Tan X, Yu T. Kinetic study of ozone photocatalytic decomposition using a thin film of TiO₂ coated on a glass plate and the CFD modelling approach. Industrial & Engineering Chemistry Research, 2014, 53(19): 7902–7909

[20] Kočí K, Obalová L, Šolcová L. Kinetic study of photocatalytic reduction of CO₂ over TiO₂. Chemical & Process Engineering, 2010, 31: 395–407

[21] Ji Y, Luo Y. Theoretical study on the mechanism of photoreduction of CO₂ to CH₄ on the anatase TiO₂(101) surface. ACS Catalysis, 2016, 6(3): 2018–2025

[22] Campbell C T, Sellers J R V. Enthalpies and entropies of adsorption on well-defined oxide surfaces: experimental measurements. Chemical Reviews, 2013, 113(6): 4106–4135

[23] Meng X, Yun N, Zhang Z. Recent advances in computational photocatalysis: a review. Canadian Journal of Chemical Engineering, 2019, 97(7): 1982–1998

[24] Dilla M, Schlögl R, Strunk J. Photocatalytic CO₂ reduction under continuous flow high-purity conditions: quantitative evaluation of CH₄ formation in the steady-state. ChemCatChem, 2017, 9(4): 696–704

[25] Tseng I, Chang W, Wu J C S. Photoreduction of CO₂ using sol-gel derived titania and titania-supported copper catalysts. Applied Catalysis B: Environmental, 2002, 37(1): 34–48

[26] Giammar D E, Maus C J, Xie L. Effects of particle size and crystalline phase on lead adsorption to titanium dioxide nanoparticles. Environmental Engineering Science, 2007, 24(1): 85–95

[27] Mao W, Wilde M, Ogura S, Chen J, Fukutani K, Matsuzaki H, Terai T. Hydrogen-accelerated phase transition and diffusion in TiO₂ thin films. Journal of Physical Chemistry C, 2018, 122(40): 23026–23033

[28] Tan L L, Ong W J, Chai S P, Mohamed A R. Photocatalytic reduction of CO₂ with H₂O over graphene oxide-supported oxygen-rich TiO₂ hybrid photocatalyst under visible light irradiation: process and kinetic studies. Chemical Engineering Journal, 2017, 308: 248–255

[29] Lopes F V S, Grande C A, Ribeiro A M, Loureiro J M, Evaggelos O, Nikolakis V, Rodrigues A E. Adsorption of H₂, CO₂, CH₄, CO, N₂ and H₂O in activated carbon and zeolite for hydrogen production. Separation Science and Technology, 2009, 44(5): 1045–1073

[30] Bazan R E, Bastos-Neto M, Moeller A, Dreisbach F, Staudt R. Adsorption equilibria of O₂, Ar, Kr, and Xe on activated carbon and zeolites: single component and mixture data. Adsorption, 2011, 17(2): 371–383

[31] Delavari S, Amin N A S. Photocatalytic conversion of CO₂ and CH₄ over immobilized titanic nanoparticles coated on mesh optimization and kinetic study. Applied Energy, 2016, 162: 1171–1185

[32] Lo C C, Hung C H, Yuan C S, Hung Y L. Parameter effects and reaction pathways of photoreduction of CO₂ over TiO₂/SO₄^2– photocatalyst. Chinese Journal of Catalysis, 2007, 28(6): 528–534

[33] Khalilzadeh A, Shariati A. Photoreduction of CO₂ over heterogeneous modified TiO₂ nanoparticles under visible light irradiation: synthesis, process and kinetic study. Solar Energy, 2018, 164: 251–261

[34] Sebastia-Saez D, Gu S, Ranganathan P, Papadikis K. Meso-scale CFD study of the pressure drop, liquid hold-up, interfacial area and mass transfer in structured packing materials. International Journal of Greenhouse Gas Control, 2015, 42: 388–399

[35] Anpo M. Photocatalytic reduction of CO₂ with H₂O on highly dispersed Ti-oxide catalysts as a model of artificial photosynthesis. Journal of CO₂ Utilization, 2013, 1: 8–17

[36] Yin W, Wen B, Ge Q, Li X, Teobaldi G, Liu L. Activity and selectivity of CO₂ photoreduction on catalytic materials. Dalton Transactions (Cambridge, England), 2020, 49(37): 12918–12928

[37] Ola O, Maroto-Valer M M. Transition metal oxide based TiO₂ nanoparticles for visible light induced CO₂ photoreduction. Applied Catalysis A, General, 2015, 502: 114–121

[38] Thompson T L, Yates J T Jr. TiO₂-based photocatalysis: surface defects, oxygen and charge transfer. Topics in Catalysis, 2005, 35(3-4): 197–210

[39] Li D, Huang Y, Li S, Wang C, Li Y, Zhang X, Liu Y. Thermal coupled photoconductivity as a tool to understand the photothermal catalytic reduction of CO₂. Chinese Journal of Catalysis, 2020, 41(1): 154–160

[40] Tan L, Xu S, Wang Z, Xu Y, Wang X, Hao X, Bai S, Ning C, Wang Y, Zhang W, et al. Highly selective photoreduction of CO₂ with suppressing H₂ evolution over monolayered double hydroxide under irradiation above 600 nm. Angewandte Chemie, 2019, 131(34): 11986–11993

[41] Kohno Y, Hayashi H, Takenaka S, Tanaka T, Funabiki T, Yoshida S. Photo-enhanced reduction of carbon dioxide with hydrogen over Rh/TiO₂. Journal of Photochemistry and Photobiology A Chemistry, 1999, 126(1-3): 117–123

[42] Zhou W, Guo J, Shen S, Pan J, Tang J, Chen L, Au C, Yin S. Progress in photoelectrocatalytic reduction of carbon dioxide. Acta Physica Sinica, 2020, 36(3): 1906048 (in Chinese)

[43] Xie F, Chen R, Zhu X, Liao Q, Ye D, Zhang B, Yu Y, Li J.CO₂ utilization: direct power generation by a coupled system that integrates photocatalytic reduction of CO₂ with photocatalytic fuel cell. Journal of CO₂ Utilization, 2019, 32: 31–36

[44] Chen M, Wu J, Lu C, Luo X, Huang Y, Jin B, Guo H, Zhang X, Argyle M, Liang Z. Photoreduction of CO₂ in the presence of CH₄ over g-C₃N₄ modified with TiO₂ nanoparticles at room temperature. Green Energy & Environment, 2021, in press

Acknowledgments

The authors thank the financial support provided by the Engineering and Physical Sciences Research Council (Grant No. EP/K021796/1), the Research Centre for Carbon Solutions and the James Watt Scholarship Programme at Heriot-Watt University. We are also grateful for the support provided by the Buchan Chair in Sustainable Energy Engineering.

Open Access

This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/.

RIGHTS & PERMISSIONS

2021 The Author(s) 2021. This article is published with open access at link.springer.com and journal.hep.com.cn