A Promising Cobalt Catalyst for Hydrogen Production

: In this work, a metal cobalt catalyst was synthesized, and its activity in the hydrogen production process was tested. The substrates were water and ethanol. Activity tests were conducted at a temperature range of 350–600 ◦ C, water to ethanol molar ratio of 3 to 5, and a feed ﬂow of 0.4 to 1.2 mol/h. The catalyst had a speciﬁc surface area of 1.75 m 2 /g. The catalyst was most active at temperatures in the range of 500–600 ◦ C. Under the most favorable conditions, the ethanol conversion was 97%, the hydrogen production efﬁciency was 4.9 mol (H 2 )/mol(ethanol), and coke production was very low (16 mg/h). Apart from hydrogen and coke, CO 2 , CH 4 , CO, and traces of C 2 H 2 and C 2 H 4 were formed.


Introduction
The global consumption of primary energy is increasing rapidly. From 2000 to 2019, primary energy consumption increased from 109,583 TWh to 162,194 TWh [1]. Such an increase in energy consumption causes increased fossil fuel consumption and CO 2 emissions, which increased by 11.32 billion tonnes. Renewable energy production is growing all the time, but too slowly to compensate for the increased energy demand. From 2000 to 2019, energy production from renewable sources increased from 2870 to 7017 TWh [1]. During this period, wind and solar energy were developed the most. Nowadays, hydrogen technologies are intensively researched. As hydrogen is not dependent on the weather, it may be a more stable energy source. It is thus necessary to develop effective technology for hydrogen production from the available renewable resources. Ethanol is a readily available raw material obtained from biomass. Ethanol production is perfectly controlled, and is a safe compound. The products of ethanol and water conversion should be hydrogen and carbon dioxide. These are produced in ethanol steam reforming (ESR) and water−gas shift reaction (WGSR). Unfortunately, many competing reactions are also possible. As a result, the post-reaction mixture may contain acetaldehyde, hydrocarbons, carbon monoxide, and coke. Our previous work showed the chemical mechanism in detail [2].
The aim of research on hydrogen production from ethanol and water is to minimize the formation of undesirable products. One of the research paths is the development of a selective and active catalyst. There are many different materials with catalytic properties when producing hydrogen from ethanol. Metals, e.g. cobalt, nickel, copper, and platinum, are often catalysts [3][4][5][6][7][8]. Of these, cobalt is very attractive because it is cheaper than noble metals. M. Konsolakis et al. [3] reported that cobalt catalysts achieved a higher ethanol conversion than nickel, copper, and iron catalysts. Additionally, cobalt catalysts are characterized by a lower coke production than nickel catalysts [9,10]. Various cobalt catalysts differing in the amount of cobalt, additives, support, or microstructure have been synthesized. Their stability, activity, and selectivity vary depending on the catalyst structure. For example, M. Konsolakis et al. [3] reported that the optimal cobalt content is 20%. On the other hand, M. Greluk et al. [11] reported that the conversion of ethanol and the selectivity of ethanol conversion into hydrogen increased with increasing the cobalt content. The highest cobalt content was 30%, and this catalyst was the most active. The reason for these divergent conclusions of the two research groups could be the use of different catalyst support. H. Song et al. [12] compared the properties of cobalt catalysts (10 wt%) supported on three carriers, and reported that the carrier influenced the activity of the cobalt catalyst. Cobalt supported on zirconium oxide was the most active. However, using aluminum oxide as a support caused a lower catalyst activity. The least active was cobalt supported on titanium oxide. The activity of the catalysts was related to the cobalt dispersion (ratio of exposed cobalt atoms to total cobalt atoms). It was the largest on zirconium oxide and the smallest on titanium oxide. On the other hand, the specific surface did not matter, because, for zirconium oxide, it was six times smaller than aluminum oxide and two times smaller than titanium oxide.
Y. Li et al. [13] compared the effect of 16 additives (in the amount of 5%) on aluminasupported cobalt catalysts. The authors reported that the additives influenced the microstructure and activity of the catalyst in different ways. For example, calcium reduced the specific surface area of the catalyst the most, while sodium increased it the most. The size of the cobalt particles was decreased the most by scandium, but titanium increased it the most. These studies did not show a simple relationship between the microstructure and the activity of the catalyst. Na, K, Ni, Cu, Zn, Zr, Ce, La, and Fe increased the ethanol conversion, while Mg, Ca, Ti, Sc, V, Cr, and Mn decreased it. Copper accelerated the coking process the most, while sodium inhibited this process.
It is worth noting that the methodology of the catalyst synthesis is also crucial for the properties of the catalyst. Y. Liu et al. [14] reported that catalyst synthesis methods influenced its activity. The most active catalyst was synthesized by homogeneous precipitation using urea, enabling the highest specific surface area and the smallest cobalt particles. Ch. Wang et al. [15] reported that the addition of various surfactants to the metal precursor solution influenced the structure of the catalyst and its activity. Two surfactants were compared, polyvinylpyrrolidone (PVP) and cetyltrimethyl ammonium bromide (CTAB). Both chemical compounds increased the specific surface area of the catalyst and the porosity, with PVP causing a more significant surface development than CTAB. The specific surface area and porosity increased with increasing surfactant concentration. However, the same correlation between the surfactant concentration and the catalyst activity was not found. The use of CTAB at a lower concentration resulted in the most significant increase in ethanol conversion. On the other hand, for PVP, a greater increase in ethanol conversion was achieved with a lower PVP content. The influence of the catalyst synthesis method on its activity was not only observed for cobalt catalysts. E.V. Matus et al. [16] reported that the method of nickel catalyst synthesis also influenced their activity. Even the change in calcination temperature was significant as it changed the process yield. In addition, in other catalytic processes, the methodology of catalyst synthesis influences its activity, as reported by P. Vacharapong et al. [17].
Generally, despite testing a wide variety of materials, further work is needed to develop an effective and durable ethanol steam reforming catalyst. In this study, a pure metallic cobalt catalyst with no additives was used. The metal catalyst is mechanically resistant, does not crumble, and does not dust. Excellent mechanical properties facilitate all operations with the catalyst, e.g., loading, unloading, and regeneration. The possibility of regeneration is an important factor because the activity of catalysts decreases during operation [3,9,10,13,[15][16][17][18][19][20][21][22]. The reason for the decreased activity may be coking [3,10,13,[16][17][18][19][20][21], sintering [3,10,13,15,20], and migration of the active phase [9,22]. Sintered catalysts and catalysts from which the active phase migrated must be replaced. In contrast, coke-coated catalysts can be regenerated. The metal catalyst does not sinter, and metal migration does not change its content on the surface. The metallic cobalt catalyst can be regenerated at high temperatures, which allows the deposited coke to be oxidized rapidly. The metal has an excellent thermal conductivity. Therefore, the temperature in a reactor is easy to control. In reactors with supported catalysts, there are differences between the set-point temperature and the temperature of the catalytic bed. V. Palma et al. [23] reported that these differences depend on the type of carrier and the size of the grains.
In this work, a metal catalyst with a high specific surface area was synthesized. As it was pure cobalt, metal migration did not change its concentration on the catalyst surface. The metal did not sinter, and the melting point of cobalt is much higher than the process temperature. A very good thermal conductivity of cobalt resulted in a very short reactor start-up time. The catalyst bed temperature stabilized within 12 min

Catalyst Characterization
The obtained metal catalyst had a specific surface area of 1.75 m 2 /g determined by nitrogen adsorption isotherm analysis using the BET isotherm (Brunauer−Emmett−Teller). This method was presented in our earlier articles [2,24]. The catalyst topography was analyzed by scanning electron microscopy (SEM), and its elemental composition was analyzed by energy-dispersive X-ray spectroscopy (EDX). These methods were presented in our earlier article [24]. Figures 1-4 show the catalyst surface and the distribution of elements before and after 20 h of operation. Coke was deposited on the catalyst, as evidenced by the presence of carbon on the catalyst surface after use. Oxygen was also present in the used catalyst. Despite the long-term reduction, there was also oxygen on the surface of the fresh catalyst. However, a comparison of the signal intensities from individual elements shows that oxygen was evenly distributed over the surface of the fresh catalyst. On the contrary, the distribution of elements for the used catalyst indicates that oxygen was present in a greater amount on the surface where carbon was also present. E. L. Viljoen and E. van Steen [25] reported that cobalt is easily oxidized by water. The oxidation process starts at 300 • C. The process conditions for producing hydrogen from ethanol and water enable cobalt oxidation by water. On the other hand, the presence of hydrogen makes it possible to reduce the cobalt oxide. Therefore, it is difficult to conclude whether the correlation between the simultaneous presence of oxygen and carbon results from the facilitated deposition of coke on the cobalt oxide or the difficult access of hydrogen to the surface covered with coke. The carbon structures of the whisker structures, previously reported by G. Słowik et al. [22] and by our group for the plasma-catalytic process [24], were not observed.

Gaseous Products
The gaseous products were H 2 , CO, CO 2 , and CH 4 (Table 1). Moreover, C 2 H 2 and C 2 H 4 were present in minimal amounts (<0.1%). There was also water vapor (~2%) and ethanol in the cooled gas. The concentration of hydrogen was from 54 to 69%. The concentration of hydrogen increased with the increasing temperature. As the concentration of hydrogen increased, the concentration of CO 2 increased, and the concentrations of CO and CH 4 decreased. Apart from temperature, an important parameter influencing the concentration of CO, CO 2 , and CH 4 was the molar ratio of water to ethanol. As the molar ratio of water to ethanol increased, the concentration of CO 2 increased. With a stoichiometric ratio of water to ethanol of 3, for the reaction (1): only at a temperature of 450 • C and above was the concentration of CO 2 greater than the concentration of CO and CH 4 . Increasing the water to ethanol molar ratio caused the concentration of CO 2 , even at 350 • C, to be higher than the concentration of CO and CH 4. CO and CH 4 are formed in the following reactions (reactions (2) and (3) The use of water excess increases the production of H 2 and CO 2 and decreases the production of CO and CH 4 .

Coke Formation
Coke production increases rapidly with increasing the feedstock flow ( Table 1). The increase in coke production was greater than the increase in substrate flow. This indicates that coke was formed in large amounts and then consumed. The most likely sequence of reactions leading to the formation of coke was initiated by ethanol dehydration (reaction (6)) [26,28]:  (8) when coke was consumed in the hydration reaction (reaction (9)) [26]: The high molar ratio of water to ethanol inhibited the ethanol dehydration reaction and accelerated the coke hydration reaction. Therefore, coke production decreased as the molar ratio of water to ethanol increased.
The coke production reached minimum values at high temperatures, ranging from 450 to 550 • C. Similarly, V. Palma et al. [19] reported that a low temperature promotes coking, whereas B. Banach et al. [29] reported that coke production peaked at temperatures in the range of 560-570 • C, and it depended on the type of carrier. The different results obtained on different catalysts indicate that the properties of the catalyst have a significant influence on the coking process.

Ethanol Conversion
The metal cobalt catalyst was active from 350 • C. The activity of the catalyst increased rapidly when the temperature increased up to 500 • C (Figures 5 and 6). The further increase in temperature had little effect on the ethanol conversion. Similar changes in the activity of various cobalt catalysts were observed by Y. Li et al. [13].  The ethanol conversion decreased with increasing the feed flow ( Figure 5). This resulted from the shortening of the contact time of the reactants with the catalyst. An increase in the molar ratio of water to ethanol increased the ethanol conversion ( Figure 6). Ethanol is more expensive than water, so it is worth using the excess water to obtain a higher ethanol conversion. Additional benefits of using excess water are reducing the production of coke, CH 4 , and CO and increasing the production of H 2 and CO 2 . On the other hand, the disadvantageous effect of using excessive water is the greater energy consumption necessary to heat and evaporate the water. Due to the energy cost in the industrial process of producing hydrogen from natural gas, the H 2 O/C molar ratio is 2.5 [27], and efforts have been made to reduce it in new installations [30]. Q. Shen et al. [31] reported that the highest ethanol conversion was achieved with a water to ethanol molar ratio of 5 (H 2 O/C molar ratio 2.5). The use of more water did not increase the ethanol conversion. Figures 7 and 8 show the effect of temperature, feed flow rate, and feed stream composition on the hydrogen yield. The hydrogen yield increased with increasing the temperature, but already at 500 • C, it reached a high value. The increase in temperature to 600 • C had little effect on the hydrogen yield. The hydrogen yield decreased with increasing the feed flow rate. In contrast, the increase in the molar ratio of water to ethanol increased the hydrogen yield. The highest hydrogen production efficiency was 4.9 for high temperatures, the lowest flow, and the highest molar ratio of water to ethanol. Q. Shen et al. [31] also reported that the highest hydrogen yield was achieved at a high temperature and a high molar ratio of water to ethanol. The highest hydrogen yield was~4.8 and was achieved for a higher molar ratio of water to ethanol compared to that used in this work.

Catalyst Preparation
The catalyst preparation scheme is shown in Figure 9. Firstly, powdered metallic cobalt was dissolved in nitric acid.
Ammonia liquor was added to the obtained cobalt nitrate, and the cobalt hydroxide precipitated.
The obtained cobalt oxide was wetly granulated in a disk granulator using a 20% sucrose solution as a granulating liquid. Next, the 0.8-2 mm fraction was separated on the sieves. The granules were dried for 24 h. Then, the granules were fired at 1250 • C for 5 h. After cooling down the granules, the 0.8-2 mm fraction was separated and reduced with hydrogen at a temperature of 300 • C for 6 h.

Catalyst Activity Tests
The installation for conducting the catalyst activity tests is shown in Figure 10. A catalyst sample (2 g) was placed in a tubular quartz reactor with an internal diameter of 11 mm. After the substrates were introduced into the reactor, the heating started. The catalyst could not be preheated in the air as cobalt oxidizes at 80 • C. Catalyst activity tests were carried out at a temperature range of 350-600 • C. The temperature of 350 • C was the lowest temperature at which the catalyst showed activity. The maximum test temperature was set at 600 • C, as the energy cost of hydrogen production increased with increasing the temperature. Currently, research is being carried out on new solid oxide fuel cells (SOFCs) that may operate at temperatures up to 600 • C [32,33]. The molar ratio of water to ethanol ranged from 3 to 5. The molar feed flow rate ranged from 0.4 to 1.2 mol/h. The methods and apparatus used in the catalyst activity tests were described in detail in our previous works [2,24,34,35].

Conclusions
The cobalt catalyst is a promising catalyst for hydrogen production from a water and ethanol mixture. The metal catalyst is resistant to sintering and active phase migration. It enables high ethanol conversion and high hydrogen production efficiency. The conditions conducive to obtaining high conversion and yield values are high temperature, low feed flow, and excess of water. The highest ethanol conversion and hydrogen production efficiency were achieved for a temperature of 550 • C, substrate flow of 0.4 mol/h, and water to ethanol molar ratio of 5. Under these conditions, the coke production was low, at 16 mg/h. The use of water in excess was beneficial, as increasing the water to ethanol molar ratio increased the concentration of H 2 and CO 2 . In contrast, the concentration of CH 4 and CO decreased. The high activity of the cobalt catalyst at a temperature of 500-600 • C means that it can be used in the production of hydrogen to supply solid oxide fuel cells operating at low temperatures.