Upgrading of Bio ‐ Syngas via Steam ‐ CO 2 Reforming Using Rh/Alumina Monolith Catalysts

: Steam ‐ CO 2 reforming of biomass derived synthesis gas (bio ‐ syngas) was investigated with regard to the steam concentration in the feed using Rh ‐ loaded alumina foam monolith catalysts, which was also accompanied by thermodynamic equilibrium calculation. With 40 vol % steam addition, steam methane reforming and water gas shift reaction were prevailed at the temperature below 640 °C, above which methane dry reforming and reverse ‐ water gas shift reaction were intensified. Substantial change of activation energy based on the methane conversion was observed at 640 °C, where the reaction seemed to be shifted from the kinetic controlled region to the mass transfer controlled region. At the reduced steam of 20 vol %, the increase in the gas velocity led to the increase in the contribution of steam reforming. Comparing to the absence of steam, the addition of steam (40 vol%) resulted in the increase in the production of H 2 and CO 2 , which in turn increased the H 2 /CO ratio by 95% and decreased the CO/CO 2 ratio by 60%. Rh ‐ loaded alumina monolith was revealed to have a good stability in upgrading of the raw bio ‐ syngas.


Introduction
High dependence of energy supply on fossil fuels is leading the world to critical challenges like the issues associated with greenhouse gas emission and the depletion of an energy source [1,2]. According to the recent report from International Energy Agency (IEA) [3], carbon emission has increased by 1.5% in 2018 to a record high of 33.1 billion tonnes, where fossil fuel still accounts for 70% of the rise. This occurs despite renewable energy from solar and wind increases a double-digit growth. Thus, it seems to be obvious that the international regulation and policies for a low carbon footprint will be stringent along with the global target of net zero emission by 2050 [4].
In recent years, as a renewable energy resource the utilization of biomass has gained significant attention, which has been reviewed in terms of overall strategies to produce biofuels [5][6][7]. Together with other thermochemical routes of biomass conversion such as pyrolysis and liquefaction, the gasification of biomass has been also studied intensively [8] as it produces versatile synthesis gas for the downstream fuel production processes like the Fisher-Tropsch process, methanol and dimethyl ether (DME) synthesis. However, prior to being fed into the downstream, the raw biomass-derived synthesis gas (hereafter, bio-syngas) needs to be conditioned in two aspects: removal of impurities and optimising the H2/CO ratio as upstream and downstream upgrading processes, respectively.
Removal of typical impurities such as tar and ammonia from the raw bio-syngas has been investigated [9][10][11] and due to less impact on environments compared to physical separation, hot catalytic gas clean-up process using Ni-based catalysts has been suggested [12]. It is relatively well elucidated how to effectively eliminate those impurities from biosyngas via physical and thermal pathways [13,14]. For upgrading the bio-syngas to high purity synthesis gas, Chattanathan et al. [15] reported the simultaneous steam and dry reforming of methane using a commercial Ni catalyst in which the maximum methane conversion with the least coke formation was achieved at 800 °C when both ratios of CH4/CO2 and CH4/steam were 1. Similarly, Wang et al. [16] proposed that the addition of steam in the reforming of bio-syngas enhances the activity and stability of the catalyst dramatically. As an analogous approach to the production of syngas optimized for synthetic fuel, the combined steam and carbon dioxide reforming of methane have been also investigated [17][18][19]. However, there is still a lack of knowledge on the reaction pathways during the upgrading of bio-syngas that typically contains hydrogen, methane, carbon monoxide, carbon dioxide and steam [20]. Furthermore, considering the recent attention on the utilization of anthropogenic carbon dioxide [21], bio-syngas containing carbon dioxide needs to be properly upgraded using the optimized catalytic system for the downstream fuel manufacturing process for example, which will also contribute to the negative carbon foot print [22].
Monoliths have been extensively investigated as catalyst supporting material in many different applications [23,24]. Recently, monolithic catalysts are gaining more attention as the process miniaturization and intensification becomes the core part in the microreactor technology [25,26]. As mentioned above, raw synthesis gas from biomass gasification usually contains tar and solid particulates, by which the process lines in the downstream are plugged. It was found that monoliths can be operated with gases containing particulates [23]. Monoliths made of refractory materials were found to provide higher activity than normal pelletized catalysts in reforming reactions [24,27]. In view of these advantages, monolith catalysts have been evaluated for the elimination of tar [12,28,29] and gaseous components [16].
The purpose of this study is to investigate the reaction pathways during the reforming of the model bio-syngas that contains typical main components found from the raw biogas. The model bio-syngas in this study does not contain any other particulate impurities and tar. Thus, any effect of these components will not be discussed in this study. In view of their outstanding performance in reforming reactions, aluminium oxide foam monolith and rhodium were chosen as the supporting material and active compound, respectively [30,31]. We have performed the reforming process under the conditions with and without the presence of steam, by which the effect of steam on overall reactions was dealt with in this study.

Catalysts and a Model Bio-Syngas
Catalyst used in the present study was Rh/alumina monolith. Rh nitrate (Sigma Aldrich Cat. No. 83750) was dissolved in deionised water to give 0.04 M solution. Alumina foam monoliths (Ceramatec, Inc. Denver, CO, USA, 30 ppi) of 60 mm in length and of 10 mm in diameter were dipped into the solution for 15 min, followed by being dried in an oven at 110 °C for 3 h. These procedures were repeated three times. After being calcined at 550 °C for 4 h in atmospheric condition, it was found that 1.5 wt % Rh2O3 (via ICP-OES analysis) has been deposited inside the channel of the monoliths. The methods for the deposition of catalyst on structured material have been well reviewed in the literature [32]. Thus-prepared Rh/alumina monolith was placed in a SiC reactor (Hexoloy SA, Saint-Gobain Ceramics, Latrobe, PA, USA) with two other fresh alumina monoliths stacked up and down. The premixed bio-syngas of CO/CO2/CH4/H2/He (28/25/12/33/2 vol %, Coregas Pty. Ltd. Thomastown, VIC, Australia) was used as a model syngas, which referred to the previous study [20]. Helium was used as an internal standard in this study. Figure 1 shows the schematic diagram of the experimental setup for the catalytic upgrading of bio-syngas in this study. Once the Rh/Al2O3 catalyst was stacked with two other fresh alumina monoliths, the reactor was heated to the set temperature in an electric furnace with the rate of 1.67 K min −1 under argon flowing. The premixed bio-syngas was fed into the reactor by means of a mass flow controller (Brooks Instrument, model 5890E, Hatfield, PA, USA) without any dilution. Steam addition was operated by a syringe pump (ISCO Inc., model 260D, Lincoln, NE, USA). Once the temperature was stabilised, argon was replaced by the feed gas and steam. The product gas was analysed using an on-line micro gas chromatograph (Varian, microGC-490, Mulgrave, VIC, Australia), in which a 10 m long PoraPLOT U column and a 10 m long Molsieve 5 Å column were equipped with thermal conductivity detectors. XRD (X-ray diffraction) analyses were performed by using a Bruker AXS D8 advance wide angle X-Ray diffraction instrument with a CuKα radiation source (λ = 1.54056 Å) on the fresh and spent samples together with a blank alumina sample. In the following sections, the net ratio of hydrogen to carbon monoxide represents the net changes in mole fractions as shown in the following equation:

Catalytic Upgrading of Bio-Syngas
Net ratio H2/CO = , where x is the mole fraction.
(1) Non-catalytic operation, which was still with blank monoliths (i.e., without any rhodium addition), has been also evaluated for ensuring that gas-phase contribution was minimised in this study Table 1 shows the materials and reaction parameters investigated in the present study.

Thermodynamic Analysis
In the present study, the steam CO2 reforming (SCR) of bio-syngas was simulated by Aspen HYSYS™ software (Aspen Technology Australia Pty Ltd., Melbourne, VIC, Australia). A Gibbs reactor under isothermal conditions was used as the SCR reactor to simulate a set of reactions including R1-R3 at given reaction conditions by solving the heat and material mass balances, based on minimizing the Gibbs free energy of the components in the reaction set. The Gibbs reactor predicted the product distribution at the chemical equilibrium for the reaction system. The reaction gauge pressure was set at 1 bar. Flow rates of 1000, 1250 and 1667 mL/min were set for the ratio of steam to total feed stream as 0, 20 and 40 vol% respectively. The reaction temperature varied from 500 to 1000 °C with an interval of 50 °C. Figure 2 shows the changes in bio-syngas composition at the equilibrium state in the temperature between 500 and 1000 °C. Feed composition considered for this simulation corresponds to the model synthesis gas as described in the experimental section with the presence of 40 mol% steam. With elevating the temperature, hydrogen composition was seen to steeply increase to reach its maximum at 750 °C and then gradually decreased. Carbon monoxide also increased to 750 °C, where the increase rate became less afterwards. Methane and carbon dioxide decreased with temperature, in which methane was almost completely converted at the temperature higher than 800 °C. Variation in CO2 fraction was found to be divided into three regions, depending on the slope of the CO2 removal rate. Below the temperature of 600 °C, CO2 removal rate was slower than that obtained from the temperature of 600-750 °C where the slope became the highest. Then, the removal rate of CO2 became minimized at the temperature region higher than 750 °C. The concentration of water decreased rapidly to the temperature of 750 °C, followed by the gradual increase afterwards. Combining all these findings, it is expected that the temperature affects significantly the main reactions. Starting from the lower temperature region, where the compositions of H2 and CO were at the minimum level, the reforming of methane with steam and carbon dioxide were simultaneously occurred. This leads to the decrease in methane, water and carbon dioxide. However, in the temperature region lower than 600 °C, the consumption rate of H2O was found to be higher than that of CO2 due to the water-gas shift (WGS) reaction (R3). The rate of methane reforming diminished as most methane was consumed at the temperature higher than 750 °C. The molar fraction of H2O reached its minimum at around 750 °C, where H2 reached the maximum level. After that, it is believed that endothermic reverse WGS (r-WGS) reaction (-R3) was intensified to increase H2O fraction and decrease H2 composition.

Change in the Main Reaction Pathways with Reaction Temperature
Experimental results from catalytic reforming reactions using Rh-loaded alumina monolith catalysts were present in Figure 3. In the temperature between 530 and 850 °C, the methane conversion from catalytic reforming was compared with that from noncatalytic reforming using the alumina monolith without Rh (Figure 3a). While the methane conversion reached more than 80% from the catalytic reforming, negligible CH4 conversion was achieved from the non-catalytic (i.e., gas phase) reaction. This allowed us to consider that the contribution of non-catalytic reactions during the catalytic reforming reactions was very little. After the reaction, the catalyst surface did contain little carbon on its surface, which was also consistent with the previous study [33]. This further verifies the carbon mass balance in this study.  Figure 3b shows the changes in the molar fraction of CO and CO2 during the reactions. Positive and negative values at Y-axis represent the production and consumption rates, respectively. It is interesting to note that with increasing the temperature the production and consumption rate of CO and CO2 changed oppositely. At 640 °C, the consumption rate of CO and production rate of CO2 were both maximised. As further increasing the temperature, the production rate of CO2 was gradually decreased whereas the consumption rate of CO was decreased and eventually became positive at 850 °C. At the lower temperature range (i.e., < 640 °C), the reactions seem to occur in a way to produce CO2 and consume CO in overall, where CO2 conversion through dry reforming is assumed to be very low and WGS reaction is favoured. Consequently, this will increase the overall mole fraction of CO2 in the effluent. It is also believed that the consumption rates of CO by the WGS reaction (R3) and possibly CO methanation at some extent is higher than the production rate of CO via the reforming reactions (R1 and R2). At the higher temperature range (i.e., >640 °C), as the CO2 reforming reactions (R1) and endothermic r-WGS reaction (-R3) are both enhanced, the mole fraction of CO2 started to decrease and simultaneously the consumption rate of CO was decreased. This results in the increase in the mole fraction of H2O, as also predicted in Figure 2.
To further investigate the reaction routes, the production rates of H2 were plotted in Figure 3c. Here, the consumption rate of CH4 was also presented after doubling the original values. Given that all hydrogen atoms produced from methane decomposition are consumed to produce H2 molecules, the doubled consumption rate of CH4 will be the maximum production of H2 from the decomposition of CH4. From Figure 3c, the overall production rate of H2 was higher than the maximum production from methane, indicating the contribution from steam. Thus, the production rate of H2 from steam ( ) can be calculated by subtracting the doubled value of consumed rate CH4 moles ( ) from the overall production rate of H2 ( ). To know the contribution of steam for H2 production, at least qualitatively, we plotted the ratio of / against the reaction temperature in Figure 3d. The ratio decreased exponentially with temperature, indicating the decrease in the contribution of steam for overall H2 production. In addition, the ratio from the experiments was found to be converged on that from thermodynamic calculation. Combining all these results, it is believed that the reactions of methane with steam prevailed over dry reforming under the condition of 40% steam concentration. This is in good agreement with recent studies [34,35], from which the reaction pathways in biosyngas in relation to steam were investigated. However, it is also obvious that the conversion of methane via dry reforming is intensified with temperature.
The ratio of H2/CO is a critical factor for the downstream fuel synthesis in that it often determines the overall efficiency of the process [36]. Changes in the H2/CO molar ratio upon the SCR of bio-syngas is plotted in Figure 4a. The H2/CO ratio from the experiments increased from 1.2 to 1.9 at the temperature of 640 °C, followed by the gradual decrease to 1.7 at 850 °C, which was close to those calculated from thermodynamic equilibrium. The ratio of H2/CO approaching to 2, which is desirable to FT synthesis and methanol production, demonstrates that the reforming of bio-syngas under steam addition can produce the tailored syngas ratio for the downstream fuel production. As expected from Figure 3b, the ratio of CO/CO2 decreased up to 640 °C and then gradually increased towards the thermodynamic values in Figure 4b. The ratio of CO/CO2 became higher than 1 at 850 °C, which was also known to be beneficial for methanol synthesis using a CO2-rich syngas [37].

Activation Energy and Transition Temperature
In Figure 5, the turnover frequency (TOF) number based on the conversion of CH4 over the reciprocal temperature is plotted. It is noteworthy that the slope of the plot sharply changed at 640 °C. In other words, the reaction rate was less affected by the temperature above 640 °C. Such a transition temperature where the slope of Arrhenius plot changes was often observed during the methane conversion reactions in the literature [38][39][40]. The reported transition temperatures were varied with the conditions employed, being around 350-450 °C when alumina-supported Pd and Pt catalysts were used for the oxidation of methane [38] and 550 °C for Rh/α-Al2O3 catalysts used in the partial oxidation of methane [39]. A similar transition was also observed at 500 °C when the Pd/γ-Al2O3 wash-coated monolith was used for the combustion of methane [40]. The transition temperature seems to depend largely on the reaction conditions such as the type of catalysts and the feed composition. Appearance of the transition temperature was mainly attributed to the diffusion of the reactant to the catalyst surface. From Figure 5, activation energies revealed from the plot for the low and high temperature region were 299.6 and 87.3 kJ mol −1 , respectively. If only internal mass transfer limitation is dominant, the activation energy for the temperature above transition point is half of that for the kinetics controlled region. However, 87.3 kJ mol −1 is apparently far less than half of the activation energy (i.e., 150 kJ mol −1 ) from the temperature below the transition point. It has been suggested that the external mass transfer limitation is attributed to such a sharp drop in activation energy [38,40]. Later experiments combined with the modelling study also suggested that external and heat transfer limitations were responsible for the activation energy lower than half of one from the kinetics regime in exothermic methane combustion [41]. However, in this study the effect of heat transfer limitation would be little as the main reactions are endothermic. Therefore, it is believed in this study that the external mass transfer limitation is mainly responsible for the decline in activation energy above the transition point. Figure 5. Variation of turnover frequency (TOF) for CH4 conversion with temperature during the SCR of bio-syngas with 40% steam over the Rh/Al2O3 monolith catalyst.
As discussed above, methane reacted mainly with steam in particular for the low temperature range (e.g., 3 < r / r ). The activation energy of 299.6 kJ mol −1 derived for this temperature range was found to be much higher than those reported from steam reforming of methane in literature as 109 kJ mol −1 for 0.4 wt % Rh/Al2O3 [42], 169.5 kJ mol −1 for 10 wt % Rh/MgO/Al2O3 [43] and 240 kJ mol −1 for 2-5 wt % Rh/Al2O3 [44]. High activation energy from this study is possibly due to the two aspects: methanation of CO/CO2 and retarding effect due to the copresence of H2O and CO2. Panagiotopoulou et al. [45] reported that Al2O3-supported Rh and Ru catalysts provided the high activity towards the methanation of CO and CO2. They have found that when CO and CO2 were present together, the predominant reaction was methanation of CO at the temperature below 500 °C. Above this temperature, the conversion of CO was found to be decreased due to the onset of reverse-WGS reaction. Panagiotopoulou and coworkers have also found that the addition of steam did not practically affect the catalytic activity of Rh catalyst but led to the increase in CO conversion due to the WGS reaction (R3). Regarding the retarding effect on the conversion of CH4, Qin et al. [46] reported that the coexistence of H2O and CO2 strongly blocked the active sites for CH4 activation. These two aspects combined to reduce the conversion of CH4, which led to the high activation energy in the lower temperature region.

Effect of Gas Velocity
As described in the previous Section 3.3, the reaction seems to be controlled by the external mass transfer at a high temperature region, which is facilitated by the reduction in the boundary film layer on the catalyst surface. In view of this, the influence of gas velocity on the SCR of bio-syngas was also investigated at 850 °C with 20% steam in this study. Figure 6a shows the consumption rate and conversion of methane as a function of GHSV. Consumption rate of CH4 was found to be increased as the thickness of film layer was reduced by increasing GHSV. However, due to the fixed length of alumina monoliths in this study, the conversion of methane decreased as a result of the reduction in residence time with the increase in GHSV. As illustrated in Figure 6b, it was found that CO was produced whereas CO2 was consumed at the condition of 850 °C with 20% steam. The production rate of CO and consumption rate of CO2 were both enhanced with increasing the GHSV. In comparison to the fact that CO2 was produced at the same temperature with 40% steam (see Figure 3b), the consumption of CO2 at 20% steam represents that the dry reforming (R1) was intensified whereas the WGS reaction (R3) was reduced. Simultaneously, the r-WGS reaction (-R3) will be thermodynamically favoured at this high temperature. This also led to the decrease in the production of H2, as shown in Figure 6c, where the production rate of H2 was very closer to the doubled amount of consumed CH4 compared to Figure 3c. Consequently, the overall H2/CO ratio was reduced (1.7→1.2) in comparison with that from the condition of 40% steam with GHSV of 7.6 × 10 4 h −1 as shown in Figure 4a. It is also noticed in Figure 6b,c that when GHSV increased, more CO and H2 were produced in consideration of the consumed amount of CO2 and CH4 (indicated by arrows). This implies that more steam is involved in the reforming reactions to produce CO and H2 with increasing GHSV. It seems that the contribution of steam in the reforming of methane was intensified and simultaneously that of the r-WGS reaction was decreased. According to the previous studies [30,47], it is generally agreed that the consumption rate of CH4 through steam reforming is faster than that through dry reforming. It was also reported that the CO2 conversion was considerably retarded when steam was added as a co-oxidant [48]. This phenomenon seems to be intensified with increasing the gas velocity, based on the result in this study. In other words, steam reforming is less affected than dry reforming by increasing the gas velocity. It is believed that more CH4 tends to convert by means of oxidising species (i.e., O* or OH*) generated from H2O than that from CO2 with the reduction in the residence time.
It was found that the overall H2/CO ratio has not been affected by changing the gas velocity in this study. As shown in Figure 6d, overall H2/CO ratio remained constant at around 1.2, which is close to that from the equilibrium value. However, with increasing GHSV, the net H2/CO ratio increased with being diverged from the equilibrium value, which is in good agreement with the previous report [49]. Table 2 shows the effect of steam concentration on the SCR reaction of bio-syngas. Methane conversion was enhanced with the addition of steam in the feed. Considering the changes in mole fraction for each component, the main pathway to convert methane seemed to substantially change. With the absence of steam, CO2 consumption was almost twice higher than that of methane, indicating methane conversion mainly occurs via dry reforming. In addition, CO production was maximised due to the prevailed R-WGS. As a result, the change in H2 mole fraction was only 6.6, which was half of the expected value (12.3) from CH4 decomposition. With 20% steam addition, where methane steam reforming was intensified and the R-WGS was suppressed, the production of H2 increased (6.613.8) while the consumption of CO2 decreased (−11.9−4.8). At the highest steam concentration of 40% in this study, the net change of CO2 eventually became positive, i.e., the production. With the change of steam concentration, the ratios of H2/CO and CO/CO2 have substantially changed. It was seen in Table 2 that H2/CO ratio has increased from 0.39 to 3.78 on the basis of their net changes in mole fractions. For the overall ratio, it has increased from the feed ratio of 1.18-1.70 at 40% steam concentration. While the overall ratio of CO/CO2 was always higher than that from the value from the feed (1.12), the ratio decreased with steam addition, indicating the suppressed dry reforming pathway. In Figure 7, the changes in the ratio of products upon steam addition were plotted together with the thermodynamic equilibrium values at the temperature of 850 °C. In overall, the addition of steam led to the increase in the production of H2 and CO2, which in turn increased the H2/CO ratio and decreased the CO/CO2 ratio. Due to the nonequilibrium characteristics in this condition, there was a relatively large discrepancy between the experimental and the thermodynamic values for the overall ratio of CO/CO2 and net ratio of H2/CO at the absence of steam and 40% steam addition, respectively. For the ratio of CO/CO2, it becomes less significant with increasing steam concentration whereas the net ratio of H2/CO increased the discrepancy with steam addition. In the absence of steam, the dry reforming (R1) was more controlled by mass transfer to reduce the ratio of CO/CO2. With 40% steam addition, the steam involved reactions (i.e., R2 and R3) were more influenced by mass transfer to limit hydrogen production, which in turn decreased the CO/CO2 ratio. In other words, depending on the steam addition, some reactions tended to be more significantly influenced by diffusion limitation, which will affect the ratios between products.  Figure 8 shows the XRD patterns for the fresh and spent Rh alumina samples together with gamma-alumina for comparison. According to the visual observation, freshly calcined Rh2O3/Al2O3 sample has no appreciable change compared to a blank α-Al2O3. This may indicate that Rh particles have been well dispersed on the surface of alumina. After being used in the SCR of bio-syngas at 850 °C, the detection of the peaks assigned for Rh (111), (200) and (220) planes [50,51] indicated the presence of metallic Rh on the surface of alumina. However, the exposure to high temperature led to the sintering of Rh particles on the surface after the reaction.

Conclusions
This study investigated the influence of steam addition on the reforming of biosyngas using Rh-loaded alumina foam monolith catalysts, with which was compared with thermodynamic equilibrium simulation. At 40% steam addition, steam reforming and WGS reaction were prevailed at the temperature below 640 °C, above which dry reforming and r-WGS reaction were gradually intensified. Correspondingly, the ratios of H2/CO and CO/CO2 approached the thermodynamic values with temperature. Transition temperature was observed at 640 °C, where Ea was substantially changed. At this transition, the methane conversion reaction during the reforming of bio-syngas seems to shift from the kinetic controlled region to the diffusion controlled region. At 20% steam addition, the increase in the gas velocity intensified the contribution of steam reforming, which in turn increased the ratio of H2/CO. Without steam, dry reforming is the main pathway for methane conversion and CO production was maximised due to prevailed R-WGS. According to the comparison of the product ratios from the upgrading process with those from thermodynamic values, main reaction pathways dependent on steam concentration seem to be more affected by diffusion control. This study demonstrated that alumina foam monolith supported metal catalyst could be used in the upgrading of raw bio-syngas with a good stability. It is also recommended that other aspects of foam monoliths catalyst such as different morphology or porosity are investigated towards the upgrading of bio-syngas.
Author Contributions: WJ, CL and JP participated in designing of the study and conducting experiments. WJ contributed to conduct experiments, interpret the outcomes and summarized it. WJ led the preparation of the manuscript. CL was partially involved in operating experiments and CL conducted the simulation of bio-syngas reforming using ASPEN analysis. JP initiated this study and supervised the overall process. CL and JP also contributed in preparing the manuscript. All authors have read and agreed to the published version of the manuscript.