Journal of Cleaner Production 200 (2018) 588e597 Contents lists available at ScienceDirect Journal of Cleaner Production journal homepage: www.elsevier.com/locate/jclepro Hydrogen production from vegetable oil via a chemical looping process with hematite oxygen carriers Guo-Qiang Wei a, b, c, d, Wei-Na Zhao e, Jun-Guang Meng b, c, d, Jie Feng a, *, Wen-Ying Li a, **, Fang He b, c, d, Zhen Huang b, c, d, Qun Yi a, Zhen-Yi Du a, Kun Zhao b, c, d, Zeng-Li Zhao b, c, d, Hai-Bin Li b, c, d a Training Base of State Key Laboratory of Coal Science and Technology Jointly Constructed by Shanxi Province and Ministry of Science and Technology, Taiyuan University of Technology, Taiyuan 030024, China Guangzhou Institute of Energy Conversion, Chinese Academy of Sciences (CAS), Guangzhou 510640, China c CAS Key Laboratory of Renewable Energy, Guangzhou 510640, China d Guangdong Provincial Key Laboratory of New and Renewable Energy Research and Development, Guangzhou 510640, China e Guangdong Mechanical & Electrical College, Guangzhou 510515, China b a r t i c l e i n f o a b s t r a c t Article history: Received 12 May 2018 Received in revised form 24 July 2018 Accepted 27 July 2018 Available online 30 July 2018 Hydrogen production from vegetable oil via a chemical looping process with hematite oxygen carriers has been carried out in a ﬁxed-bed reactor at 1023.15e1173.15 K. The lattice oxygen release process and crystalline transformation for the oxygen carrier in the reaction process are investigated. Results indicate that the maximum H2 composition achieved in the chemical looping hydrogen stage is over 91.72% when hematite oxygen carriers consisting of Fe2O3, Al2O3 and SiO2 are used. The active component Fe2O3 in oxygen carriers represents three reduced peaks corresponding to the crystalline form transition from Fe2O3 to Fe3O4. Up to 96.13% of the lattice oxygen in Fe2O3 can be consumed in the reduction process. The deep reduction of FeO to Fe represents a lower reaction rate with the exhaustion of the lattice. The maximum carbon conversion efﬁciency of vegetable oil of 79.10% when the ratio of water to oil is 1.2 and hydrogen-rich gas with a highest concentration of 91.72% are achieved in the chemical looping reforming stage and H2 production stage, respectively. The impurities in the gas are ascribed to the carbon deposition and steam reaction in the chemical looping process. The micrographs of fresh and used oxygen carriers present irregular and blocky structures, sintering is not observed after the multi-cycle reaction process. The crystalline form and reaction activity remain stable after 20-cycle experiments, suggesting that the hematite oxygen carrier is a promising candidate for hydrogen generation from vegetable oil via the chemical looping process. © 2018 Elsevier Ltd. All rights reserved. Keywords: Vegetable oil Oxygen carriers Chemical looping reforming Chemical looping hydrogen High purity H2 production 1. Introduction The output of waste cooking oil (WCO) increases rapidly with the development of the economy and society, which will cause environmental pollution and threaten human physical health if these oils are discharged unreasonably (Mehrasbi et al., 2017). On the other hand, energy crises and emissions are besetting many countries in the world. The WCO containing triglyceride and fatty acid fractions is a promising feedstock alternative to produce energy or chemicals, which can relieve the crisis and emissions to some extent * Corresponding author. ** Corresponding author. E-mail addresses: email@example.com (J. Feng), firstname.lastname@example.org (W.-Y. Li). https://doi.org/10.1016/j.jclepro.2018.07.263 0959-6526/© 2018 Elsevier Ltd. All rights reserved. (Morais et al., 2010; Senthur Prabu et al., 2017). There are various applications for the WCO as a raw material in producing soap, glycerin and synthetic plasticizers (Zheng et al., 2018). Additionally, another application for the WCO is to synthesize biodiesel though esteriﬁcation and transesteriﬁcation (Fereidooni et al., 2018; Stephen and Periyasamy, 2018). This process presents special advantages such as renewability, environmental properties and compatibility with engines and fuel standards (Moecke et al., 2016; Nabi et al., 2017). However, there are technical problems that must be overcome before further industrialization. In view of the water, free fatty acids, diglycerides, triglycerides and other impurities in the WCO, complex preprocessing is needed to meet the requirements of hydrodeoxygenation, decarbonylation, decarboxylation and thermal cracking reactions (Wang et al., 2017b). G.-Q. Wei et al. / Journal of Cleaner Production 200 (2018) 588e597 589 CaHbCO$ ¼ CaHb$ þ CO$ (5) CH$ þ 5H$ ¼ CH4 þ H2 (6) Furthermore, some homogenous catalysts or acidic catalysts must be adopted in the catalytic thermal cracking of WCO and biodiesel production process (Chuah et al., 2017; Tran et al., 2016; Ueki et al., 2018). In addition, the conductivity and thermal efﬁciency of the WCO are insufﬁcient, which may damage the reaction apparatus (7) due to overheating (Li et al., 2017). Improving the thermal conductivity of the deﬁcient WCO becomes another priority before biodiesel production (Borugadda and Goud, 2016). So, some complex equipment and technologies are needed to separate, purify and catalyze WCO raw material in the biodiesel production process (Asikin-Mijan et al., 2017; de Mello et al., 2017; Lani et al., 2017), which increase the cost and technical complexity, as well as the environmental risk (Eguchi et al., 2015). Additionally, the storage stability of biodiesel must also be a focus due to the high amount biodiesel gas blended in the production. It is very necessary to use further amounts of antioxidant additives to minimize biodiesel €lczmann et al., 2016). Consequently, the conversion degradation (Po path of the WCO with low cost, high efﬁciency and low environmental risk needs to be developed as a more promising alternative. In regard to H2 production, there are various research efforts examining steam reforming for fossil fuels, such as petroleum, coal and natural gas, because this is a simple and mature technique that has a better economic beneﬁt (Shahbaz et al., 2017; Zhao et al., 2017). Nevertheless, some disadvantages are also exhibited such as the high cost associated with hydrogen separation, greenhouse gas emissions and non-regeneration. To solve the energy, environmental and sustainable development problems, chemical looping reforming (CLR) coupled with chemical looping hydrogen (CLH) was proposed to produce syngas and high purity hydrogen from WCO at a lower cost with less pollution. This process uses WCO as the feedstock instead of fossil fuels, such as coal, petroleum and CH4, revealing advantages for solving environmental problems and producing clean energy simultaneously. Additionally, it provides a new industrial application of WCO. With the consistently growing global demand for fats and vegetable oils over the past decade, it is evaluated that the worldwide consumption of fats and vegetable oils will increase by 25% and reach 178 Mt in 2025. Approximately 20% of these oils is disposed of after the cooking process, which is the source of waste cooking oil (Alqaralleh et al., 2016; Lombardi et al., 2018). Consequently, the waste cooking oil CLR coupled with CLH process could have great potential in industrial application for syngas and high purity hydrogen production due to its lower cost, less pollution and convenient scale-up. The CLR and CLH processes are performed separately in the integrated system, which is composed of a steam reactor (SR), a fuel reactor (FR) and an air reactor (AR). The CLR process is carried out initially in the FR, where the WCO is pyrolyzed into small molecular intermediates following reactions (1) to (7). CmHnCOOCaHb ¼ CmHnO$ þ CaHbCO$ (1) CaHbCO$ ¼ CaHb-1CO$ þ H$ (2) CaHCO$ ¼ CaH$ þ CO$ (3) CaH$ ¼ Ca-1H$ þ C$ (4) WCO þ MemOn ¼ H2 þ CH4 þ CO þ CO2 þ … þ MemOn-d (8) MemOn-d þ H2O ¼ MemOn-dþε þ H2 (9) MemOn-dþε þ air ¼ MemOn þ N2 (10) Then, the intermediate products further react with oxygen carriers to generate the ﬁnal syngas. Apparently, the oxygen carriers (MemOn) supply active lattice oxygen to promote the macromolecular intermediate cracking to generate syngas and shift the reaction equilibrium to higher carbon conversion. The oxygen carriers are correspondingly reduced to lower valence states, as exhibited in equation (8). Following the above steps, the reduced oxygen carriers are transported into the SR to perform the CLH reaction, accompanied by highly puriﬁed H2 production. The oxygen carriers are oxidized partially to recover lattice oxygen, described by equation (9). To recover lattice oxygen completely after the CLH process, an air oxidized process is implemented in the AR according to equation (10). The oxygen carriers recover their original form, and the reactions continuously proceed with the oxygen carriers circulating in three reactors. Obviously, a key role associated with heat and oxygen transfer in the CLR and CLH processes is assigned to oxygen carriers. The lattice oxygen transport and heat conductivity capacity of the oxygen carries show signiﬁcant effects on the chemical looping process. Thus far, oxygen carriers including metal oxides, perovskite and nonmetallic oxides have been investigated in chemical looping combustion (CLC), CLR, CLH, etc. (Liu et al., 2017; Wei et al., 2017). Though Ni- and Cu-based oxygen carriers indicate better reactivity, Fe-based oxygen carriers are much preferred in the chemical looping process due to their moderate reactivity, higher melting point, non-toxicity and low cost (Huang et al., 2016; Sajen et al., 2016; Wang et al., 2017a). Furthermore, Fe-based oxygen carriers are suitable for the CLR and CLH processes, which has been demonstrated in CH4 reforming and H2 production (Xiao et al., 2014; Zeng et al., 2015). A CLH experiment for H2 generation was performed in a 25 kWth reactor by Fan's group with synthetic Febased oxygen carriers, and more than 99.99% purity H2 with 100% carbon capture was achieved in the reaction process (Fan et al., 2008). Additionally, the mixed oxygen carriers indicate better reactivity, anti-sintering and anti-carbonization due to the metallic synergistic effect, whereas the higher cost restricts further largescale application (Huang et al., 2014). Hematite oxygen carriers are ideal candidates for CLR and CLH processes due to their lower application cost. In addition, hematite displays a high performance in other similar processes such as tar catalytic cracking or bio-oil CLH processes. Therefore, instead of using fossil fuels, such as coal, petroleum and CH4, the chemical looping coupling process was proposed to produce syngas and high purity hydrogen from waste cooking oil by adopting a hematite oxygen carrier with a lower cost and less 590 G.-Q. Wei et al. / Journal of Cleaner Production 200 (2018) 588e597 pollution. 2. Experiment and apparatus 2.1. Materials To exclude interference factors and investigate the reaction mechanism, vegetable oil from the local market in Guangzhou (China) was used as a model of WCO. The proximate and ultimate analysis results of the vegetable oil were as follows (mass fraction): 99.94%volatile matter, 0.06%ash, 76.32%C, 10.91%H, 12.63%O, 0.1%N, and 0.05%S, which were based on a dry basis. The average molecular formula can be calculated as CH1.22O0.12 according to the analysis results, ignoring N and S. The lower heating value of the vegetable oil was 38.59 MJ/(kg, db). Natural hematite samples from China were ﬁrst pulverized and screened to 0.18e0.25 mm particles, which were next calcined at 950 C in a mufﬂe furnace for 4 h to obtain the oxygen carriers for testing in the chemical process. The mass composition of different elements in hematite oxygen carrier analyzed by X-ray ﬂuorescence (XRF) were shown to be as follows: 62.78%Fe, 32.27%O, 2.83% Si, 1.49%Al, 0.18%Mg, 0.1%P, 0.17%Ca, 0.07%Zn, 0.04%Na, 0.04%Mn, 0.03%Ti, 0.02%K, 0.02%Ni, etc. 2.2. Method and apparatus Thermogravimetric analysis (TGA) measurements (STA409C/PC, NETZSCH) combining with temperature-programed reduction (TPR) analysis were used to investigate the reaction activity and transfer rate of active lattice oxygen in hematite particles in a H2 atmosphere. The hematite particles were ﬁrst ground into powder, and then, the TGA test was performed with a 25 mg hematite powder sample over the temperature range of 303.15e1273.15 K. The heating rate of the reaction zone was kept at 20 K/min, and the reducing gas mixture with volume fraction of 5%H2 and 95%N2 was introduced at a ﬂow rate of 100 mL/min. In addition, 30 mL/min of N2 was used as a protective gas. To further study the reaction mechanism of the CLR coupled with CLH process of vegetable oil and oxygen carriers, ﬁxed-bed reactor experiments were carried out after the TGA test. Before the experiments, the design and selection of experimental parameters needed to be performed. Temperature plays a key role in improving the reaction activity and rate of oxygen carriers for the chemical looping process. Importantly, the oxygen carriers are prone to sintering at high temperature. Additionally, high temperature produces a negative effect on the CLH reaction due its exothermic characteristic according to thermodynamic analysis. Therefore, a temperature range of 1023.15e1173.15 K is adopted in this experiment to achieve a good fuel conversion rate while avoiding sintering. In addition, reaction time is another important factor affecting fuel conversion in the chemical looping process. It is obvious that a longer reaction time can promote fuel conversion efﬁciency; however, the reaction rate of oxygen carriers will decrease with the lattice oxygen consistently being consumed. Based on our previous research, the reaction time of 46 min is suitable for the CLR or CLH process with hematite oxygen carriers in a ﬁxed-bed reactor. A schematic drawing of the experiment is displayed in Fig. S1 in the supporting information. It consisted of a ﬁxed-bed reactor, a gas condensing and cleaning unit, a gas analysis unit and a computer control unit. The ﬁxed-bed reactor included an electric furnace, a quartz tube with a porous distributor and a thermocouple. The reaction temperature and gas ﬂow rate were controlled by thermocouple and mass ﬂow controllers, respectively. Two liquid chromatography (HPLC) micropumps were adopted to inject deionized water and vegetable oil, separately. The experimental procedure is shown in the ﬂowchart of Fig. 1. Before the experiments, a moderate amount of silica wool was spread out on the distributor of the quartz tube, and then, 2 g of oxygen carriers were put into the quartz tube. The reaction temperature was set at 1023.15e1173.15 K, and N2 was used to replace the air atmosphere at a ﬂow rate of 100 mL/min. The oil was injected at a ﬂow rate of 0.1 mL/min, and deionized water was introduced at a variable ﬂow rate. Liquid production collected from the condensing unit was small and therefore ignored. The reduced oxygen carriers were recovered in the original crystal form by introducing an air atmosphere after the CLR and CLH reactions. Finally, the production gas was puriﬁed and dried with isopropanol and an allochroic silica gel desiccant. An Agilent 7890A gas chromatograph was used to analyze the composition of gas production. 2.3. Characterization of oxygen carriers X-ray diffraction (XRD, X'Pert PRO MPD) with Cu Ka radiation (40 kV, 40 mA) was used to characterize the crystallinity of the oxygen carriers in the range of 10e80 . The scan rate and step size were kept at 2 /min and 0.0167, respectively. To analyze the elemental composition of samples, X-ray ﬂuorescence (XRF, AXIOSMAXPETRO) was employed. Additionally, the evolution of the microstructure and BET surface area for N2 physisorption were investigated by ﬁeld emission scanning electron microscopy (Hitachi S4800) and a Micromeritics ASAP 2010, respectively. Before the experiments, the samples were kept at 573 K under a vacuum condition for 3 h to eliminate the air in the oxygen carrier particles. In addition, an oxygen carrier sample of 35 mg was analyzed by H2 temperatureprogrammed reduction (TPR, CPB-1, Quantachrome) to investigate the reactivity and lattice oxygen release performance of the hematite oxygen carrier. The ﬂow rate of the gas mixture (5%H2þ95%N2) was 120 mL/min, and the heating rate was maintained at 10 K/min. 2.4. H2 production evaluation Based on N2 balance, the gas relative compositions (Ci) of produced syngas and H2 were calculated with Eq. (11). Z ci ¼ Z 0 t 0 t xi dt v xCO þ xCO2 þ xH2 þ xCH4 þ xC2 H4 þ xC2 H6 dt (11) where xi denotes the volume fraction of species i, i indicates the compositions and v is the volume ﬂow of syngas. The carbon conversion efﬁciency (hc) in the system was deﬁned as the ratio of the carbon element converted into gaseous products from the oil fed into the ﬁxed-bed reactor. It was calculated by Eq. (12): 12 Vco2 þ Vco þ VCH4 þ 2VC2 H4 þ 2VC2 H6 GV 100% hc ¼ 22:4 ð303=273Þ Mc (12) where Gv (m3/kg) and Mc (%) are the gas yield and carbon fraction in oil, respectively. Vco, Vco2, VCH4, VC2H4 and VC2H6 (%) represent the volume compositions of CO, CO2, CH4, C2H4 and C2H6 in the outlet gas of the ﬁxed-bed reactor. The lower heating value (LHV, kJ/Nm3) of the gas products was calculated with Eq. (13): LHV ¼ 126VCO þ 108VH2 þ 359VCH4 þ 635VC2 Hm (13) where VCO, VH2, VCH4 and VC2Hm are the volume fractions of CO, H2, G.-Q. Wei et al. / Journal of Cleaner Production 200 (2018) 588e597 591 Fig. 1. Flowchart of the CLR and CLH experimental procedures. CH4 and C2Hm in the ﬂue gas, respectively. 3. Results and discussion 3.1. Characterization of hematite oxygen carriers XRD analysis was used to characterize the crystalline form of fresh oxygen carriers, as shown in Fig. 2. Obviously, the main components of the hematite oxygen carriers were Fe2O3, Al2O3 and Fig. 2. XRD analysis of the fresh oxygen carriers. SiO2, which were veriﬁed with the Joint Committee on Powder Diffraction Standards (JCPDS) card. It is apparent that Fe2O3 in oxygen carriers can supply lattice oxygen in the CLR and CLH processes. To avoid sintering or aggregation in the redox reactions, SiO2 and Al2O3 played a positive role in the oxygen carriers (Bao et al., 2016; Zhu et al., 2017). In addition, some impurities (MgO, CaO, etc.) in the hematite oxygen carriers were detected at the same time, which were too small to be marked on the pattern. TPR analysis of oxygen carriers was performed at 303.15e1373.15 K. The results are displayed in Fig. 3. There were Fig. 3. TPR analysis of the oxygen carriers. 592 G.-Q. Wei et al. / Journal of Cleaner Production 200 (2018) 588e597 three reduction peaks in the TPR proﬁle at 671.68, 770.18 and 1154.52 K. Based on the XRD analysis, it was found that Fe2O3 was the primary active component in hematite oxygen carriers. Correspondingly, the reduction peak at 671.68 K was ascribed to the crystal form transition from Fe3þ (Fe2O3) to Fe3þ$Fe2þ (Fe3O4) in the oxygen carriers, and the adsorbed oxygen was consumed by the reaction as follows (Meshkani and Rezaei, 2015): Fe2O3 þ H2 ¼ 2FeO þ H2O (14) 2H$ þ [O$] ¼ H2O (15) Moreover, it was observed the transition peak at 770.18 K corresponded to the reaction of Fe3O4 to FeO (Fe3O4þH2 ¼ 3FeO þ H2O), and the reduction peak at 1154.52 K, which was assigned to the generation of metallic iron, followed reaction (16). FeO þ H2 ¼ Fe þ H2O (16) TPR analysis showed that the lattice oxygen and adsorbed oxygen in hematite samples exhibited better reactivity with H2. To further verify the reaction rate and limitation of the hematite sample, a TG-DTG (thermogravimetric analysis-differential thermal gravity) analysis was carried out from 303.15 to 1223.15 K with the mixture atmosphere of 5%H2 and 95%N2 (volume fraction). As displayed in Fig. 4, the total weight loss of hematite oxygen carriers reached 25.86%, which is ascribed to the reaction of active lattice oxygen [O] and hydrogen (Reaction 15). Since the total oxygen content in hematite oxygen carrier was 32.27% according to the XRF data, the weight loss analysis indicated that 80.14% of oxygen in the oxygen carriers was involved in the reduction reaction. Moreover, 96.13% active lattice oxygen in Fe2O3 was consumed in the reduction process, which suggested a larger active lattice oxygen capacity and better reactivity for the hematite oxygen carriers. Correspondingly, the DTG proﬁle was depicted as a function of reaction time. Three reduction peaks were observed in the entire reaction process at 673.15, 773.15 and 1123.15 K, which were assigned to the reaction stages of Fe2O3 / Fe3O4, Fe3O4 / FeO and FeO / Fe. At the beginning of reduction, the maximum weight loss rate of Fe2O3 / Fe3O4 was 0.0043%/s, indicating lower lattice oxygen consumption rate. As the reaction proceeded, the conversion rate of Fe3O4 increased rapidly to 0.0142%/s due to the FeO phase generation. Finally, the deep reduction of FeO / Fe presented a lower reaction rate with the exhaustion of the lattice. Fig. 4. TG-DTG analysis of the oxygen carriers. 3.2. Fixed-bed reactor investigation on the CLR coupled with CLH process 3.2.1. Gas composition evolution as a function of temperature in the CLR process The syngas composition evolution as a function of temperature is shown in Fig. 5. It was observed that the H2 component increased with rising temperature, while the CH4, C2H4, C2H6 and CO2 displayed a decreasing tendency in the researched temperature range (1023.15e1173.15 K). In particular, the CO concentration ﬁrst increased and then decreased slightly. There were some factors that can explain these phenomena. First, the high temperature was beneﬁcial to the pyrolysis reaction of vegetable oil, as shown in equations (1)e(7). This meant that more carbon/hydrogen free radicals (such as CaHb-1CO$, H$, CO$ and CaH$) were released during vegetable oil decomposition to generate intermediate pyrolysis gases (Chang et al., 2017). In this stage, the content of low molecular weight gas (such as H2, CO, CH4 and C2H4) increased to promote the CLR reaction with oxygen carriers (equation (8)). Second, the high temperature promoted the endothermic reactions in the CLR process, listed in equations (17), (20)e(22), (24) and (26), which consumed H2, CO2, CH4, CO, C2H4 and C2H6 to generate more H2 and CO. These reactions contributed to the increase of H2 and CO concentrations in the syngas. Meanwhile, other gas components decreased with the consumption in the CLR process. In addition, the high temperature depressed the exothermic reactions (equations (18), (23) and (25)) in the CLR process, which reduced the amount of CO generated. Therefore, as these reactions went to equilibrium, CO displayed an upward tendency and then tended to be stable. By contrast, the H2 composition increased with rising temperature, while that of the other gases decreased gradually. 3H2 þ Fe2O3 ¼ 2Fe þ 3H2O △H1173.15K ¼ 61.79 kJ (17) 3CH4 þ Fe2O3 ¼ 2Fe þ3CO þ 6H2 △H1173.15K ¼ 738.88 kJ (18) CO2 þ C ¼ 2CO △H1173.15K ¼ 550.42 kJ (19) CO2 þ CH4 ¼ 2CO þ2 H2 △H1173.15K ¼ 258.82 kJ (20) 1.5C2H4 þ Fe2O3 ¼ 2Fe þ 3 CO þ 3H2 △H1173.15K ¼ 413.68 kJ (21) C2H6 þ Fe2O3 ¼ 2Fe þ 3CO þ 4.5H2 △H1173.15K ¼ 629.71 kJ (22) Fig. 5. Syngas composition evolution in the CLR process of vegetable oil. G.-Q. Wei et al. / Journal of Cleaner Production 200 (2018) 588e597 CO þ H2O ¼ CO2 þ H2 △H1173.15K ¼ 33.13 kJ (23) C þ H2O ¼ CO þ H2 △H1173.15K ¼ 135.71 kJ (24) 3CO þ Fe2O3 ¼ 2Fe þ 3CO2 △H1173.15K ¼ 37.61 kJ (25) CH4 þ 2H2O ¼ CO2 þ 4H2 △H1173.15K ¼ 192.56 kJ (26) Correspondingly, the carbon conversion efﬁciency of vegetable oil in the CLR process under different temperatures is listed in Table 1. It was found that a high temperature can obviously enhance the carbon conversion in the CLR process based on the endothermic reactions between lattice oxygen and the pyrolysis products from vegetable oil, promoting reaction equilibrium toward the production direction. It seemed that a high temperature beneﬁted the CLR process, while the maximum carbon conversion efﬁciency of 57.04% was achieved at 1173.15 K, rather than 1223.15 K. This was ascribed to carbon deposition in the reaction process. Theoretically, with an increase of the temperature, the vegetable oil was prone to cracking on the surface of oxygen carriers to generate carbon deposition with the catalytic effect of the oxygen carriers at the reduced state (Wang et al., 2017b). The carbon deposition increased sharply at 1223.15 K, leading to a decrease in carbon conversion efﬁciency in the gas production. Meanwhile, exothermic reaction (19), which can consume the carbon deposition, was depressed at high temperature. Additionally, the dimerization reaction of free radical to oil promoted a decrease in the carbon conversion efﬁciency at high temperature. 3.2.2. Effect of water/oil on gas composition in the CLR process To improve the composition of the production gas, deionized water was introduced into the CLR reaction system at 1173.15 K. The effect of water/oil ratio (W/O, volume) on the gas composition is shown in Fig. 6. As indicated in the curves, the H2 and CO compositions in the syngas increased obviously with the increasing of water/oil ratio, whereas the contents of CH4, C2H4 and CO2 declined steadily in the reaction process. This was ascribed to the water-gas shift reaction, the water gas reaction and the steam reforming reaction listed in equations (23), (24) and (26). These reactions were enhanced to generate CO and H2 with the introduced steam. Moreover, the produced CO2 was further consumed by the Boudouard reaction (19), leading to a gradual decline of the CO2 concentration. Additionally, the introduced water contributed to the declining concentration of C2H4 and C2H6 via reactions (27) and (28), which produced more CO and H2. C2H4 þ 2H2O ¼ 2CO þ 4H2 △H1173.15K ¼ 56.07 kJ (27) C2H6 þ 2H2O ¼ 2CO þ 5H2 △H1173.15K ¼ 90.49 kJ (28) As shown in Table 2, the carbon conversion efﬁciency increased promptly with increasing steam-to-oil ratio in the CLR process. The maximum carbon conversion efﬁciency of 79.10% was achieved at a W/O of 1.2. It was obvious that steam contributed to the carbon conversion of vegetable oil in the CLR process. These reactions Table 1 Carbon conversion efﬁciency as a function of temperature. Temperature/K Carbon conversion/% 1023.15 1073.15 1123.15 1173.15 1223.15 30.88 53.12 54.80 57.04 52.81 593 Fig. 6. Syngas composition under different W/O ratios in the CLR process. Table 2 Carbon conversion efﬁciency under different W/O ratios. W/O ratio/mL/mL Carbon conversion/% 0.1 0.2 0.5 0.7 1.0 1.2 50.22 52.26 56.21 64.31 76.71 79.10 promoted the carbon conversion in the CLR process by enhancing the generation of CH4, C2H4, CO and CO2, similar to reactions (23), (24), (26), (27) and (28). H2O was helpful in the CLR processing of vegetable oil at a lower steam-to-oil ratio, while it showed negative effects on the following CLH process. This was ascribed to partial oxidization of oxygen carriers by the steam, which decreased the production of H2 in the next reaction stage. The reactions for the H2 production are represented in equations (29) and (30). 3Fe þ 4H2O ¼ Fe3O4 þ 4H2 △H1173.15K ¼ 96.64 kJ (29) 3FeO þ H2O ¼ Fe3O4 þ H2 △H1173.15K ¼ 46.11 kJ (30) 3.2.3. H2 generation as a function of reaction temperature in the CLH process Following the CLR process of vegetable oil, the CLH stage for H2 production was investigated by splitting deionized water with oxygen carriers in a ﬁxed-bed reactor. Fig. 7 reveals the H2 production as a function of reaction temperature. It was observed that the H2 production increased obviously from 1023.15 to 1173.15 K and then decreased at 1223.15 K. Accordingly, the composition of the gas production decreased from 91.72% to 89.05%. The highest H2 concentration of 91.72% was achieved at 1023.15 K, and the maximum yield of 501.06 mL/g H2-rich gas with 89.89% H2 concentration was obtained at 1173.15 K, corresponding to a 450.41 mL/g pure H2 yield. As described in equations (29) and (30), the H2 was generated from the reactions between reduced oxygen carriers and H2O. Moreover, the impurity gas was ascribed to the carbon deposited in the CLR stage of vegetable oil processing and subsequent reactions between H2O and deposited carbon in the CLH stage. Speciﬁcally, the carbonaceous gas was generated from reactions (24) and (31). These endothermic reactions were enhanced to produce more carbonaceous gas with increasing temperature, resulting in the 594 G.-Q. Wei et al. / Journal of Cleaner Production 200 (2018) 588e597 value-added proposition for the CLH process. 3.3. Successive cycle tests for the CLR coupled with CLH process Fig. 7. H2 production and composition as a function of reaction temperature. To investigate the multi-cycle reaction performance of hematite oxygen carriers, the CLR, CLH and air oxidation three stage reactions were alternatively performed in the ﬁxed-bed reactor at 1173.15 K. The carbon conversion efﬁciency hc in the CLR stage, H2 concentration in the CLH stage and O2 composition in the air oxidization stage are exhibited in Table 3. It was observed that the carbon conversion of vegetable oil in the CLR stage ﬂuctuated in the range of 55.14%e60.51%, indicating a stable lattice oxygen output of the oxygen carriers after a 20-cycle reaction. Accordingly, the H2 production concentration varied from 89.93% to 92.44% in the multi-cycle reaction process. To recover the lattice oxygen of oxygen carriers, the air oxidization stage was implemented after the CLH process, and the oxygen concentration was stable in the scope of 13.78%e14.68%. The reaction between oxygen carriers and O2 can be represented as reaction (32). 4Fe3O4 þ O2 ¼ 6Fe2O3 △H1173.15K ¼ 102.57 kJ decline of the H2 gas concentration. C þ 2H2O ¼ CO2 þ 2H2 △H1173.15K ¼ 102.57 kJ (31) Though these side reactions produced undesirable carbonaceous gas, they also generated H2, suggesting partial advantageous factors for H2 production. To increase the purity of H2, the carbon deposition needs to be minimized. There are some strategies to improve the carbon deposition in the vegetable oil reforming stage. On the one hand, it is better to perform the experiments in a circulating ﬂuidized bed, which can restrain the carbon deposition increase via intense collisions between OC particles (Wei et al., 2015). A challenge involved in this experiment is the energy penalty and technical complexity. On the other hand, some steam is beneﬁcial to increasing the carbon conversion efﬁciency and reducing the carbon deposition in the CLR stage. Nevertheless, the introduction of H2O has a negative impact on the reduced oxygen carriers, which leads to the decline of H2 production in the following H2 production stage. Finally, it seems that an effective method is to design and develop anti-carbon oxygen carriers, which can solve the carbon deposition fundamentally (Bimbela et al., 2017). In addition, although small amounts of carbonaceous gas impurities are mixed in the production, a higher H2 volume concentration of 91.7% is available from the chemical looping process than that from other H2 production processes, such as biomass steam gasiﬁcation (40.8%) (Xiao et al., 2017), catalytic steam reforming of bio-oil (70.8%) (Bimbela et al., 2017) and supercritical water gasiﬁcation of coal (less than 60%) (Jin et al., 2015), indicating a higher (32) Moreover, carbonaceous gas was still observed in the outlet gas, suggesting that the carbon deposition was not eliminated completely after the CLH stage. On the whole, the 20-cycle reactions displayed a better cyclic reaction performance of the hematite oxygen carriers. 3.4. XRD pattern of fresh and regenerated oxygen carriers To further investigate the crystalline form evolution of oxygen carriers in the reaction process, XRD analysis of oxygen carriers at different reaction cycles are exhibited in Fig. 8. As shown in the XRD patterns, the fresh hematite oxygen carriers were mainly composed of Fe2O3, Al2O3 and SiO2. After ﬁve multi-cycle reactions, the diffraction peak of SiO2 was minimized, while the Fe2O3 characteristic peaks were consistent with those of the fresh oxygen carriers, indicating that the lattice oxygen carriers were fully recovered. The minimization of the SiO2 peak was ascribed to mass loss in the redox cycle consumption. With the cyclic reaction running, the Fe2O3 characteristic peaks remained in line with those of the fresh oxygen carriers after 20 cyclic reactions, which displayed the stable cyclic reaction performance of hematite oxygen carriers. 3.5. Morphology changes of hematite oxygen carriers The micrograph of fresh and multi-cycle reaction hematite oxygen carriers is illustrated in Fig. 9. An irregular blocky structure Table 3 Results of successive cycle tests for treatment of vegetable oil by the coupled CLR and CLH process. Cycle 1 2 3 5 7 9 11 15 20 CLR stage CLH stage Air oxidization stage Carbon conversion % Reforming gas heating value MJ/m3 H2 concentration % O2 composition % Carbonaceous gas % 57.04 55.14 54.99 59.83 58.28 59.38 59.95 56.58 60.51 26.70 26.17 28.57 29.49 26.66 30.45 28.93 28.56 27.81 89.93 92.63 91.10 90.60 90.80 91.16 91.44 92.40 90.58 14.68 14.18 14.04 13.78 13.44 14.13 14.43 14.24 13.47 1.10 1.93 1.30 1.57 1.90 1.44 0.89 1.16 1.32 G.-Q. Wei et al. / Journal of Cleaner Production 200 (2018) 588e597 595 occurred. Correspondingly, the porous gaps evolved into porous channels. In the subsequent cyclic reaction processes, the particle size decreased, and agglomeration phenomena increased. Additionally, the porous channel structure was maintained after 20 cyclic reactions. Apparent sintering did not occur in regenerated hematite oxygen carriers. Additionally, the crystalline form and reaction activity remained stable according to experimental results given in Fig. 8 and Table 3, suggesting that the hematite oxygen carrier had good potential for the CLR coupled with CLH process. 4. Conclusions Fig. 8. XRD patterns of the oxygen carriers at different reaction cycles. with particle sizes of 5e10 mm and porous gaps was distributed in the fresh oxygen carriers. Obviously, the porous structure was beneﬁcial to reactant gas diffusion into oxygen carriers, enhancing the reaction process (Das et al., 2018). After ﬁve cycles, the oxygen carriers' particle sizes reduced to 2e3 mm, and agglomeration A chemical looping coupling process is proposed and proven to produce syngas and high purity hydrogen from vegetable oil instead of fossil fuels by adopting hematite oxygen carriers with a lower cost and less pollution. In the reduction process of hematite oxygen carries, the reaction follows the path of Fe3þ / Fe3þ$Fe2þ / Fe2þ / Fe0, achieving the maximum conversion rate in the second reaction stage. High temperature and water have a positive effect on the CLR process, while excessive steam depresses deep reduction of oxygen carriers in the CLR process, decreasing H2 production in the subsequent CLH stage. The optimized reaction temperature for the CLR stage is 1173.15 K, corresponding to 57.04% carbon conversion efﬁciency for vegetable oil. The maximum yield of 501.06 mL/g for H2-rich gas with an 89.89% H2 concentration is obtained at 1173.15 K in the CLH stage, corresponding to a 450.41 mL/g pure H2 yield. The concentration of H2 production Fig. 9. SEM analysis of the oxygen carriers for multi-cycle tests: (a) fresh oxygen carriers; (b) 5 cycles; (c) 10 cycles; and (d) 20 cycles. 596 G.-Q. Wei et al. / Journal of Cleaner Production 200 (2018) 588e597 achieved from the CLH process varies from 89.93% to 92.44% in a multi-cycle reaction process. Acknowledgements The authors gratefully acknowledge the ﬁnancial support of the National Key Research and Development Program of China (Grant No. 2016YFB0901401), the National Natural Science Foundation of China (Grant No.: U1610221; 51776133), the Science and Technology Projects of Guangdong (Grant No.: 2015A020215023), Science and Technology Projects of Guangzhou (Grant No.: 201707010202), Guangdong Provincial Key Laboratory of New and Renewable Energy Research and Development (Grant No.: Y709jj1001), and Training base of State Key Laboratory of Coal Science and Technology Jointly Constructed by Shanxi Province (Grant No.: MKX201701) and Ministry of Science and Technology, Taiyuan University of Technology (Grant No.: MKX201701). Notation list AR C2H4 C2H6 CaO CH4 CLC CLH CLR CO CO2 db DTG Fe2Fe2O3 Fe3þ Fe3O4 FR H H2 HPLC LHV MemOn MgO O O2SR TG TPR W/O WCO XRD XRF hc air reactor ethylene ethane calcium oxide methane chemical looping combustion chemical looping hydrogen chemical looping reforming carbon monoxide carbon dioxide dry basis differential thermal gravity Ferrous ion iron oxide Ferric ion ferroferric oxide fuel reactor hydrogen radical hydrogen high performance liquid chromatography lower heating value metal oxide (oxygen carrier) magnesium oxide oxygen radical oxygen ion steam reactor thermo gravimetric H2 temperature programmed reduction water/oil ratio waste cooking oil X-ray diffraction X-ray ﬂuorescence carbon conversion efﬁciency Appendix A. Supplementary data Supplementary data related to this article can be found at https://doi.org/10.1016/j.jclepro.2018.07.263. References Alqaralleh, R.M., Kennedy, K., Delatolla, R., Sartaj, M., 2016. Thermophilic and hyperthermophilic co-digestion of waste activated sludge and fat, oil and grease: evaluating and modeling methane production. J. Environ. Manag. 183 (3), 551e561. Asikin-Mijan, N., Lee, H.V., Abdulkareem-Alsultan, G., Afandi, A., Tauﬁq-Yap, Y.H., 2017. Production of green diesel via cleaner catalytic deoxygenation of Jatropha curcas oil. J. Clean. Prod. 167, 1048e1059. Bao, J.H., Chen, L.Y., Liu, F., Fan, Z., Nikolic, H.S., Liu, K.L., 2016. Evaluating the effect of inert supports and alkali sodium on the performance of red mud oxygen carrier in chemical looping combustion. Ind. Eng. Chem. Res. 55, 8046e8057. Bimbela, F., Abrego, J., Puerta, R., García, L., Arauzo, J., 2017. Catalytic steam reforming of the aqueous fraction of bio-oil using Ni-Ce/Mg-Al catalysts. Appl. Catal. B 209, 346e357. Borugadda, V.B., Goud, V.V., 2016. Improved thermo-oxidative stability of structurally modiﬁed waste cooking oil methyl esters for bio-lubricant application. J. Clean. Prod. 112, 4515e4524. Chang, J.S., Cheng, J.C., Ling, T.R., Chern, J.M., Wang, G.B., Chou, T.C., Kuo, C.T., 2017. Low acid value bio-gasoline and bio-diesel made from waste cooking oils using a fast pyrolysis process. J. Taiwan Inst. Chem. Eng. 73, 1e11. Chuah, L.F., Klemes, J.J., Yusup, S., Bokhari, A., Akbar, M.M., 2017. A review of cleaner intensiﬁcation technologies in biodiesel production. J. Clean. Prod. 146, 181e193. Das, S., Deen, N.G., Kuipers, J.A.M., 2018. Multiscale modeling of ﬁxed-bed reactors with porous (open-cell foam) non-spherical particles: hydrodynamics. Chem. Eng. J. 334, 741e759. de Mello, M., Young, A., Villardi, H., Pessoa, F., Salgado, A., 2017. Biodiesel production by the methylic-alkaline and ethylic-enzymatic routes: discussion of some environmental aspects. J. Clean. Prod. 144, 347e357. Eguchi, S., Kagawa, S., Okamoto, S., 2015. Environmental and economic performance of a biodiesel plant using waste cooking oil. J. Clean. Prod. 101, 245e250. Fan, L., Li, F., Ramkumar, S., 2008. Utilization of chemical looping strategy in coal gasiﬁcation processes. Particuology 6, 131e142. Fereidooni, L., Tahvildari, K., Mehrpooya, M., 2018. Trans-esteriﬁcation of waste cooking oil with methanol by electrolysis process using KOH. Renew. Energy 116, 183e193. Huang, Z., He, F., Feng, Y.P., Zhao, K., Zheng, A.Q., Chang, S., Wei, G.Q., Zhao, Z.L., Li, H.B., 2014. Biomass char direct chemical looping gasiﬁcation using NiOmodiﬁed iron ore as an oxygen carrier. Energy Fuels 28, 183e191. Huang, Z., Zhang, Y., Fu, J.J., Yu, L.H., Chen, M., Liu, S., He, F., Chen, D.Z., Wei, G.Q., Zhao, K., Zheng, A.Q., Zhao, Z.L., Li, H.B., 2016. Chemical looping gasiﬁcation of biomass char using iron ore as an oxygen carrier. Int. J. Hydrogen Energy 41, 17871e17883. Jin, H., Chen, Y.N., Ge, Z.W., Liu, S.K., Ren, C.S., Guo, L.J., 2015. Hydrogen production by Zhundong coal gasiﬁcation in supercritical water. Int. J. Hydrogen Energy 40, 16096e16103. Lani, N.S., Ngadi, N., Yahya, N.Y., Rahman, R.A., 2017. Synthesis, characterization and performance of silica impregnated calcium oxide as heterogeneous catalyst in biodiesel production. J. Clean. Prod. 146, 116e124. Li, W.J., Zou, C.J., Li, X.K., 2017. Thermo-physical properties of waste cooking oilbased nanoﬂuids. Appl. Therm. Eng. 112, 784e792. Liu, S., Xiang, D., Xu, Y., Sun, Z., Cao, Y., 2017. Relationship between electronic properties of Fe3O4 substituted by Ca and Ba and their reactivity in chemical looping process: a ﬁrst-principles study. Appl. Energy 202, 550e557. Lombardi, L., Mendecka, B., Carnevale, E., 2018. Comparative life cycle assessment of alternative strategies for energy recovery from used cooking oil. J. Environ. Manag. 216, 235e245. Mehrasbi, M.R., Mohammadi, J., Peyda, M., Mohammadi, M., 2017. Covalent immobilization of candida Antarctica lipase on core-shell magnetic nanoparticles for production of biodiesel from waste cooking oil. Renew. Energy 101, 593e602. Meshkani, F., Rezaei, M., 2015. Preparation of nanocrystalline metal (Cr, Al, Mn, Ce, Ni, Co and Cu) modiﬁed ferrite catalysts for the high temperature water gas shift reaction. Renew. Energy 74, 588e598. Moecke, E.H.S., Feller, R., Santos, H.A.d., Machado, M.d.M., Cubas, A.L.V., Dutra, A.R.d.A., Santos, L.L.V., Soares, S.R., 2016. Biodiesel production from waste cooking oil for use as fuel in artisanal ﬁshing boats: integrating environmental, economic and social aspects. J. Clean. Prod. 135, 679e688. Morais, S., Mata, T.M., Martins, A.A., Pinto, G.A., Costa, C.A.V., 2010. Simulation and life cycle assessment of process design alternatives for biodiesel production from waste vegetable oils. J. Clean. Prod. 18, 1251e1259. Nabi, M.N., Zare, A., Hossain, F.M., Ristovski, Z.D., Brown, R.J., 2017. Reductions in diesel emissions including PM and PN emissions with diesel-biodiesel blends. J. Clean. Prod. 166, 860e868. Hancso €lczmann, G., To th, O., Beck, A., k, J., 2016. Investigation of storage stability Po of diesel fuels containing biodiesel produced from waste cooking oil. J. Clean. Prod. 111, 85e92. Sajen, S., Singh, S.K., Mungse, P., Rayalu, S., Watanabe, K., Saravanan, G., Labhasetwar, N., 2016. Mechanically stable mixed metal oxide of Cu and Mn as oxygen carrier for chemical looping syngas combustion. Energy Fuels 30, 7596e7603. Senthur Prabu, S., Asokan, M.A., Roy, R., Francis, S., Sreelekh, M.K., 2017. Performance, combustion and emission characteristics of diesel engine fuelled with waste cooking oil bio-diesel/diesel blends with additives. Energy 122, 638e648. Shahbaz, M., Yusup, S., Inayat, A., Patrick, D.O., Pratama, A., Ammar, M., 2017. Optimization of hydrogen and syngas production from PKS gasiﬁcation by using coal bottom ash. Bioresour. Technol. 241, 284e295. Stephen, J.L., Periyasamy, B., 2018. Innovative developments in biofuels production from organic waste materials: a review. Fuel 214, 623e633. Tran, T.T.V., Kaiprommarat, S., Kongparakul, S., Reubroycharoen, P., Guan, G., Nguyen, M.H., Samart, C., 2016. Green biodiesel production from waste cooking G.-Q. Wei et al. / Journal of Cleaner Production 200 (2018) 588e597 oil using an environmentally benign acid catalyst. Waste Manag. 52, 367e374. Ueki, Y., Saiki, S., Hoshina, H., Seko, N., 2018. Biodiesel fuel production from waste cooking oil using radiation-grafted ﬁbrous catalysts. Radiat. Phys. Chem. 143, 41e46. Wang, B.W., Li, J., Ding, N., Mei, D.F., Zhao, H.B., Zheng, C.G., 2017a. Chemical looping combustion of a typical lignite with a CaSO4eCuO mixed oxygen carrier. Energy Fuels 31, 13942e13954. Wang, J., Zhong, Z.P., Zhang, B., Ding, K., Xue, Z.Y., Deng, A.D., Ruan, R., 2017b. Upgraded bio-oil production via catalytic fast co-pyrolysis of waste cooking oil and tea residual. Waste Manag. 60, 357e362. Wei, G.Q., He, F., Huang, Z., Zheng, A.Q., Zhao, K., Li, H.B., 2015. Continuous operation of a 10 kW(th) chemical looping integrated ﬂuidized bed reactor for gasifying biomass using an iron-based oxygen carrier. Energy Fuels 29, 233e241. Wei, G.Q., He, F., Zhao, W.N., Huang, Z., Zhao, K., Zhao, Z.L., Zheng, A.Q., Wu, X.S., Li, H.B., 2017. Experimental investigation of FeeNieAl oxygen carrier derived from hydrotalcite-like precursors for the chemical looping gasiﬁcation of biomass char. Energy Fuels 31, 5174e5182. Xiao, R., Zhang, S., Peng, S.H., Shen, D.K., Liu, K.L., 2014. Use of heavy fraction of bio- 597 oil as fuel for hydrogen production in iron-based chemical looping process. Int. J. Hydrogen Energy 39, 19955e19969. Xiao, Y.H., Xu, S.P., Song, Y.B., Shan, Y.Y., Wang, C., Wang, G.D., 2017. Biomass steam gasiﬁcation for hydrogen-rich gas production in a decoupled dual loop gasiﬁcation system. Fuel Process. Technol. 165, 54e61. Zeng, D.W., Xiao, R., Zhang, S., Zhang, H.Y., 2015. Bio-oil heavy fraction for hydrogen production by iron-based oxygen carrier redox cycle. Fuel Process. Technol. 139, 1e7. Zhao, K., Zheng, A.Q., Li, H.B., He, F., Huang, Z., Wei, G.Q., Shen, Y., Zhao, Z.L., 2017. Exploration of the mechanism of chemical looping steam methane reforming using double perovskite-type oxides La1.6Sr0.4FeCoO6. Appl. Catal. B 219, 672e682. Zheng, T., Wu, Z.Y., Xie, Q.L., Fang, J.J., Hu, Y.C., Lu, M.Z., Xia, F., Nie, Y., Ji, J.B., 2018. Structural modiﬁcation of waste cooking oil methyl esters as cleaner plasticizer to substitute toxic dioctyl phthalate. J. Clean. Prod. 186, 1021e1030. Zhu, M., Chen, S.Y., Ma, S.W., Xiang, W.G., 2017. Carbon formation on iron-based oxygen carriers during CH4 reduction period in chemical looping hydrogen generation process. Chem. Eng. J. 325, 322e331.