Comparison of the Reactive Adsorption Desulfurization Performance ...

30 downloads 9005 Views 1MB Size Report
Mar 22, 2016 - School of Chemical and Biomedical Engineering, Nanyang Technological University ...... (3) Kašpar, J.; Fornasiero, P.; Hickey, N. Automotive catalytic ... adsorptive desulfurization of a model diesel fuel on regenerable Ni−.
Article pubs.acs.org/EF

Comparison of the Reactive Adsorption Desulfurization Performance of Ni/ZnO−Al2O3 Adsorbents Prepared by Different Methods Rooh Ullah,† Peng Bai,*,† Pingping Wu,† Zhanquan Zhang,‡ Ziyi Zhong,§ U. J. Etim,† Fazle Subhan,†,∥ and Zifeng Yan*,† †

State Key Laboratory of Heavy Oil Processing, PetroChina Key Laboratory of Catalysis, China University of Petroleum, Qingdao, Shandong 266555, People’s Republic of China ‡ Petrochina Petrochemical Research Institute, Beijing 102206, People’s Republic of China § School of Chemical and Biomedical Engineering, Nanyang Technological University (NTU), 62 Nanyang Drive, Singapore 637459, Singapore ∥ Department of Chemistry, Abdul Wali Khan University Mardan, Khyber Pakhtunkhwa 23200, Pakistan ABSTRACT: In this study, a series of Ni/ZnO−Al2O3 mixed oxide (MO) adsorbents were prepared by the one-step homogeneous precipitation method and the cation−anion double hydrolysis (CADH) method for reactive adsorption desulfurization (RADS) using thiophene as a model fuel in a fixed bed reactor. The synthesized adsorbents were characterized by N2 sorption, X-ray diffraction (XRD), scanning electron microscopy (SEM), transmission electron microscopy (TEM), H2 temperature-programmed reduction (H2-TPR), ultraviolet−visible (UV−vis) diffuse reflectance spectroscopy (DRS), and Raman spectroscopy. Results show that both Ni loading and the preparation method have a significant effect on the RADS activities of the adsorbent. Among the studied adsorbents, 10% Ni/ZnO−Al2O3 prepared by the one-step urea precipitation method showed the best RADS performance, with a thiophene conversion up to 96% and a sulfur adsorption capacity of 86 mg of S/g, which is 34% larger than that of CADH adsorbents. In addition, upon five RADS−regeneration cycles, sample 10% Ni/ZnO−Al2O3 exhibited a drop of only 3% in thiophene conversion, indicating the high stability of the Ni/ZnO−Al2O3 adsorbent prepared by homogeneous precipitation. Characterization results show that the one-step homogeneous precipitation method could facilitate the formation of small ZnO particles while suppressing the formation of inactive ZnAl2O4. On the other hand, by decreasing the formation of NiAl2O4, the homogeneous precipitation method could also generate high concentration of Ni0 sites, which are the active centers for the hydrogenolysis of C−S bonds. These findings indicate that a high-performance adsorbent for RADS can be obtained by employing a proper preparation method with good control on the adsorbent structure. The reactive desulfurization of fluidized catalytic cracking (FCC) gasoline is achieved over Ni/ZnO-based adsorbents. Sulfur atoms in the organosulfur compounds are selectively captured by forming metal sulfides, and the remaining hydrocarbon molecules without sulfur are released back to the main stream.19−22 The RADS mechanism over the Ni/ ZnO-based adsorbent has already been investigated by several groups. Tawara et al.14 deduced that the ZnO support had a kind of strong metal−support interaction (SMSI) with Ni and could automatically regenerate Ni during the desulfurization process. Babich and Moulijin17 reported that the RADS process might proceed via three series steps. In the first step, NiS was formed upon exposure to organosulfur compounds. In the second step, NiS reacted with H2 to restore Ni active sites with the release of H2S. In the final step, resulting H2S reacted with ZnO to produce ZnS. Huang et al.9 verified that sulfur species reacted with nickel to form Ni3S2, instead of NiS. Bezverkhyy et al.23 investigated the reaction kinetics of thiophene over the Ni/ ZnO adsorbent by thermogravimetric analysis and identified a rapid sulfur chemisorption on Ni, followed by a nucleation-

1. INTRODUCTION It has been recognized that sulfur oxides contribute to the accumulation of PM2.5 and PM10 in the atmosphere as a result of the formation of sulfates. Organosulfur compounds in gasoline not only cause the formation of acid rain by emitting sulfur oxides but also deactivate catalysts in vehicle exhaust converters.1−4 Therefore, one of the major worldwide concerns over the transportation fuel quality is the sulfur content. In China and the U.S., the sulfur concentration in gasoline will be regulated to be less than 10 ppm from 2017 onward.5 To meet the more stringent environmental protection legislations and regulations, ultradeep desulfurization technologies are desired to produce ultralow sulfur fuels.6,7 Among the various technologies used for removing sulfur from gasoline, the reactive adsorption desulfurization (RADS) technology represented by the S-Zorb process developed by Conoco Phillips Co.8−10 has been implemented on an industrial scale by SINOPEC in China and is gaining increasing attention. This technology combines the advantages of both adsorption desulfurization11,12 and catalytic hydrodesulfurization technologies.13−15 In the S-Zorb process, sulfur in gasoline is removed in moderate hydrogen pressure without a severe decrease of the octane number.16−18 © 2016 American Chemical Society

Received: January 29, 2016 Revised: March 19, 2016 Published: March 22, 2016 2874

DOI: 10.1021/acs.energyfuels.6b00232 Energy Fuels 2016, 30, 2874−2881

Article

Energy & Fuels

Teflon-lined stainless-steel autoclave for hydrothermal treatment at 100 °C for 24 h. The precipitate was filtered, washed with deionized water, dried at 80 °C, and calcined at 550 °C in air for 2 h. The synthesized samples are denoted as X% Ni/ZnO−Al2O3, where X represents the Ni loading. For comparison, a sample with the same composition of 10% Ni/ZnO−Al2O3 was also synthesized by the CADH method following our previous procedure.28 Typically, 2.32 g of Pluronic P123 was dissolved in 30 mL of distilled water to obtain a clear solution, which was subsequently added with 2.97 g of Zn (NO3)2·6H2O and 20 mL of aqueous solution of NaAlO2 (3.28 g) under continuous stirring. Afterward, an appropriate amount of Ni(NO3)2·6H2O dissolved in 5 mL of distilled water was added dropwise, and the solution was further stirred for 4 h. The resultant solution was transferred to a stainless-steel Teflon-lined autoclave and processed with similar procedures used in the one-pot homogeneous approach, and the material obtained was denoted as 10% Ni−ZnO− Al2O3. 2.3. RADS Performance Evaluation. The RADS performance evaluation of Ni/ZnO−Al2O3 adsorbents was conducted in a fixed bed reactor system. A microreactor with an internal diameter of 8 mm and 250 mm in length was used for this study. Prior to the desulfurization experiment, the reactor was washed thoroughly with ethanol for 24 h, followed by purging with pure N2 for 30 min. A total of 1 g (20−40 mesh) of adsorbent was loaded into the center of the microreactor column. Prior to the RADS reaction, the adsorbent was reduced in the presence of hydrogen with a flow rate of 20 mL/min under 0.5 MPa at 400 °C for 4 h. After the reduction process, the temperature and pressure were set at 350 °C and 1.5 MPa. The model fuel was preheated to 120 °C and injected along with H2 into the column by a microinjection pump at a weight hourly space velocity (WHSV) of 4 h−1 with a H2/oil ratio of 400. The sulfur content in the liquid product was analyzed with a BRUKER 450 gas chromatograph coupled with a pulsed flame photometric detector (PFPD). 2.4. Characterization of Adsorbents. N2 sorption of the Ni/ ZnO−Al2O3 adsorbents was carried out at 77 K using a Micromeritics TriStar 3000 analyzer. Prior to adsorption analyses, all samples were degassed at 300 °C for 12 h in vacuum. The Brunaur−Emmett− Teller2 method was used to calculate the specific surface area (SBET) in the relative pressure range of 0.05−0.25. The Barrett−Joyner− Halenda (BJH) method was used to calculate the pore size distribution (PSD).29 X-ray diffraction (XRD) patterns were obtained on a PAN Analytical X’Pert PRO MPD X-ray diffractometer (XRD) coupled with nickel-filtered Cu Kα radiation (40 kV and 40 mA) in a scanning range of 2θ = 5−70°. Morphology of the calcined sample was studied with S4800 scanning electron microscopy (SEM) operating at 15 kV. Elemental distribution mapping was analyzed by a JEM-2100 microscope equipped with energy-dispersive X-ray spectroscopy (EDX) at 200 kV. The high-resolution transmission electron microscopy (HRTEM) images were obtained on a JEM-2100 microscope. The H2 temperature-programmed reduction (H2-TPR) study of samples was conducted through a Chem-BET 3000 TPD/ TPR analyzer (Quantachrome, Boynton Beach, FL). Ultraviolet− visible (UV−vis) diffuse reflectance spectroscopy (DRS) spectra were measured on a Hitachi U-4100 spectrophotometer in the range of 200−800 nm. A Raman spectrophotometer (Thermo Fisher Scientific, DXR) was employed to measure Raman spectra for adsorbents at room temperature with a backscattering geometry.

controlled conversion of ZnO to ZnS. The rate-limiting step was reported to be different during different stages of reaction. In the beginning, the thiophene decomposition was determined to be the rate-limiting step, where a highly dispersed Ni metal was favorable, while with the partial sulfidation of ZnO, the thiophene diffusion could become the rate-limiting step because of the collapse of particle voids caused by the volume expansion during the transformation of ZnO to ZnS.15 On the basis of the possible RADS mechanism described above, it is reasonable to conclude that an adsorbent with a highly porous structure and a high dispersion of metal components is the key to a high RADS activity. Besides that, the size of ZnO was found to have a significant impact on the RADS performance. ZnO particles with a smaller particle size exhibited much higher RADS activity and larger sulfur capture capacity than those with a larger particle size.1 In industry, a kneading method is commonly used to prepare the RADS adsorbent, where typically zinc oxide, nickel oxide, pseudoboehmite, and SiO2 are physically mixed together to form the adsorbent.16,24,25 Such a kind of adsorbent has a poor pore structure and low dispersions of Ni and ZnO; thus, poor RADS performance is commonly observed.25 Therefore, new approaches for the synthesis of adsorbents with superior RADS performance are desirable. As a result of the uniform composition at the molecular level, homogeneous precipitation has been demonstrated to be effective to achieve uniformly dispersed particles. Especially, when urea is used as the precipitation agent, a metal carbonate hydroxide product is usually obtained. By decomposition of the product at an elevated temperature, CO2, H2O, and even NH3 will be released, leading to the formation of a porous metal oxide with abundant structural defects,26,27 which may be suitable for the application as RADS adsorbents. In addition, a versatile cation− anion double hydrolysis method (CADH) was reported for the preparation of mesoporous mixed oxide, where high dispersions of metal oxides were achieved even at high loadings.28 Motivated by the above-mentioned progress in adsorbent synthesis, we aim to develop effective and feasible approaches for preparing efficient adsorbents for the RADS process and to understand the real structural factors that contribute to the good RADS performance. In this work, Ni/ZnO−Al2O3 adsorbents were prepared using the one-pot homogeneous precipitation approach and the CADH method. The adsorbents were evaluated in RADS reaction using a model fuel as the feedstock. The RADS performance of adsorbents prepared by the two methods was compared under identical conditions, and the samples were characterized with a number of techniques.

2. EXPERIMENTAL SECTION 2.1. Chemicals and Feedstock. In this study, a model fuel with a sulfur concentration of 3000 ppm was prepared by introducing thiophene (analytical grade, Aldrich) as a sulfur source in n-octane (analytical grade, Aldrich). ZnCl2, Zn(NO3)2·6H2O, AlCl3·6H2O, NaAlO2, Pluronic P123, and urea were purchased from Sinopharm Chemical Reagent Co., Ltd. NiCl2 and Ni(NO3)2·6H2O were supplied by Shanghai Chemical Reagent Hanson. All chemicals were used without further purification. 2.2. Preparation of Adsorbents. For the one-pot homogeneous precipitation, typically 30 mL of ZnCl2 (1.36 g) aqueous solution was mixed with 20 mL of AlCl3·6H2O (2.41 g) aqueous solution with continuous stirring for 1 h. Then, 7.2 g of urea was added at room temperature. Subsequently, an aqueous solution of NiCl2 with different concentrations was added dropwise to the final mixture, followed by stirring for another 3 h. Finally, the mixture was transferred to a

3. RESULTS AND DISCUSSION 3.1. RADS Performances of Adsorbents. The RADS performances of Ni/ZnO−Al2O3 adsorbents are shown in Figure 1. The corresponding sulfur adsorption capacities (mg of S/g) are summarized in Figure 2. The 10% Ni/ZnO−Al2O3 adsorbent exhibits a high RADS activity with the highest thiophene conversion of 99% for the first 15 mL of fuel flow, which gradually decreases to 96% with increasing the oil volume to 30 mL, corresponding to a cumulative sulfur adsorption capacity of 78 mg of S/g at 30 mL of oil flow 2875

DOI: 10.1021/acs.energyfuels.6b00232 Energy Fuels 2016, 30, 2874−2881

Article

Energy & Fuels

Figure 1. RADS profiles of adsorbents using the model fuel as feedstock.

Figure 3. XRD patterns of (a) fresh and (b) spent adsorbents. Figure 2. Sulfur adsorption capacity of adsorbents at 30 mL of fuel flow.

favored the formation of this spinel in either dispersed or nondispersed form. Furthermore, after reduction and before the RADS test, the diffraction pattern of reduced 10% Ni−ZnO− Al2O3 is similar to that of unreduced 10% Ni−ZnO−Al2O3. The absence of reflections assignable to the metallic Ni phase indicates that nickel particles were highly dispersed in the sample. The XRD patterns of spent adsorbents are shown in Figure 3b, where characteristic peaks of the hexagonal zinc sulfide phase [ZnS, Joint Committee Powder Diffraction File (JCPDF) number 01-089-2191] are observed. This observation suggests the presence of ZnO, because ZnAl2O4 is well-recognized to be inactive in the RADS process. As reported,16,32−35 ZnO is inactive toward thiophene but effectively reactive toward H2S to form ZnS, which is responsible for the sulfur capacity of Ni/ ZnO−Al2O3 adsorbents (Figure 1). Therefore, the appearance of a higher ratio of the inactive ZnAl2O4 phase in the 10% Ni− ZnO−Al2O3 sample will reduce the ZnO amount in the sample,34 thus decreasing the sulfur capacity. In comparison to 10% Ni−ZnO−Al2O3, evidently, 10% Ni/ZnO−Al2O3 shows higher peak intensities for ZnS with lower intensities for the ZnAl2O4 phase, implying that a larger amount of ZnO was converted to ZnS in 10% Ni/ZnO−Al2O3,23 which is in good agreement with the results shown in Figures 1 and 2, where a larger sulfur capacity is observed on 10% Ni/ZnO−Al2O3 than on 10% Ni−ZnO−Al2O3. The presence of the NiAl2O4 phase in sample 12% Ni/ZnO−Al2O3 after H2 reduction and the RADS process indicates that NiAl2O4 could not be reduced to its active form of Ni036,37 but remained intact during the RADS process, resulting in a very low amount of the active Ni0 phase available for the hydrogenolysis of C−S bonds of thiophene. This is in well agreement with the results shown in Figures 1 and 2, where the poorest RADS performance of sample 12% Ni/ZnO−Al2O3 is observed. It is well-accepted that the reaction of nickel with the organosulfur compound is the rate-limiting step in the RADS technique, which breaks C−S bonds and generates H2S.33,34 Huang et al.21 verified the same phenomena by sulfur K-edge X-ray absorption near-edge

(Figure 2). It is worth noting that this sulfur adsorption capacity is 3-fold higher than that of the Ni/ZnO-based adsorbents prepared by the conventional kneading method catalyst.16 In the case of 8 and 12% Ni/ZnO−Al2O3 adsorbents, the thiophene conversions after 15 mL of fuel flow are reduced to 85 and 62%, along with the sulfur adsorption capacities of 61 and 41 mg of S/g, respectively, at 30 mL of fuel flow, implying the significant influence of Ni loading on the RADS performance. In contrast, under the same RADS conditions, the 10% Ni−ZnO−Al2O3 adsorbent synthesized using the cation−anion double hydrolysis method displays a lower RADS performance than that of the 10% Ni/ZnO−Al2O3 adsorbent. As revealed in Figure 1, the initial thiophene conversion at 15 mL of fuel is 86%, which is reduced to less than 71% after treating 30 mL of fuel with a corresponding capacity of 56 mg of S/g, which is about two-thirds of that of 10% Ni/ZnO− Al2O3. In short, the measured sulfur adsorption order is as follows: 10% Ni/ZnO−Al2O3 > 8% Ni/ZnO−Al2O3 > 10% Ni−ZnO−Al2O3 > 12% Ni/ZnO−Al2O3. 3.2. XRD Analysis. The XRD patterns of the adsorbents are shown in Figure 3a. Only the gahnite phase [ZnAl2O4, Joint Committee on Powder Diffraction Standards (JCPDS) file number 00-001-1146] is identified in all adsorbents, suggesting either existence of amorphous ZnO in the highly dispersed form or non-existence of the ZnO phase. The existence of ZnAl2O4, which is inactive in the RADS process,16 indicates the strong interaction between ZnO and Al2O3.30 This interaction seems to be more prominent in the 10% Ni−ZnO−Al2O3 sample. In sample 12% Ni/ZnO−Al2O3, besides the ZnAl2O4 phase, an additional nickel aluminum oxide phase (NiAl2O4, JCPDS card number 00-001-1299) with reflections at 2θ of 45.1°, 59.5°, and 65.4° is also observed, indicating the strong interaction of Ni with Al 2 O 3 . Gorewit and Tsutsui31 investigated Ni species in Ni/Al2O3-based adsorbents and concluded that the amount of inactive NiAl2O4 spinel was dependent upon the nickel loading and a high loading of Ni 2876

DOI: 10.1021/acs.energyfuels.6b00232 Energy Fuels 2016, 30, 2874−2881

Article

Energy & Fuels structure (XANES) study and concluded that the C−S bond cleavage is the rate-limiting step in the RADS technique, which strongly relies on chemically free Ni0. Hence, to obtain a good RADS performance, Ni should be uniformly distributed over small ZnO particles.1 Therefore, any effective synthetic approach should be able to reduce the possibility of the interaction between ZnO and Ni with Al2O3, preventing the formation of inactive NiAl2O4 and ZnAl2O4 but increasing the loading of highly dispersed active species, such as ZnO and Ni0. These results indicate that the formation of well-dispersed active ZnO and Ni0 species in Ni/ZnO−Al2O3 adsorbents may account for their superior RADS performance. 3.3. UV−vis DRS Analysis. To reveal the differences in phase composition,9 UV−vis DRS spectra were recorded for samples 10% Ni/ZnO−Al2O3 and 10% Ni−ZnO−Al2O3 (Figure 4). As shown in Figure 4, an absorbance peak at

Figure 5. Raman spectra of samples 10% Ni/ZnO−Al2O3 and 10% Ni−ZnO−Al2O3.

Ni−ZnO−Al2O3. Besides, a strong E1 longitudinal optical (LO) phonon mode is also detected at 580 cm−1 for sample 10% Ni/ ZnO−Al2O3, attributed to the phonon structure of the ZnObased materials.45−48 In contrast, this E1 (LO) phonon for sample 10% Ni−ZnO−Al2O3 is of higher intensity than that of 10% Ni/ZnO−Al2O3 and exhibits a red shift to 588 cm−1, demonstrating the ZnO agglomeration in the bulky phase with intrinsic structure defects49 or larger ZnO particles in 10% Ni− ZnO−Al2O3 than that in 10% Ni/ZnO−Al2O3.50 This is in accordance with the UV−vis DRS analysis (Figure 4) and probably related to the templating effect of the copolymer on the inorganic framework of 10% Ni−ZnO−Al2O3. Therefore, the Raman spectra verified the existence of small ZnO particles over the 10% Ni/ZnO−Al2O3 adsorbent and provide a reasonable explanation for the RADS results shown in Figures 1 and 2 that the presence of small ZnO particles as an independent phase is one of the key factors contributing to the high RADS performance.1 3.5. Textual Properties. Figure 6a exhibits N2 adsorption− desorption isotherms of adsorbents. All Ni/ZnO−Al2O3

Figure 4. UV−vis DRS spectra of 10% Ni/ZnO−Al2O3 and 10% Ni− ZnO−Al2O3 samples.

around 370 nm is observed for both samples, which are assigned to ZnO,30,38,39 demonstrating the presence of ZnO particles in the two samples. In comparison to that of 10% Ni− ZnO−-Al2O3, the higher intensity of the ZnO peak at ca. 371 nm for 10% Ni/ZnO−Al2O3 indicates the larger amount of ZnO in the later, which can explain their differences well in RADS activities (Figures 1 and 2). However, the absorbance band reveals a red shift from 371 to 381 nm for sample 10% Ni−ZnO−Al2O3 synthesized by the CADH method, implying the ZnO crystal growth and coagulation with the effect of the copolymer at high temperatures.38,40 In addition, another peak is observed in the UV region below 300 nm, which is assigned to ZnAl2O4.30,41,42 Obviously, the higher intensity of the absorbance peak below 300 nm in 10% Ni−ZnO−Al2O3 suggests the larger concentration of inactive ZnAl2O4, which can explain its poorer thiophene conversion activity and lower sulfur adsorption capacity well. This result is also consistent with the above XRD findings (Figure 3). Thus, it can be concluded that the higher RADS performance for 10% Ni/ ZnO−Al2O3 may partially originate from its higher concentration of well-dispersed ZnO. 3.4. Raman Spectroscopy Analysis. The Raman spectra measured at room temperature for 10% Ni/ZnO−Al2O3 and 10% Ni−ZnO−Al2O3 adsorbents are shown in Figure 5. The spectra exhibit a E2 active Raman mode at 99 cm−1, which mainly involves ZnO motion and is recognized as the lattice vibration of ZnO.43,44 As observed, 10% Ni/ZnO−Al2O3 shows a higher ZnO intensity compared to 10% Ni−ZnO−Al2O3, probably because of the weaker interaction between ZnO and aluminates in 10% Ni/ZnO−Al2O3, which is in agreement with the XRD results (Figure 3), where a lower intensity of ZnAl2O4 peaks is observed for 10% Ni/ZnO−Al2O3 than that of 10%

Figure 6. (a) N2 adsorption−desorption isotherms and (b) PSD of fresh adsorbents.

adsorbents demonstrate type-IV adsorption isotherms with a H-2 hysteresis loop, proving the existence of an ink-bottle mesopore in these samples.51 From the PSD curves and Table 1, it is observed that the pore size and pore volume of all Ni/ ZnO−Al2O3 adsorbents have a decline trend with increasing Ni loading (Figure 6b). As seen from Table 1, Ni/ZnO−Al2O3 2877

DOI: 10.1021/acs.energyfuels.6b00232 Energy Fuels 2016, 30, 2874−2881

Article

Energy & Fuels Table 1. Physical Properties of Different Adsorbents

a

adsorbent

SBET (m2/g) BRa

SBET (m2/g) ARb

V (cm3/g) BRa

V (cm3/g) ARb

D (nm) BRa

D (nm) ARb

8% Ni/ZnO−Al2O3 10% Ni/ZnO−Al2O3 12% Ni/ZnO−Al2O3 10% Ni−ZnO−Al2O3

177.6 174.5 166.5 119.8

145.9 131.8 147.3 98.9

0.43 0.37 0.28 0.25

0.29 0.24 0.20 0.17

9.81 7.66 5.86 8.29

7.55 5.73 4.89 6.67

BR = before reaction. bAR = after reaction.

morphology of the 10% Ni/ZnO−Al2O3 sample, SEM, HRTEM, and surface elemental distribution mapping analyses were conducted, and the results are shown in Figure 8. It is

with 8% Ni loading has a higher pore volume and a larger average pore size than that with 12% Ni loading. A similar trend was also observed for the specific surface area. This is possibly attributed to the occupation of mesoporous channels by Ni species.52 In contrast, sample 10% Ni−ZnO−Al2O3 has a lower surface area compared to the other adsorbents, indicating that the templating effect of the copolymer at high crystallization temperatures induces the crystal growth, consistent with studies showing that crystalline materials usually have a low surface area than the amorphous counterparts.53,54 To understand the structural change of adsorbents during the RADS process, the N2 sorption characterization for spent adsorbents was performed. Figure 7 displays N2 adsorption−

Figure 8. (a) SEM image, (b) TEM image, and elemental distribution mapping of (c) oxygen, (d) aluminum, (e) nickel, and (f) zinc of sample 10% Ni/ZnO−Al2O3.

clearly seen from the SEM image (Figure 8a) that sample 10% Ni/ZnO−Al2O3 displays a spongy-like morphology composed of small particles in the range of about 500−1000 nm. In addition, mesopores in the matrix of the 10% Ni/ZnO−Al2O3 sample are clearly observed from the TEM image in Figure 8b, confirming the porous features of 10% Ni/ZnO−Al2O3, as shown in Table 1. From the SEM−EDX images of the surface elemental distribution mapping for 10% Ni/ZnO−Al2O3 (panels c−f of Figure 8), it is verified that the ZnO and NiO particles are homogeneously dispersed over ZnAl2O4 and/or Al2O3 distributed inside the wide-ranging pores. The excellent dispersion of Ni and ZnO contributes to the higher RADS activity, which is responsible for hydrogenolysis of C−S bonds in thiophenic compounds and capturing of H2S, respectively.25,56 3.7. H2-TPR Characterization. In the Ni/ZnO−Al2O3based samples, Ni0 plays an active role in the cleavage of C−S bonds in the organosulfur compound with the release of H2S.33,34 The cleavage of C−S bonds is considered to be the rate-limiting step in the RADS process, which strongly depends upon the reducibility of Ni2+.21 To understand the reducibility of Ni2+ in the adsorbents, the H2-TPR technique was used. As shown in Figure 9, two TCD signals were observed. The peak at around 300 °C is attributed to the reduction of Ni2+ species weakly interacted with aluminate ions. The other peak in the range of 500−750 °C is assigned to the reduction of strongly interacted Ni2+ species at the tetrahedral/octahedral vacancies of aluminate spinels and the decomposition of the NiAl2O4 spinel phase.52,57−59 Because the pre-reduction temperature used prior to the RADS test was 400 °C, the weakly interacted

Figure 7. (a) N2 sorption isotherms and (b) PSD of adsorbents after RADS evaluation.

desorption isotherms and PSD curves of spent Ni/ZnO−Al2O3 samples. The adsorbed amounts of N2 on the spent adsorbents become smaller, revealing the altering of the pores to a lower level by sulfur adsorption and carbon deposition,55 in accordance with the XRD results for the spent adsorbents (Figure 3). With regard to 10% Ni/ZnO−Al2O3, its surface area, pore volume, and pore size were reduced by 24.5, 37.1, and 25.2% (Table 1), respectively, comparatively larger than all other adsorbents, consistent with its highest desulfurization activity among the investigated adsorbents, as shown in Figure 2. In comparison to 10% Ni−ZnO−Al2O3, 10% Ni/ZnO− Al2O3 possesses a higher surface area and proper pore structure, which can accommodate more phase transition and coke deposition, thus enhancing the RADS activity with a longer lifetime. 3.6. Morphology and Microstructure of the Adsorbent. To obtain deep insights into the textural structure and 2878

DOI: 10.1021/acs.energyfuels.6b00232 Energy Fuels 2016, 30, 2874−2881

Article

Energy & Fuels

Because RADS is an incessant process, the used adsorbents should be repeatedly regenerated. Only a stable porous material with well-dispersed active components will effectively reduce the mass-transfer resistance of reactant molecules to contact with Ni/ZnO-based materials and improve regeneration stability and RADS performance. It seems from the above discussion that the homogeneous precipitation by urea can lead to the formation of highly dispersed small ZnO particles. That is because, when urea was decomposed at 100 °C, the uniform increase of solution pH caused the homogeneous precipitation of Zn2+, Al3+, and Ni2+ ions, resulting in the formation of fine particles. The proposed RADS mechanism for NiO/ZnO− Al2O3 is illustrated in Scheme 1. As seen, the organosulfur

Figure 9. TPR profiles of Ni/ZnO−Al2O3 adsorbents.

Ni2+ ions could be reduced to Ni0 atoms during the prereduction, which are active centers for the cleavage of C−S bonds. In contrast, strongly interacted Ni2+ species are inactive in the RADS process. As observed from Figure 9, the intensity of the peak for the weakly interacted Ni2+ species increases with the increase of Ni loading, indicating the same trend of RADS activity of the adsorbents. However, considering the high content of the ZnAl2O4 phase in sample 12% Ni/ZnO−Al2O3, produced H2S could not be effective stored by ZnO. Thus, the highest sulfur capacity of sample 10% Ni/ZnO−Al2O3 should be attributed to its relatively high content of easily reducible Ni2+ and low content of the ZnAl2O4 phase. 3.8. Regenerations of the Catalyst Sample and RADS Mechanism. Adsorbent regeneration is an indispensable process in the RADS process, where regeneration stability and long longevity of the catalyst are vitally important for their industrial use.25,60,61 The key factors for deactivation of adsorbents include carbon deposition, ZnS formation, and sintering of the active metal particles, ZnO particles, and formation of impurity. The activity loss by carbon deposition and ZnO sulfurization can be recovered in the regeneration process in lean oxygen/air gas flow at elevated temperatures, whereas the sintering of the catalyst is considered as the perpetual deactivation. The 10% Ni/ZnO−Al2O3 adsorbent was selected to evaluate its stability during the multicycle regeneration. Regeneration of the spent adsorbent was carried out in the tubular furnace in an air atmosphere at 550 °C for 4 h. Figure 10 shows the stability and RADS activity after 5

Scheme 1. Proposed RADS Mechanism of Thiophene over the Ni/ZnO−Al2O3 Adsorbent

molecules can interact first with Ni particles and form Ni3S2, as revealed by Huang et al.21 in their sulfur K-edge XANES investigation. Meanwhile, the sulfur-free fuel goes back to the main stream, while Ni3S2 reacts with H2 and generate H2S, which is very reactive toward ZnO to form ZnS through nucleation control sulfidation,9,21 as also shown in our XRD study (Figure 3b). In this incessant RADS process, small and well-dispersed ZnO particles are favorable for the continuous H2S interaction and sulfur mass transfer to ZnS formation, which is believed as a critical factor to achieve high sulfur adsorption capacity. After complete deactivation, the adsorbent can be used again for another RADS cycle after regeneration in air/O2, followed by reduction in H2.

4. CONCLUSION In this work, a one-pot homogeneous precipitation approach toward the preparation of Ni/ZnO−Al2O3 is investigated and the resulting samples are employed for RADS in high-sulfurconcentration model fuel. The 10% Ni/ZnO−Al2O3 sample achieves the highest thiophene conversion up to 96% with a processing amount of 30 mL of model fuel, corresponding to a 78 mg of S/g sulfur adsorption capacity, which is 27% higher than that of the counterpart 10% Ni−ZnO−Al2O3 prepared by the cation−anion double hydrolysis method. By tailoring the loading of Ni on these samples, it is found that too high Ni loading intends to form inactive NiAl2O4, decreasing the expected activity. By comparison of 10% Ni/ZnO−Al2O3 to 10% Ni−ZnO−Al2O3, it is revealed that there are several factors that can contribute to the higher RADS performance, namely: (1) The formation of small particles of ZnO and NiO with a higher surface area and larger pore volume via

Figure 10. Regeneration performance of sample 10% Ni/ZnO−Al2O3.

consecutive regeneration cycles under conditions similar to the fresh adsorbent. The adsorbent achieves a desulfurization activity up to 96% and a slight drop by 3% in thiophene conversion compared to the initial activity, after the fifth cycle regeneration, indicating its good RADS stability. As such, Ni/ ZnO−Al2O3 is relatively stable without a significant change on Ni and ZnO particles upon calcination and Ni reduction. 2879

DOI: 10.1021/acs.energyfuels.6b00232 Energy Fuels 2016, 30, 2874−2881

Article

Energy & Fuels

(8) Xu, G.; Diao, Y.; Zou, K.; Zhang, Z. Cause analysis of sorbent deactivation in S-Zorb unit for gasoline desulfurization. Pet. Process. Petrochem. 2011, 42 (12), 1−6. (9) Huang, L.; Wang, G.; Qin, Z.; Dong, M.; Du, M.; Ge, H.; Li, X.; Zhao, Y.; Zhang, J.; Hu, T.; Wang, J. In situ XAS study on the mechanism of reactive adsorption desulfurization of oil product over Ni/ZnO. Appl. Catal., B 2011, 106 (1), 26−38. (10) Huan, H.; Salissou, M. N.; Dezhi, Y.; Xuan, M.; Li, S. Study on reactive adsorption desulfurization of model gasoline on Ni/ZnO−HY adsorbent. Chin. Pet. Process. Petrochem. Technol. 2013, 15 (3), 57−64. (11) Kim, J. H.; Ma, X.; Zhou, A.; Song, C. Ultra-deep desulfurization and denitrogenation of diesel fuel by selective adsorption over three different adsorbents: a study on adsorptive selectivity and mechanism. Catal. Today 2006, 111 (1), 74−83. (12) Srivastav, A.; Srivastava, V. C. Adsorptive desulfurization by activated alumina. J. Hazard. Mater. 2009, 170 (2), 1133−1140. (13) Tawara, K.; Nishimura, T.; Iwanami, H.; Nishimoto, T.; Hasuike, T. New hydrodesulfurization catalyst for petroleum-fed fuel cell vehicles and cogenerations. Ind. Eng. Chem. Res. 2001, 40 (10), 2367−2370. (14) TAWARA, K.; NISHIMURA, T.; IWANAMI, H. Ultra-deep hydrodesulfurization of kerosene for fuel cell system.(Part 2): Regeneration of sulfur-poisoned nickel catalyst in hydrogen and finding of auto-regenerative nickel catalyst. Sekiyu Gakkaishi 2000, 43 (2), 114−120. (15) Ryzhikov, A.; Bezverkhyy, I.; Bellat, J.-P. Reactive adsorption of thiophene on Ni/ZnO: Role of hydrogen pretreatment and nature of the rate determining step. Appl. Catal., B 2008, 84 (3), 766−772. (16) Meng, X.; Huang, H.; Shi, L. Reactive Mechanism and Regeneration Performance of NiZnO/Al2O3-Diatomite Adsorbent by Reactive Adsorption Desulfurization. Ind. Eng. Chem. Res. 2013, 52 (18), 6092−6100. (17) Babich, I. V.; Moulijin, J. A. Science and technology of novel processes for deep desulfurization of oil refinery streams: A review. Fuel 2003, 82 (6), 607−631. (18) Khare G. P. Desulfurization process and novel bimetallic sorbent systems for same. U.S. Patent 6,274,533 B1, Aug 14, 2001. (19) Bezverkhyy, I.; Ryzhikov, A.; Gadacz, G.; Bellat, J.-P. Kinetics of thiophene reactive adsorption on Ni/SiO2 and Ni/ZnO. Catal. Today 2008, 130 (1), 199−2055. (20) Ito, E.; Van Veen, J. On novel processes for removing sulphur from refinery streams. Catal. Today 2006, 116 (4), 446−460. (21) Huang, L.; Wang, G.; Qin, Z.; Du, M.; Dong, M.; Ge, H.; Wu, Z.; Zhao, Y.; Ma, C.; Hu, T.; Wang, J. A sulfur K-edge XANES study on the transfer of sulfur species in the reactive adsorption desulfurization of diesel oil over Ni/ZnO. Catal. Commun. 2010, 11 (7), 592−596. (22) Ge, H.; Tang, M.; Wen, X.-D. Ni/ZnO Nano Sorbent for Reactive Adsorption Desulfurization of Refinery Oil Streams. In Applying Nanotechnology to the Desulfurization Process in Petroleum Engineering; Saleh, T. A., Ed.; IGI Global: Hershey, PA, 2016; pp 216− 239, DOI: 10.4018/978-1-4666-9545-0.ch007. (23) Bezverkhyy, I.; Ryzhikov, A.; Gadacz, G.; Bellat, J.-P. Kinetics of thiophene reactive adsorption on Ni/SiO2 and Ni/ZnO. Catal. Today 2008, 130 (1), 199−205. (24) Mesters, C. M. A. M. Catalyst and its use in desulphurisation. U.S. Patent 7,297,655 B2, Nov 20, 2007. (25) Wen, Y.; Wang, G.; Wang, Q.; Xu, C.; Gao, J. Regeneration Characteristics and Kinetics of Ni/ZnO−SiO2−Al2O3 Adsorbent for Reactive Adsorption Desulfurization. Ind. Eng. Chem. Res. 2012, 51 (10), 3939−3950. (26) Bai, P.; Su, F.; Wu, P.; Wang, L.; Lee, F. Y.; Lv, L.; Yan, Z.-f.; Zhao, X. Copolymer-controlled homogeneous precipitation for the synthesis of porous microfibers of alumina. Langmuir 2007, 23 (8), 4599−4605. (27) Bai, P.; Wu, P.; Yan, Z.; Zhou, J.; Zhao, X. Self-assembly of clewlike ZnO superstructures in the presence of copolymer. J. Phys. Chem. C 2007, 111 (27), 9729−9733.

homogeneous precipitation using urea as the precipitation agent, as compared to that of the template-assisted CADH method, may be the reason for their higher performance. However, the later usually leads to a relatively higher crystallinity with the sacrifice of the surface area. (2) The reduced formation of ZnAl2O4 and NiAl2O4 phases by the homogeneous urea precipitations, which are inactive for RADS, may be the reason for their higher performance. This study proves that the facile homogeneous precipitation approach is effective to prepare the adsorbents with enhanced RADS performance and obtains a clear structure−performance relationship for the adsorbents.



AUTHOR INFORMATION

Corresponding Authors

*Telephone: +86-532-86981856. Fax: +86-532-86981295. Email: [email protected]. *Telephone: +86-532-86981296. Fax: +86-532-86981295. Email: [email protected]. Notes

The authors declare no competing financial interest.



ACKNOWLEDGMENTS This work was financially supported by the Joint Funds of the National Natural Science Foundation of China and the China National Petroleum Corporation (U1362202), the Natural Science Foundation of China (21206195), the Fundamental Research Funds for the Central Universities (14CX02050A and 14CX02123A), the Shandong Provincial Natural Science Foundation (ZR2012BM014), and the project sponsored by the Scientific Research Foundation for Returned Overseas Chinese Scholar. Ziyi Zhong ([email protected]) works at the Institute of Chemical Engineering in Singapore and also holds an adjunct associate professor position in the School of Chemical and Biomedical Engineering, NTU, Singapore.



REFERENCES

(1) Zhang, Y.; Yang, Y.; Han, H.; Yang, M.; Wang, L.; Zhang, Y.; Jiang, Z.; Li, C. Ultra-deep desulfurization via reactive adsorption on Ni/ZnO: The effect of ZnO particle size on the adsorption performance. Appl. Catal., B 2012, 119−120, 13−19. (2) Dasgupta, S.; Gupta, P.; Aarti; Nanoti, A.; Goswami, A. N.; Garg, M. O.; Tangstad, E.; Vistad, Ø. B.; Karlsson, A.; Stöcker, M. Adsorptive desulfurization of diesel by regenerable nickel based adsorbents. Fuel 2013, 108, 184−189. (3) Kašpar, J.; Fornasiero, P.; Hickey, N. Automotive catalytic converters: current status and some perspectives. Catal. Today 2003, 77 (4), 419−449. (4) Farrauto, R. J.; Heck, R. M. Catalytic converters: state of the art and perspectives. Catal. Today 1999, 51 (3), 351−360. (5) Kwasniewski, V.; Blieszner, J.; Nelson, R. Petroleum refinery greenhouse gas emission variations related to higher ethanol blends at different gasoline octane rating and pool volume levels. Biofuels, Bioprod. Biorefin. 2016, 10 (1), 36−46. (6) Mansouri, A.; Khodadadi, A. A.; Mortazavi, Y. Ultra-deep adsorptive desulfurization of a model diesel fuel on regenerable Ni− Cu/γ-Al2O3 at low temperatures in absence of hydrogen. J. Hazard. Mater. 2014, 271, 120−130. (7) Ma, X.; Velu, S.; Kim, J. H.; Song, C. Deep desulfurization of gasoline by selective adsorption over solid adsorbents and impact of analytical methods on ppm-level sulfur quantification for fuel cell applications. Appl. Catal., B 2005, 56 (1), 137−147. 2880

DOI: 10.1021/acs.energyfuels.6b00232 Energy Fuels 2016, 30, 2874−2881

Article

Energy & Fuels

(48) Decremps, F.; Pellicer-Porres, J.; Saitta, A. M.; Chervin, J.-C.; Polian, A. High-pressure Raman spectroscopy study of wurtzite ZnO. Phys. Rev. B: Condens. Matter Mater. Phys. 2002, 65 (9), 092101. (49) Alim, K. A.; Fonoberov, V. A.; Balandin, A. A. Origin of the optical phonon frequency shifts in ZnO quantum dots. Appl. Phys. Lett. 2005, 86 (5), 053103. (50) Calizo, I.; Alim, K. A.; Fonoberov, V. A.; Krishnakumar, S.; Shamsa, M.; Balandin, A. A.; Kurtz, R. Micro-Raman spectroscopic characterization of ZnO quantum dots, nanocrystals and nanowires. Proceedings of the Integrated Optoelectronic Devices 2007; San Jose, CA, Jan 22−24, 2007; 64810N, DOI: 10.1117/12.713648. (51) Gregg, S. J.; Sing, K. S. W. Adsorption, Surface Area and Porosity, 2nd ed.; Academic Press: London, U.K., 1982. (52) Aslam, S.; Subhan, F.; Yan, Z.; Xing, W.; Zeng, J.; Liu, Y.; Ikram, M.; Rehman, S.; Ullah, R. Rapid functionalization of as-synthesized KIT-6 with nickel species occluded with template for adsorptive desulfurization. Microporous Mesoporous Mater. 2015, 214, 54−63. (53) Bai, P.; Wu, P.; Yan, Z.; Zhao, X. Cation−anion double hydrolysis derived mesoporous γ-Al2 O3 as an environmentally friendly and efficient aldol reaction catalyst. J. Mater. Chem. 2009, 19 (11), 1554−1563. (54) Zhang, Z.; Pinnavaia, T. J. Mesostructured γ-Al2O3 with a lathlike framework morphology. J. Am. Chem. Soc. 2002, 124 (41), 12294−12301. (55) Ju, F.; Liu, C.; Meng, C.; Gao, S.; Ling, H. Reactive Adsorption Desulfurization of Hydrotreated Diesel over a Ni/ZnO−Al2O3−SiO2 Adsorbent. Energy Fuels 2015, 29 (9), 6057−6067. (56) Wei, Q.; Chen, J.; Song, C.; Li, G. HDS of dibenzothiophenes and hydrogenation of tetralin over a SiO2 supported Ni-Mo-S catalyst. Front. Chem. Sci. Eng. 2015, 9 (3), 336−348. (57) Hu, D.; Gao, J.; Ping, Y.; Jia, L.; Gunawan, P.; Zhong, Z.; Xu, G.; Gu, F.; Su, F. Enhanced investigation of CO methanation over Ni/ Al2O3 catalysts for synthetic natural gas production. Ind. Eng. Chem. Res. 2012, 51 (13), 4875−4886. (58) Bazyari, A.; Mortazavi, Y.; Khodadadi, A. A.; Thompson, L. T.; Tafreshi, R.; Zaker, A.; Ajenifujah, O. T. Effects of alumina phases as nickel supports on deep reactive adsorption of (4, 6-dimethyl) dibenzothiophene: Comparison between γ, δ, and θ-alumina. Appl. Catal., B 2016, 180, 312−323. (59) Barroso, M. N.; Gomez, M. F.; Arrúa, L. A.; Abello, M. C. Hydrogen production by ethanol reforming over NiZnAl catalysts. Appl. Catal., A 2006, 304, 116−123. (60) Rohr, D. F.; Vogel, J. A. System and method of regenerating desulfurization beds in a fuel cell system. U.S. Patent 20140260964 A1, Sept 18, 2014. (61) Liu, M.; Ye, X.; Liu, Y.; Wang, X.; Wen, Y.; Sun, H.; Li, B. Highly selective epoxidation of propylene in low-pressure continuous slurry reactor and the regeneration of catalyst. Ind. Eng. Chem. Res. 2015, 54 (20), 5416−5426.

(28) Bai, P.; Wu, P.; Yan, Z.; Zhao, X. A generalized approach to the synthesis of mesoporous mixed metal oxides using the cation-anion double hydrolysis method. Sci. Adv. Mater. 2011, 3 (6), 994−1003. (29) Barrett, E. P.; Joyner, L. G.; Halenda, P. P. The determination of pore volume and area distributions in porous substances. I. Computations from nitrogen isotherms. J. Am. Chem. Soc. 1951, 73 (1), 373−380. (30) Sampath, S. K.; Cordaro, J. F. Optical properties of zinc aluminate, zinc gallate, and zinc aluminogallate spinels. J. Am. Ceram. Soc. 1998, 81 (3), 649−654. (31) Gorewit, B.; Tsutsui, M. σ−π Rearrangements and Their Role in Catalysis. In Advances in Catalysis; Eley, D. D., et al., Eds.; Academic Press: New York, 1979; Vol. 27, pp 227−263, DOI: 10.1016/S03600564(08)60057-7. (32) TAWARA, K.; IMAI, J.; IWANAMI, H. Ultra-deep hydrodesulfurization of kerosene for fuel cell system.(Part 1): Evaluations of conventional catalysts. Sekiyu Gakkaishi 2000, 43 (2), 105−113. (33) Tawara, K.; Nishimura, T.; Iwanami, H.; Nishimoto, T.; Hasuike, T. Ultra-deep hydrodesulfurization of kerosene for fuel cell system.(Part 3) Development and evaluation of Ni/ZnO catalyst. Sekiyu Gakkaishi 2001, 44 (1), 43−51. (34) Fan, J.; Wang, G.; Sun, Y.; Xu, C.; Zhou, H.; Zhou, G.; Gao, J. Research on Reactive Adsorption Desulfurization over Ni/ZnO− SiO2−Al2O3 Adsorbent in a Fixed-Fluidized Bed Reactor. Ind. Eng. Chem. Res. 2010, 49 (18), 8450−8460. (35) Wang, G.; Wen, Y.; Fan, J.; Xu, C.; Gao, J. Reactive Characteristics and Adsorption Heat of Ni/ZnO−SiO2−Al2O3 Adsorbent by Reactive Adsorption Desulfurization. Ind. Eng. Chem. Res. 2011, 50 (22), 12449−12459. (36) Ashrafi, M.; Pfeifer, C.; Pröll, T.; Hofbauer, H. Experimental study of model biogas catalytic steam reforming: 2. Impact of sulfur on the deactivation and regeneration of Ni-based catalysts. Energy Fuels 2008, 22 (6), 4190−4195. (37) Damyanova, S.; Dotceva, A.; Eliyas, A.; Vladov, C.; Petrov, L. Recent development of the hydrodesulfurization catalysts. J. Environ. Prot. Ecol. 2000, 1 (4), 478−484. (38) Monticone, S.; Tufeu, R.; Kanaev, A. Complex nature of the UV and visible fluorescence of colloidal ZnO nanoparticles. J. Phys. Chem. B 1998, 102 (16), 2854−2862. (39) Xie, W.; Li, Y.; Sun, W.; Huang, J.; Xie, H.; Zhao, X. Surface modification of ZnO with Ag improves its photocatalytic efficiency and photostability. J. Photochem. Photobiol., A 2010, 216 (2), 149−155. (40) Bai, P.; Xing, W.; Zhang, Z.; Yan, Z. Facile synthesis of thermally stable mesoporous crystalline alumina by using a novel cation−anion double hydrolysis method. Mater. Lett. 2005, 59 (24), 3128−3131. (41) Rafla-Yuan, H.; Cordaro, J. Optical reflectance of aluminumdoped zinc oxide powders. J. Appl. Phys. 1991, 69 (2), 959−964. (42) Ragul, G.; Sumathi, S.; Nehru, M.; KC, S.; Kumar, S.; Bajpai, S.; Goel, C.; KC, S.; Ray, S. S.; Mandal, A. Synthesis, Characterization and Photocatalytic Study of Zinc Aluminate Nanopowders against Rhodamine−B and Crystal Violet Dyes. Int. J. Appl. Eng. Res. 2013, 8 (19), 2175−2178. (43) Chen, D.; Wang, Z.; Ren, T.; Ding, H.; Yao, W.; Zong, R.; Zhu, Y. Influence of Defects on the Photocatalytic Activity of ZnO. J. Phys. Chem. C 2014, 118 (28), 15300−15307. (44) Cuscó, R.; Alarcón-Lladó, E.; Ibanez, J.; Artús, L.; Jiménez, J.; Wang, B.; Callahan, M. J. Temperature dependence of Raman scattering in ZnO. Phys. Rev. B: Condens. Matter Mater. Phys. 2007, 75 (16), 165202. (45) Guo, S.; Du, Z.; Dai, S. Analysis of Raman modes in Mn-doped ZnO nanocrystals. Phys. Status Solidi B 2009, 246 (10), 2329−2332. (46) Bundesmann, C.; Ashkenov, N.; Schubert, M.; Spemann, D.; Butz, T.; Kaidashev, E.; Lorenz, M.; Grundmann, M. Raman scattering in ZnO thin films doped with Fe, Sb, Al, Ga, and Li. Appl. Phys. Lett. 2003, 83 (10), 1974−1976. (47) Wei, X.; Man, B.; Liu, M.; Xue, C.; Zhuang, H.; Yang, C. Blue luminescent centers and microstructural evaluation by XPS and Raman in ZnO thin films annealed in vacuum, N2 and O2. Phys. B 2007, 388 (1), 145−152. 2881

DOI: 10.1021/acs.energyfuels.6b00232 Energy Fuels 2016, 30, 2874−2881