Journal of Power Sources 274 (2015) 619e628
Contents lists available at ScienceDirect
Journal of Power Sources journal homepage: www.elsevier.com/locate/jpowsour
Synthesis, characterization and activity pattern of carbon nanofibers based copper/zirconia catalysts for carbon dioxide hydrogenation to methanol: Influence of calcination temperature Israf Ud Din a, Maizatul S. Shaharun a, *, Duvvuri Subbarao b, A. Naeem c a b c
Department of Fundamental and Applied Sciences, Universiti Teknologi PETRONAS, Malaysia Department of Chemical Engineering, Universiti Teknologi PETRONAS, Malaysia National Centre of Excellence in Physical Chemistry, University of Peshawar, Pakistan
h i g h l i g h t s A new CNFs based Cu/ZrO2 catalyst for CO2 hydrogenation to methanol was investigated. CNFs served well as a catalyst support in slurry reactor for the title reaction. Phase transformation of zirconia with calcination temperature was observed by XRD analysis. The efficiency of the catalyst was influenced by calcination temperature. N2O chemisorptions and activity data revealed dependency of CO2 conversion on Cu surface area.
a r t i c l e i n f o
a b s t r a c t
Article history: Received 5 September 2014 Received in revised form 8 October 2014 Accepted 15 October 2014 Available online 22 October 2014
A series of novel carbon nanofibers (CNFs) supported bimetallic copper/zirconia catalysts are synthesized by deposition precipitation method and calcined at different temperatures. Calcined catalysts are characterized by various techniques like X-ray diffraction, N2 adsorptionedesorption, N2O chemisorption, high resolution transmission electron microscopy, temperature programmed reduction, X-ray photoelectron spectroscopy and temperature programmed desorption (CO2 & NH3). The structureeactivity correlation is discussed in details. The results demonstrate 450 C as optimum calcination temperature for methanol synthesis rate with CO2/H2 feed volume ratio of 1:3. CO2 conversion is found to be directly proportional to copper metallic surface area (SCu), while a linear relationship is observed between methanol synthesis rate and fraction of dispersed Cu. © 2014 Elsevier B.V. All rights reserved.
Keywords: Methanol synthesis Slurry reactor Carbon nanofibers Copper based catalysts Carbon dioxide conversion
1. Introduction Global warming has been a serious threat to the natural environment and carbon dioxide is the main contributor to the phenomenon [1]. Nature has maintained the balance of natural CO2 in the atmosphere by its natural fixation, where it is converted to organic compounds. However, it is the human induced CO2 that has created a state of imbalance of carbon dioxide. Several ways like reduction of carbon dioxide emissions, CO2 capture and chemical transformation to valuable products were proposed to mitigate carbon dioxide. However, chemical transformation of this human induced CO2 will not only mitigate this problem but will also produce valuable products. A * Corresponding author. E-mail address:
[email protected] (M.S. Shaharun). http://dx.doi.org/10.1016/j.jpowsour.2014.10.087 0378-7753/© 2014 Elsevier B.V. All rights reserved.
number of products like salicylic acid, urea, plastics and methanol have been synthesized from carbon dioxide [2,3]. Currently, methanol has been produced industrially in gas phase from mixture of syngas and CO2 over CueZnO/Al2O3 catalyst at 220e300 C temperature and 5e10 MPa pressure. It is believed that methanol formation takes place mainly from CO2 whereas CO acts as scavenger of surface oxygen [4e6]. Nevertheless, application of this current methanol synthesis catalyst CuO/ZnO/Al2O3 does not look promising for CO2 hydrogenation. This is due to the production of CO as a sequence of parallel reverse water gas shift reaction (RWGS) observed with CO2 hydrogenation to methanol. This triggered the investigations to find a new catalyst system that can effectively hydrogenate CO2 to methanol. CNFs have special physiochemical characteristics like higher surface area, higher mechanical strength and surface defects for
620
I. Ud Din et al. / Journal of Power Sources 274 (2015) 619e628
holding catalyst particles. These properties make them a suitable choice as catalyst support. Moreover, CNFs are more advantageous in slurry phase reactors due to their mesoporous nature. Furthermore, due to lack of any 2D porosity in CNFs, accessibility of the reacting molecules to the active sites is enhanced, which in turn improves the activity profile of catalyst [7,8]. Consequently, CNFs based catalysts have shown better activity and selectivity as compared to traditional supports based catalysts [9]. Moreover, a special hexagonal thin morphology adopted by nickel particles based on CNFs resulted in higher metal support interaction. Subsequently, nickel deposited CNFs catalysts exhibited better performance for hydrogenation reactions as compared to classical supported catalysts like silica and alumina [10,11]. Similarly, higher metal support interaction produced in CNFs based platinumeruthenium catalysts also resulted in better dispersion of Pd metal [12]. Furthermore, Ledoux et al. reported that greater thermal conductibility of CNFs enabled a rapid evacuation of heat generated during the course of reaction [13]. This additional ability of CNFs will further support their application as catalyst support in exothermic reactions. Several factors like starting material and preparation method affect the catalytic properties of catalyst. Among such factors calcination temperature is a vital factor to alter the catalytic profile for certain reactions [14]. Irmawati et al. studied effect of calcination temperature on physiochemical properties of antimony vanadium mixed oxide catalysts [15]. In their work, escalation in calcination temperature not only led to increase agglomeration of catalyst particles but surface area was also adversely affected. Similar observation was also reported by Al-Zeghayer et al. where rise in calcination temperature declined activity of catalyst towards cyclohexyl benzene formation [16]. Moreover, intensification of calcination temperature also enhanced interaction between the active components of the catalyst. However, sintering of active sites was observed at elevated calcination temperature [17]. In this work, CNFs based Cu/ZrO2 catalysts were synthesized by deposition precipitation method. In order to scrutinize effect of calcination temperature, catalysts were calcined at different temperatures. Physiochemical parameters as well as activity studies for liquid phase methanol synthesis were investigated in slurry reactor. 2. Experimental
nitrate hydrate (SIGMA-ALDRICH, USA) was added to the stirring solution. After complete dissolution of both nitrate salts, required quantity of CNFs-O was added to the solution. Temperature of the solution was increased to 90 C. A 10 ml of urea solution having 1 g of urea was added to the slurry solution as a precipitating agent. The slurry was stirred for 20 h, cooled and filtered by vacuum filtration. The precipitates were dried in oven at 110 C for overnight. The dried catalysts were calcined in N2 flow at four different temperatures (350, 450, 500 and 550 C) for 3 h and labeled as CZC350, CZC450, CZC500 and CZC550, respectively. 2.3. Characterization PANalytical model Empyrean X-ray diffractometer was used for crystallographic analysis of catalyst components. Phase identification was performed by PANalytical HighScore Plus. The XRD data were measured at room temperature from 20 to 80 at 2q Bragg angle. Surface area and pore volume of CZC catalysts were analyzed by N2 adsorptionedesorption isotherms technique, using Micrometrics ASAP 2020 device. BET method was utilized for the evaluation of adsorption isotherms while BJH method was used to evaluate pore size distribution [20]. Metallic surface area of copper (SCu) was determined by N2O chemisorption technique [21]. Catalysts were first reduced with a flow of H2 at 500 C. After reduction, samples were cooled to 60 C in He flow and purged for 30 min. Then N2O was introduced at 60 C for 1 h. Residual N2O was removed by He flow for 1 h. Finally Temperature Programmed Reduction (TPR) analysis was conducted for the second time at 500 C. Surface area and dispersion of Cu were calculated by assuming surface atomic density of 1.46 1019 Cuat m2 and stoichiometry of Cu:N2O ¼ 2, respectively. Average particle size (dCu) was obtained by a conventional formula as follows
dCu ðnmÞ ¼
X 2Y
where X is the H2-uptake for first reduction and Y is H2-uptake for second reduction. Likewise, distribution of Cu content was estimated by the following equation [22].
Cu0 surface area Cu content BET surface area
2.1. Functionalization of carbon nanofibers (CNFs)
RCu ¼
The CNFs with herringbone type morphology (GNF-100) were purchased from Carbon Nano-material Technology Co. Ltd., Korea. The diameter, length and specific surface area are 50e100 nm, 10e30 mm and 100e300 m2 g1, respectively. CNFs were modified by treating with 35 volume % nitric acid solution. The refluxing was performed for 16 h at elevated temperature (90 C). During this process carbon nanofibers (CNFs) were oxidized and converted to oxidized carbon nanofibers (CNFs-O). After cooling to room temperature, CNFs-O were filtered by vacuum filtration, washed several times with distilled water and then dried overnight in oven at 100 C.
Morphology and particle size measurement of the catalysts were conducted by using transmission electron microscopy (Zeiss LIBRA 200TEM), Accelerating Voltage: 200 kV. Samples were dispersed in isopropanol and sonicated in ultrasonic device for about 1 h and the suspensions were dropped onto a copper grid [23]. Reduction behavior and metal support interaction were investigated by Temperature Programmed Reduction (TPR) in temperature range of 30e800 C with heating rate of 10 C min1, using TPDRO1100 MS equipped with thermal conductivity detector (TCD). TPR analysis were performed in 5 vol.% H2/N2 flow with a flow rate of 20 cm3 min1. X-ray photoelectron spectroscopy was utilized to investigate chemical nature and surface composition of copper. X-ray photoelectron spectroscope (XPS, Thermo-Fisher K-Alpha) equipped with monochromitised AlK source having ultimate energy resolution of 0.5 eV was employed for this purpose. Avantage software was used for peak fitting and chemical state identification. Surface acidity and basicity were examined by NH3 temperature programmed desorption (NH3-TPD) and CO2 temperature
2.2. Synthesis of Cu·ZrO2/CNFs-O (CZC) catalysts by deposition precipitation method A series of Cu$ZrO2/CNFs-O catalysts with constant loading of Cu (15 wt%), ZrO2 (10 wt%) and CNFs-O (75 wt%) has been synthesized by deposition precipitation method [18,19]. Firstly, a required amount of Cu (NO3)2$3H2O (R&M Chemicals, UK) was dissolved in 400 ml distilled water. Similarly a known quantity of zirconyl
I. Ud Din et al. / Journal of Power Sources 274 (2015) 619e628
programmed desorption (CO2-TPD), respectively. Prior to TPD analysis, catalysts were subjected to higher temperature of 500 C for one hour in inert atmosphere to desorb the surface moisture and other adsorbed molecules. Pre-reduced catalysts were cooled to room temperature and were saturated with pure NH3 and CO2 in case of NH3-TPD and CO2-TPD, respectively. The adsorption of gases was continued for one hour at 90 C and physiosorbed molecules were fluxed out with He flow. The adsorbed gases in both cases were desorbed in temperature range of 40e800 C. Desorption of gases at relative degree of temperature were quantified by calibrated TCD. 2.4. Activity studies Activity of catalysts in CO2 hydrogenation to methanol was evaluated in autoclave slurry reactor (Parr 4593 with a regulator Parr 4848). Ethanol was used as a slurry solvent for the liquid phase hydrogenation reaction. A 0.5 g of reduced sample was suspended in 25 ml of reaction solvent placed in reactor vessel [24]. The reactor was purged with reactant gases at room temperature to remove any air or other gases. A mixture of H2/CO2 gases with 3:1 was used as reactant gases. Reactor was pressurized with reactant gases to 30 bars and reaction temperature was raised to 180 C. Catalytic hydrogenation of CO2 to methanol was carried out for 2 h. Liquid and gas products were analyzed by flame ionization detector (FID) and thermal conductivity detector (TCD), respectively. Turnover frequency of methanol was calculated by following formula [25,26].
Number of molecules of methanol produced TOFMeOH s1 ¼ TimeðsÞ$number of mettalic copper atoms ¼
A$N a 3600$SCu $Na 1
1
where, A represents methanol activity in mol g h , Na is Avogadro's number (6.023 1023), SCu denotes metallic copper surface area in m2 g1 and Na designates number of Cu atoms in a monolayer (Na ¼ 1.469 1019 atoms m2) 3. Results and discussions 3.1. Influence of calcination temperature on physicochemical properties of CZC catalysts 3.1.1. X-ray diffraction study Phase analysis of catalyst components were investigated by XRD technique. XRD profile of catalysts calcined at different temperatures is shown in Fig. 1. For comparison an XRD spectrum of bare oxidized CNFs was also included. Two prominent peaks were detected at 2q values of 26 and 44 indicating the hexagonal graphitic planes of CNFs (JCPDS No. 41-1487). Similarly, diffraction pattern with peaks at 32.6 , 35.5 , 38.7, 48.8 , 53.6 , 58.3 , 61.67, 66.4 , 68.1, 72.3 and 75.1 on 2q scale was found which is indexed as monoclinic phased tenorite CuO with JCPDS card files No. 481548 (a ¼ 4.62 Å, b ¼ 3.43 Å, and c ¼ 5.06 Å). Phase distribution was affected by the degree of calcination temperature. Catalysts calcined at lower temperature were less crystalline as disclosed by low intensity peaks of CZC350 and CZC450 catalysts. The lower crystallinity indicates a well dispersed and highly amorphous phase of CuO component at lower calcination temperature. Furthermore, no peak was observed for ZrO2, suggesting that ZrO2 is either in amorphous phase or exists in very fine crystalline form which could not be detected by XRD. However, further rise in calcination temperature not only increased CuO crystallinity but it also led to
621
develop two additional reflections in CZC500 and CZC550 catalysts around 2q, 24 and 56 , indicating the presence of monoclinic zirconia [27]. Similarly, existence of tetragonal polymorph of zirconia was also identified by sharp reflections at 30.3 and 50.4 on 2q scale at higher calcinations [28]. 3.1.2. BET surface area and pore size distribution Surface area is an important parameter for determining the catalyst activity. N2 adsorptionedesorption isotherms of the catalysts calcined at different temperature are shown in Fig. 2. Each catalyst exhibited typical type-IV isotherms with H4 type hysteresis loops having sharp inflection between p/p0 ranges of 0.75e0.94. This discloses mesoporous nature of catalyst support. BET surface area was significantly affected by variation in calcination temperature as shown in Table 1. A sharp increase in BET surface area was observed by increasing calcination temperature from 350 to 450 C. Consequently, total pore volume was increased from 0.29 to 0.39 cm3 g1. This increase in surface area can also be justified by a remarkable increase of total adsorbed gas from 225 to 293 cm3 g1 for CZC350 and CZC450 catalysts, respectively. However, this trend could not lasted for catalysts calcined at more elevated temperature. Contrary to the earlier trend, BET surface area was observed to be reduced with further rise in calcination temperature. This could be ascribed to sintering phenomenon and growth of catalyst particles at higher calcination temperature [29,30]. Similarly, decline in surface area could also be attributed to sintering of zirconia associated with the mobility of Zr ions on the surface of the support at elevated temperature [31]. Sintering of catalyst particles at elevated temperature was also evident from TEM images of the respective catalysts. Nevertheless, average pore diameter was not affected and remains almost constant throughout the temperature range of calcinations (Table 1). 3.1.3. Size and dispersion of Cu Investigations of Cu surface area, variation of Cu size as well as dispersion of active sites are crucial as they have a pivotal role in the performance of catalysts for structure-sensitive reactions like CO2 reduction to methanol. Tabulated values of Cu surface area (SCu), Cu crystallite size (dCu), Cu dispersion (DCu) and fraction of surface Cu (RCu) are documented in Table 1. Although magnitudes of SCu, DCu, and RCu were less affected by increasing calcination temperature from 350 to 450 C, however values of all three parameters were significantly dropped with further rise. Influence of calcination
Fig. 1. X-ray diffraction patterns of (a) CNFs-O, (b) CZC350, (c) CZC450, (d) CZC500 and (e) CZC550 catalysts.
622
I. Ud Din et al. / Journal of Power Sources 274 (2015) 619e628
Fig. 2. N2 adsorptionedesorption of (a) CZC350, (b) CZC450, (c) CZC500 and (d) CZC550 catalysts.
temperature on variation of dCu and subsequent SCu was also studied. Generally, homogenous distribution of copper on catalyst support resulted in higher surface area of copper. In current study, SCu was witnessed to be decreased by raising degree of calcination temperature. This indicates that homogeneity of Cu is adversely affected by increase in calcination temperature. This may also be attributed to growth of Cu particles at elevated temperature. Contrary, size of Cu was twofold increased when calcination temperature was raised from 350 to 550 C. This is mainly due to agglomeration of Cu at higher degree of temperature as shown by TEM images. This trend is also in agreement with the previous reported data [32e34]. Nevertheless, dispersion of Cu was adversely effected by intensification of calcination temperature. This could be due to sintering of Cu particles at elevated temperature. As shown in Table 1, distribution of Cu has been affected by variation in calcination temperature. RCu was progressively decreased with increasing calcination temperature and lowest RCu value was recorded for catalysts calcined at maximum calcination temperature. This implies that Cu is partially embedded in the ZrO2 phase and the process accelerates with further rise in calcination temperature. Moreover, it could be further inferred that intensification of calcination temperature and subsequent Cu embedding provide less exposed Cu surface and more CueZrO2 interfacial surface area. Furthermore, due to large interfacial surface area this
feature also reveals better CueZrO2 interactions as a function of increasing calcination temperature. In the end of this discussion, an interesting correlation was developed by these three parameters in a sense that larger dCu exhibited lower SCu which resulted in reduction of DCu. 3.1.4. Morphology studies Transmission electron microscopy (TEM) was utilized to study the morphology of the catalysts. TEM images of studied catalysts with magnification of 200 K are presented in Fig. 3. Dark black spherical shaped particles were identified as copper particles whereas tetragonal shaped light colored particles were recognized as zirconia particles. TEM studies revealed well distributed particles for CZC350 and CZC450 with average particle size of 3 and 5 nm, respectively. However, agglomeration as well as growth of catalysts particles were observed when calcination temperature was increased beyond 450 C. Consequently, a twofold increase was observed in average particle size for CZC500 and CZC550 catalysts. This growth of particles with increasing calcination temperature is due to agglomeration of particles and has been consistently reported in literature [35]. By a closer look at Fig. 3, one can see that intensification of calcination temperature beyond 450 C has resulted sintering phenomenon, as evident from sub-images of CZC500 and CZC550 catalysts. TEM investigations confirmed the
Table 1 Copper metal and catalyst surface area of the calcined samples. Catalyst
Ads. gas (cm3 g1)
SBET (m2 g1)
Pore dia. (nm)
Pore vol. (cm3 g1)
SCu (m2 g1) N2O
DCu (%) N2O
dCu (nm) N2O
RCu N2O
CZC350 CZC450 CZC500 CZC550
225 293 250 250
109 155 118 114
10.7 10.1 11.2 11.2
0.29 0.39 0.33 0.32
8.6 8.0 4.2 3.6
23 22 14 18
4 5 8 10
0.52 0.34 0.23 0.21
I. Ud Din et al. / Journal of Power Sources 274 (2015) 619e628
623
Fig. 3. TEM images of (a) CZC350, (b) CZC450, (c) CZC500 and (d) CZC550 catalysts.
N2O chemisorption agglomeration.
results
regarding
Cu
dispersion
and
3.1.5. Reduction behavior and metal support interaction In order to get insight into metal support interaction, TPR profile of all studied catalysts is presented in Fig. 4. TPR bands are deconvoluted into three reduction peaks, denoted as peak (a), peak (b) and peak (g). The appearance of two reduction peaks for CuO is previously reported and attributed for stepwise reduction of copper (Cuþ2eCuþ1eCu). Similarly, it could also be assigned to highly dispersed and bulk-like copper. Reduction peak (a) at around 260 C is ascribed to highly dispersed copper [36]. Peak (b) is assigned to reduction of bulk-like CuO while peak (g) shows gasification of carbon nanofibers in higher temperature range [37e39].
In fact by close inspection, the reduction peaks of both dispersed and clustered copper were observed at relatively higher temperature as compared to those observed for Cu containing catalysts based on traditional catalysts supports like alumina and zirconia. This is because of the electronegativity gradient between copper and corresponding catalyst support (cCu cs) as shown in Table 2. In case of alumina and zirconia, copper accepts electrons from the support whereas in case of CNFs electrons are donated by copper, thus making the reduction of CuO difficult in this case [40]. In turn, this also suggests strong interaction of copper and CNFs as compared to other traditional catalyst supports. Shifting of reduction peak to higher temperature can also be attributed to interaction of copper and zirconia at the surface of CNFs. Because of high work function of ZrO2 than Cu, interaction between the two creates electron deficiency in Cu. Hence reduction of copper is observed at higher temperature [41]. Total hydrogen uptake, extent or degree of reduction (H2/Cu), position of reduction peak, number of dispersed and bulk phase Cu and their relative abundance are documented in Table 3. Total hydrogen uptake was observed to be in the order: CZC350 > CZC450 > CZC500 > CZC550 which indicates that H2 uptake is suppressed by increase in calcination temperature. Moreover, reduction peak (a) was shifted to higher temperature with increasing the degree of calcination temperature beyond 450 C. This is because of the catalyst particle size variation associated with the calcination temperature as shown by TEM and N2O Table 2 Electronegativity gradient of common catalyst supports with copper.
Fig. 4. TPR profile of (a) CZC350, (b) CZC450, (c) CZC500 and (d) CZC550 catalysts.
Element
Electronegativity (c)
cCu cs
Al Zr C Cu
1.61 1.33 2.55 1.90
0.29 1.57 0.65
624
I. Ud Din et al. / Journal of Power Sources 274 (2015) 619e628
Table 3 TPR results of calcined samples. Sample
CZC350 CZC450 CZC500 CZC550 a
H2 consumed (mmoles g1)
2225 2172 1772 1212
H2/Cu
0.94 0.91 0.74 0.51
H2 consumed (mmoles g1)
Red. temp. (oC) Peak a
Peak b
Peak a
Peak b
269 266 276 245
423 383 362 327
1312 1542 1007 643
912 692 765 568
A(a)/(A(a) þ A(b))a (%)
59 71 57 53
Relative abundance of Cu.
measurements. Copper with small size is easily reduced at the surface of the support as compared to the large one [42]. The trend can be attributed to growth of particles with degree of calcination temperature. In contrast, position of peak (b) was shifted to lower temperature range with increasing the degree of calcination temperature (Table 2). Moreover, fraction of reduction peak (a) was observed to be intensified when calcination temperature was raised from 350 to 450 C, but subsequently reduced with further increase in calcination temperature. Almost similar tendency was recorded for reduction peak (b). Since highly dispersed copper is the main active part in Cu-based catalysts, hence relative abundance of highly dispersed was calculated and reported in Table 2. Calcination temperature affected the distribution of different copper phases on the surface of catalysts. Relative abundance of dispersed copper was incremented by increasing the calcination temperature from 350 to 450 C. However, further rise in calcination temperature adversely affected the relative abundance of dispersed Cu. This could also be due to agglomeration and growth of Cu at elevated temperature. 3.1.6. Electronic state of catalyst components The chemical states of catalyst components were evaluated by XPS (Fig. 5). XPS spectra of calcined catalysts exhibited peaks for Cu2p3/2 at 934 eV and 942 eV as parent and satellite peak, respectively [43]. Similarly, XPS peaks appeared at 954 eV and 962 eV for Cu2p1/2 core electrons and satellite excitations. The emergence of the satellite structure is due to the charge transfer transitions from the ligands (O2 ions in case of CuO) into the unfilled (d9) valence orbital of Cuþ2 ion. This transition of ligands electrons is not possible in case of Cuþ or Cu0 species due to the completely filled (d10) shells. Hence, the satellite structure is a characteristic peak of Cuþ2 ion. In the current study, the appearance
Fig. 5. XPS spectra of (A) Cu component of (a) CZC350, (b) CZC450, (c) CZC500, (d) CZC550 and (B) Zr component of (a) CZC350, (b) CZC450, (c) CZC500, (d) CZC550 catalysts.
of satellite peaks with concomitant parent peak confirms Cuþ2 as a predominant Cu oxidation state in all studied catalysts: irrespective of the magnitude of calcination temperature. Furthermore, occurrence of Cu2p3/2 at 934 eV with consistent satellite peak also suggests a contribution of tenorite CuO which support the XRD findings. Similarly, a spineorbit doublet was observed for Zr3d5/2 and 3d3/2 core levels centered at 182.2 eV and 184.6, respectively. The appearance of two different binding energy peaks with energy gap of 2.4 eV suggests two different kinds of zirconium ions. The higher binding energy peaks (184.6 eV) represent Zrþ2 species while lower binding energy peaks (182.2 eV) indicate the existence of pure zirconia in Zrþ4 form. In order to identify the different Cu species the Cu2p3/2 peak was resolved in two different peaks, as shown in Fig. 6. Cupric ion can easily be identified on the basis of coupling phenomenon between unpaired electrons observed in Cuþ2 while this is lacking in case of metallic and cuprous ion. However, it is quite difficult to differentiate between Cu0 and Cuþ1 because their binding energies and full width at half maximum (FWHM) values are so close that they overlap each other. Cu2p3/2 peak observed at 932.7 eV was attributed to Cuþ1 ion while the one observed at 933.7 eV was assigned to Cuþ2 ion. Furthermore, the existence of two different CuO species is also consistent with the H2-TPR findings, where CuO with two different kinds of reducibility were also recognized. The binding energies of Cu2p3/2 andZr3d5/2, their FWHM values and Cu/Zr ratio has been documented in Table 4. The tabulated data shows a slight shift of Cu2p3/2 towards higher binding energies with increasing calcination temperature, suggesting a decline in Cu dispersion. According to the XPS results, Cu dispersion is adversely affected by calcination temperature until 500 C. However, it is increased with further intensification. This is exactly in accordance to the N2O chemisorption results. In contrast, no significant variation was observed in binding energies of Zr3d5/2. The FWHM value of Cu2p3/2 was increased for CZC450 as compared to CZC350 catalyst suggesting the distortion in coordination symmetry of Cuþ2 to highly distorted octahedral symmetry by making additional Cuþ2eCu2 bond with neighbor O2 ions. Unlike Cu2p3/2, FWHM values of Zr3d5/2 remained also invariant for all catalysts. So in this case no particular information regarding nature and bonding of zirconium ion could be obtained. Furthermore, distribution of surface Cu was also influenced by variation in calcination temperature. Cu/Zr ratio was increased for CZC450 as compared to CZC350 catalyst. Nevertheless, the ratio was adversely affected by further rise in calcination temperature indicating depletion of surface Cu and subsequent enrichment of surface Zr content. Agglomeration of Cu as function of increasing calcination temperature as indicated by TEM results could be one the reasons for Cu/Zr decline.
3.1.7. Acidic sites investigations During the course of this investigation, the number and nature of acidic sites were probed by TPD-NH3. Fig. 7 shows the NH3desorption profile of CZC catalysts calcined at different temperature.
I. Ud Din et al. / Journal of Power Sources 274 (2015) 619e628
625
Fig. 6. Peak fitting curves of Cu2p3/2 peak of (a) CZC350, (b) CZC450, (c) CZC500, (d) CZC550 catalysts.
Number of total acidic sites were evaluated from intensity of desorption peaks while number of total acidic sites divided by surface area produced the density of total acidic sites [44]. The degree of temperature at which ammonia is desorbed, is an indicator for strength of acidic sites in the catalysts. Desorption signals lower than 200 C represent weak acidic sites, desorption between 200 and 450 C (peaks a and b) is due to the presence of acidic sites with medium strength while ammonia desorption at more than 450 C (peaks g and d) is attributed to strong acidic sites. In the current study, the acidic sites were found over temperature range of 250e700 C which implies that strength of acidic sites was distributed from medium to strong with no weak acidic site. Although by a closer look at TPD profile of CZC550 a small deflection of desorption can be visualized. However, in some cases NH3 desorption from weak acidic sites is reported to be the result of hydrogen bonded NH4þ rather than desorption of ammonia from the surface of acidic sites of catalyst. Moreover, weak acidic sites are of less interest due to their negligible role in selective catalytic reduction [45]. Furthermore, acidic sites with moderate strength
have been declared as an adequate parameter for evaluation of acidities of catalysts [46,47]. The values of total number and density of acidic sites are documented in Table 5. The magnitude of acidic sites recorded in this study was found very close to the recently published work of Ning et al. where NH3-desorption was ascribed to the presence of both Lewis and Bronsted acids sites on surface of zirconia [47]. Increase of calcination temperature affected the acidic profiles of
Table 4 XPS data of catalysts calcined at different temperature. Sample
CZC350 CZC450 CZC500 CZC550
Binding energy (eV)
FWHM (eV)
Cu2p3/2
Zr3d5/2
Cu2p3/2
Zr3d5/2
Atomic ratio Cu/Zr
934.14 934.24 934.45 934.12
182.43 182.22 182.28 182.21
4.37 4.55 4.15 4.36
4.69 4.74 4.61 4.73
0.79 1.20 1.05 0.71
Fig. 7. NH3-TPD profiles of (a) CZC350, (b) CZC450, (c) CZC500 and (d) CZC550 catalysts.
626
I. Ud Din et al. / Journal of Power Sources 274 (2015) 619e628
Table 5 Acidic and basic properties of calcined samples. Catalyst Number of total Density of total Number of total Density of total basic sites acidic sites acidic sites basic sites (mmol g1 cat) (mmol m2) (mmol g1 cat) (mmol m2) CZC350 CZC450 CZC500 CZC550
9.40 7.13 8.71 6.02
0.08 0.04 0.07 0.05
0.85 0.72 0.92 0.53
0.007 0.004 0.007 0.004
the catalysts. Generally the acidity as well as basicity of catalyst is reduced by increasing the calcination temperature. Nevertheless, concentration of acidic sites was increased with increasing calcination temperature from 450 to 500 C and was declined on further rise in calcination temperature. The tendency of monoclinic zirconia (m-ZrO2) to have more acidic sites than tetragonal counterpart (t-ZrO2) has consistently been reported in the literature [48,49]. Based on these observations, the discrepancy could be due to presence of m-ZrO2 polymorph. This was further justified by XRD results where existence of monoclinic zirconia was observed for CZC500 and CZC550 catalysts. The reduction of acidic sites on further increase of calcination temperature could be attributed to the growth of zirconia, as indicated by TEM measurement. However, owing to the different quantities of total acidic sites and surface area, no clear trend was observed in density of total acidic sites of all studied catalysts. 3.1.8. Basic sites investigations The CO2-TPD profile of CZC catalysts is shown in Fig. 8. The desorption quantities with relative desorption peaks are listed in Table 5. The study disclosed significant quantities of basic sites on the surface of all catalysts. Indeed, Lewis basic sites associated with the surface of zirconia are responsible for adsorbing CO2 [4,50e52]. The appearance of desorption peaks are due to the decomposition of carbonate species [52]. Each catalyst exhibited four desorption peaks in the temperature range of 300e690 C. Desorption peaks below 500 C (peaks a and b) were attributed to weak and medium basic sites whereas peaks above 500 C (peaks g and d) were ascribed to strong basic sites. Comparative study of CO2-TPD spectra at different calcination temperatures revealed some interesting results. By increasing calcination temperature, the strongly bound CO2 peaks were shifted to slightly higher temperature, indicating strength of the CO2 bindings as a function of calcination temperature. On the other hand, no established trend was recorded for TPD peaks ascribed to weakly basic sites. In the literature,
monoclinic zirconia is described to display CO2 desorption peak at higher temperature as compared to tetragonal polymorph. The stronger binding of CO2 molecules to m-ZrO2 have been elucidated by different factors like higher concentration of OH groups and carbonate ions on the surface of monoclinic zirconia relative to tetragonal form [52,53]. Based on these observations, shifting of strong basic sites peak to higher temperature may also be due to occurrence of monoclinic polymorph as indicated by XRD patterns. This could also be supported by increased magnitude of total uptake of CO2 for CZC500 than CZC450 (Table 5). However, total quantity of CO2 desorbed for catalyst processed at highest calcination temperature of the current study was declined to almost 2/3 of the catalyst calcined at lowest temperature. The drop of CO2 uptake could be attributed to the growth of zirconia, associated with rise of calcination temperature. This also indicates that smaller size zirconia holds relatively higher concentration of basic sites. Similar observations of depressed CO2 adsorption as a function of calcination temperature were also reported by Li et al. [54]. Density of total acidic sites like density of basic sites was less affected by variation in calcination temperature. In conclusion, the overall results indicate that rise in calcination temperature less affected the strength of basic sites. Nevertheless, distribution of basic sites has been altered significantly. Moreover, reduction in basicity also indicates lowering of electron-donating potential of electron-rich Zr4þ, as higher basicity leads to boost the electron-donating properties of cation [53e55]. 3.2. Influence of calcination temperature on catalytic performance of CZC catalysts CNFs based Cu/ZrO2 catalysts calcined at different temperature were tested for methanol synthesis as well as CO2 conversion. The reaction data of all studied catalysts are listed in Table 6. Methanol synthesis rate was improved by raising calcination temperature from 350 to 450 C. However, it was depressed with further intensification in calcination temperature. Likewise, TOF value of methanol was progressively increased with increasing calcination temperature up to 500 C and was dropped with maximum degree of calcination temperature in this study. In contrast, CO2 hydrogenation was adversely affected by the increasing calcination temperature. The activity pattern of the catalysts calcined at different temperature could be justified by the physiochemical investigations, discussed in the earlier sections. A comparative study of the activity data of this novel catalyst with the reported literature revealed that the current CZC catalysts showed higher activity for methanol yield and CO2 conversion as compared to that recorded by Sloczynski et al. over Ag/ZnO/ZrO2 and Au/ZnO/ZrO2 catalysts [56]. Similarly, the results obtained in this study were very much comparable in terms of methanol yield and CO2 conversion to the work of Liu et al., where they carried out CO2 hydrogenation over Cu/Ga2O3/ZrO2 catalysts [57]. Likewise, magnitude of comparable activity data were reported for carbon nanotube-supported Pd/ZnO catalysts [58]. There has been a general consensus about the involvement of two active sites namely copper (Cu) and zirconia (ZrO2) for CO2 Table 6 Catalytic activity of calcined samples in slurry reactor.
Fig. 8. CO2-TPD profiles of (a) CZC350, (b) CZC450, (c) CZC500 and (d) CZC550 catalysts.
Catalyst
Meth. yield (g Kg1 cat. h1)
CO2 conversion (%)
TOFMeOH 103 (s1)
CZC350 CZC450 CZC500 CZC550
27 34 24 17
15 14 10 4
1.12 1.52 2.04 1.69
I. Ud Din et al. / Journal of Power Sources 274 (2015) 619e628
627
hydrogenation to methanol over Cu/ZrO2 catalysts [4,51,59]. More active form of copper is characterized to be the one with higher degree of dispersion and easily reducible. To demonstrate the relationship between methanol synthesis rate and fraction of dispersed Cu, a plot is presented in Fig. 9. As depicted in the figure, rate of methanol formation progressively increased with increasing fraction of dispersed Cu. Obviously, more dispersed Cu provides more number of active sites to the reacting molecules. Consequently, higher degree of dispersed Cu will lead to higher activity towards methanol formation. The second active site i.e. ZrO2 has been reported in two different polymorphic forms namely tetragonal and monoclinic form. Rhodes and Bell conducted a comparative study between tetragonal and monoclinic polymorphs of ZrO2 for methanol synthesis and concluded that latter polymorph of ZrO2 is 20 times more active than the former one [60]. In this work, activity of catalysts was also affected by phase of zirconia. Rate of methanol formation was enhanced from 27 to 34 g kg1 cat. h1 when calcination temperature was raised from 350 to 450 C. This could be due to monoclinic ZrO2 polymorph. Nevertheless, the trend of catalyst activity with variation in calcination temperature was not persisted any longer and rate of methanol formation was depressed on further increase in calcination temperature. This could be attributed to the growth and sintering of both Cu and ZrO2 active sites, as indicated by TEM observations. On the other hand, CO2 conversion was gradually decreased by intensifying degree of calcinations temperature, indicating lower activity of CZC catalysts for CO2 conversion as a function of increasing calcinations temperature. Crystallite size of Cu is also an important parameter for CO2 hydrogenation over Cu based catalysts. To elaborate the influence of Cu crystallite size on the rate of methanol formation and CO2 conversion, a graph is plotted and is shown in Fig. 10. As it is evident from the graph that rate of methanol formation is gradually reduced with increasing crystallite size of Cu with exception of CZC450 catalyst. On the other hand, CO2 conversion decreases linearly with the growth of Cu crystallite. A similar trend of lower activity with increase Cu crystallite size has also been reported elsewhere [61,62]. This is quite understandable as crystals with large size have less surface area and fewer defect sites as compared to small sized crystals. Moreover, smaller Cu particles tend to increase the interfacial area with the neighboring metal oxide [62]. By this mechanism small sized Cu particles promote a synergic effect, leading to increase in the overall performance of the catalyst. The variation of catalysts activities with respect to changes in Cu crystallite size also indicates that CO2 hydrogenation to methanol is a structure-sensitive reaction.
Based on the reported mechanistic studies, surface area of metallic copper (SCu) is considered as an important parameter in CO2 hydrogenation to methanol. Methanol yield has been reported to increase linearly with increasing SCu [63e65]. However, in the work of Sun et al., although methanol yield was increased with the increase of SCu but the relationship was not linear [26]. Nevertheless, the conflicting results of catalysts activity with respect to SCu are reported [25,66]. Despite all these discrepancies, there is a general consensus that higher value of SCu is favorable for overall activity of CO2 hydrogenation. In the recent study, the trend of methanol formation was not completely aligned with variation of SCu, as highest methanol formation 34 (g Kg1 cat. h1) was recorded for SCu ¼ 8.0 as compared to 27 (g Kg1 cat. h1) for SCu ¼ 8.6, as depicted by Fig. 11. On the other hand, a linear relationship was observed between SCu and CO2 conversion. The decline in methanol synthesis rate and continuous enhancement in CO2 conversion for catalyst with highest SCu clearly indicates decrease in methanol selectivity. CO formation by reverse water gas shift reaction could be the alternative pathway for CO2 conversion. Similar observations were also reported elsewhere for Cu/ZnO/ Al2O3 catalysts [26]. To further elaborate the relationship between catalysts activity and SCu, turnover frequency (TOF) for methanol production was calculated and is displayed in Table 6. According to Boudart's theory, a straight line between TOF and SCu should be obtained if the catalyst activity is only dependent on SCu [67,68]. However, in the current study there is a general trend of decreasing TOF as the SCu is
Fig. 9. Relationship between fraction of dispersed Cu and methanol synthesis rate.
Fig. 11. Relationship between Cu surface area and catalysts activity.
Fig. 10. Relationship between Cu size and catalysts activity.
628
I. Ud Din et al. / Journal of Power Sources 274 (2015) 619e628
increased. This further indicates that catalyst activity is not only related to SCu but other factors like degree of dispersion in this case can play a decisive role in determining the activity profile of catalyst. Recently, similar observations were also documented elsewhere [25,26]. Nevertheless, the overall activity of the catalyst i.e. CO2 conversion was found to be directly proportional to SCu. The same trend of CO2 conversion with respect to SCu was observed in the recent work of Guo et al. [44]. The increase in CO2 conversion with increasing SCu can be described as follows. Increase in SCu produces more atomic hydrogen on the surface of Cu which is supplied to ZrO2 sites by spillover effect. Subsequently with higher concentration of atomic hydrogen, more molecules of CO2 are utilized for hydrogenation [44]. 4. Conclusion The effect of calcination temperature on the physiochemical properties of CNFs based Cu/ZrO2 catalysts was investigated. Variation in degree of calcination temperature affected both physiochemical properties as well as activity profiles of catalyst remarkably. Surface area and fraction of dispersed Cu were enhanced when calcination temperature was raised from 350 to 450 C, which resulted in better performance of the catalyst. Nevertheless, growth and sintering of catalyst particles and reduction of both SCu and SBET surface area at elevated calcination temperature led to depress the overall activity of Cu.ZrO2/CNFs catalyst for methanol synthesis in slurry reactor. A linear relationship between catalyst activity towards CO2 conversion and its SCu was observed. Acknowledgment This work was supported by Ministry of Higher Education Malaysia, FRGS No:FRGS/1/2011/SG/UTP/02/13 and Universiti Teknologi PETRONAS, Malaysia. References [1] M. Tamura, K. Noro, M. Honda, Y. Nakagawa, K. Tomishige, Green Chem. 15 (2013) 1567e1577. [2] H. Zhan, F. Li, P. Gao, N. Zhao, F. Xiao, W. Wei, L. Zhong, Y. Sun, J. Power Sources 251 (2014) 113e121. [3] S. Shironita, K. Karasuda, M. Sato, M. Umeda, J. Power Sources 228 (2013) 68e74. [4] F. Arena, G. Italiano, K. Barbera, S. Bordiga, G. Bonura, L. Spadaro, F. Frusteri, Appl. Catal. A 350 (2008) 16e23. [5] M.S. Israf Ud Din, Duvvuri Subarao, Abdul Naeem, Appl. Mech. Mater. 246 (2014) 83e87. [6] J. Weigel, R.A. Koeppel, A. Baiker, A. Wokaun, Langmuir 12 (1996) 5319e5329. [7] C. Pham-Huu, N. Keller, L.J. Charbonniere, R. Ziessel, M.J. Ledoux, Chem. Commun. (2000) 1871e1872. [8] C. Pham-Huu, N. Keller, G. Ehret, L.C.J. Charbonniere, R. Ziessel, M.J. Ledoux, J. Mol. Catal. A Chem. 170 (2001) 155e163. [9] C.P.-H. Marc-Jacques Ledoux, Catal. Today 102e103 (2005) 2e14. [10] F. Salman, C. Park, R.T.K. Baker, Catal. Today 53 (1999) 385e394. [11] A. Chambers, T. Nemes, N.M. Rodriguez, R.T.K. Baker, J. Phys. Chem. B 102 (1998) 2251e2258. [12] T. Mahmood, J.O. Williams, R. Miles, B.D. McNicol, J. Catal. 72 (1981) 218e235. [13] M.-J. Ledoux, C. Pham-Huu, Catal. Today 102e103 (2005) 2e14. [14] S.W. Oh, H.J. Bang, Y.C. Bae, Y.-K. Sun, J. Power Sources 173 (2007) 502e509. [15] M.A.R.R. Irmawati, Y.H. Taufiq-Yap, Z. Zainal, Catal. Today 93e95 (2004) 631e637. [16] B.Y.J.Y.S. Al-Zeghayera, Appl. Catal. A 292 (2005) 287e294. [17] Y.Z. Yiwei Zhanga, Yian Lia, Yu Wangb, Yi Xub, Peicheng Wub, Catal. Commun. 8 (2007) 1009e1016. [18] M.L. Toebes, M.K. van der Lee, L.M. Tang, M.H. Huis, T. Veld, J.H. Bitter, A.J. van Dillen, K.P. de Jong, J. Phys. Chem. B 108 (2004) 11611e11619. [19] M.K. van der Lee, J. van Dillen, J.H. Bitter, K.P. de Jong, J. Am. Coll. Surg. 127 (2005) 13573e13582. [20] F. Arena, K. Barbera, G. Italiano, G. Bonura, L. Spadaro, F. Frusteri, J. Catal. 249 (2007) 185e194. [21] X.-L. Du, Q.-Y. Bi, Y.-M. Liu, Y. Cao, H.-Y. He, K.-N. Fan, Green Chem. 14 (2012) 935e939.
n-Cabrera, J.L.G. Fierro, Appl. Catal. A 253 [22] J. Agrell, M. Boutonnet, I. Melia (2003) 201e211. nez, R. Knapp, X. Lin, T.E. Müller, A. Türler, B. Wierczinski, [23] C. Sievers, O. Jime J.A. Lercher, J. Mol. Catal. A Chem. 279 (2008) 187e199. [24] S. Barradas, E.A. Caricato, P.J. van Berge, J. van de Loosdrecht, Elsevier, Stud. Surf. Sci. Catal. (2000) 55e65. [25] H. Ahouari, A. Soualah, A. Le Valant, L. Pinard, P. Magnoux, Y. Pouilloux, React. Kinet. Catal. Lett. 110 (2013) 131e145. [26] Q. Sun, Y.-L. Zhang, H.-Y. Chen, J.-F. Deng, D. Wu, S.-Y. Chen, J. Catal. 167 (1997) 92e105. [27] B. Neppolian, Q. Wang, H. Yamashita, H. Choi, Appl. Catal. A Gen. 333 (2007) 264e271. [28] B. Vahid, N. Saghatoleslami, H. Nayebzadeh, A. Maskooki, Chem. Biochem. Eng. Q. 26 (2012) 71e77. rska, A. Zaleska, E. Kowalska, T. Klimczuk, J.W. Sobczak, E. Skwarek, [29] P. Go W. Janusz, J. Hupka, Appl. Catal. B 84 (2008) 440e447. [30] S.T. Choo, Y.G. Lee, I.-S. Nam, S.-W. Ham, J.-B. Lee, Appl. Catal. A Gen. 200 (2000) 177e188. [31] P.D.L. Mercera, J.G. van Ommen, E.B.M. Doesburg, A.J. Burggraaf, J.R.H. Ross, Appl. Catal. 71 (1991) 363e391. [32] D.Y. Yoon, J.-H. Park, H.-C. Kang, P.S. Kim, I.-S. Nam, G.K. Yeo, J.K. Kil, M.-S. Cha, Appl. Catal. B 101 (2011) 275e282. ska[33] A. Cybula, J.B. Priebe, M.-M. Pohl, J.W. Sobczak, M. Schneider, A. Zielin Jurek, A. Brückner, A. Zaleska, Appl. Catal. B 152e153 (2014) 202e211. [34] B. Valle, B. Aramburu, A. Remiro, J. Bilbao, A.G. Gayubo, Appl. Catal. B 147 (2014) 402e410. [35] M.L.W. Victoria, J.S. Kevin, The Effect of Calcination Temperature on the Properties and Hydrodeoxygenation Activity of Ni2P Catalysts Prepared Using Citric Acid, Novel Materials for Catalysis and Fuels Processing, American Chemical Society, 2013, pp. 287e300. [36] W.P. Dow, T.J. Huang, J. Catal. 147 (1994) 322e332. [37] G.L. Bezemer, P.B. Radstake, U. Falke, H. Oosterbeek, H.P.C.E. Kuipers, A.J. van Dillen, K.P. de Jong, J. Catal. 237 (2006) 152e161. [38] C.D. Taboada, J. Batista, A. Pintar, J. Levec, Appl. Catal. B 89 (2009) 375e382. nchez, A. Romero, J.L. Valverde, Catal. Commun. [39] J.A. Díaz, A.R. de la Osa, P. Sa 44 (2014) 57e61. [40] X. Yao, Q. Yu, Z. Ji, Y. Lv, Y. Cao, C. Tang, F. Gao, L. Dong, Y. Chen, Appl. Catal. B 130e131 (2013) 293e304. [41] H.-D. Zhuang, S.-F. Bai, X.-M. Liu, Z.-F. Yan, J. Fuel Chem. Technol. 38 (2010) 462e467. [42] C.-M. Hung, Powder Technol. 191 (2009) 21e26. [43] Y. Matsumura, J. Power Sources 238 (2013) 109e116. [44] X. Guo, D. Mao, G. Lu, S. Wang, G. Wu, J. Mol. Catal. A Chem. 345 (2011) 60e68. [45] E.-Y. Choi, I.-S. Nam, Y.G. Kim, J. Catal. 161 (1996) 597e604. [46] K.T.K. Basavaiah, N. Rajendra Prasad, S.G. Hiriyanna, K.B. Vinay, Indian J. Chem. Technol. 16 (2009) 272e277. [47] N. Li, G.A. Tompsett, T. Zhang, J. Shi, C.E. Wyman, G.W. Huber, Green Chem. 13 (2011) 91e101. [48] Z.-Y. Ma, C. Yang, W. Wei, W.-H. Li, Y.-H. Sun, J. Mol. Catal. A Chem. 227 (2005) 119e124. [49] Y. Zhao, W. Li, M. Zhang, K. Tao, Catal. Commun. 3 (2002) 239e245. [50] I.A. Fisher, A.T. Bell, J. Catal. 178 (1998) 153e173. [51] I.A. Fisher, A.T. Bell, J. Catal. 172 (1997) 222e237. [52] K. Pokrovski, K.T. Jung, A.T. Bell, Langmuir 17 (2001) 4297e4303. [53] B. Bachiller-Baeza, I. Rodriguez-Ramos, A. Guerrero-Ruiz, Langmuir 14 (1998) 3556e3564. [54] N.T. Juan Li, Wei Song, Ensheng Zhan, Wenjie Shen, Gold Bull. 42 (2009) 48e60. €nkko € nen, H. Bradshaw, H. Stephenson, S. Airaksinen, [55] T. Viinikainen, H. Ro M. Reinikainen, P. Simell, O. Krause, Appl. Catal. A 362 (2009) 169e177. ski, R. Grabowski, A. Kozłowska, P. Olszewski, J. Stoch, J. Skrzypek, [56] J. Słoczyn M. Lachowska, Appl. Catal. A 278 (2004) 11e23. [57] X.-M. Liu, G.Q. Lu, Z.-F. Yan, Appl. Catal. A 279 (2005) 241e245. [58] X.-L. Liang, X. Dong, G.-D. Lin, H.-B. Zhang, Appl. Catal. B Environ. 88 (2009) 315e322. [59] D. Bianchi, T. Chafik, M. Khalfallah, S.J. Teichner, Appl. Catal. A 123 (1995) 89e110. [60] M.D. Rhodes, A.T. Bell, J. Catal. 233 (2005) 198e209. €vecker, F. Abild-Pedersen, [61] M. Behrens, F. Studt, I. Kasatkin, S. Kühl, M. Ha S. Zander, F. Girgsdies, P. Kurr, B.-L. Kniep, M. Tovar, R.W. Fischer, J.K. Nørskov, € gl, Science 336 (2012) 893e897. R. Schlo [62] S. Natesakhawat, J.W. Lekse, J.P. Baltrus, P.R. Ohodnicki, B.H. Howard, X. Deng, C. Matranga, ACS Catal. 2 (2012) 1667e1676. [63] M. Saito, T. Fujitani, M. Takeuchi, T. Watanabe, Appl. Catal. A Gen. 138 (1996) 311e318. [64] R.A. Koeppel, A. Baiker, C. Schild, A. Wokaun, Effect of preparation variables on catalytic behaviour of copper/zirconia catalysts for the synthesis of methanol from carbon dioxide, in: P.A.J.P.G.G. Poncelet, B. Delmon (Eds.), Stud. Surf. Sci. Catal., Elsevier, 1991, pp. 59e68. [65] C. Baltes, S. Vukojevi c, F. Schüth, J. Catal. 258 (2008) 334e344. [66] H. Berndt, V. Briehn, S. Evert, Appl. Catal. A 86 (1992) 65e69. [67] X. Guo, D. Mao, G. Lu, S. Wang, G. Wu, J. Catal. 271 (2010) 178e185. [68] S. Bailey, K.C. Waugh, Catal. Lett. 17 (1993) 371e374.
