Chemical looping glycerol reforming for hydrogen

i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y x x x ( 2 0 1 8 ) 1 e1 2

Available online at www.sciencedirect.com

ScienceDirect journal homepage: www.elsevier.com/locate/he

Chemical looping glycerol reforming for hydrogen production by Ni@ZrO2 nanocomposite oxygen carriers Bo Jiang a,b, Lin Li a, Zhoufeng Bian b, Ziwei Li b, Yang Sun a, Zhehao Sun a, Dawei Tang a,*, Sibudjing Kawi b,**, Binlin Dou c, Maria A. Goula d a

Key Laboratory of Ocean Energy Utilization and Energy Conservation of Ministry of Education, Dalian University of Technology, Dalian 116023, China b Department of Chemical & Biomolecular Engineering, National University of Singapore, Singapore 117585, Republic of Singapore c School of Energy and Power Engineering, University of Shanghai for Science and Technology, Shanghai 200093, China d Laboratory of Alternative Fuels and Environmental Catalysis, Department of Environmental and Pollution Control Engineering, Western Macedonia University of Applied Sciences, Kozani 50100, Greece

article info

abstract

Article history:

The research describes the synthesis of nanocomposite Ni@ZrO2 oxygen carriers (OCs) and

Received 20 March 2018

lanthanide doping effect on maintaining the platelet-structure of the nanocomposite OCs.

Received in revised form

The prepared OCs were tested in chemical looping reforming of glycerol (CLR) process and

8 May 2018

sorption enhanced chemical looping reforming of glycerol (SE-CLR) process. A series of

Accepted 13 May 2018

characterization techniques including N2 adsorption-desorption, X-ray diffraction (XRD),

Available online xxx

inductively coupled plasma optical emission spectrometry (ICP-OES), high resolution transmission electron microscopy (HRTEM), H2 temperature-programmed reduction (H2-

Keywords:

TPR), H2 pulse chemisorption and O2 temperature-programmed desorption (O2-TPD) were

Chemical looping reforming

used to investigate the physical properties of the fresh and used OCs. The results show that

Sorption enhanced

the platelet-stack structure of nanocomposite OCs could significantly improve the metal

Oxygen carrier

support interaction (MSI), thus enhancing the sintering resistance. The effect of lanthanide

Nickel

promotion on maintaining this platelet-stack structure increased with the lanthanide

Lanthanide doping

radius, namely, La3þ > Ce3þ > Pr3þ > Yb3þ. Additionally, the oxygen mobility was also

Zirconia

enhanced because of the coordination of oxygen transfer channel size by doping small radius lanthanide ions. The CeNi@ZrO2 showed a moderate ‘dead time’ of 220 s, a high H2 selectivity of 94% and a nearly complete glycerol conversion throughout a 50-cycle CLR test. In a 50-cycle SE-CLR stability test, the CeNi@ZrO2eCaO showed high H2 purity of 96.3%, and an average CaCO3 decomposition percentage of 53% without external heating was achieved. © 2018 Hydrogen Energy Publications LLC. Published by Elsevier Ltd. All rights reserved.

* Corresponding author. ** Corresponding author. E-mail addresses: [email protected] (D. Tang), [email protected] (S. Kawi). https://doi.org/10.1016/j.ijhydene.2018.05.065 0360-3199/© 2018 Hydrogen Energy Publications LLC. Published by Elsevier Ltd. All rights reserved. Please cite this article in press as: Jiang B, et al., Chemical looping glycerol reforming for hydrogen production by Ni@ZrO2 nanocomposite oxygen carriers, International Journal of Hydrogen Energy (2018), https://doi.org/10.1016/j.ijhydene.2018.05.065

2


Introduction In recent years, hydrogen has been widely applied as a clean energy source for electrical power generation and transportation fuel [1,2]. According to many projections, hydrogen plants are the major source of CO2 emissions to the atmosphere, and this trend will continue in the foreseeable future [3]. However, the conventional steam reforming process for H2 production has some disadvantages such as catalysts deactivation resulted from coke deposition and high energy consumption [4,5]. Researchers have proposed a chemical looping reforming (CLR) process, where the catalyst is repeatedly regenerated by air, as one of the possible routes to solve these problems [6,7]. The CLR is performed by cycling the oxygen carrier (OC) between a fuel feed step, where the OC contacts fuel, and an air feed step, where the OC and deposited coke are oxidized by air [8,9]. Additionally, the oxidation of oxygen carrier in air feed step could provide heat for the calcination of the sorbent in a sorption enhanced chemical looping reforming (SE-CLR) process [10e13]. A distinguished OC should accomplish some characteristics such as excellent redox property, high catalytic reactivity and improved resistance to coke and sintering [8]. Extensive efforts have been dedicated to design OCs adaptable for the CLR process. Zafar et al. [14] have reported that Ni-based oxygen carrier shows highest redox reactivity compared with other oxygen carriers. Additionally, Ni is renowned for its excellent ability for CeC and CeH bond rupture, resulting in high catalytic reactivity in fuel feed step [15,16]. Nevertheless, the major obstacle to Ni-based OCs application is the fast deactivation derived from active phase sintering and coke formation [17,18]. Stabilization of Ni nanoparticles by solid-phase crystallization and enhancing the removal of coke deposition by high oxygen mobility oxides are two typical strategies to improve stability of OCs [19,20]. Zirconia is widely employed as an OC support because of its amphoteric characteristic and oxygen storage capacity (OSC) [21,22]. Xu et al. [23] have demonstrated that the metalsupport interaction (MSI) could significantly improve when the particle size of ZrO2 decreased to an extent that it was comparable to that of Ni particles. Because the amount of ZrO2 nanoparticles are dominant in the catalyst, each Ni nanoparticle would be confined by several ZrO2 nanoparticles, generally showing a platelet-stack structure in transmission electron microscopy (TEM) images. This kind of catalysts with size-comparable Ni and ZrO2 particles (denoted as Ni@ZrO2) would be named as a metal/oxide nanocomposite catalyst better than a conventional metal-supported catalyst (denoted as Ni/ZrO2) [23]. Additionally, it has been demonstrated that smaller particle size of OCs could shorten the distance of oxygen releasing from bulk to surface, and thereby, more oxygen vacancies are beneficial to improve coke resistance [24]. Compared with the conventional metal-support Ni/ZrO2 OC, the Ni@ZrO2 OC with nanocomposite structure is assumed to possess higher coke resistance and sintering resistance. Although the platelet-structure of nanocomposite OCs shows a potential to provide better performance in CLR, the thermal stability of this structure is low due to the phase transformation. Li et al. [25] have compared the conventional

Ni/ZrO2 catalyst with the nanocomposite Ni@ZrO2 catalyst, and they discovered that the Ni/ZrO2 is mainly comprised of monoclinic zirconia (m-ZrO2) while the Ni@ZrO2 is composed of the metastable tetragonal zirconia (t-ZrO2). It has been reported the phase transformation from t-ZrO2 to m-ZrO2 is complete at about 650e700 C, leading to the loss of the unique platelet-stack structure in nanocomposite OCs [26]. Lanthanide oxide as a stabilizer, which could dissolve into zirconia and slow down or prevent the process, has been widely n et al. [27] have studied the stabilization investigated. Ryde effect of Ca, Mg and Ce on nanocomposite ZrO2 in CLC process and concluded that ZrO2 stabilized with Ce performs well. Apart from being stable, the stabilized zirconia with lanthanide would contribute to the formation of defect sites, thus improving the oxygen mobility [28]. In this paper, we synthesize a series lanthanide promoted Ni@ZrO2 nanocomposite OCs and investigate the stabilization effect of lanthanide promotion on nanocomposite Ni@ZrO2 OCs in CLR. The physical-chemical properties of fresh and used OCs are characterized by N2 adsorption-desorption, Xray diffraction (XRD), inductively coupled plasma optical emission spectrometry (ICP-OES), high resolution transmission electron microscopy (HRTEM), H2 temperatureprogrammed reduction (H2-TPR), H2 pulse chemisorption and O2 temperature-programmed desorption (O2-TPD). Moreover, the activity and stability tests of CLR and SE-CLR are performed to evaluate these OCs.

Experimental Preparation of OCs NiO and lanthanide oxide (CeO2, LaO2, YbO2 and PrO2) nanoparticles were prepared by a homogeneous precipitation method. In a typical synthetic process, NiCl2$6H2O (Aladdin, 99.9%), sodium dodecyl sulfate (SDS, Aladdin, 99.9%) and urea (Aladdin, 99.5%) were first dissolved in de-ionized water with molar ration of 1:2:30:60. This solution was continuously stirred to transparent at 40 C and then kept for 20 h at 80 C. The insoluble substance was washed with the de-ionized water and ethanol to remove the chloride species. The resultant was dried at 90 C for 9 h. Then it was further calcined for 3 h at 300 C. As for the preparation of nanocomposite oxygen carriers, the NiO and lanthanide oxide nanoparticles were dispersed in the ZrOCl2$8H2O (Aladdin, 98%) solution with the help of PEG (Aladdin, 3 wt% solution) under sonication treatment. The Ni and lanthanide loading were fixed to 20 wt% and 5 wt%, respectively. The precipitant agent (Aladdin, NH4OH, 1 M) was dropped into the mixture till a pH of 9.5. The obtained solid was then transferred to an autoclave and kept at 140 C for 10 h. Subsequently, it was washed with ethanol and deionized water, dried for 9 h at 90 C and calcined for 3 h at 800 C. The synthesized nanocomposite OCs, which were promoted by Yb, Pr, Ce and La, were denoted as YbNi@ZrO2, PrNi@ZrO2, CeNi@ZrO2 and LaNi@ZrO2, respectively. As for the conventional Ni/ZrO2 oxygen carrier, ZrO2 particles were prepared as the same way as the nanocomposite oxygen carrier except without addition of Ni and lanthanide

Please cite this article in press as: Jiang B, et al., Chemical looping glycerol reforming for hydrogen production by Ni@ZrO2 nanocomposite oxygen carriers, International Journal of Hydrogen Energy (2018), https://doi.org/10.1016/j.ijhydene.2018.05.065

3


precursors. And then, the supported conventional Ni/ZrO2 oxygen carrier was synthesized by impregnating the asprepared ZrO2 powder with a Ni(NO3)2$6H2O (Aladdin, 98%). The Ni loading was also fixed to 20 wt%. Subsequently, the product was dried at 90 C for 9 h and calcined at 800 C for 3 h. In SE-CLR tests, the prepared oxygen carriers were physically mixed with the CaO sorbent with a molar ratio of 1:1. The CaO sorbent was prepared by calcinating commercial nano CaCO3 at 800 C over 4 h. The CO2 absorption capacity of this CaO has been reported in our previous work [29].

used to detect the O2, and the TDX-01 column was used to separate CO2, CO and N2. Between the fuel and air feed step, a N2 purge was performed to evacuate residual gas in the reactor. The carbon balance was within ±5 wt% for both CLR. In the SE-CLR tests, the fuel feed step was performed at 650 C while the air feed step was performed at 800 C because of the regeneration of the sorbent. The gas hourly space velocity (GHSV) in fuel feed step is 250 h1 and in air feed step is 400 h1. The glycerol conversion, steam conversion and coke deposition weight ratio were defined as follows:

Characterization of OCs Xgly or st ¼ The textural properties were investigated by N2 adsorptiondesorption (Micromeritics Tri-Star 3000) at 77 K. The specific surface areas were determined by the Brunauer-EmmettTeller (BET) method. The average pore diameters and the cumulative volumes of pores were obtained using BarrettJoyner-Halenda (BJH) method. The nickel amount was measured by ICP-OES (Optima2000DV, Perkinelmer). XRD were performed by ShimadzuXRD-6000 with the Cu Ka radiation (40 Kv/30 mA) over a 2q range of 10 e85 . Morphologies of the fresh and used OCs were observed by HRTEM (Tecnai G2 F20) at 200 kV. H2-TPR was carried out on a Micromeritics Autochem II 2920 instrument. Generally, the samples (100 mg) were pretreated at 400 C for 1 h in Ar flow (30 ml/min) and then cooled to 100 C. After the pretreatment, a flow mixture of 10 mol% H2eAr (30 ml/min) was injected with the temperature rising from 100 C to 800 C (10 C/min). H2 chemisorption was also performed by the same device to calculate the Ni surface area. O2-TPD measurements were conducted by Micromeritics Autochem II 2920. In a typical procedure, the fresh OC (100 mg) was treated at for 1 h 400 C Ar flow (30 ml/ min), followed by O2 adsorption for 60 min at 200 C. Then, the sample was cooled down to ambient temperature. After 3 h Ar purging, the temperature was increased from 100 C to 1000 C at a heating rate of 10 C/min.

CLR and SE-CLR tests All the tests were performed in a fixed-bed reactor, which was shown in our previous publications [30e32]. The oxygen carrier (40e80 meshes) of 1 g was placed in the middle of the reactor. Prior to tests, the prepared OCs were reduced for 30 min in a 5% H2eN2 flow at 500 C. For the accuracy of the measurement of OC oxidation and reduction data, both the internal and external mass transfer resistance was removed. In the fuel feed step, a glycerol solution was introduced into a heating chamber (300 C) with a steam to carbon ratio (S/C) of 3 to evaporate the mixture in N2 flow (300 ml/min, 20 C, 1.01 105 Pa). The effluents were analyzed by a gas chromatograph (GC, Agilent 7890A) with two detectors. One is a thermal conductivity detector (TCD) with a TDX-01 column using helium as a carrier gas to analyze the species including H2, CO, CO2 and CH4, and the other is a flame ionization detector (FID) with a Porapak-Q column using nitrogen as a carrier gas to measure the concentration of C3H8O3, H2O and CH3CHO. In the air feed step, the OC was oxidized in air flow (50 ml/min). The effluents were detected by another GC (Agilent 7890A) with a TCD. The 5A molecular sieve column was

Fin Fout 100% Fin Z

Coke deposition ðwt%Þ ¼

(1) 0:79 Fair yco2 Arc dt yN2 100% moc

(2)

where the F refers to the C3H8O3 or H2O flux of the inlet and outlet, respectively, ArC represents the relative atomic mass of carbon and mOC is the mass of OCs. The NiO conversion in the ‘dead time’ or the Ni oxidation conversion in the air feed step was defined as the integration of the reduction or oxidation rate [33,34]: FNiO/Ni ¼ Fout; dry yCO þ yCo2 þ yCH3 CHO Fst Xst Fgly Xgly (3) FNi/NiO ¼ 2Fo2 Fout; dry 2yo2 þ yCO þ 2yCO2

(4)

where the dry molar flow rate Fout, dry was calculated from nitrogen balance, and the y represents the concentration of the product. The H2 and C-containing species selectivity were defined as follows: SH2 ¼

3 FH2 ; produced 100% 7 Fgly Xgly

Fj i 100% Sj ¼ P Fj i

(5)

(6)

where j represents the carbon containing species including CH4, CO, CO2, and CH3CHO, and i represents the number of the carbon atoms in this species.

Results and discussion Characterization of fresh OCs Table 1 summarizes the textural properties of fresh OCs. Obviously, the nanocomposite OCs possessed large surface areas, which were almost 2.5 times larger than that of Ni/ZrO2, indicating the smaller ZrO2 particles of nanocomposite OCs. The nanocomposite OCs possessed larger pore volume up to 0.32 cm3/g while the pore volume of the supported OC was only 0.15 cm3/g. The average pore size of all the samples was about 10 nm, suggesting the formation of mesoporous structure in the prepared OCs. Despite the similar pore size of these OCs, the pore volume between the nanocomposite OCs and the Ni/ZrO2 showed remarkable difference. This could result


4


Table 1 e Textural properties of the oxygen carriers. Samples Ni/ZrO2 Ni@ZrO2 YbNi@ZrO2 PrNi@ZrO2 CeNi@ZrO2 LaNi@ZrO2 a b c

Ni dispersion (m2/gNi)a

SBET (m2/g)

Vpore (cm3/g)

Dpore (nm)

NiO particle size (nm)b

Ni content (wt.%)c

Lanthanide content (wt.%)c

2.4 7.6 8.7 8.0 9.3 8.9

36 90 94 92 97 99

0.15 0.39 0.35 0.38 0.32 0.37

10.2 11.4 10.7 11.5 10.3 9.9

24.2/29.1 10.2/11.2 8.2/9.1 9.0/10.3 7.1/7.9 8.5/9.3

19.6 20.1 19.3 19.0 19.2 18.9

e e 4.8 4.9 5.1 5.0

Determined by H2 chemisorption. Determined by Scherrer's equation from the NiO (220) plane (fresh/used). Determined by ICP-OES.

from the partial substitution of NiO nanoparticles for ZrO2 nanoparticles in nanocomposite OCs. Additionally, the lanthanide promotion has a significant impact on the physical properties, as the slight increase of the surface areas and Ni dispersion were observed. This is mainly due to the decrease of the crystallite size by the lanthanide doping. Fig. 1 presents the XRD profiles of the fresh OCs. The formation of different crystalline phases of ZrO2 could be observed in nanocomposite and supported OCs. The peaks at 28.2 and 31.5 are the reflection peaks of the monoclinic zirconia, while the peak at 30.3 could be indexed to the reflection peak of the metastable tetragonal zirconia. The two different crystalline phase compositions could also account for the texture discrepancy (listed in Table 1). It has been demonstrated that the increase of the proportion of t-ZrO2 would enlarge the surface area of the catalyst [35]. The diffraction peaks of lanthanide were not observed in all the nanocomposite samples, revealing that lanthanide oxide is either incorporated in the t-ZrO2 or well dispersed in OCs [26]. The NiO crystal sizes decreased in the following sequence: Ni/ ZrO2 (24.2 nm) > Ni@ZrO2 (10.2 nm) > PrNi@ZrO2 (9.0 nm) > LaNi@ZrO2 (8.5 nm) > YbNi@ZrO2 (8.2 nm) > CeNi@ZrO2 (7.1 nm). It implies that the lanthanide oxide promotion significantly improves the NiO dispersion in the prepared OCs. As shown in Fig. 2, the reduction peaks at 580 K and 680 K were observed for Ni/ZrO2 and Ni@ZrO2. The peak at 580 K is assigned to the reduction of bulk NiO, and the peak at 680 K is ascribed to the reduction of NiOx species which interact with ZrO2 strongly. The enhanced MSI in Ni@ZrO2 could probably result from the increased interfacial area of the

nanocomposite structure. The other two reduction peaks were present for all the lanthanide promoted OCs at low temperature (LT, around 525 K) and high temperature (HT, 680 K). Obviously, the reduction temperature of LT peak of lanthanide promoted OCs was lower than that of bulk NiO (580 K) in Ni/ZrO2. The peak at 525 K is due to the reduction of trace amounts of NiO clusters, suggesting that smaller NiO clusters are formed on the surface of ZrO2 [36]. Navarro et al. [37] have demonstrated that the incorporation of La or Ce to Al2O3 support could increase the reduction intensity of NiO species with a weak MSI. Therefore, the reduction behavior for most of the NiO in lanthanide doping nanocomposite OCs would act as Ni@ZrO2, but trace amounts of NiO could act as Ni/ZrO2 due to the weak MSI. The active surface areas of different OCs are listed in Table 1. Compared with Ni/ZrO2, all nanocomposite OCs showed five times higher Ni dispersion due to the increased interfacial areas. Additionally, lanthanide promoter further increased the Ni dispersion in all nanocomposite OCs. The formation of the proposed nanocomposite structure is evidenced by the TEM profiles (shown in Fig. 3). The conventional Ni/ZrO2 shows a fixed-frame structure, while the nanocomposite OCs are mainly composed of platelet-stack structure. The fixed-frame structure is characterized by bulk matrix in the TEM profiles whereas the platelet-stack structure exhibits sharp-edge particle morphology. According to the XRD results, we could infer that the m-ZrO2 corresponds to the fixed-frame structure, and that the t-ZrO2 is linked with the platelet-stack structure [25]. As illustrated in Fig. 3a, the aggregated NiO nanoparticles in Ni/ZrO2 dispersed on its

Fig. 1 e XRD profiles of fresh oxygen carriers.

Fig. 2 e TPR profiles of fresh oxygen carriers.



5

Fig. 3 e TEM profiles of fresh oxygen carriers. (a) Ni/ZrO2 (b) Ni@ZrO2 (c) YbNi@ZrO2 (d) PrNi@ZrO2 (e) CeNi@ZrO2 (f) LaNi@ZrO2.

m-ZrO2 matrix. In contrast, most NiO nanoparticles in Ni@ZrO2 were confined by size-comparable t-ZrO2 nanoparticles (Fig. 3b). One of the most important factors of CLR is the oxygen supply from OCs [38]. Thus, the O2-TPD was performed to evaluate the oxygen mobility of fresh OCs. As shown in Fig. 4a, there is only one peak for Ni/ZrO2 and Ni@ZrO2 while there are two peaks for lanthanide promoted samples. The first peak at about 351 C is derived from the chemical adsorption oxygen (a O species) [39]. The second peak at about 750 C is associated with oxygen defect, regarding as the partial crystal oxygen (b O species) [39]. Therefore, the amount of b O specie represents the oxygen mobility. Munoz et al. [40] have demonstrated that the promotion of Ce or Pr, which is recognized for their

excellent oxygen storage capacity, would improve the active phase redox properties. Additionally, as shown in Fig. 4a, the b O species desorption amount decreases in the following sequence: YbNi@ZrO2 > PrNi@ZrO2 > CeNi@ZrO2 > LaNi@ZrO2. Liu et al. [41] have investigated FeeCe solution OC and proposed that two types of oxygen transfer could exist in solid solution: (1) by lattice diffusion and (2) by vacancy diffusion, in which an oxygen atom could jump to a neighboring vacancy. Therefore, oxygen vacancies adjacent to each other would form an oxygen transfer channel, and oxygen could transfer from the inside to outside only from the oxygen transfer channel at a high rate. Fig. 4b reveals the b O species desorption amount variation with the channel radius. The channel radii were calculated by the equation in Ref. [42]. Therefore,

Fig. 4 e (a) O2-TPD profiles of oxygen carriers' supports, and (b) the relationship between channel radius of oxygen carrier's support and O2 desorption amount of b oxygen. Please cite this article in press as: Jiang B, et al., Chemical looping glycerol reforming for hydrogen production by Ni@ZrO2 nanocomposite oxygen carriers, International Journal of Hydrogen Energy (2018), https://doi.org/10.1016/j.ijhydene.2018.05.065

6


the improved oxygen mobility of lanthanide promoted OCs should be ascribed to the coordination of oxygen transfer channel size by doping small radius lanthanide ions. The desorption temperature of b oxygen shifted to low temperature with the increase in doping radius: YbNi@ZrO2 > PrNi@ZrO2 > CeNi@ZrO2 > LaNi@ZrO2. The desorption temperature is related with the activation energy of oxygen diffusion process. Krishnamurthy et al. [43] have demonstrated that the activation energy for oxygen diffusion decreases monotonously with increasing cation size and is nearly equal to the energy barrier for oxygen ion hopping across ZreZr edges for large dopants.

CLR activity tests As shown in Fig. 5, H2 production is delayed for all OCs, indicating that the steam reforming could not entirely happen till NiO is sufficiently reduced to Ni. All the samples showed the negative steam conversion, resulting from the H2O production in the NiO reduction period. The ‘dead time’ decreased in this order: Ni@ZrO2 > YbNi@ZrO2 > LaNi@ZrO2 > CeNi@ZrO2 > PrNi@ZrO2 > Ni/ZrO2. This sequence corresponds with the TPR results shown in Fig. 2. Given the proposed mechanism of ‘dead time’ [44], the duration correlates with MSI and active phase dispersion. At the steady stage, the glycerol conversion decreases in the following order: CeNi@ZrO2 > YaNi@ZrO2 > LaNi@ZrO2 > PrNi@ZrO2 > Ni@ZrO2 > Ni/ZrO2. It would be associated with the larger active surface areas (listed in Table 1), since Ni is recognized as the active site for CeC and CeH bond rupture. Compared with Ni/Al2O3 OC in our previous publication, all the samples possessed higher glycerol conversion and shorter reduction period [45].

Fig. 6 e H2 and C-containing species selectivity during the steam reforming state of fuel feed step at 650 C.

The product distribution at 650 C is illustrated in Fig. 6. All the nanocomposite OCs showed higher H2 selectivity and lower aldehyde selectivity than the Ni/ZrO2, which is ascribed to the high Ni dispersion derived from the platelet-stack structure. The CH4 selectivity was low for all the OCs, suggesting that methanation reaction was inhibited under 650 C on account of its extremely exothermic nature. Additionally, a significant improvement in H2 selectivity was achieved in lanthanide promoted OCs, accompanied with higher CO2 selectivity as well as lower CO selectivity. As the CO2/CO ratio is an indicator for water gas shift reaction (WGS), the improved H2 selectivity should be attributed to the enhanced WGS activity. Campos et al. [46] have demonstrated that the strong MSI provides an easier dissociation of OeH bonds from H2O and makes it difficult to cleave CeO bonds during WGS reaction. Therefore, the enhanced MSI from the confinement

Fig. 5 e NiO conversion, glycerol conversion and steam conversion during the fuel feed step at 650 C. (a) Ni/ZrO2 (b) Ni@ZrO2 (c) YbNi@ZrO2 (d) PrNi@ZrO2 (e) CeNi@ZrO2 (f) LaNi@ZrO2. Please cite this article in press as: Jiang B, et al., Chemical looping glycerol reforming for hydrogen production by Ni@ZrO2 nanocomposite oxygen carriers, International Journal of Hydrogen Energy (2018), https://doi.org/10.1016/j.ijhydene.2018.05.065


effect of plate-stack structure contributes to the WGS activity. Moreover, the improved oxygen mobility due to lanthanide promotion, which is evidenced by TPR results, is also beneficial to activate water [47]. The CO2 concentration and Ni oxidation conversion during air feed step at 650 C are exhibited in Fig. 7. Because of the exothermal nature of coke and Ni oxidation, it would produce a significant amount of heat in air feed step, which can be used to regenerate the CO2 sorbent in a SE-CLR process. Meantime, the lattice oxygen consumed in the fuel feed step would be replenished with air [48]. The CO2 peak corresponds to the coke deposition amount in fuel feed step. The coke deposition weight ratio calculated by integrating the CO2 peak are shown in Fig. 8. The nanocomposite OCs showed low coke deposition. The Ni crystal size is strongly related to coke formation. It has been elucidated that small nickel particles show high saturation concentration of coke, thus leading to a low driving force for coke diffusion over Ni particles [49]. Moreover, the lanthanide promotion would further decrease the coke deposition amount. It has been proposed that the coke formation could be remarkably suppressed by the interaction between the Ni crystals and Ce or La species [50].

7

Fig. 8 e Coke deposition in fuel feed step.

CLR stability tests As shown in Fig. 9, the nanocomposite OCs show higher stability than Ni/ZrO2, and that stability of the promoted OCs is better than the Ni@ZrO2. The CeNi@ZrO2 exhibited a nearly complete glycerol conversion throughout the testing period, while the glycerol conversion of Ni/ZrO2 continually decreased from 91% to 80%. The stability of the OCs followed an order of LaNi@ZrO2 > CeNi@ZrO2 > PrNi@ZrO2 > YbNi@ZrO2 > Ni@ZrO2 > Ni/ZrO2. The decline of the glycerol conversion may be related to the loss of the active sites of OCs. Generally, the OC deactivation in CLR process is mainly due to the active

Fig. 9 e Glycerol conversion during fuel feed step at 650 C in 50-cycle stability tests. metal sintering. Two major factors would contribute to the improved sintering resistance: (1) the geometric confinement of the platelet-stack structure which prevents the Ni nanoparticles from sintering, and (2) the inhibitory effect of phase

Fig. 7 e CO2 concentration and Ni oxidation conversion during air feed step at 650 C. (a) Ni/ZrO2 (b) Ni@ZrO2 (c) YbNi@ZrO2 (d) PrNi@ZrO2 (e) CeNi@ZrO2 (d) LaNi@ZrO2. Please cite this article in press as: Jiang B, et al., Chemical looping glycerol reforming for hydrogen production by Ni@ZrO2 nanocomposite oxygen carriers, International Journal of Hydrogen Energy (2018), https://doi.org/10.1016/j.ijhydene.2018.05.065

8


transformation (from t-ZrO2 to m-ZrO2) which maintains the platelet-stack structure. And the inhibitory effect improved with the increase of the lanthanide ion radius, namely, La3þ > Ce3þ > Pr3þ > Yb3þ. A comparison of several investigations on CLR using Nibased OCs is shown in Table 2. The comparison includes CLR of glycerol, ethanol and acetic acid because of the limited references of CLR of glycerol. The glycerol possesses the most complicated molecular structure and lowest solubility; therefore, CLR of glycerol is considered as the most tough one among the three processes. Nevertheless, the CeNi@ZrO2 still showed best performance with an average H2 yield of 11 wt% and an average fuel conversion of 99% in a 50-cycle test. Therefore, it can be inferred that CeNi@ZrO2 is a prospective OC candidate.

Characterization of used OCs As shown in Fig. 10, after the stability tests, only two inconspicuous peaks of m-ZrO2 were displayed in CeNi@ZrO2 and the LaNi@ZrO2, indicating that the t-ZrO2 phase (related with the platelet-stack structure) was stabilized by La and Ce doping. The m-ZrO2 phase (related with the fixed-frame structure) was remarkable in both Pr and Yb promoted OCs, and the m-ZrO2 even became a main phase in YbNi@ZrO2. Additionally, the praseodymium zirconium oxide (Pr2Zr2O7) and ytterbium zirconium oxide (Yb4Zr3O12) were detected. The inhibitory effect on phase transformation from the t-ZrO2 (platelet-stack structure) to m-ZrO2 (fixed-frame structure) decreased as the following sequence: La3þ > Ce3þ > Pr3þ > Yb3þ. The discrepancy of inhibitory effect for different lanthanide doping is mainly associated with the ion contraction. The lanthanide ionic radii follow the sequence of La3þ (0.103 nm) > Ce3þ (0.102 nm) > Pr3þ (0.099 nm) > Yb3þ (0.086 nm). When the phase transformation occurs, because of the mismatch of the ion radii of zirconium and lanthanide, the lanthanide ions substituting for Zr4þ are possibly ejected into the interstitials and formed amorphous oxide, reacting with tZrO2 to further generate lanthanide zirconium oxide. The larger radius lanthanide ion has more chance to be expelled and is detrimental to the formation of lanthanide zirconium oxide. It is suggested that larger radius lanthanide is liable to form amorphous oxide in the interstitials, which would impede the nucleation of m-ZrO2 and thus suppress the phase transformation. By comparison, the easier formation of lanthanide zirconium oxides such as Pr2Zr2O7 and Yb4Zr3O12 tends to undergo the phase transformation.

Fig. 10 e XRD profiles of used oxygen carriers. The TEM profiles of the used OCs are shown in Fig. 11. As shown in Fig. 11aef, after the 50-cycle stability tests, the Ni@ZrO2 shows the loss of the platelet-stack structure, indicating the phase transformation occurs. Because of the phase transformation, the platelet-stack structure is converted to the fixed-frame structure, resulting in the loss of the confinement effect. The different ZrO2 phases were also identified using lattice fringe on the HRTEM images. The mZrO2 particles with d-spacing of 0.283 nm and 0.315 nm were associated with (1 1 1) and (1 1 1) planes, respectively, while the t-ZrO2 with a d-spacing of 0.296 nm was assigned to (1 1 1) plane. The platelet-stack area sizes in the TEM images correspond to the t-ZrO2 peak intensity in the XRD profiles of the used OCs.

SE-CLR activity and stability tests In a SE-CLR process, the OC plays two important roles, which are (1) providing heat for the sorbent calcination in the air feed step and (2) catalyzing the steam reforming reaction in the fuel feed step [10]. The heat released in the air feed step is derived from the Ni and coke oxidation. Therefore, the enhanced oxygen mobility contributes to the more heat release. As illustrated in Fig. 12, during the pre-breakthrough period, NiO reduction, reforming and in-situ CO2 capture occur simultaneously, leading to high H2 purity and low CH4 and CO purity. Compared with Ni/ZrO2eCaO (94.0%), the CeNi@ZrO2eCaO showed higher H2 purity (96.5%). The CO2 produced in NiO reduction period would be absorbed by CaO, which would shift the NiO reduction equilibrium, resulting in shorter ‘dead time’. After the first reduction period

Table 2 e Comparison of CLR using different OCs. OCs CeNi/Al-MCM-41 Ni/Al2O3 0.5CeNi/PSNT Pd/NieCo HT Ni/CaAlOx Ni/SBA Ni/MMT CeNi@ZrO2

Ni loading (wt%) T ( C) S/C Tested cycles Reactant Average conversion Average H2 yield (wt%) 6 13.3 24.7 ~20 ~15 ~12 ~19.9 19.2

650 600 650 575 650 650 650 650

1.5 1.5 3 3 3 3 4 3

10 5 10 6 10 14 20 50

Glycerol Glycerol Glycerol Acetic acid Acetic acid Ethanol Ethanol Glycerol

~90% ~99% ~100% e 99%a ~84% ~78% 99%

6.2 10.4 12.5 ~10.5 ~10.5 e e 11.2

Ref. [8] [45] [51] [44] [52] [32] [30] This work



9

Fig. 11 e TEM profiles of used oxygen carriers. (a) (g) Ni/ZrO2, (b) (h) Ni@ZrO2, (c) (i) YbNi@ZrO2, (d) (j) PrNi@ZrO2, (e) (k) CeNi@ZrO2 and (f) (l) LaNi@ZrO2.

(for Ni/ZrO2eCaO about 8 min, and for CeNi@ZrO2 about 5 min), the metallic nickel is gradually exposed to the reactant, and then the steam reforming is dominant. In the prebreakthrough period, the carbonation of CaO results in the shift of the reforming and WGS reactions. With the saturation of the CaO, the sorption enhanced effect would gradually be abated. In the breakthrough region, the rate of CaO carbonation begins to slow because of the diffusion limitation. In the postbreakthrough region, the CaO is fully saturated, and the previous sorption enhanced effect disappears, resulting in a

reformate gas composition that approaches that of CLR condition. At that time, the feed was turned to air for reoxidation of Ni and regeneration of the sorbent, and no additional external heating lasted for about 20 min for both samples. The reactor temperature was increased to 800 C, and the sorbent regeneration was driven by the strong endothermic reaction as indicated by the CO2 profile. About 45% of the captured CO2 was released for Ni/ZrO2eCaO without external heating, while the CeNi@ZrO2eCaO desorbed 50% captured CO2. This is due to the strong oxygen mobility of the CeNi@ZrO2eCaO, which


10


Fig. 12 e Sorption enhanced chemical looping reforming using Ni/ZrO2eCaO and CeNi@ZrO2eCaO. (a) fuel feed step for Ni/ ZrO2eCaO, (b) air feed step for Ni/ZrO2eCaO, (c) fuel feed step for CeNi@ZrO2eCaO, (d) air feed step for CeNi@ZrO2eCaO. (1dead time, 2-pre-breakthrough, 3-breakthrough and 4-post-breakthrough).

enhances the Ni oxidation reaction. Although the oxidation of the coke deposition would also release CO2, it is much less when compared with CO2 generated from reforming process [51]. As the Ni oxidation was gradually completed, the temperature decreased to about 650 C. Meantime, the external heating was switched on to regenerate the remaining CaCO3. In a large-scale production process, SECLR would be employed in moving bed reactors, and the mole ratio of OC to sorbent would be optimized to make sure the heat generated from OC oxidation could completely compensate that for CaO regeneration. The multi-cycle SE-CLR stability tests were employed on the CeNi@ZrO2eCaO. In contrast, Ni/ZrO2eCaO as the reference OC was also evaluated. As shown in Fig. 13a, even

though the H2 purity in the prebreakthrough period of the two different OCs appears close at the beginning (96.2% for CeNi@ZrO2eCaO vs 95.1% for Ni/ZrO2eCaO), the H2 purity gap gradually increases with the cycles proceeding. At the end of the stability test, the H2 purity of CeNi@ZrO2eCaO was still around 95.0%, while the Ni/ZrO2eCaO had a purity of 87.4%. The high stability of CeNi@ZrO2eCaO is derived from the enhanced inhibitory effect of phase transformation and the improved sintering resistance originated from nanocomposite structure. As shown in Fig. 13b, the percentage of CaCO3 regeneration without external heating in the air feed step maintains stable despite some fluctuations in the 50-cycle test, suggesting that the OC is continuously reduced and reoxidized in the same manner and

Fig. 13 e Sorption enhanced chemical looping reforming stability tests of CeNi@ZrO2eCaO and Ni/ZrO2eCaO. Please cite this article in press as: Jiang B, et al., Chemical looping glycerol reforming for hydrogen production by Ni@ZrO2 nanocomposite oxygen carriers, International Journal of Hydrogen Energy (2018), https://doi.org/10.1016/j.ijhydene.2018.05.065


therefore generates constant amount of heat during the air feed step. In contrast, the percentage of Ni/ZrO2eCaO dropped from 45.2% to 35.3% after the stability test, implying the loss of the oxygen mobility ability due to the sintering of NiO and ZrO2 support.

Conclusion In conclusion, we designed the nanocomposite Ni@ZrO2 OCs and investigated the lanthanide promotion effect on Ni@ZrO2 in CLR. Specifically, the platelet-stack structure of nanocomposite Ni@ZrO2 could significantly improve the Ni dispersion and sintering resistance, which leads to an enhanced MSI in OCs. Although the platelet-stack structure of the Ni@ZrO2 showed a strong confinement effect to prevent sintering, the thermal stability of this structure was erratic on account of the phase transformation. The t-ZrO2 to m-ZrO2 phase transformation of nanocomposite OCs was remarkably suppressed by lanthanide doping and the inhibitory effect was strengthened with the lanthanide radius, namely, La3þ > Ce3þ > Pr3þ > Yb3þ. Moreover, the oxygen mobility was also enhanced because of the coordination of oxygen transfer channel size by doping small radius lanthanide ion. Considering the activity and stability of different nanocomposite OCs, the CeNi@ZrO2 showed a moderate ‘dead time’ of 220 s, a high H2 selectivity of 94% and a nearly complete glycerol conversion throughout a 50-cycle CLR test. In SE-CLR tests, the CeNi@ZrO2 OC also presented a high activity with a high H2 purity of 96.5% in the prebreakthrough period, and the heat release generated by OC oxidation decomposed about 50% CaCO3 in the air feed step because of the improved oxygen mobility by lanthanide doping. Moreover, the CeNi@ZrO2 OC showed a high stability in a 50-cycle multicycle SE-CLR test with an average 96.3% H2 purity in prebreakthrough period and an average CaCO3 decomposition percentage of 53% without external heating.

Acknowledgments This work was supported by National Science Foundation of China [grant number 51476022]; China Scholarship Council [CSC number 201706060128]; and Capacity Building Plan for Some Non-military Universities and Colleges of Shanghai Scientific Committee [grant number 18060502600].

references

[1] Wang C, Chen Y, Cheng Z, Luo X, Jia L, Song M, et al. Sorption-enhanced steam reforming of glycerol for hydrogen production over a NiO/NiAl2O4 catalyst and Li2ZrO3-based sorbent. Energy Fuel 2015;29:7408e18. lu TN. “Green” path from fossil-based to [2] Muradov N, Vezirog hydrogen economy: an overview of carbon-neutral technologies. Int J Hydrogen Energy 2008;33:6804e39. [3] Muradov N. Low to near-zero CO2 production of hydrogen from fossil fuels: status and perspectives. Int J Hydrogen Energy 2017;42:14058e88.

11

[4] Muradov N. Hydrogen via methane decomposition: an application for decarbonization of fossil fuels. Int J Hydrogen Energy 2001;26:1165e75. [5] Muradov N. Low-carbon production of hydrogen from fossil fuels. In: Subramani Basile A, Veziroglu TN, editors. Compendium of hydrogen energy. Oxford: Woodhead Publishing; 2015. p. 489e522. [6] Keller M, Leion H, Mattisson T. Chemical looping tar reforming using La/Sr/Fe-containing mixed oxides supported on ZrO2. Appl Catal B Environ 2016;183:298e307. [7] Dou B, Zhang H, Cui G, Wang Z, Jiang B, Wang K, et al. Hydrogen production and reduction of Ni-based oxygen carriers during chemical looping steam reforming of ethanol in a fixed-bed reactor. Int J Hydrogen Energy 2017;42:26217e30. [8] Jiang B, Dou B, Wang K, Song Y, Chen H, Zhang C, et al. Hydrogen production from chemical looping steam reforming of glycerol by Ni based Al-MCM-41 oxygen carriers in a fixed-bed reactor. Fuel 2016;183:170e6. [9] Adiya SG, Ibrahim Z, Dupont V, Mahmud T. Steam reforming of shale gas in a packed bed reactor with and without chemical looping using nickel based oxygen carrier. Int J Hydrogen Energy 2018;43:6904e17. [10] Antzara A, Heracleous E, Lemonidou AA. Energy efficient sorption enhanced-chemical looping methane reforming process for high-purity H2 production: experimental proofof-concept. Appl Energy 2016;180:457e71. [11] Ryden M, Ramos P. H2 production with CO2 capture by sorption enhanced chemical-looping reforming using NiO as oxygen carrier and CaO as CO2 sorbent. Fuel Process Technol 2012;96:27e36. [12] Pimenidou P, Rickett G, Dupont V, Twigg MV. High purity H2 by sorption-enhanced chemical looping reforming of waste cooking oil in a packed bed reactor. Bioresour Technol 2010;101:9279e86. [13] Hafizi A, Rahimpour MR, Hassanajili S. High purity hydrogen production via sorption enhanced chemical looping reforming: application of 22Fe2O3/MgAl2O4 and 22Fe2O3/ Al2O3 as oxygen carriers and cerium promoted CaO as CO2 sorbent. Appl Energy 2016;169:629e41. [14] Zafar Q, Mattisson T, Gevert B. Redox investigation of some oxides of transition-state metals Ni, Cu, Fe, and Mn supported on SiO2 and MgAl2O4. Energy Fuel 2006;20:34e44. [15] Jiang B, Zhang C, Wang K, Dou B, Song Y, Chen H, et al. Highly dispersed Ni/montmorillonite catalyst for glycerol steam reforming: effect of Ni loading and calcination temperature. Appl Therm Eng 2016;109(Part A):99e108. [16] Muradov NZ. CO2-free production of hydrogen by catalytic pyrolysis of hydrocarbon fuel. Energy Fuel 1998;12:41e8. [17] Tang M, Xu L, Fan M. Progress in oxygen carrier development of methane-based chemical-looping reforming: a review. Appl Energy 2015;151:143e56. [18] Li D, Zeng L, Li X, Wang X, Ma H, Assabumrungrat S, et al. Ceria-promoted Ni/SBA-15 catalysts for ethanol steam reforming with enhanced activity and resistance to deactivation. Appl Catal B Environ 2015;176e177:532e41. [19] Zheng YE, Wei YG, Li KZ, Zhu X, Wang H, Wang YH. Chemical-looping steam methane reforming over macroporous CeO2-ZrO2 solid solution: effect of calcination temperature. Int J Hydrogen Energy 2014;39:13361e8. [20] Zheng Y, Li K, Wang H, Tian D, Wang Y, Zhu X, et al. Designed oxygen carriers from macroporous LaFeO3 supported CeO2 for chemical-looping reforming of methane. Appl Catal B Environ 2017;202:51e63. [21] Nichele V, Signoretto M, Menegazzo F, Gallo A, Dal Santo V, Cruciani G, et al. Glycerol steam reforming for hydrogen production: design of Ni supported catalysts. Appl Catal B Environ 2012;111e112:225e32.


12


[22] Silvester L, Antzara A, Boskovic G, Heracleous E, Lemonidou AA, Bukur DB. NiO supported on Al2O3 and ZrO2 oxygen carriers for chemical looping steam methane reforming. Int J Hydrogen Energy 2015;40:7490e501. [23] Xu BQ, Wei JM, Yu YT, Li Y, Li JL, Zhu QM. Size limit of support particles in an oxide-supported metal catalyst: Nanocomposite Ni/ZrO2 for utilization of natural gas. J Phys Chem B 2003;107:5203e7. [24] Li K, Wang H, Wei Y, Yan D. Partial oxidation of methane to syngas with air by lattice oxygen transfer over ZrO2-modified CeeFe mixed oxides. Chem Eng J 2011;173:574e82. [25] Li SR, Zhang CX, Huang ZQ, Wu GW, Gong JL. A Ni@ZrO2 nanocomposite for ethanol steam reforming: enhanced stability via strong metal-oxide interaction. ChemComm 2013;49:4226e8. [26] Montoya JA, Romero-Pascual E, Gimon C, Del Angel P, Monzon A. Methane reforming with CO2 over Ni/ZrO2eCeO2 catalysts prepared by solegel. Catal Today 2000;63:71e85. n M, Cleverstam E, Johansson M, Lyngfelt A, [27] Ryde Mattisson T. Fe2O3 on Ce-, Ca-, or Mg-stabilized ZrO2 as oxygen carrier for chemical-looping combustion using NiO as additive. AIChE J 2010;56:2211e20. [28] Zheng Y, Li K, Wang H, Zhu X, Wei Y, Zheng M, et al. Enhanced activity of CeO2eZrO2 solid solutions for chemicallooping reforming of methane via tuning the macroporous structure. Energy Fuel 2015;30:638e47. [29] Jiang B, Dou B, Wang K, Zhang C, Li M, Chen H, et al. Sorption enhanced steam reforming of biodiesel by-product glycerol on Ni-CaO-MMT multifunctional catalysts. Chem Eng J 2017;313:207e16. [30] Jiang B, Dou B, Wang K, Zhang C, Song Y, Chen H, et al. Hydrogen production by chemical looping steam reforming of ethanol using NiO/montmorillonite oxygen carriers in a fixed-bed reactor. Chem Eng J 2016;298:96e106. [31] Wang K, Dou B, Jiang B, Zhang Q, Li M, Chen H, et al. Effect of support on hydrogen production from chemical looping steam reforming of ethanol over Ni-based oxygen carriers. Int J Hydrogen Energy 2016;41:17334e47. [32] Wang K, Dou B, Jiang B, Song Y, Zhang C, Zhang Q, et al. Renewable hydrogen production from chemical looping steam reforming of ethanol using xCeNi/SBA-15 oxygen carriers in a fixed-bed reactor. Int J Hydrogen Energy 2016;41:12899e909. [33] Pimenidou P, Rickett G, Dupont V, Twigg MV. Chemical looping reforming of waste cooking oil in packed bed reactor. Bioresour Technol 2010;101:6389e97. [34] Cheng F, Dupont V, Twigg MV. Direct reduction of nickel catalyst with model bio-compounds. Appl Catal B Environ 2017;200:121e32. [35] Yamaguchi T. Application of ZrO2 as a catalyst and a catalyst support. Catal Today 1994;20:199e217. [36] Natesakhawat S, Oktar O, Ozkan US. Effect of lanthanide promotion on catalytic performance of solegel Ni/Al2O3 catalysts in steam reforming of propane. J Mol Catal Chem 2005;241:133e46. [37] Navarro RM, Guil-Lopez R, Ismail AA, Al-Sayari SA, Fierro JLG. Ni- and PtNi-catalysts supported on Al2O3 for

[38]

[39] [40]

[41]

[42]

[43]

[44]

[45]

[46]

[47]

[48]

[49]

[50]

[51]

[52]

acetone steam reforming: effect of the modification of support with Ce, La and Mg. Catal Today 2015;242:60e70. Zhao K, He F, Huang Z, Wei G, Zheng A, Li H, et al. Perovskitetype oxides LaFe1xCoxO3 for chemical looping steam methane reforming to syngas and hydrogen co-production. Appl Energy 2016;168:193e203. Zhu L, Yu J, Wang X. Oxidation treatment of diesel soot particulate on CexZr1xO2. J Hazard Mat 2007;140:205e10. ~ oz M, Moreno S, Molina R. Promoting effect of Ce and Pr Mun in Co catalysts for hydrogen production via oxidative steam reforming of ethanol. Catal Today 2013;213:33e41. Liu F, Chen LY, Neathery JK, Saito K, Liu KL. Cerium oxide promoted iron-based oxygen carrier for chemical looping combustion. Ind Eng Chem Res 2014;53:16341e8. Fornasiero P, Dimonte R, Rao GR, Kaspar J, Meriani S, Trovarelli A, et al. Rh-loaded CeO2-ZrO2 solid-solutions as highly efficient oxygen exchangers: dependence of the reduction behavior and the oxygen storage capacity on the structural-properties. J Catal 1995;151:168e77. K R, SD J, KK N, C R. Effects of lanthanide dopants on oxygen diffusion in yttria-stabilized zirconia. J Am Ceram Soc 2005;88:2143e51. Fermoso J, Gil MV, Rubiera F, Chen D. Multifunctional Pd/NiCo catalyst for hydrogen production by chemical looping coupled with steam reforming of acetic acid. ChemSusChem 2014;7:3063e77. Jiang B, Dou B, Song Y, Zhang C, Du B, Chen H, et al. Hydrogen production from chemical looping steam reforming of glycerol by Ni-based oxygen carrier in a fixedbed reactor. Chem Eng J 2015;280:459e67. lez N, Fierro JLG, Campos CH, Osorio-Vargas P, Flores-Gonza Reyes P. Effect of Ni loading on lanthanide (La and Ce) promoted g-Al2O3 catalysts applied to ethanol steam reforming. Catal Lett 2016;146:433e41. Ma H, Zeng L, Tian H, Li D, Wang X, Li X, et al. Efficient hydrogen production from ethanol steam reforming over Lamodified ordered mesoporous Ni-based catalysts. Appl Catal B Environ 2016;181:321e31. Neal L, Shafiefarhood A, Li F. Effect of core and shell compositions on MeOx@LaySr1yFeO3 coreeshell redox catalysts for chemical looping reforming of methane. Appl Energy 2015;157:391e8. Christensen KO, Chen D, Lødeng R, Holmen A. Effect of supports and Ni crystal size on carbon formation and sintering during steam methane reforming. Appl Catal A Gen 2006;314:9e22. nchez-Sa nchez M, Navarro R, Fierro J. Ethanol steam Sa reforming over Ni/MxOyeAl2O3 (M ¼ Ce, La, Zr and Mg) catalysts: influence of support on the hydrogen production. Int J Hydrogen Energy 2007;32:1462e71. Jiang B, Li L, Bian Z, Li Z, Othman M, Sun Z, et al. Hydrogen generation from chemical looping reforming of glycerol by Ce-doped nickel phyllosilicate nanotube oxygen carriers. Fuel 2018;222:185e92. Omoniyi OA, Dupont V. Chemical looping steam reforming of acetic acid in a packed bed reactor. Appl Catal B Environ 2018;226:258e68.
