Low temperature steam reforming of ethanol over advanced carbon nanotube-based catalysts

2.1 Catalyst preparation

Multiwalled CNTs (Sigma-Aldrich, inner diameter×outer diameter×length=2–6 nm×10–15 nm×0.1–10 μm, USA) were pretreated in nitric acid in order to remove amorphous carbon and to introduce polar groups on the structure [43]. The pretreatment was done in the following way: CNTs were sonicated and refluxed in 70% HNO₃ (70%, Sigma-Aldrich, ACS reagent, Germany) for 3 h and 8 h, respectively, followed by rinsing, centrifuging and decanting several times with deionized water, and drying. The nickel (Ni(acac)2, 95%, Aldrich, USA) decorated CNT samples were prepared by a wet impregnation method with 10 wt.% Ni nominal loading as described elsewhere [10]. Reduction was made under 15% H₂/Ar flow at 250°C for 15 min. Preparation of Pt promoted Ni₁₀Pt_x/CNTs (10 wt.% Ni and x=1 wt.%, 1.5 wt.% and 2 wt.% Pt) were done by mixing the reduced Ni₁₀/CNT catalyst with 40 cm³of ethanol (99.5%, Altia, Finland) and the required amount of Pt(acac)₂ (99.99%, Aldrich, USA) solution in 200 cm³ of water. The ZnO promoted Ni/CNTs were also prepared with 10 wt.% ZnO (Zinc acetate, 99.99%, Aldrich, USA) in the same procedure by using basic water solution. All other carbon supported catalysts, i.e. activated carbon [AC (FLUKA, purum, Belgium)] and graphitic carbon black (GCB, <20 µm, Aldrich, Switzerland) were pretreated and synthesized in a similar procedure as the CNT-based catalysts with 2 wt.% Ni loadings. Commercial 16.6 wt.% Ni/Al₂O₃ catalyst pellets (HTC400, Crosfield, Warrington, England) crushed and sieved to a powder form (<250 μm particle size) was used as a reference catalyst.

2.2 Catalyst characterization

The specific surface area (S_BET), pore size and pore volume were analyzed by N₂ adsorption-desorption isotherms at -196°C using Micromeritics ASAP2020, USA. The surface area of catalyst samples and the pore size/volume were calculated by the Brunauer-Emmett-Teller (BET) and Barrett-Joyner-Halenda methods, respectively.

Catalyst materials were characterized with X-ray powder diffraction (XRD, Siemens D5000, USA, CuK_α –radiation) to determine the phase and volume averaged crystal size of the catalyst particles (from the fitting parameters of Lorentzian functions over the reflected intensity peaks). The metal content of the samples was measured by energy-dispersive X-ray analysis (EDX, Inca installed on Zeiss Ultra plus FESEM, Germany). Elemental concentrations were measured from at least three locations of each specimen. The average particle size and particle size distribution were determined from energy filtered transmission electron microscopy (EFTEM, LEO 912 OMEGA, Germany, acceleration voltage 120 kV) images by measuring at least 100 particles from each sample.

2.3 Catalyst activity test

In the activity experiment, 100 mg of a metal decorated CNT-based catalyst was packed with quartz wool in a tubular quartz reactor with an inner diameter of 8 mm. The catalyst sample was pretreated and reduced in 20% H₂(in N₂) with 80 cm³ min^-1 flow rate by heating from room temperature up to 350°C with a rate of 15°C min^-1 followed by reduction at 350°C for 30 min, and cooled down to room temperature under the same gas flow. The SRE reaction was tested from 150°C to 450°C with a heating rate of 10°C min^-1. The liquid ethanol-water mixture (molar ratio of 1:3) of 0.091 cm³ min^-1 was fed into the reactor by a peristaltic pump with a total flow rate of 600 cm³ min^-1 (N₂ as carrier gas). All of the experiments were carried out under atmospheric pressure with a constant gas hourly space velocity (38,000 h^-1). The catalyst activity was evaluated based on the outlet concentrations measured by a Fourier transform infrared (FTIR) spectrometer (GASMET) and XMTC H₂ analyzer. The performance of the tested catalysts in SRE was evaluated in terms of ethanol conversion, hydrogen production rate, product compositions and hydrogen selectivity.

3.1 Structural properties

The surface characteristics i.e. surface area (S_BET), pore size and pore volume of the Ni-based CNT catalysts and also the reference catalyst (Ni/Al₂O₃) are presented in Table 1. The S_BET values of 209–269 m² g^-1 for Ni supported CNT-based catalysts are considerably higher than that of the pristine CNTs (75 m² g^-1), indicating a structural change of the support during the catalyst synthesis procedure.

During the purification and carboxylation steps, structural changes, e.g. creation of defects and scavenging of impurities, might probably have occurred on CNTs. This can result in an increase in the surface area from 75 m² g^-1 to 218 m² g^-1and the pore volume from 0.22 cm³ g^-1 to 0.67 cm³ g^-1 for MWCNTs and CNT-COOH (carboxylated CNTs), respectively. Moreover, after the acidic pretreatment steps, the tip openings of tubes also contribute to the change of the overall pore structure [26, 37].

The surface area of the functionalized CNT support increased from 218 m² g^-1 to 269 m² g^-1 after the incorporation of Ni. The pore sizes of CNT-based catalysts varied between 9 nm and 12 nm, i.e. the samples resemble mesoporous materials. There are significant changes in the pore volume and surface area values after the pretreatment steps. By contrast, the pore volume changed dramatically for the carboxylated and Ni decorated CNTs in comparison with the pristine sample. In the textural properties of the Ni₁₀Pt_x/CNT catalysts, no differences were found among the samples. There is a slight decrease in surface area in NiPt-based catalysts due to Pt dispersed on the exterior surface of the tubes. However, in the case of ZnO promoted catalysts, i.e. (ZnO)₁₀Ni₁₀/CNT, the surface area is decreased from 269 m² g^-1 to 209 m² g^-1 due to bigger ZnO particles dispersed on the tube surface, which also blocks the tube openings. For Ni₁₀(ZnO)₁₀ and (ZnO)₁₀Ni₁₀/CNT-based catalysts, a bigger difference appears in the S_BET and pore characteristics compared to Ni₁₀/CNT, due to the presence of relatively large ZnO particles in the samples. The reference catalyst Ni/Al₂O₃ exhibits lower S_BETand pore volume values compared to the Ni-decorated CNT-based catalysts.

The TEM analyses of the Ni₁₀ and Ni₁₀Pt_x (x=1–2 wt.%) CNT-based catalysts (Figure 1) show that the average Ni nanoparticle size varies from 3.5 nm to 4.8 nm for all Ni₁₀Pt_x/CNT catalysts. As reported by Halonen et al. [26], due to the tube confinement effects in CNTs, small particles with a narrow size distribution are obtained. It is difficult to distinguish between the Ni and Pt nanoparticles (Figure 1) as the Ni loading (10 wt.%) is much higher than that of Pt (1–2 wt.%). Some of the nanoparticles might be present in the interior of the nanotubes [27]. Since the inner diameter of the nanotubes is around 2–6 nm and the kinetic diameter of the reactant and product molecules is less than 2 nm [38], the reaction might also occur in the inner cavity of the nanotubes [38]. In the Ni₁₀Pt_x/CNT (x=1–2 wt.%) catalysts, the Pt(111) metallic phase was observed at 2θ=40^o(Figure 1). The peak intensity of the NiO phase might have been reduced because of the formation of new crystal planes in the Ni₁₀Pt_x/CNT and that might have enhanced the reducibility of Ni²⁺ to Ni⁰ for promoted catalysts. This phenomenon has been reported over various NiPt catalysts, e.g. in [35, 36, 42, 45] and thus, it is expected that Ni will be reduced more easily on CNT when Pt and ZnO are present [41, 42, 45].

Figure 1:

X-ray powder diffraction (XRD) patterns and energy filtered transmission electron microscopy (EFTEM) images of (A) Ni₁₀/carbon nanotube (CNT), (B) Ni₁₀Pt₁/CNT, (C) Ni₁₀Pt_1.5/CNT and (D) Ni₁₀Pt₂/CNT catalysts.

Before the reduction step, X-ray diffraction analyses indicate that the oxide phases, i.e. NiO and ZnO, were present (Figure 2). The bigger particles observed in the TEM images were assigned as ZnO according to the very narrow reflection peaks in the corresponding X-ray diffraction patterns. For Ni₁₀Pt_x/CNT catalysts, no reflections were detected from the alloy, thus indicating that no alloy was formed.

Figure 2:

X-ray powder diffraction (XRD) pattern and energy filtered transmission electron microscopy (EFTEM) micrograph of (ZnO)₁₀Ni₁₀/carbon nanotube (CNT) catalyst.

From the EDX analyses, small amounts of Al and Fe impurities are detected, indicating that the catalyst used for growing the CNTs was not completely removed during the support preparation. The elemental composition of fresh and used CNT-based catalysts was analyzed by EDX and summarized in Table 2. The Ni and Pt contents (wt.%) in the NiPt-based catalysts were higher than the nominal loadings due to the partial oxidation of the carbon content.

As the amount of added Pt increased (from 1 wt.% to 2 wt.%), the heterogeneity of the sample increased due to its non-uniform distribution of Pt (Table 3). The average crystal size measured by XRD (using the Scherrer equation) was found to have larger values compared to those received from the TEM analysis.

3.2 Influence of carbon support type

Three different carbon-type supports with 2 wt.% Ni as the nominal loading were prepared and denoted as Ni₂/CNT, Ni₂/GCB and Ni₂/AC catalysts. The nature of the carbon support (AC and GCB) was studied in SRE and the samples were prepared with a similar procedure and the same Ni metal content as CNTs. Based on TEM, the Ni nanoparticles are finely dispersed and decorated over the surface of the CNTs compared to other carbon supports (figures not presented). The average particle size for Ni₂/CNT was 2 nm which was lower than that for Ni₂/GCB (3.3 nm) and Ni₂/AC (4.9 nm). Over the Ni₂/CNT catalyst, it was found that a more narrow distribution of particle sizes was achieved than on other carbon supports [44].

As presented in Figure 3A and B, the ethanol conversion and hydrogen production rate is higher for Ni₂/CNT in comparison with the other carbon supported Ni-based catalysts. The results correlated with the highly dispersed Ni particles, size and the narrow distribution on the support [31]. The Ni₂/CNT catalyst exhibits the highest activity and selectivity in the SRE reaction due to tube confinement effects of the CNTs. The results are in good agreement with the previous studies on CNTs [29, 31].

Figure 3:

Nature of the type of carbon supported Ni-catalyst (Ni/carbon nanotube [CNT], Ni/graphitic carbon black [GCB], and Ni/activated carbon [AC]) in steam reforming of ethanol (SRE) on (A) ethanol conversion and (B) hydrogen production rate (in terms of mol_H2 min^-1 g_cat^-1) as a function of reaction temperature.

3.3 Influence of Pt addition on Ni/CNT

Overall, the Ni₁₀Pt_x-based catalysts resulted in a slight improvement in the ethanol conversion (above 300°C) and hydrogen production (around 300–400°C) was seen compared to the reference catalyst (Figure 4A and B). The Ni particle size increased slightly after the Pt addition and had a positive influence on the SRE activity. There might be two main effects induced by adding Pt, i.e. first, the Ni nanoparticles size increased due to successive thermal treatments, and second, the reducibility of NiO particles probably increased at low temperatures, thus forming more active Ni⁰ nanoparticles [36, 42, 45, 46]. Furthermore, the Pt promotion of the Ni catalyst probably prevented the oxidation of Ni with steam [35, 36]. The ethanol conversion and hydrogen production increased with temperature for all the tested catalysts. The reaction temperature had a significant effect on the reforming activity. At low temperatures, i.e. 150–275°C, the ethanol conversion in the case of the Pt promoted Ni/CNT-based catalysts was comparable to that of the commercial catalyst. Between 225°C and 300°C, ethanol conversion and H₂ production remained quite constant. It is probable that in the temperature range of 150–225°C, adsorption of ethanol is occurring and above that up to ~300°C slight desorption results in a flat plateau. Over the Ni_16.6/Al₂O₃ reference catalyst, the H₂ production is slightly higher at temperatures below 325°C, most probably due to high ethanol conversion and strong acidity character. However, the H₂ production rate per gram of catalyst values for Ni₁₀ and Ni₁₀Pt_x on CNT-based catalysts are higher compared to Ni_16.6/Al₂O₃ catalysts between 350°C and 450°C.

Figure 4:

The effect of Pt addition on (A) ethanol conversion and (b) hydrogen production rate (in terms of mol_H2 min^-1 g_cat^-1) as a function of reaction temperature for Ni₁₀Pt_x/carbon nanotube (CNT) (x=0 wt.%, 1 wt.%, 1.5 wt.% and 2 wt.%) and the reference catalysts in steam reforming of ethanol (SRE).

At 350°C, the maximum ethanol conversion decreased as follows: Ni₁₀Pt₂/CNT=Ni₁₀Pt_1.5/CNT>Ni₁₀Pt₁/CNT=Ni₁₀/CNT>>Ni_16.6/Al₂O₃ (Figure 4A). Al₂O₃ is known to be a highly acidic support, and decomposition reactions that were occurring produced intermediates such as methane and other hydrocarbons which are coke precursors, as reported in [8]. In the range from 150°C to 325°C, over the Ni_16.6/Al₂O₃ catalyst, H₂ production was more or less similar to that of the CNT-based catalysts. Above 350°C, a significantly higher hydrogen production rate was achieved over CNT-based catalysts than Ni_16.6/Al₂O₃ (Figure 4B). Over the CNT-based catalysts, a considerable amount of CH₃CHO was formed up to 400°C. In the case of the CNT-based catalysts, dehydrogenation of ethanol [Eq. (2)] was the first step to take place at 150–300°C followed by subsequent CH₃CHO reforming and disproportionation [Eqs. (6) and (10)–(12)]. The decreasing order of CH₃CHO concentration at 350°C was detected to be as follows: Ni₁₀/CNT>Ni₁₀Pt₁/CNT>Ni₁₀Pt_1.5/CNT=Ni₁₀Pt₂/CNT>Ni_16.6/Al₂O₃(Figure 5A).

Figure 5:

The concentration of products (A) CH₃CHO, (B) CH₄, (C) CO and (D) CO₂as a function of reaction temperature over Ni₁₀Pt_x/carbon nanotube (CNT) (x=0 wt.%, 1 wt.%, 1.5 wt.% and 2 wt.%) and the reference catalyst in steam reforming of ethanol (SRE).

Over the Ni_16.6/Al₂O₃reference catalyst, the formation of CH₃CHO was found to be very low compared to the Ni₁₀/CNT and Ni₁₀Pt_x/CNT (x=1–2 wt.%) catalysts between 150°C and 400°C. At temperatures above 400°C, complete decomposition and reforming of acetaldehyde was detected over the CNT-based catalysts. The formation of CO and CH₄ was very high over the Ni_16.6/Al₂O₃ catalyst, thus decomposition of acetaldehyde might be the dominant reaction compared to reforming at low temperatures [8]. Over the CNT-based catalysts, the formation of methane showed a more or less similar trend for all the catalysts (Figure 5B). Henceforth, at low temperatures, it is difficult to control methane formation; only at high temperatures (i.e.>600°C), methane can be converted completely to H₂ and carbon oxides [11, 15]. Over Ni_16.6/Al₂O₃, the methane production was very high at 150–400°C compared to the CNT-based catalysts. It is also worth mentioning that the formation of ethylene was found to be high over the Ni_16.6/Al₂O₃ (figures not shown), whereas over the CNT-based catalysts, the amount of ethylene was relatively lower than 0.2 vol.%. The hydrocarbons formation detected over the Ni on alumina-based catalyst is due to the dehydration and decomposition reactions [8].

In Figure 5C and D, it can be seen that over the Ni₁₀Pt_x/CNT catalysts, the formation of CO and CO₂ increased at temperatures above 300°C. The CO formation increased up to 400°C, having the maxima at around 375°C, and then decreased as presented in Figure 5C. Above 400°C, the concentration of CO was detected to decline, which might be explained by two reactions; firstly, due to the water-gas-shift (WGS) reaction [Eq. (10)], and secondly, the carbon formation reactions [Eqs. (9) and (11)]. The decomposition reactions are highly pronounced over the Ni_16.6/Al₂O₃, whereas the dehydrogenation, reforming and disproportionation reactions are the possible reaction network over the CNT-based catalysts. The promotional effect of Pt was found to be slightly beneficial in CO and CH₄ reduction; a similar phenomenon was reported in [39]. Over the Ni₁₀Pt₂/CNT catalyst, the formation of CO₂ is observed to be the highest at 400°C compared to other Ni₁₀Pt_x/CNT catalysts, which can be justified by the decreasing trend in the CO concentration via WGS reaction [Eq. (10)]. Furthermore, over the Ni_16.6/Al₂O₃, CO₂ was probably formed via the Boudouard reaction [Eq. (11)], and started to increase with temperature. In the case of CNT-based catalysts, CO₂ increased with temperature due to reforming and the WGS reactions. The addition of noble metals such as Pt to Ni/CNT can probably improve the reducibility, and particle agglomeration is thus avoided as reported in some studies [34–36, 39]. Both the Ni₁₀Pt₁/CNT and Ni₁₀Pt_1.5/CNT catalysts behave similarly and exhibit similar values in the product composition and ethanol conversion. Overall, no significant improvement was observed in the catalytic activity after adding Pt to the Ni/CNT catalyst.

3.4 Influence of ZnO addition to Ni/CNT

ZnO was added into the Ni/CNT catalysts in two ways, i.e. ZnO was incorporated either after the Ni precursor impregnation or before the impregnation step. It can be seen from the textural properties that both the catalysts exhibit different surface characteristics. In the case of the (ZnO)₁₀Ni₁₀/CNT catalyst, the S_BET value is lower than that of the Ni₁₀(ZnO)₁₀/CNT catalyst (Table 1). ZnO addition modifies the surface structure of the Ni/CNT catalyst [41]. In Figure 6A, the ethanol conversion over Ni₁₀(ZnO)₁₀/CNT is high compared to (ZnO)₁₀Ni₁₀/CNT in the studied temperature range. Already, ~50% ethanol conversion was achieved over Ni₁₀(ZnO)₁₀/CNT at 225^oC (T₅₀=225°C), and over (ZnO)₁₀Ni₁₀/CNT at 300°C. A complete ethanol conversion was achieved over Ni₁₀(ZnO)₁₀/CNT at 350°C and for (ZnO)₁₀Ni₁₀/CNT at 425°C. Over the Ni₁₀(ZnO)₁₀/CNT catalyst, the Ni particles are dispersed on the bigger ZnO particles, which indicates that more ethanol is consumed due to the proximity of Ni/NiO and ZnO particles on CNTs. Both the ZnO promoted catalysts achieved higher H₂ production (Figure 6B) than the other catalysts, i.e. Ni and NiPt-based catalysts. For both the NiZnO-based catalysts, the hydrogen production steadily increased with temperature. The Ni₁₀(ZnO)₁₀/CNT catalyst produced slightly higher amounts of H₂ between 150°C and 350°C compared to (ZnO)₁₀Ni₁₀/CNT due to the enhanced steam reforming activity and also higher conversions. At 450°C, (ZnO)₁₀Ni₁₀/CNT produced much higher hydrogen production rate values compared to Ni₁₀(ZnO)₁₀/CNT and Ni₁₀/CNT catalysts. Over the (ZnO)₁₀Ni₁₀/CNT catalyst we obtained the highest H₂ yield of 3.3 H₂ moles per mole of ethanol reacted compared to all other catalysts. Thermodynamic calculations made (using HSC Chemistry 7) on ethanol conversion and H₂ yield in the studied temperature range are presented in Supplemental Table 1.

Figure 6:

The effect and the mode of ZnO addition on the Ni₁₀/carbon nanotube (CNT) catalyst in steam reforming of ethanol (SRE): (A) ethanol conversion and hydrogen production as a function of the reaction temperature and (B) hydrogen production rate (in terms of mol_H2 min^-1 g_cat^-1) over the ZnO promoted Ni₁₀/CNT catalysts.

By adding ZnO to the Ni₁₀/CNT catalyst, a significant improvement in catalyst activity was observed. Over the (ZnO)₁₀Ni₁₀/CNT catalyst, the undesirable byproducts formation was significantly lower than over the Ni₁₀(ZnO)₁₀/CNT catalyst (Figure 7). The concentration of CH₃CHO was relatively higher over ZnO promoted catalysts (Figure 7A) and this indicates the ethanol dehydrogenation is more pronounced than the other CNT-based catalysts. Methane is formed above 350°C, and its formation starts to increase as a function of temperature, as shown in Figure 7B. The methane formation presents a similar trend for all of the CNT-based catalysts except for the reference catalyst. Over the (ZnO)₁₀Ni₁₀/CNT catalyst, the CO concentration is very low in the temperature region between 200°C and 350°C (Figure 7C). For Ni₁₀(ZnO)₁₀/CNT, the CO concentration increases until the temperature reaches 400°C and decreases then to <0.1 vol.%. Both CO and CH₄ free regions are achieved over the (ZnO)₁₀Ni₁₀/CNT catalyst at temperatures below 350°C.

Figure 7:

The concentrations of products (A) CH₃CHO, (B) CH₄, (C) CO and (D) CO₂as a function of reaction temperature over ZnO promoted Ni₁₀/carbon nanotube (CNT) catalysts.

ZnO on the Ni₁₀/CNT catalyst probably forms bigger clusters/clumps around the Ni particles, which may enhance the oxidation of CO to CO₂ (Figure 7D) and also the WGS reaction [18, 24]. This is because of the redox property of ZnO which plays a vital role in catalyst activity for SRE. Moreover, the promotional effects of ZnO might affect the reducibility of NiO at lower temperatures [20, 24, 30, 41]. Henceforth, over the ZnO promoted catalysts, the formation of CO is relatively lower compared to all other tested catalysts in SRE. The probable synergistic effects of Ni is to cleave the C-C bond in the ethanol molecule and to promote the segregated ZnO particles which probably reduce the undesirable CO formation. Moreover, the oxygen content in the ZnO promoted catalysts was higher (Table 2) than in the case of the Ni₁₀/CNT catalyst.

The variation in selectivities towards the desired products (i.e. H₂ and CO₂) was observed based on the mode of ZnO addition (Table 4). Over (ZnO)₁₀Ni₁₀/CNT catalysts, slightly higher H₂ selectivity was achieved than the Ni₁₀(ZnO)₁₀/CNT due to the fact that the ZnO particles are segregated from the Ni counterparts over the CNTs and this might have a positive influence on steam activation by ZnO. In the case of the Ni₁₀(ZnO)₁₀/CNT catalyst, Ni might have strong interactions with ZnO, which is indicated by more side reactions, i.e. decomposition and methanation.

Over ZnO promoted catalysts, CH₃CHO might undergo complete reforming and decomposition above 350°C [Eqs. (6) and (7)]. Consequently, the low concentrations of CO and methane are due to the WGS, disproportionation and decomposition reactions. Both the ZnO promoted catalysts exhibit the highest H₂ selectivity of ~60–76% between 150°C and 450°C. The CH₃CHO formation and decomposition reactions are more pronounced over the Ni₁₀(ZnO)₁₀/CNT catalysts, and also the WGS activity was lower compared to the (ZnO)₁₀Ni₁₀/CNT catalysts. This phenomenon can be seen from the byproducts CH₄ and CO selectivities (Table 4). In the case of the ZnO promoted Ni/CNT catalyst, the bigger ZnO particles on the CNT surface have a greater role in CO reduction to form CO₂ due to the surface oxidation and this facilitates the shift reaction; these phenomena are in good agreement with the previous studies [36, 41, 47]. The byproduct formation rate is very low over (ZnO)₁₀Ni₁₀/CNT and thus it is the more ideal catalyst to be developed since it is producing H₂ with very low CO concentration (see Figure 8). By incorporating ZnO to Ni₁₀/CNT, changes in structural and surface properties of the catalyst occur, probably due to the electronic transfer between ZnO and Ni and also with CNT [41].

Figure 8:

Reaction scheme and possible products and carbon formation during steam reforming of ethanol (SRE) on (A) Ni₁₀/carbon nanotube (CNT) and (B) (ZnO)₁₀Ni₁₀/CNT catalysts (carbon amount: C_ads>C_ads^*).

3.5 Deactivation of CNT-based catalysts

The stability was compared between the Ni₁₀/CNT and (ZnO)₁₀Ni₁₀/CNT catalysts at 400°C by analyzing ethanol conversion and H₂production as a function of time-on-stream (TOS) (Table 5). The (ZnO)₁₀Ni₁₀/CNT catalyst was selected because of its low CO and CH₄selectivity at 400°C.

SRE over the Ni₁₀/CNT catalyst achieves ethanol conversion of 95% at the TOS value of 10 min and over the (ZnO)₁₀Ni₁₀/CNT catalyst with ~90% conversion. The hydrogen production, i.e. ~9 vol.%, was higher over the ZnO promoted catalyst during the first 90 min. Over the (ZnO)₁₀Ni₁₀/CNT catalyst, the ethanol conversion drops from 90% to 78% at the TOS value of 100–120 min and reaches ~70% at 240 min. Further, (ZnO)₁₀Ni₁₀/CNT is more stable after 150 min, conversion stabilizes over the period of time, and the decreasing trend is slower compared to the initial drop between 60 min and 75 min. Over the Ni₁₀/CNT catalyst, the ethanol conversion and H₂ production decline significantly from 95% to 52% and 7.8 vol.% to 6.5 vol.%, respectively.

The decreasing trend in the ethanol conversion and hydrogen production was due to the catalyst deactivation. The carbon balance was calculated and it was found that it remains below 100% and indicates carbon deposition during the reaction test. Initially, the carbon balance was close to ~97% at temperatures below 200°C, and the carbon balance decreases with temperature due to the carbon formation via undesirable reactions. Moreover, by EDX analysis, the increment in carbon wt.% (Table 6) in used catalysts was confirmed.

The deactivation of the studied catalysts is attributed to the carbon deposition over the catalyst particles and a significant amount of carbon formation was confirmed by the TEM images (Figure 9). The nature of carbon coating around metal catalysts was as-grown carbon nanofibers/nanotubes and soot. The carbon deposition might occur during the CO disproportionation and decomposition reactions. Further, the catalyst mass before and after the reaction test was measured, and it was found out that the catalyst mass was increased (Table 6). The physical appearance of the used catalysts was different and the formation of more dense and gummy carbon was observed. According to [12], with the steam-to-carbon ratio higher than 1.8 there is no coke formation; in this study, the ratio was 1.5 and the carbon/coke deposition is high. The abundant filamentous carbon is observed in addition to different types of carbons formed at the proximity of catalyst particles during the reaction (Figure 9). Significant soot and amorphous carbon formation was found over the Ni₁₀/CNT catalysts (Figure 9A and B). The values presented in Table 6 indicate that minor changes in the textural properties of the used Ni₁₀/CNT and (ZnO)₁₀Ni₁₀/CNT catalysts were found after the reaction test for 240 min. The S_BET value increased slightly for the Ni₁₀/CNT sample (fresh) from 269.2 m² g^-1 to 283.9 m² g^-1(used); this increment is due to the amorphous carbon deposition over the catalyst surface (Table 6). The amount of adsorbed carbon is relatively high in the case of Ni/CNT (from Figure 8: C_ads>C*_ads) compared to the ZnO promoted catalyst.

Figure 9:

Different types and shapes of carbon formed over carbon nanotube (CNT)-based catalysts during steam reforming of ethanol (SRE). Energy filtered transmission electron micrographs of used catalysts at different magnifications (A, B) Ni₁₀/CNT and (C, D) (ZnO)₁₀Ni₁₀/CNT.

The pore volume and size decreased from 0.66 cm³ g^-1 to 0.44 cm³ g^-1 and 9.4 nm to 6.2 nm, respectively, most probably caused by the soot and carbon nanofibers that covered the nanotube surfaces. The stability of the ZnO promoted catalyst in terms of ethanol conversion and hydrogen production is better compared to Ni₁₀/CNT during the TOS value of 240 min, and the CO and CH₄concentrations are found to be low. A rapid deactivation was observed over the Ni₁₀/CNT catalyst which might be due to the detected large amounts of encapsulated soot and carbon formation.