Improved H₂ Production by Ethanol Steam Reforming over Sc₂O₃-Doped Co-ZnO Catalysts

Abstract

H₂ production by catalytically ethanol steam reforming (ESR) is an effective and prospective method for the application of fuel cells. However, the catalysts’ desirable activity and stability remains an unprecedented challenge. Herein, a type of Sc₂O₃-doped Co-ZnO catalyst was developed by a co-precipitation method. The so-constructed Co₂Zn₁Sc_0.3 catalyst exhibited a superb catalytic performance compared with Co-ZnO, giving a STY(H₂) as high as 1.099 mol·h⁻¹·g_-cat⁻¹ (data taken 100 h after the reaction started). In comparison, the pristine Co-ZnO catalyst only afforded a STY(H₂) of 0.684 mol·h⁻¹·g_-cat⁻¹ under identical reaction conditions. Characterization results revealed that the Sc₂O₃ dopant strengthened the electronic interaction between Co species and ZnO, which was in favour of elevating the reduction temperature of Co oxides and boosting the dispersion of the Coⁿ⁺ (n = 1 or 2). The introduction of Sc₂O₃ induced the formation of O²⁻ and OH⁻. All of these effects effectively inhibited the sintering of active Co species and markedly improved the activity and operating stability of the catalyst.

1. Introduction

Ethanol is an important candidate as a chemical carrier of hydrogen for fuel cell applications. Not only is it less hazardous than methanol, it can be also produced from a variety of biomass sources. Therefore, much attention is recently focused on the development of suitable catalysts for steam reforming of ethanol (ESR; C₂H₅OH + 3H₂O → 2CO₂ + 6H₂) [1,2]. A number of reported catalysts have demonstrated good performance in terms of activity and selectivity for the ethanol steam reforming (ESR) reaction, i.e., Rh (or Pt or Pd)/CeO₂–ZrO₂ [3,4], Ni–Rh/CeO₂ [5], Ir/CeO₂ [6], Pt/CeZrO₂ [7], Co/ZnO [8,9], Co/Al₂O₃ (/SiO₂ or /MgO) [10], Co/ZrO₂ (or /CeO₂) [11,12,13], Pt–Co/ZnO [14], Ni/Y₂O₃ (or /ZrO₂) [15,16,17], Ni/Mg–Al mixed oxide [18,19], and Ni/ZrO₂–Yb₂O₃ [20]. Although noble metal-based catalysts showed excellent catalytic activities for ESR, their application on a large scale was limited due to scarcity and expense.

In this regard, the development of non-precious metal-based catalysts is imperative for chemists. ZnO possesses the eminent catalytic reactivity of ESR. The incorporation of Co to ZnO apparently improves its performance [21], and the low cost and excellent performance of the Co-ZnO catalyst has attracted considerable attention and been widely reported. However, although some promising results have been obtained on the ESR reaction, there are still several problems to be solved, such as low stability and high reaction temperature. The results of Song et al. [11,12] indicated that the catalyst deactivation was due to the carbon deposition and cobalt sintering. Introduction of ceria improved activity and durability because of the higher oxygen mobility of ceria. The addition of sodium could evidently enhance the catalytic performance of Co-ZnO, which was attributed to reduction in coke formation [9]. In addition, the reactivity of Co-ZnO catalysts could be further increased by impregnation of Pt [14].

Based on the aforementioned summarization, catalyst stability is one of the most crucial challenges in the development of catalysts for producing hydrogen from ESR. Catalyst deactivation is generally attributed to the carbon deposition, sintering, and oxidation of metal particles [1,13,22,23,24,25,26,27]. Carbon deposition in the form of filaments or whiskers is often found in addition to amorphous carbon.

It is well known that the radius of Sc³⁺ (0.073 nm) is close to that of Zn²⁺ (0.074 nm). Thus, Sc₂O₃ may be readily embedded into the ZnO crystal lattice by ion-doped method in order to modify the crystal structure of ZnO and achieve more stable composites. With these ideas in mind, a series of Sc₂O₃-doped Co-ZnO catalysts were prepared. The physicochemical properties of the catalysts were examined by Transmission electron microscopy (TEM), X-ray Powder Diffraction (XRD), X-ray photoelectron spectroscopy (XPS), Brunauer-Emmett-Teller (BET), and H₂-temperature-programmed reduction (H₂-TPR). The ESR was performed to investigate the effect of Sc₂O₃ doping. After that, we discussed the possible promotional effects of Sc₂O₃ on the ESR performance of the catalysts.

2. Results and Discussion

2.1. Catalyst Performance

The reactivity of ESR (data taken 24 h after the reaction started, similarly hereinafter) over a series of Co_1.25Zn₁Sc_k (molar ratio) catalysts was optimized. The results (Figure 1) showed that CO₂, CO, and CH₄ were the main carbon-containing products and the yield of ethyl ether was negligible. Under the reaction conditions of 0.5 MPa, 723 K, feed-gas composition of C₂H₅OH/H₂O/N₂ = 15/45/40 (molar ratio), and GHSV = 150,000 mL·h⁻¹·g_-cat⁻¹, the five catalysts Co_1.25Zn₁Sc_k with varied k values at 0.2, 0.3, 0.4, 0.6, and 0.8 afforded X(EtOH) of 15.8%, 16.6%, 14.9%, 8.6%, and 5.6% with corresponding STY(H₂) of 0.951, 1.037, 0.931, 0.668 and 0.513 mol·h⁻¹·g_-cat⁻¹, respectively. It is clear that when the Sc/Zn molar ratio (k) was set at 0.3, the catalyst presented high X(EtOH) and STY(H₂), and low S(CO) and S(CH₄).

Next, we investigated the Co/Zn molar ratio (i) effect on the reactivity of ESR. The results are showed in Figure 2. The four catalysts Co_iZn₁Sc_0.3 with varied i values at 1.25, 2, 2.5, and 3 gave X(EtOH) of 16.6%, 19.1%, 18.5%, and 15.9% with corresponding STY(H₂) of 1.037, 1.151, 1.107 and 0.918 mol·h⁻¹·g_-cat⁻¹, respectively, under the aforementioned reaction conditions. Obviously, the Co_iZn₁Sc_0.3 catalyst with i = 2 could show high X(EtOH) and STY(H₂), and low S(CO) and S(CH₄).

The ESR performance of optimal Co₂Zn₁Sc_0.3 catalyst was then evaluated under the aforementioned optimized reaction conditions: 0.5 MPa, 723 K, feed-gas composition EtOH/H₂O/N₂ = 15/45/40, and GHSV = 150,000 mL·h⁻¹·g_-cat⁻¹. The results (data taken 100 h after the reaction started) showed that the X(EtOH) reached 16.6%, with S(CO₂), S(CO) and S(CH₄) of 80.6%, 7.9% and 11.5%, and the corresponding STY(H₂) of 1.099 mol·h⁻¹·g_-cat⁻¹. This STY(H₂) value was 1.6 times that (0.684 mol·h⁻¹·g_-cat⁻¹) of the Sc₂O₃-free counterpart Co₂Zn₁ under identical reaction conditions (

Table 1. Reactivity of ESR for H₂ production over Co₂Zn₁Sc_0.3 and Co₂Zn₁ and the other compared catalysts.

Catalyst	SBET (m2·g−1)	X-EtOH (%)	STY-H2 (mol·g−1·h−1)	Selectivity of C-Containing Products (%)			Reference
				CO2	CO	CH4
Co2Zn1Sc0.3 a	134.2	16.6	1.099	80.6	7.9	11.5	This work
Co2Zn1 a	61.7	10.8	0.684	73.5	8.1	18.4
Ni1.25Zr1 b	134.6	11.0	0.247	70.0	4.7	25.3	[20]
Ni1.25Zr1Yb0.8 b	161.8	18.2	0.396	83.0	11.4	5.6
RhFe/Ca–Al2O3 c	94.3	100	0.052	60.3	0.0	39.7	[28]

^a Reaction conditions: 0.5 MPa pressure, 723 K temperature, 15/45/40 molar ratio of EtOH/H₂O/N₂, and 150,000 mL·h⁻¹·g_-cat⁻¹ GHSV. The activity data was taken 100 h after the reaction started; ^b Reaction conditions: 0.5 MPa pressure, 723 K temperature, 12.5/37.5/50 molar ratio of EtOH/H₂O/N₂, and 90,000 mL·h⁻¹·g_-cat⁻¹ GHSV. The activity data was taken 100 h after the reaction started; ^c Reaction conditions: 623 K temperature, 1/10 molar ratio of EtOH/H₂O, and 34,000 mL·h⁻¹ GHSV. The activity data was taken 1.5 h after the reaction started.

). It has been reported that a 18.2% X(EtOH) was obtained over the Ni_1.25Zr₁Yb_0.8 catalyst, affording 83.0% S(CO₂), 11.4% S(CO), and 5.6% S(CH₄) [20]. These values were slightly superior to those on the Co₂Zn₁Sc_0.3 catalyst, but the STY(H₂) on Ni_1.25Zr₁Yb_0.8 was much lower than that on Co₂Zn₁Sc_0.3. The catalyst containing noble metals such as RhFe/Ca-Al₂O₃ exhibited attracting X(EtOH). However, it gave the byproduct of S(CH₄) up to 39.7% [28]. More importantly, the STY(H₂) was far lower than the aforementioned catalysts. Generally speaking, the Co₂Zn₁Sc_0.3 catalyst possessed the superior reactivity of ESR.

Heat treatment at higher temperatures was performed to investigate the thermal resistance of Co₂Zn₁ and Co₂Zn₁Sc_0.3 catalysts. Typically, the thermal resistance of the catalyst was investigated under the reaction conditions of 0.5 MPa, EtOH/H₂O/N₂ = 15/45/40, GHSV = 150,000 mL·h⁻¹·g_-cat⁻¹, and T ranging at 723–873 K. During the 90 h on stream, the catalysts were treated at high T at 823 and 873 K for 12 h, and then cooled down to 723 K for the ESR. The results are shown in Figure 3. To our surprise, the X(EtOH) fell to 0% over the Co₂Zn₁ catalyst. In contrast to this, the X(EtOH) on the catalyst Co₂Zn₁Sc_0.3 still maintained at a level of 12.6% after the whole process of heat treatment. The finding indicated that the catalyst Co₂Zn₁Sc_0.3 considerably improved the thermal resistance for ESR compared to the pristine Co-ZnO catalyst.

2.2. TEM Characterization

Figure 4 shows the TEM images and the corresponding particle distribution of as-reduced and spent catalysts of Co₂Zn₁Sc_0.3 and Co₂Zn₁. The spent catalysts were run through the ESR for 24 h under the reaction conditions of 0.5 MPa, 823 K, feed-gas composition of C₂H₅OH/H₂O/N₂ = 15/45/40, and GHSV = 150,000 mL·h⁻¹·g_-cat⁻¹. It can be seen that the Co₂Zn₁Sc_0.3 catalyst exhibited excellent dispersion, with a wide size distribution from 8 to 20 nm and a mean size of 13.5 nm. For the reduced Co₂Zn₁ catalyst, the particle size ranged between 9 nm and 30 nm, with a mean size of 17.2 nm. It could be deduced that introducing the Sc₂O₃ to Co-ZnO catalyst was in favor of boosting the dispersion of the active species. On the spent Co₂Zn₁Sc_0.3 catalyst, the observed nanoparticles were fairly uniform in shape and size. The particle size was estimated to be in the range of 12–28 nm, with a mean size of 19.1 nm. In contrast, the particle size of the spent Co₂Zn₁ catalyst was in the range of 20–75 nm, accompanied with the formation of an appreciable amount of carbon-nanotubes or carbon-nanofibers. It could be concluded that the addition of Sc₂O₃ to Co-ZnO inhibited the growth of the particles to some extent, and thus effectively suppressed the sintering and deactivation of the catalyst. These results were in line with the surface area values of the two spent catalysts, as well as the ESR reactivity over the two catalysts (Table 1 and Figure 3).

2.3. XRD and XPS Characterization

The XRD patterns, which are used to identify the phase structure and crystallite size of the materials of the reduced and spent catalysts of Co₂Zn₁ and Co-ZnO doped with Sc₂O₃, are shown in Figure 5. No matter whether the samples were reduced or spent, pure Co-ZnO catalyst or doped with Sc₂O₃, the characteristic diffraction of the ZnO appeared at 2θ = 31.7°, 34.4°, 36.2°, 47.5°, 56.5°, 62.8°, and 67.9° in all the samples. Except for several characteristic diffraction peaks corresponding to ZnO, the peaks located at 2θ = 44.2°, 51.5° and 2θ = 36.8°, 59.4°, 65.2° were ascribed to metallic Co and Co₃O₄ for the reduced Co₂Zn₁ catalyst, respectively. As for the Co₂Zn₁Sc_0.3 catalyst, the main Co species are in the form of CoO with the diffraction lines at 2θ = 42.4° and 61.5°. The results indicated that the Sc₂O₃-doping had a positive effect on the crystal phase of Co-ZnO catalyst, which would be favorable to stabilize Coⁿ⁺ (n = 1 or 2) with high oxidation states, thus effectively refraining from sintering. On the other hand, the XRD patterns of the spent catalysts were similar to those of as-reduced ones, except for the sharper peak shape. The Co and Zn components existed mainly in the forms of metallic Co and ZnO for the catalysts after reaction. The particle size of Co and ZnO in the two spent catalysts was estimated according to the Scherrer’s equation. The average particle sizes of Co and ZnO were 3.5 and 26.9 nm for Co₂Zn₁Sc_0.3, and 19.6 and 37.2 nm for Co₂Zn₁, respectively. The results showed a significant effect on the dispersion of Co species with the addition of small amounts of Sc₂O₃.

XPS analysis is a powerful way of distinguishing the chemical environment and the change in content of surface species. The XPS profile of the spent catalysts displayed that the signals assigned to the Co, Zn, Sc, and O atoms were observed on the surface of the Sc₂O₃-containing and Sc₂O₃-free catalysts. In the Co 2p XPS profile (Figure 6A) of the Sc₂O₃-doped Co-ZnO catalyst, the main peak shifted to a lower binding energy compared with the Sc₂O₃-free counterpart. A similar situation was happened on the Zn 2p XPS profile (Figure 6B). These findings implied that the electronic density on Co atoms increased after Sc₂O₃ doping [29,30], which was in favor of the adsorption and activation of the reactants or intermediates [31,32,33]. Sc species in the Sc₂O₃-doped Co-ZnO catalysts existed in the form of Sc₂O₃ based on the Sc 2p XPS profile (Figure 6C).

Figure 6A displays the Co 2p XPS profile taken on the used catalysts of Co₂Zn₁Sc_0.3 and Co₂Zn₁. It displays a main line and a satellite for Co 2p XPS profile. The observed Co(2p_3/2) and Co(2p_1/2) peaks appeared at about 780.9 eV and 796.6 eV, respectively. According to the literature [34], the Co 2p XPS profile could be divided into three peaks. They were Co_x⁰, CoO and Co(OH)₂ with the corresponding Coⁿ⁺(2p_3/2) B.E. values at 778.0 eV, 780.4 eV and 782.0 eV, respectively [35,36,37]. The binding energies of Coⁿ⁺(2p_3/2) for the reference compounds, along with their relative intensities, are given in

Table 2. XPS binding energy and relative content of the Co species with different valence states at the surface of the tested catalysts.

Catalyst	B.E. (Co 2p3/2)/eV			Relative Content/mol %
	Co0	CoO	Co(OH)2	Co0	CoO	Co(OH)2
Co2Zn1	778.5	780.4	781.8	7.7	15.5	76.8
Co2Zn1Sc0.3	778.4	780.4	781.9	1.2	48.2	50.6

. From the above analysis, it could be concluded that the doping of small amounts of Sc₂O₃ into the Co₂Zn₁ catalyst led to lessening significantly the mol % of Co⁰ species (Figure 6A and Table 2). This is in agreement with the XRD results shown in Figure 5.

As seen in the O 1s XPS profile depicted in Figure 6D, the peaks located at 530.4 eV, 531.6 eV and 532.3 eV were ascribed to oxide anion (O²⁻), adsorbed oxygen (O_ads) and hydroxide anion (OH⁻), successively, on the surface of the spent catalysts of Co₂Zn₁Sc_0.3 and Co₂Zn₁ [31,38]. The content of the O species was calculated and presented in

Table 3. XPS binding energy and relative content of the O-species with different valence states at the surface of the catalysts.

Catalyst	B.E. (Co 2p3/2)/eV			Relative Content/mol %
	O2−	Oads	OH−	O2−	Oads	OH−
Co2Zn1	530.6	531.5	532.4	27.6	36.5	35.9
Co2Zn1Sc0.3	530.5	531.6	532.5	33.9	21.5	44.6

. The mol % of O²⁻ and OH⁻ species arrived 33.9% and 44.6% on the surface of Co₂Zn₁Sc_0.3 respectively, being 1.22 and 1.44 times of the corresponding values (27.6% and 35.9%) of the compared Co₂Zn₁ catalyst. However, the content of O_ads (21.5%) on the surface of the spent Co₂Zn₁Sc_0.3 was evidently lower than the corresponding values (36.5%) of the compared Co₂Zn₁ catalyst. These results indicated that O²⁻ and OH⁻ species favored the H₂ production based on our experimental results, which needs further research in the future.

2.4. H₂-TPR

The H₂-TPR profiles of Co₂Zn₁Sc_0.3 and Co₂Zn₁ catalysts are illustrated in Figure 7. The H₂-TPR curve of Co₂Zn₁ catalyst (Figure 7b) showed the two peaks at 593 K and 791 K, successively. The former may be due to the reduction of Co₂O₃ to Co₃O₄, or further to CoO. The latter is likely attributed to the reduction of Co₃O₄ or CoO to Co⁰ [39,40], which interacted strongly with ZnO surface, and, thus, were difficult to be reduced at lower temperatures. The addition of the appropriate amount of Sc₂O₃ to the Co₂Zn₁ host further strengthened the interaction between Coⁿ⁺ species and the Zn_x-Sc_y-O_z compound-oxide surface, leading to an elevating reduction-temperature of those Coⁿ⁺ species (with the main TPR peak elevating to 933 K from 791 K) (Figure 7a), and stabilizing and anchoring Coⁿ⁺ particles. This result was in line with the aforementioned results of XPS observation and the thermal stability test of the catalysts.

3. Discussion

It is well known that most of the catalyst deactivation of ESR is attributed to the active metal sintering and the deposition of various type of carbon on the surface of the catalyst [41,42]. How to solve these problems is the main challenge in this field. In this work, we apply the ion-doped method to the catalyst preparation. Based on the ionic radius of Sc³⁺ and Zn²⁺ being similar to each other, Sc³⁺ could easily enter into the ZnO lattice. Schottky defects in the form of cationic vacancies would be generated simultaneously for reaching charge compensation. Hence, Coⁿ⁺ can be immobilized at the surface cation-sites. This result is in accordance with the former work of our teams [43]. This will be conducive to suppressing the agglomeration of the Coⁿ⁺ species and maintaining the high Coⁿ⁺ dispersion between the metal Co and ZnO though electronic effect, thus markedly improving the activity and thermal durability of the catalyst. The transformation of O species was also responsible for the excellent catalytic performance, which needs in-depth discussion. These points are proved by the results of the characterizations, and all of the above factors will be responsible for the enhancement effect of the Sc₂O₃ introduction.

4. Experimental

4.1. Catalyst Preparation

The catalysts Co-ZnO and Sc₂O₃-doped Co-ZnO were prepared by a co-precipitation method. Typically, an appropriate amount of Co(NO₃)₂·6H₂O and Zn(NO₃)₂·6H₂O (Sinopharm Chemical Reagent Co. Ltd., Shanghai, China), and Sc(NO₃)₃·6H₂O (Diyang Chemical Reagent Co. Ltd., Shanghai, China) were dissolved in deionized water together. The required amount of K₂CO₃ (Sinopharm Chemical Reagent Co. Ltd., Shanghai, China) was made to aqueous solution in a beak and put it into a water bath at 353 K. The former solution was rapidly added into the latter until the pH ≈ 7 was achieved. The suspension was continuously stirred for another 30 min and then filtered. The precipitate was washed five times, then dried at 383 K for 12 h and calcined in a muffle oven at 623 K for 2 h. Co-ZnO was prepared in a similar way. All samples were pressed, crushed, and sieved to a size of 20–40 mesh for the activity testing.

4.2. Catalytic Evaluation

ESR reactions were tested in a fixed-bed continuous-flow reactor combined with gas chromatograph (GC). The catalyst was reduced at 623 K for 5 h under purified H₂ stream. The ESR reaction was performed under the reaction condition of 0.5 MPa, 723 K, and a feed gas composition of C₂H₅OH/H₂O/N₂ = 15/45/40 (molar ratio) (Sinopharm Chemical Reagent Co. Ltd., Shanghai, China), and GHSV = 150,000 mL·h⁻¹·g_-cat⁻¹. The feed contained C₂H₅OH, and H₂O was introduced into the reactor by a syringe pump (Series II Pump, 10 mL Heads) with N₂ as the internal standard and dilution gas. The C₂H₅OH and H₂O mixed liquid was vaporized at 473 K before access to the reactor.

Prior to entering the sampling valve of the GC, the products passed through a low constant temperature bath (DC-2006) to separate liquid products. An online GC (GC-2014C, Shimadzu, Kyoto, Japan), equipped with a thermal conductivity detector (TCD) (TDX-01 column) and with He as carrier gas, was used to separate N₂, CO, CH₄ and CO₂. H₂ was detected online by TCD using another GC-2014C (TDX-01 column) and with Ar as carrier gas.

Analysis of the liquid products showed that ethyl ether could be ignored and CO₂, CO, and CH₄ were the main products. Thus, ethanol conversion (symbolized as X(EtOH)) (Equation (1)), selectivity of CO₂, CO, and CH₄ (symbolized as S(CO₂), S(CO), and S(CH₄)) (Equations (2)–(4)), and H₂ space-time-yield (symbolized as STY(H₂)) (Equation (5)) could be calculated through a N₂-internal standard method according to the following equations:

$$X_{C{H}_{3}C{H}_{2}OH}=\frac{n_{C{O}_2}+n_{CO}+n_{C{H}_4}}{2n_{C{H}_{3}C{H}_{2}OH}^{in}}=\frac{F_{N_2}{f}_{C{O}_2}^{\prime }{A}_{C{O}_2}+F_{N_2}{f}_{CO}^{\prime }{A}_{CO}+F_{N_2}{f}_{C{H}_4}^{\prime }{A}_{C{H}_4}}{44800A_{N_2}{n}_{C{H}_{3}C{H}_{2}OH}^{in}}$$

(1)

$$S_{C{O}_2}=\frac{f_{C{O}_2}^{\prime }{A}_{C{O}_2}}{f_{C{O}_2}^{\prime }{A}_{C{O}_2}+f_{CO}^{\prime }{A}_{CO}+f_{C{H}_4}^{\prime }{A}_{C{H}_4}}$$

(2)

$$S_{CO}=\frac{f_{CO}^{\prime }{A}_{CO}}{f_{C{O}_2}^{\prime }{A}_{C{O}_2}+f_{CO}^{\prime }{A}_{CO}+f_{C{H}_4}^{\prime }{A}_{C{H}_4}}$$

(3)

$$S_{C{H}_4}=\frac{f_{C{H}_4}^{\prime }{A}_{C{H}_4}}{f_{C{O}_2}^{\prime }{A}_{C{O}_2}+f_{CO}^{\prime }{A}_{CO}+f_{C{H}_4}^{\prime }{A}_{C{H}_4}}$$

(4)

$$ST{Y}_{H_2}=\frac{n_{H_2}}{m_{cat.}}=\frac{60F_{N_2}{f}_{H_2}^{\prime }{A}_{H_2}}{22400A_{N_2}{m}_{cat.}}$$

(5)

where, $F_{N_2}$ represents the flow of N₂ (mL·min⁻¹), $n_{C{H}_{3}C{H}_{2}OH}^{in}$ represents the molar of liquid feed (ethanol and H₂O) inputting to the reactor, n represents the molar of the production (such as CO, CO₂, CH₄, etc.), f’ represents the molar calibration factor, A represents the peak area in GC; m_cat represents the catalyst weight loaded.

4.3. Characterizations

Transmission electron microscopy (TEM) measurements were performed on the Technai F30 and F20 electron microscope (FEI Corp., Hillsboro, OR, USA). XRD measurements were carried out on an X’Pert PRO X-ray diffractometer (Ultima IV, Rigaku, Tokyo, Japan) with Cu K_α (λ_α1 = 0.15406 nm, λ_α2 = 0.15443 nm) radiation. X-ray photoelectron spectroscopy (PHI Quantum 2000 Scanning ESCA Microprobe, PHI, Eden Prairie, MN, USA) measurements were done on a VG ESCA LAB MK-2 apparatus with Al-K_α radiation (15 kV, 25 W, hν = 1486.6 eV) under ultrahigh vacuum (10⁻⁷ Pa), calibrated internally by the carbon deposit C(1s) (E_b = 284.6 eV). The specific surface area was determined by N₂ adsorption using a Micromeritics ASAP 2020 system (Micromeritics, Norcross, GA, USA). Tests of H₂-temperature-programmed reduction (H₂-TPR) and H₂-temperature-programmed desorption (H₂-TPD) of the catalysts were conducted on a fixed-bed continuous-flow microreactor, and change of hydrogen-signal was monitored by an on-line GC (Shimadzu GC-8A, Kyoto, Japan) with a TCD.

5. Conclusions

Sc₂O₃-doped Co-ZnO catalyst for H₂ production by ESR was prepared. The optimized Co₂Zn₁Sc_0.3 exhibited the highest EtOH conversion and STY of H₂. The H₂ STY was 1.099 mol·h⁻¹·g_-cat⁻¹, 1.6 times higher than that on the Co-ZnO catalyst without Sc₂O₃ doping under the optimal reaction conditions. XPS and TPR results disclosed that the electronic interaction between Co and ZnO after Sc₂O₃ addition was strengthened, which was beneficial to the dispersion of the Coⁿ⁺. The introduction of Sc₂O₃ to Co-ZnO also promoted the formation of O²⁻ and OH⁻, which would be conducive to H₂ production, but it needs the further investigation. The present work may contribute to the precise design of catalysts for ESR to produce H₂ with desirable activity and better durability.

It must be admitted that the improved performance has been achieved over the present Sc₂O₃-doped Co-ZnO catalyst, but the selectivity to CO and CH₄ is still high, especially when the ESR is run at the reaction temperature higher than 723 K. It is well known that CO is poison to the anode of the fuel cell and the formation of CH₄ sacrifices the selectivity of H₂. Therefore, more efforts are necessary to fabricate novel catalysts that can efficiently generate more H₂ and minimize CO selectivity. In particular, we must consider the catalysts that minimize undesirable products at higher temperature operations in future work.