Hydrogen Production through Oxidative Steam Reforming of Acetic Acid over Ni Catalysts Supported on Ceria-Based Materials

Abstract

Oxidative steam reforming allows higher energy efficiency and lowers coke deposition compared to traditional steam reforming. In this work, CeO₂-based supports have been prepared with Ni as the active phase, and they were tested in the oxidative steam reforming of acetic acid. The influence of the O₂/AcOH molar ratio (0–0.3) has been evaluated over Ni/CeO₂. The results stated that by increasing oxygen content in the feeding mixture, acetic acid conversion increases, too, with a decrease in coke deposition and hydrogen yield. To have a proper balance between the acetic acid conversion and the hydrogen yield, an O₂/AcOH molar ratio of 0.075 was selected to study the catalytic performance of Ni catalysts over different supports: commercial CeO₂, a novel mesostructured CeO₂, and CeO₂-SBA-15. Due to higher Ni dispersion over the support, the mesostructured catalysts allowed higher acetic acid conversion and hydrogen yield compared to the nonporous Ni/CeO₂. The best catalytic performance and the lowest coke formation (120.6 mg_coke·g_cat⁻¹·h⁻¹) were obtained with the mesostructured Ni/CeO₂. This sample reached almost complete conversion (>97%) at 500 °C, maintaining the hydrogen yield over 51.5% after 5 h TOS, being close to the predicted value by the thermodynamic equilibrium that is due to the synergistic coordination between Ni and CeO₂ particles.

1. Introduction

Substantial growth in population and economic development are the main reasons for the energy demand rise [1,2]. This demand is mainly covered by fossil fuels, resulting in about 830 million tons of CO₂ emissions per year [3]. Against fossil fuels, hydrogen, which is considered the energy vector of the future, could replace them and reduce the related carbon emissions [4]. In 2020, 90 Mt of hydrogen was produced worldwide, mainly from thermochemical processes using fossil fuels [5]. As the sustainability of hydrogen is related to the cleanness of the hydrogen production pathway and the energy used during its obtention, biomass-to-hydrogen processes have gained attention in the last few years [6,7]. More specifically, the combination of the thermochemical conversion of biomass, followed by the reforming process of the produced oil, is a promising alternative [8,9,10]. Bio-oil is a complex mixture of different oxygenated hydrocarbons (carboxylic acids, ketones, phenols, and aldehydes among others [11]) that could be separated into an organic fraction and an aqueous fraction. The aqueous fraction is composed of different water-soluble compounds with variable concentrations, where acetic acid is usually one of the major compounds [9,12,13]. Bio-oil aqueous fraction steam reforming has been studied as a viable alternative for producing green hydrogen from biomass [6,14,15]. However, this process has some limitations that are due to the catalyst deactivation primarily produced from coke deposition and because the primary reaction is highly endothermic [16,17]. On the contrary, oxidative steam reforming, which implies the co-feeding of oxygen, has been less reported on despite its higher energy efficiency, given that the oxidation reactions can supply part of the energy required for the endothermic reactions. Moreover, it also allows delaying catalyst deactivation by favoring the combustion of carbon deposits [18,19]. Notwithstanding the benefits of adding oxygen, the main drawbacks of oxidative steam reforming are the lower hydrogen selectivity and catalyst deactivation by the metallic, active phase oxidation; thus, the oxygen/steam ratio should be controlled [20]. The overall reaction that described this process is summarized in Equation (1).

$$C_n{H}_m{O}_K+p{O}_2+\left(2n-k-2p\right)H_{2}O\to nC{O}_2+\left(2n+\frac{m}{2}-k-2p\right)H_2$$

(1)

Apart from that, it is well-known that the catalyst also plays a crucial role in the reforming processes [21]. This is why many researchers focus on catalyst formulation, being that Ni-based catalysts are the most studied [19,22] given their ability to crack C-C and C-H bonds [23], which is necessary according to the mechanism described in the literature that is represented in Scheme 1.

As support, ceria has been reported to have high oxygen mobility, which can reduce the formation of carbon deposits over the active sites [25]. Unfortunately, the nonporous structure with a low surface of ceria area has limited its application as a support in reforming reactions [26]. However, by modifying its physical properties related to geometrical features including size, dimension, orientation, and morphology, it is possible to enhance its catalytic role by decreasing the coke formation and inhibiting the deactivation caused by metal sintering. Therefore, the control of the support morphology plays a key crucial role in improving the catalytic activity since it can facilitate the formation of oxygen vacancies [26] and also influence the formation of metal species ascribed to the Ce³⁺/Ce⁴⁺ redox properties of CeO₂ [27]. Moreover, the high surface area of ceria allows better Ni-species dispersion because of the smaller crystallite size, thus strengthening the interaction between Ni and CeO₂ and promoting Ni reducibility, which may result in a better catalytic performance [26]. Despite this, a ceria-based porous structure with a high surface area used as support for reforming purposes is rarely reported. Preparing mesoporous ceria would allow taking advantage of the oxygen mobility of CeO₂ but also the reaching of higher metal dispersion on a support with a higher surface area, as reported before [28]. To the best of our knowledge, no papers have been published on acetic acid oxidative steam reforming using ceria-based mesoporous structures as support. Thus, this work aims to study the influence of different CeO₂-based supports: a novel mesostructured CeO₂ prepared by nanocasting, a commercial CeO₂, and CeO₂-SBA-15 on the activity and hydrogen selectivity during the oxidative steam reforming of acetic acid as a bio-oil aqueous fraction model compound using Ni as the active phase.

2. Results and Discussion

2.1. Catalysts Characterization

Nitrogen adsorption–desorption isotherms of the prepared porous catalysts and the supports are displayed in Figure 1, except for the commercial CeO₂ given its nonporous structure. As can be discerned, all the porous samples showed a type IV isotherm according to the IUPAC classification characteristic of mesoporous materials with an H1 and H3-hysteresis loop for SBA-15-based samples and mesoporous cerium-based materials, respectively. While the H1 hysteresis loop indicates the presence of a well-ordered hexagonal mesoporous structure characteristic of SBA-15 materials [15], H3-type hysteresis loops are known for not having any limitation adsorption at high P/P₀, typical of layered particles that give rise to porosity [29]. These results agree with the structure observed by TEM for the SBA-15 and CeO₂-m supports displayed in Figure 2a,b, respectively.

The physicochemical properties of the supports and the catalysts used in this work are summarized in

Table 1. Physicochemical properties of synthesized supports and catalysts.

Sample	Nia (wt.%)	Cea (wt.%)	SBET (m2/g)	Vpb (cm3/g)	Dpc (nm)	DNiOd (nm)	DNie (nm)
CeO2	-	-	-	-	-	-	-
Ni/CeO2	6.2	-	-	-	-	32.5	28.4
CeO2-m	-	-	114	0.29	7.3	-	-
Ni/CeO2-m	6.7	-	98	0.26	8.1	12.5	12.2
SBA-15	-	-	601	0.91	7.7	-	-
CeO2-SBA-15	-	9.1	490	0.72	7.2	-	-
Ni/CeO2-SBA-15	6.3	8.5	432	0.63	7.3	11.2	11.0

^a Determined by ICP-AES in calcined samples; ^b pore volume measured at P/P₀ = 0.97; ^c BJH desorption average pore diameter; ^d mean crystallite size determined from XRD pattern of calcined catalysts using the Scherrer equation in the (200) diffraction line of NiO; ^e mean crystallite size determined from XRD pattern of reduced catalysts using the Scherrer equation in the (111) diffraction line of Ni.

. The metal loadings, determined by ICP-AES analysis, were close to the nominal values fixed during the catalyst synthesis. Attending to BET surface area values, SBA-15-based materials achieved the highest values despite this value decreasing when Ni is incorporated into the support during the synthesis step as reported before [30]. Ce-based catalysts, more specifically the CeO₂-m sample, exhibit similar BET surfaces and pore volume rates than those described in the literature for similar materials prepared by nanocasting using this template [31,32]. However, this support presents a lower surface area when compared to the template used during the synthesis (SBA-15). Roggenbuck et al. [31] ascribed this reduction to the differences between the density of the bulk phases (SiO₂ and CeO₂) and the microporosity present in SBA-15, which significantly contributes to its specific surface area. Instead, commercial CeO₂ has a nonporous structure [26].

Figure 3 shows the XRD diffractograms of the calcined catalysts. On the one hand, peaks arise in all the samples at 28.5, 33, 47.3, 56.1, 69.4, 76.7, and 79.1°, corresponding to cubic CeO₂ (JCPDS 01-089-8436), these peaks being more intense in the commercial CeO₂ support given its higher crystallinity. In the case of Ni/CeO₂-SBA-15, these peaks evidence that CeO₂ is in the form of crystallites distributed over the SBA-15 not being completely dispersed, with unlikely interaction with the Ni phase [33]. On the other hand, peaks at 37.3, 43.4, and 63° can also be discerned from the presence of cubic NiO (JCPDS 01-075-0197). The low relative of NiO peaks in Ni/CeO₂ is due to the high intensity of CeO₂ diffraction lines making them difficult to be clearly observed. Broadly, in the Ni/CeO₂-m and Ni/CeO₂-SBA-15 samples, these peaks are broader when compared to Ni/CeO₂, indicating smaller crystallite size. This is ascribed to the high surface area of the mesostructured supports, which allows better distribution of Ni species on the carrier. Therefore, it is reasonable that the mean crystallite sizes calculated by applying the Scherrer equation on the (200) diffraction plane of cubic NiO pattern for these samples were around 60% smaller than that obtained for Ni/CeO₂ (see Table 1).

The reducibility of the prepared catalysts was evaluated through H₂-TPR analysis, whose profiles are represented in Figure 4. In all the samples, two main features could be distinguished at temperatures below 500 °C, ascribed to the reduction in NiO species with different degrees of interaction with the support. Samples supported on ceria show similar profiles where another peak at higher temperatures (740–805 °C) can also be discerned. According to what is described in the literature, this peak is assigned to the reduction of Ce⁴⁺ to Ce³⁺ in bulk ceria [34,35]. The differences between the reduction temperatures of NiO particles are related to lower crystallite sizes found over Ni/CeO₂-m than over the commercial CeO₂. On the other side, the reduction profile of the Ni/CeO₂-SBA-15 sample is similar to that of Ni/SBA-15 described elsewhere [36] but shifted to lower temperatures ascribed to the presence of cerium in the sample [37].

After the activation step to reduce the NiO species, all the catalysts were again characterized by XRD whose diffractograms are displayed in Figure 5. As can be observed, no diffraction lines of the NiO pattern could be distinguished. On the contrary, diffraction lines of cubic Ni⁰ (JCPDS 01-071-3740) arise at 44.6 and 51.8, stating that the reduction step was successful, being that all the Ni-oxides were reduced to the metallic state. No diffraction peaks corresponding to metallic Ce could be observed since Ce oxides need higher reduction temperatures, above 1000 °C [38]. As happened with calcined samples, the Ni/CeO₂-m and Ni/CeO₂-SBA-15 samples present broader peaks than the nonporous Ni/CeO₂, resulting in lower mean crystallite size calculated using the Scherrer equation from the (111) diffraction plane of Ni⁰ (see Table 1), as a consequence of the higher surface area in the supports.

2.2. Catalytic Performance

The influence of the O₂/AcOH molar ratio in hydrogen production by acetic acid steam reforming was evaluated with the Ni/CeO₂ catalyst at 500 °C. This temperature was selected to work in mild temperature conditions, more favorable from an energetic point of view, in which coke formation is favored to observe apparent differences between the O₂/AcOH molar ratios used. The O₂/AcOH molar ratios were 0, 0.03, 0.075, 0.15, and 0.3. The results obtained after 5 h time-on-stream are shown in Figure 6. When oxygen content increases, acetic acid conversion increases along with a decrease in coke deposition and hydrogen yield. Compared to the conventional steam reforming (O₂/AcOH = 0), the reaction using the highest O₂/AcOH molar ratio (0.3) achieved a 16% higher acetic acid conversion value and a decrease of about 90 mg_coke·g_cat⁻¹·h⁻¹, which implies a 40% reduction in the coke deposition rate. These results underline the benefits of increasing the oxygen ratio during the reaction. Still, it should also be highlighted that the hydrogen yield decreases by 8%, which is one of the main drawbacks of including oxygen in the feed mixture. The obtained results are entirely in line with those reported by Batista da Silva et al. [39], in which they study the effect of the O₂/AcOH molar ratio from a thermodynamic point of view. To maximize hydrogen production, it is mandatory to control the oxygen addition to take advantage of oxidative reforming while maintaining the hydrogen yield as high as possible. Despite the acetic acid conversion following a rising trend, it could be appreciated how the amount of coke deposited remains almost constant at an O₂/AcOH molar ratio ≥ 0.075. At the same time, with a 0.075 O₂/AcOH molar ratio, hydrogen yield decreases by only 2.4% compared to conventional steam reforming. For that reason, to have a proper balance between the oxygen fed and the hydrogen produced, this ratio (O₂/AcOH = 0.075) was fixed from now on.

Then, the catalytic performance of Ni/CeO₂, Ni/CeO₂-SBA-15, and Ni/CeO₂-m was evaluated using an O₂/AcOH molar ratio = 0.075. In this sense, results in terms of acetic acid conversion and hydrogen yield along TOS are displayed in Figure 7. On the one hand, attending to acetic acid conversion (Figure 7a), the Ni/CeO₂ sample reached lower conversion values than the other samples, and deactivation with time was evidenced by the decrease in conversion from 81.3% to 70.2% after 5 h TOS. Despite the presence of CeO₂ in the catalyst formulation reported to promote coke gasification being active in water-gas shift reaction [40], this behavior could be ascribed to the Ni/CeO₂ nonporous structure, resulting in a low active phase dispersion over the support as a consequence of low surface area. This statement agrees with the highest Ni mean crystallite size determined by the Scherrer equation (see Table 1).

Ni/CeO₂-SBA-15 and Ni/CeO₂-m reached almost complete conversion at short TOS, but the Ni/CeO₂-SBA-15 sample decreased conversion values when TOS is above 3 h. However, Ni/CeO₂-m presents high stability, maintaining the acetic acid conversion at around 97%, at least during the 5 h of TOS. Since Ni mean crystallite sizes were similar for both samples, this behavior could be ascribed to differences in the support behavior. The lower CeO₂ content in the SBA-15-based sample does not have as much oxygen mobility as Ni/CeO₂-m and thus does not promote the coke gasification to a greater extent. In this line, Bereketidou et al. [41] studied the effect of CeO₂ content (0, 10, and 20 wt.%) in the support of Ni-based catalysts for biogas reforming. They reported that doping a 20 wt.% of CeO₂ to Ni/Al₂O₃ provided higher reforming activity and hydrogen yield compared to the unpromoted sample while suppressing carbon formation attributed to the redox properties of ceria. Equally, Santos et al. [42] reported a similar behavior in the partial oxidation of methane using Pt/CeO₂/Al₂O₃ catalysts. They evaluated the effect of CeO₂ addition ranging from 1 to 20 wt.%, leading to the conclusion that catalysts with higher CeO₂ loadings were more resistant to coke deposition given the oxygen exchange capacity of ceria to store/release oxygen because of the Ce⁴⁺/Ce³⁺ redox couple. On the other hand, hydrogen yield followed a similar trend to conversion for the structured catalysts, being higher than the hydrogen yield achieved with the nonporous sample. Ni/CeO₂-SBA-15 and Ni/CeO₂-m reached higher hydrogen yields compared to Ni/CeO₂, ascribed to smaller Ni crystallites, which may lead to more available active sites. Whereas the Ni/CeO₂-m sample showed higher stability given the almost constant value of hydrogen yield obtained, being close to the value expected in the thermodynamic equilibrium, Ni/CeO₂-SBA-15 revealed clear deactivation with TOS.

Coke formation has been reported to be one of the main causes of deactivation in reforming catalysts [14,43,44], not only referring to the amount of coke deposited but also related to its nature. Thus, coke deposited during the oxidative steam reforming of acetic acid was characterized by means of TGA analysis to the used samples, performed under an air-flowing atmosphere. In this context, the derivative thermogravimetric profiles are displayed in Figure 8 where the amount of coke deposited in terms of mg_coke·g_cat⁻¹·h⁻¹ has been included. Broadly, all the profiles have a maximum in the range of 500–600 °C, which could be ascribed to the formation of carbon nanofibers with different ordering degrees given that the amorphous carbon is more reactive in reactions with O₂ oxidizing at temperatures below 400 °C [45]. Concerning all the curves, the maximum in the oxidizing temperature increases in the following order: Ni/CeO₂-m < Ni/CeO₂ < Ni/CeO₂-SBA-15, denoting that more condensed coke deposits have been formed with Ni/CeO₂-SBA-15, being thus more difficult to remove by air combustion [14]. This could also explain the higher coke production obtained for this sample. On the contrary, Ni/CeO₂-m reached the lowest temperature, near 500 °C, related to the more defective carbon nanofilaments, thus being more easily gasified during the reforming.

As stated before, apart from coke formation as a cause of deactivation, oxygen in the feed mixture could lead to active phase oxidation, leading to non-active sites [20]. In this regard, XRD analyses were performed on the used catalysts, whose diffractograms are represented in Figure 9. Peaks corresponding to graphitic carbon (JCPDS 00-041-1487) at 26.5° can be distinguished as a consequence of carbon deposition during the acetic acid oxidative steam reforming. This peak is more pronounced in the Ni/CeO₂-SBA-15 sample, according to higher coke deposition. These peaks could not be clearly discerned in the case of Ni/CeO₂, given the intensity of CeO₂ peaks. Peaks ascribed to cubic Ni⁰ (JCPDS 01-071-3740) arise in all the samples at 44.6 and 51.8°, as most Ni species are under a reduced state during the whole reaction. Unlike the other samples, Ni/CeO₂-SBA-15 also exhibits peaks at 37.3, 43.4, and 63° because of the presence of cubic NiO (JCPDS 01-075-0197), indicating that apart from having the highest coke deposition value, the oxidation of the Ni⁰ sites is taking place, which also contributes to catalyst deactivation, thus explaining the conversion and hydrogen yield drops observed in Figure 7.

In summary, higher conversions and hydrogen yields were achieved with the catalysts with porous structures ascribed to the better dispersion of Ni species given their higher surface area to allow smaller crystallites size. When referred to coke, the samples with higher CeO₂ content present lower coke formation because of the promoting effect of CeO₂ toward carbon gasification. It is noteworthy that the carbon deposition over Ni/CeO₂-SBA-15 was almost two times higher compared to the other samples, which could explain the acetic acid conversion drop since coke is known to be one of the main causes of deactivation in Ni catalyst along with the active phase oxidation. In light of the above, mesostructured Ni/CeO₂-m was the catalyst with the best catalytic performance reaching the highest conversion and hydrogen yield and the lowest carbon deposition.

3. Materials and Methods

3.1. Supports and Catalysts Preparation

A series of Ni catalysts were prepared using different cerium-based supports. Three supports were employed: commercial cerium oxide (Scharlau), CeO₂-modified SBA-15, and mesostructured CeO₂. The commercial CeO₂ sample was treated by air calcination at 550 °C for 5 h with a heating rate of 1.8 °C/min. The SBA-15 material was prepared following a hydrothermal method described by Zhao et al. [46] using Pluronic 123 and TEOS surfactant and silica precursor, respectively. The as-synthesized sample was then calcined at 550 °C for 5 h at a heating rate of 1.8 °C/min to remove the templating agent. This SBA-15 silica was used to prepare the CeO₂-modified SBA-15 support (denoted as CeO₂-SBA-15) and the mesostructured ceria (CeO₂-m). In the first case, SBA-15 was subjected to incipient wetness impregnation using Ce(NO₃)₃·6H₂O as the precursor for achieving a 10 wt.% loading of Ce in the calcined support. On the other hand, the ordered mesoporous ceria (CeO₂-m) was prepared by nanocasting, as described elsewhere [47,48,49,50,51,52]. More specifically, in this work, SBA-15, as a mesoporous hard template, was filled with Ce(NO₃)₃·6H₂O as the metal precursor. For this purpose, 4.85 g of cerium nitrate was dissolved in 50 mL of pure ethanol after progressively adding 6 g of SBA-15 to achieve a homogeneous mixture. The mixture was stirred at room temperature until complete ethanol evaporation. Then, the SBA-15 filled with cerium nitrate was calcined at 600 °C for 6 h with a heating rate of 1.8 °C/min to get cerium oxide. This procedure was repeated three times to ensure the complete filling of the SBA-15 pores. To remove the SBA-15 used as the template, the sample was washed with an aqueous NaOH solution and kept under stirring for 4 h at 60 °C. Then, mesoporous cerium particles were recovered by filtration and subsequently calcined at 550 °C.

The incorporation of Ni onto the three supports mentioned above was accomplished by using an aqueous solution of Ni(NO₃)₂·6H₂O to get Ni loadings around 7 wt.% in the final catalysts following the incipient wetness impregnation technique that was followed for CeO₂-SBA-15 [37]. However, with commercial CeO₂ and CeO₂-m, the wet impregnation method was used because of their low total pore volumes. Wet impregnation was done by adding 1.4 g of CeO₂ or CeO₂-m to the precursor solution (0.52 g of nickel nitrate dissolved in 5 mL of water) under stirring. Next, the mixture was aged in a rotary evaporator for 4 h. Finally, all samples were calcined at 550 °C for 5 h.

3.2. Supports and Catalysts Characterization

Textural properties of the prepared materials were determined using nitrogen adsorption–desorption isotherms at 77 K on a Micromeritics TRISTAR 3000 sorptometer. Before analysis, the samples were outgassed under vacuum at 200 °C for 4 h. Whereas surface area was calculated according to the Brunauer–Emmett–Teller method (BET), pore diameter was derived from the desorption isotherm applying the Barrett–Joyner–Halenda method (BJH). The calcined samples were dissolved by acidic digestion and measured in a Varian Vista-PRO AX CCD-Simultaneous ICP-AES spectrophotometer to have information about the metallic content. TEM micrographs were acquired on a high-resolution transmission electron microscope JEOL JEM-F200 (200 kV). The XRD technique was used to determine the crystalline phases in the samples according to the JCPDS index from the ICSD database using POWD-12++ software and the mean crystallite sizes were determined by the Scherrer equation. The corresponding diffractograms were acquired using a Philips X’pert PRO diffractometer using Cu Kα radiation. The samples’ reducibility and metal-support interactions were studied using TPR analysis on a Micromeritics Autochem 2910 system. The experiments were carried out in situ under a reducing gas mixture (10% H₂/Ar) by flowing 35 N mL/min through the sample with a heating rate of 5 °C/min up to 980 °C. Samples were previously outgassed under Ar flow at 110 °C for 30 min. Thermogravimetric analyses were performed to evaluate the amount of carbon deposited during catalytic tests. These tests were done in air flow with a heating rate of 5 °C/min up to 1000 °C on a TGA-DSC Mettler Toledo thermobalance.

3.3. Catalytic Tests

The prepared catalyst activity in the acetic acid oxidative steam reforming was accomplished on a Microactivity-Pro unit provided by PID Eng & Tech. S.L. The equipment consists of a stainless steel fixed-bed tubular reactor (i.d. = 9.2 mm, L = 300 mm) placed inside an electric oven, as schematized in Figure 10. The temperature inside the reactor is measured using a K-thermocouple. All the components are located inside a stainless-steel hot box with a convector of hot air at 200 °C to avoid condensation of volatile products in the pipes and to allow the preheating of the reactants. The reactor outlet has a Peltier effect unit to condense and separate condensable vapors.

The reactions were performed isothermally at 500 °C and atmospheric pressure. Before, all the catalysts were in situ reduced under flowing pure hydrogen (30 mL/min) at 600 °C for 5.5 h with a heating rate of 2 °C/min according to H₂-TPR results. The reaction feed is a mixture containing acetic acid (AcOH) and water (S/C =2), which is pumped at a flow rate of 0.075 mL/min, vaporized at 200 °C, and mixed with different oxygen–nitrogen mixtures while maintaining the total gas flow at 60 mL/min to achieve a GHSV of 11,000 h⁻¹. The oxygen flow rate was fixed to get the O₂/AcOH molar ratios between 0 (steam reforming) and 0.3 (autothermal reforming of acetic acid [53,54]).

The outlet gas composition is measured online using an Agilent 490 Micro-GC equipped with a thermal conductivity detector (TCD), a Pora Plot U column (10 m), and a Molecular Sieve 5A column (20 m) using He and Ar as the carrier gas, respectively. Condensable vapors were trapped in the condenser and subsequently analyzed in a Varian CP-3900 chromatograph equipped with a CP-WAX 52 CB column and flame ionization detector (FID), using 1,4-butanediol as an internal standard.

The catalytic performances were compared in terms of acetic acid conversion (X_AcOH) and hydrogen yield (Y_H2) using Equations (2) and (3), respectively, where n represents the molar flow of acetic acid or hydrogen at the inlet or outlet stream of the reactor.

$${\mathrm{X}}_{\mathrm{AcOH}}=\frac{{\mathrm{n}}_{\mathrm{AcOH},\mathrm{in}}{-\mathrm{n}}_{\mathrm{AcOH},\mathrm{out}}}{{\mathrm{n}}_{\mathrm{AcOH},\mathrm{in}}}·100$$

(2)

$${\mathrm{Y}}_{{\mathrm{H}}_2}=\frac{{\mathrm{n}}_{{\mathrm{H}}_2,\mathrm{out}}}{{4·\mathrm{n}}_{\mathrm{AcOH},\mathrm{in}}}·100$$

(3)

The value of carbon deposited during the reaction is reported according to the following expression:

$$Coke\left(m{g}_{coke}·g_{cat}{}^{-1}·h^{-1}\right)=\frac{m_{coke}\left(mg\right)}{m_{catalyst}\left(g\right)·TOS\left(h\right)}·100$$

(4)

where m_coke is the amount of coke formed during the reaction determined by TGA, and m_catalyst is the amount of fresh catalyst introduced into the reactor.

4. Conclusions

Hydrogen production using different Ni catalysts supported on CeO₂-based materials through the oxidative steam reforming of acetic acid as a bio-oil aqueous fraction model compound has been studied for the first time.

First, the effect of the O₂/AcOH molar ratio was studied over Ni/CeO₂. The results evidenced that by increasing the oxygen content, the acetic acid conversion increased along with a decrease in coke deposition and hydrogen yield. To get high acetic acid conversion and hydrogen yield with lower coke deposition, an O₂/AcOH molar ratio = 0.075 was selected.

Three nickel catalysts supported on Ce-based material (commercial CeO₂, a mesostructured CeO₂, and CeO₂-SBA-15) were used in the oxidative steam reforming of acetic acid. As a consequence of better metallic dispersion on the mesostructured CeO₂ and the CeO₂-modified SBA-15 samples, the corresponding catalysts reached conversions 27% and 17% higher than the nonporous Ni/CeO₂, respectively, with an increase of about 25 and 15%, respectively, of hydrogen yield after 5 h TOS. On the other hand, coke deposition was lower with the catalysts prepared over CeO₂ (both commercial and mesostructured) because of the intimate contact of Ni with CeO₂, which favors the oxygen mobility toward carbon gasification. As observed by XRD, in Ni/CeO₂-SBA-15, ceria was deposited as particles, not being completely dispersed over the support; thus, the contact between Ni particles and CeO₂ is less likely to take place. For that reason, coke deposition over Ni/CeO₂-SBA-15 was almost two times higher compared to the other samples, which, along with the active phase oxidation detected by XRD, could explain the conversion drop when using this sample.

In conclusion, the mesostructured Ni/CeO_2-m catalyst achieved almost complete conversion of acetic acid (>97%) and values of hydrogen yield (~52%) close to the predicted by the thermodynamic equilibrium along with the lowest carbon deposition (120.6 mg_coke·g_cat⁻¹·h⁻¹) after 5 h TOS at 500 °C. These results evidence that the mesostructured ceria obtained from SBA-15 by nanocasting can be an interesting support for oxidative steam reforming catalysts.