Contents lists available at ScienceDirect

Chemical Engineering Journal

journal homepage: www.elsevier.com/locate/cej

Carbon nanofibers based copper/zirconia catalysts for carbon dioxidehydrogenation to methanol: Effect of copper concentration

Israf Ud Dina,d,⁎, Maizatul S. Shaharunb,⁎, A. Naeemc, S. Tasleemd, Mohd Rafie Johana

a Nanotechnology & Catalysis Research Centre (NANOCAT), Institute of Postgraduate Studies, University of Malaya, 50603 Kuala Lumpur, Malaysiab Department of Fundamental and Applied Sciences, Universiti Teknologi PETRONAS, Malaysiac National Centre of Excellence in Physical Chemistry, University of Peshawar, Pakistand Chemistry Department, Kohat University of Science and Technology, Kohat, Pakistan

A R T I C L E I N F O

Keywords:Methanol synthesisSlurry reactorCu loadingCNFZirconia

A B S T R A C T

A series of novel bimetallic copper/zirconia carbon nanofibers supported catalysts with different Cu contents(5–25 wt%) were synthesized via deposition precipitation method. The physicochemical characterization of thecalcined catalysts was carried out by X-ray diffraction, inductively coupled plasma optical emission spectro-scopy, N2 adsorption–desorption, N2O chemisorption, temperature programmed reduction, X-ray photoelectronspectroscopy, high resolution transmission electron microscopy and temperature programmed CO2 desorption.Structure-reactivity correlation for catalytic hydrogenation of CO2 to methanol was discussed in details. Reactionstudies revealed 15 wt% as optimum Cu concentration for CO2 conversion to methanol with CO2/H2 feed volumeratio of 1:3. Cu surface area was found to play a vital role in methanol synthesis rate. CO2 conversion wasobserved to be directly proportional to the number of total basic sites. A comparative study of this novel catalystwith the recently reported data revealed the better CO2 conversion at relatively low reaction temperature.

1. Introduction

Lifestyle quality has been improved by incessant technologicalprogress over the last century. At the same time, it has produced pro-gressive depletion of natural resources. Consequently, natural en-vironment has witnessed numerous catastrophic phenomenon likeglobal warming [1,2]. Being a major contributor to the phenomenon,many countries around the globe has pursued to reduce carbon dioxideconcentrations in the environment. Numerous ways have been pro-posed to address the problem. CO2 recycling is one of the viable stra-tegies that will reduce CO2 concentrations and at the same time pro-duce value-added products such as methanol [3–5].

Currently, methanol synthesis is largely run by a mixture of syngasand CO2 over on Cu–ZnO/Al2O3. However, recent mechanistic in-vestigations revealed that CO2 is the major contributor to methanolsynthesis while CO acts as an inhibitor to the active metal sites [6,7].Furthermore, CO2 hydrogenation has been declared as superior routethan conventional syngas method in terms of carbon utilization andreaction rates [7–9]. Current methanol synthesis catalysts have draw-backs like low mechanical strength, poor thermal stability, toxicity andshort lifetime [10,11]. Similarly, strong hydrophilic character of Al2O3

is one of the reasons for lower activity of current methanol synthesis

catalyst towards CO2 hydrogenation route [12–14]. Until now, an ap-propriate catalyst could not be designed due to lack of understanding ofthe CO2 hydrogenation reaction mechanism [15]. More recently at-tention has focused on several aspects of the catalyst system. Catalystsupport is one of the components that were extensively investigated inCO2 hydrogenation [16,17]. However, catalyst support that couldwithstand the reaction conditions and effectively disperse active metalshas not been established.

Carbon materials such as carbon nanotubes, graphene oxide andcarbon nanofibers are good candidates as the catalyst supports due totheir attractive physical and thermal properties. Carbon nanotubes andgraphene oxides have been studied as catalyst supports for CO2 hy-drogenation to methanol [18,19]. The economic studies on applicationof carbon nanofibers also revealed that carbon nanofibers is cost ef-fective than carbon nanotubes, as the market price for carbon nano-tubes is 100 times higher as compared to carbon nanofibers [20].Carbon nanofibers (CNFs) due to their special physiochemical char-acteristics such as tuneable and controllable texture are considered asexcellent catalyst support [21]. Furthermore, CNFs being pure, chemi-cally inert and mechanically strong surpass other traditional catalystsupports like silica and alumina [22]. Moreover, in liquid phase reac-tions mass transfer plays a major role in determining the performance

http://dx.doi.org/10.1016/j.cej.2017.10.087Received 5 July 2017; Received in revised form 10 October 2017; Accepted 16 October 2017

⁎ Corresponding authors at: Nanotechnology & Catalysis Research Centre (NANOCAT), Institute of Postgraduate Studies, University of Malaya, 50603 Kuala Lumpur, Malaysia (Israf UdDin).

E-mail addresses: drisraf@yahoo.com (I.U. Din), maizats@petronas.com.my (M.S. Shaharun).

Chemical Engineering Journal 334 (2018) 619–629

Available online 18 October 20171385-8947/ © 2017 Elsevier B.V. All rights reserved.

MARK

of the catalyst. In this regard, CNFs due to the higher porosity, lowertortuosity and absence of micro pores reduce mass transfer limitations[23,24]. CNFs being a combination of both active carbon and stablegraphite, are considered as excellent material for catalyst support.Moreover, the appreciable stability of carbon materials in acidic as wellas in basic conditions makes CNFs as suitable candidate for catalystssupport. High external surface area and special morphology of CNFsallow rapid access of reactant molecules to the active metal sites. Si-milarly, complete absence of bottled pores in CNFs, empower 3D ap-proach of the reactants to the entire volume of catalyst. Furthermore,CNFs due to higher thermal conductivity, evacuate heat generatedduring reaction. Consequently, CNFs based catalysts have recordedbetter performance as compared to traditional supported based cata-lysts [25].

Although there is a general consensus on the role of Cu as an activesite for H2 dissociation in CO2 hydrogenation to methanol [26–28], theoptimum amount of Cu is not univocal. Being an active site, increase inCu content improves catalyst performance but there is always an op-timum amount above which catalyst activity starts to decay. Variousoptimum concentrations of Cu have been reported for different types ofcatalyst. Bell studied the effect of Cu loadings on Cu/ZrO2 catalyst forCO2 hydrogenation to methanol and found that catalyst activity wasinitially increased with increasing Cu content up to 15 wt% however,further addition of Cu reduced the activity profile [29]. The decrease inactivity with higher Cu loading was attributed to the reduction in metaldispersion.

Choice of preparation method not only affects the physiochemicalparameters of catalyst but also marks the overall activity profile ofcatalyst. Catalyst prepared by Deposition Precipitation (DP) methodshowed better surface area and better performance towards methanolsynthesis as compared to catalysts prepared by co-precipitation (CP)and impregnation method (IM) [30]. Moreover, catalyst prepared by DPmethod displayed greater yield of methanol with much higher se-lectivity and higher conversion of CO2 as compared to catalysts pre-pared by conventional co-precipitated method [31]. In the currentstudy DP method has been employed for the synthesis of catalysts.

CO2 hydrogenation has generally been conducted in slurry reactorsand fixed bed reactors. However due to exothermic nature of the re-action, the reaction is thermodynamically favored in slurry reactor.Similarly, a comparative study of slurry reactor and fixed bed reactorfor CO2 hydrogenation was conducted by Kim et al. [32]. They con-cluded that application of slurry reactor increased more than two foldsconversion of CO2 as compared to fixed bed reactor.

The aim of the current study, is to investigate the influence of Culoading on the physicochemical and activity profile of CNF supportedCu/ZrO2 catalyst. The physicochemical parameters of catalysts werestudied using various characterization techniques and the catalyticactivity was investigated in three-phase slurry reactor.

2. Experimental

2.1. Material and methods

In the current work, Carbon nanofibers were purchased from Korea,Cu (NO3)2·3H2O from R&M Chemicals. Similarly, zirconyl nitrate hy-drate (H2N2O8Zr) was procured from Sigma-Aldrich.

2.2. Activation of CNFs

Prior to use as catalyst support, CNFs were activated by treatingwith 35 vol% solution of nitric acid. The as-received CNFs were re-fluxed with nitric acid at 90 °C for 16 h. After the oxidation, the solutiontemperature was brought to room temperature and filtered by vacuumfiltration. Treated CNFs were washed and dried overnight at 110 °C.The treatment resulted in conversion of CNFs to oxidized CNFs (CNFs-O).

2.3. Catalysts synthesis

CNFs-O based Cu/ZrO2 catalysts with different Cu loadings weresynthesized by deposition precipitation method (DP method). Requiredquantity of Cu (NO3)2·3H2O was first dissolved in distilled water. Thiswas followed by addition of required amount of zirconyl nitrate hydrate(H2N2O8Zr) to the solution. When both precursors were completelydissolved, oxidized CNFs-O was added to the precursor’s solution.Solution temperature was raised to 90 °C. Urea solution (1 g of ureadissolved in 10ml distilled water) was poured to the stirring solution asa precipitating agent. The process was continued for 20 h. After cooling,solution was filtered by vacuum filtration. Then precipitates were driedovernight at 110 °C. The dried catalysts were calcined in N2 flow at350 °C for 3 h. The prepared catalysts were assigned as CZCx, where “x”is the wt% of Cu in the CNFs-O based Cu/ZrO2 catalysts. Catalyst weresynthesized with variant Cu content of 5, 10, 15, 20 and 25wt% at afixed ZrO2 content (10 wt%) for each catalyst.

2.4. Catalysts characterization

Concentration of bulk phase Cu was confirmed by inductively cou-pled plasma optical emission spectroscopy (ICP-OES). First the sampleswere dissolved in aqua regia. To ensure complete dissolution of samplematrix, samples were digested in high performance microwave diges-tion system (Milestone).

Crystallographic properties of synthesized catalysts were in-vestigated by X-ray diffractometer (PANalytical model Empyrean).PANalytical HighScore Plus software was used for evaluation of XRDdata. The XRD analysis was conducted at room temperature from 20° to80° at 2θ Bragg angle.

Morphology of the catalysts was investigated by transmission elec-tron microscopy. The Zeiss LIBRA 200TEM with accelerating Voltage:200 kV was employed for this purpose.

Nitrogen adsorption–desorption technique was utilized to determinethe pore volume and surface area of prepared catalysts. MicrometricsASAP 2020 was used to evaluate the textural properties while BJHmethod was employed to investigate pore size distribution [9].

The N2O chemisorption technique was carried out to assess surfacearea (SCu), particle size (dCu), dispersion (DCu), and distribution (RCu) ofmetallic copper [33]. Initially, catalysts were reduced in H2 flow at500 °C. The reduced samples were cooled in He flow to 60 °C andpurged for 30min. The N2O gas was introduced for 1 h at temperatureof 60 °C. Samples were flushed with He flow for 1 h to remove residualN2O. The samples were reduced for the second time at 500 °C. Themagnitudes of SCu and DCu were calculated by assuming stoichiometryof Cu:N2O=2 with surface atomic density of 1.46×1019 Cuat/m2. ThedCu was obtained by a formula given below [9,34]:

=d nmD

( ) 104(%)Cu

Cu

The values of RCu were estimated by the following relationship [35].

=

×

RMetallic Cu surface area

Total Cu content BET surface areaCu

Metal support interaction and reduction behavior of catalysts wereidentified by Temperature Programmed Reduction (TPR). For thispurpose, TPDRO1100 MS equipped with thermal conductivity detector(TCD) was utilized for the reduction studies. TPR analysis was con-ducted from 30 to 800 °C temperature range with 10 °Cmin−1 heatingrate. Samples were reduced in flow of 5 vol% H2/N2.

Surface composition and chemical state of copper was investigatedby X-ray photoelectron spectroscopy (XPS). X-ray photoelectron spec-troscope (XPS, Thermo-Fisher K-Alpha) was employed for this study.Peak fitting was performed by advantage software for identification ofdifferent chemical state of catalyst components.

Basicity profile of catalysts was evaluated by CO2 temperature

I.U. Din et al. Chemical Engineering Journal 334 (2018) 619–629

620

programmed desorption. Before TPD studies, samples were treated attemperature of 500 °C in inert atmosphere for 60min. The catalystswere saturated with pure CO2 at room temperature. Adsorption of CO2

was continued at 90 °C for one hour. The desorption studies of CO2 wereconducted from 40 to 800 °C temperature range.

2.5. Activity studies

Catalysts activity was evaluated for CO2 hydrogenation to methanolin autoclave slurry reactor Parr 4593. The detailed procedure of me-thanol synthesis has been reported elsewhere [25]. Briefly, a reducedcatalyst 0.5 g was placed in in reactor vessel contain 25ml ethanol asreaction solvent. Prior to reaction studies, reactor was purged at roomtemperature with reactant gases. The reactor was pressurized to 30 barswith a mixture of H2/CO2 gases with 3:1 M ratio. Reaction was con-tinued for 2 h at 180 °C. Liquid contents were examined by flame io-nization detector (FID) while gas products were quantified by thermalconductivity detector (TCD).

Rate of methanol synthesis was calculated as follows.

= ×Rate of methanol yield g of methanol produced Kg of catalyst h/

Methanol Turnover frequency (TOF) was calculated as follows[36,37]:

=

×

=×

× ×

−TOF sNumber of methanol molecules produced

Time s Number of metallic Cu atomsA N

S N

( )( )

3600

MeOH

A

Cu a

1

where A represents activity in mol/g.h, NA is Avogadro's number, SCudenotes metallic Cu surface area in m2/g and Na designates atomiccupper surface density (Na =1.469× 1019 atoms/m2)

3. Results and discussions

3.1. Influence of Cu content on physicochemical properties of CZC catalysts

3.1.1. ICP-OES investigationsCu loadings were quantified via ICP-OES and the data were pre-

sented in Table 1. As indicated by tabulated data, the magnitudes ofbulk Cu measured by ICP, were in very close agreement to the targetedvalues. This showed that the majority of active catalyst componentswere successfully deposited on the catalyst support. This in turn alsosupported the efficiency of DP method for synthesis of catalysts withhigh loadings of active metals.

3.1. 2. Phase determinationFig. 1 displays the XRD patterns of calcined CNFs-O based Cu/ZrO2

catalysts with different Cu loadings. For comparison, the diffractionpattern of bare CNFs-O was also incorporated. The two major peakswere detected on 2θ scale at 26° and 44°, indicating the diffractions of(0 0 2) and (1 0 0) planes of graphitic carbon nanofibers [38–41].

A small diffraction peak visualized at 2θ=30.3° was indexed astetragonal zirconia (t-ZrO2, JCPDS 88-1007) whereas a small XRD peakat 2θ=24.4° was recognized as monoclinic zirconia (m-ZrO2) [42]. X-ray diffraction pattern at 32.6°, 35.5°, 38.7°, 48.8°, 58.3°, 61.67°, 66.4°,68.1°, 72.3° and 75.1° was found and indexed as monoclinic phased

tenorite CuO with JCPDS card files No. 48-1548 (a= 4.62 Å,b=3.43 Å, and c=5.06 Å).

These peaks were intensified when Cu loading was increased from 5to 10wt%. This implied that the degree of CuO crystallization wasenhanced by increasing the content of Cu. Another interesting ob-servation was the intensification of t-ZrO2 with the increase of Culoadings. On the other hand, XRD peak of m-ZrO2 (2θ=24.4°) observedin CZC5 catalyst disappeared in CZC10 catalysts. The disappearance ofm-ZrO2 and subsequent increment of t-ZrO2 indicated that polymorphictransformation of zirconia had been facilitated by increased of Culoading. In contrast, on further increment of Cu (10–15 wt%) the in-tensity of t-ZrO2 peak was weakened and the XRD peak of m-ZrO2 re-emerged. This lowering intensity of t-ZrO2 peak and subsequent re-appearance of m-ZrO2 suggested the back transformation of t-ZrO2 to m-ZrO2 form. Furthermore, the degree of CuO crystallization increasedwhen Cu content exceeded 15wt%, while both zirconia phases almostdisappeared. This could be due to stabilization of zirconia by Cu, aspreviously reported in the literature [27]. Such observations were alsorecorded by Ko et al. for Cu/ZrO2 catalyst with increasing Cu content[27]. Nevertheless, intensity of CuO peak was reduced for CZC catalystwith maximum Cu content. The decrease in CuO peak intensity withincreasing in Cu content shows the conversion of crystalline CuO toamorphous form. Such observations were also recorded by Rajaramet al. [43]. More importantly, the CuO peak was slightly broadenedwith the increase of Cu from 5 to 10wt%, nevertheless, the peak widthwas progressively narrowed down as a function of increasing Cu load-ings. This behavior of CuO, as a consequent of Cu loadings advocatedagglomeration of CuO, which is in line with the literature [44].

3.1.3. TEM morphologyTEM images of CZC catalysts with different CuO loadings are shown

in Fig. 2. The spherical shaped dark black particles were recognized ascopper, while zirconia particles were identified as the one with

Table 1Bulk compositions, BET surface areas and N2O chemisorption data of CZC catalysts.

Sample Targeted Cu loading Cu loadings by ICP-OES SBET (m2/g) pore volume (cm3/g) pore diameter (nm) SCu (m2/g)N2O DCu (%)N2O dCu (nm) N2O RCuN2O

CZC5 5 4.54 109 0.29 10.7 5.7 17.7 5.9 1.0CZC10 10 9.5 123 0.32 10.4 9.8 13.25 7.8 0.8CZC15 15 14.1 133 0.36 10.3 13.1 16.0 6.5 0.6CZC20 20 19.2 92 0.29 12.9 12.3 11.0 9.5 0.6CZC25 25 24.8 88 0.27 12.4 11.0 8.5 12.2 0.5

Fig. 1. XRD pattern of (a) CNFs-O, (b) CZC5, (c) CZC10, (d) CZC15, (e) CZC20 and (f)CZC25 catalysts.

I.U. Din et al. Chemical Engineering Journal 334 (2018) 619–629

621

tetragonal in shape [45–47]. Besides, TEM investigations revealed well-dispersed particles for both Cu and ZrO2.

Furthermore, a uniform particle distribution was observed for cat-alysts with lower loading of Cu and most of the CNFs surface laid bare,indicating a higher potential of catalyst support for anchoring catalystsparticles. This uniformity and dispersion of catalysts components hadbeen persisted with further Cu loadings, while an optimum dispersionwas observed in the case of CZC15 catalysts. However, both dispersionand particle size were adversely affected with the excess of Cu loadings.This is described by the densely populated and poorly dispersed catalystparticles with further increase in Cu content.

3.1.4. Textural propertiesThe isotherms of CZC catalyst with different Cu showed type-IV

isotherms along with H4 type hysteresis loops with a sharp inflection inp/po ranges of 0.83–0.96 (Fig. 3).

This reveals that catalyst support is mesoporous in nature. The BETsurface area and pore size volume are listed in Table 1.

Both BET surface areas, as well as pore volume, increased with theincrease of Cu loading from 5 to 15wt%. The enlargement of surfacearea with the increase of Cu content was also observed by Grift et al.[48]. However, both BET surface areas and pore volume reduced pro-gressively with further Cu loadings and had the lowest values of 88m2/g and 0.27 cm3/g, respectively with maximum Cu content of 25 wt%.The decline in surface area and pore volume indicated a partialblockage of the support pores at higher Cu loadings. The trend of

surface area and pore volume as a function of increasing Cu contentcould also be due to copper agglomeration at Cu content> 15wt%[44]. This conclusion was further demonstrated by XRD results. Con-trary to the BET surface area, the average pore diameter graduallydecreased with the increase of Cu loadings and increased when Cucontent is greater than 15wt%.

3.1.5. Copper surface area and dispersionThe metallic copper surface area (SCu), average particle size (dCu),

dispersion (DCu), and relative distribution (RCu) were measured via N2Ochemisorption (Table 1). As the tabulated data indicates, increasing Cucontent had influenced surface properties of Cu significantly. Besides,the SCu was progressively increased by incorporating Cu content inparent catalyst and an optimum SCu was obtained for CZC15 catalyst.However, the metallic surface area was depressed with further rise inconcentration of Cu content. These observations clearly suggested ag-glomeration of CuO at Cu content≥15wt%. This was also supported byTEM and XRD results of the respective catalysts. Such a correlationbetween Cu concentration and SCu was also reported elsewhere [33].On the other hand, DCu initially decreased, and then, increased with theincrease in Cu loading from 5 to 10wt% and 10–15wt%, respectively.Generally, a linear relationship is observed for surface area and dis-persion of copper. In the current case, the correlation remained in linewith exception of CZC5 catalyst, whereby low Cu surface area waswitnessed despite of higher Cu dispersion. The greater interaction be-tween Cu and ZrO2 components could be one of the reasons. More

50 nm 50 nm 50 nm

50 nm 50 nm

)c()b()a(

(d) (e)

Zirconia Particle

Copper Particle

Copper Particle

Zirconia Particle

Zirconia Particle

Zirconia Particle

Copper Particle

Copper Particle

Copper Particle

Zirconia Particle

Fig. 2. TEM images of (a) CZC5, (b) CZC10, (c) CZC15, (d) CZC20 and (e) CZC25 catalysts.

I.U. Din et al. Chemical Engineering Journal 334 (2018) 619–629

622

recently, similar observations were also recorded by Freitas et al. forCu/ZrO2 catalyst with varying Cu content [44]. In contrast, the mag-nitude of dCu showed an opposite trend with corresponding catalysts.This had been quite obvious as smaller particles exhibited higher dis-persion and vice versa. The growth in particle size and the concomitantdecline in dispersion with Cu content beyond 15 wt% was a clearmanifestation of Cu agglomeration with higher loadings. These resultswere further justified by TEM images. Similarly, distribution of metallicCu (RCu) was affected by Cu content in the parent catalyst. In addition,value of unity was observed for catalyst with the lowest Cu con-centration and it gradually decreased with the increase of Cu content.This implied that Cu was embedded partially in the ZrO2 phase and thephenomenon continued until the lowest RCu value was recorded forcatalysts with maximum Cu loadings. Besides, it could be further in-ferred that increasing Cu content and consequent embedding providedless exposed surface copper and more Cu–ZrO2 interfacial surface area.Moreover, this trend also revealed better interactions in Cu–ZrO2 as aconsequent of increasing Cu content.

3.1.6. ReducibilityH2-TPR analysis was conducted to identify and quantify copper

species in the catalysts. H2-TPR profile of Cu loaded catalysts is given inFig. 4, while total H2-uptake, extent of reduction (H2/Cu), position and

Fig. 3. N2 adsorption-desorption isotherms of (a) CZC5, (b) CZC10, (c) CZC15, (d) CZC20 and (e) CZC25 catalysts.

Temperature ( oC)

H2

Con

sum

ptio

n(a

.u.)

Fig. 4. H2-TPR profile of (a) CZC5, (b) CZC10, (c) CZC15, (d) CZC20 and (e) CZC25catalysts.

I.U. Din et al. Chemical Engineering Journal 334 (2018) 619–629

623

distribution of reduction peaks and relative abundance of dispersedcopper are listed in Table 2.

Two reduction peaks were observed in the TPR profile of CZC cat-alysts, which had been consistently reported for Cu/ZrO2 system. Thereduction peak at lower temperature range of 250–280 °C (peak α) wasascribed to dispersed copper, while the reduction peak observed intemperature range of 350–500 °C (peak β) was due to bulk phase copper[49]. Furthermore, the presence of two reduction peaks has also beenrecorded and described as stepwise reduction of copper (Cu2+→Cu+→ Cu) [50]. Likewise, lower reduction peak indicated strongerinteraction between Cu and ZrO2, while the reduction peak at highertemperature was designated to weaker interaction between the twointeracting metal oxides. In the current study, both reduction peakswere slightly shifted to relatively lower temperature by increasing Culoading up to 15wt% (Table 2). The trend of reduction peaks shifting tolower temperature with increasing Cu content was due to the formationof dimeric Cu-species [51].

Moreover, shifting of reduction peaks with respect to increasing Culoading clearly showed that Cu reduction was facilitated by promotionof Cu content. On the other hand, reduction maxima of both peaks wereshifted to higher temperature when Cu content exceeded above 15wt%.

This in turn, indicated diminishing interaction between Cu and ZrO2 asa consequent of increasing Cu content.

Obviously, the magnitude of total H2-uptake increased with in-creasing Cu loading. However, the extent of Cu reduction (H2/Cu) is animportant parameter to assess the fraction of reducible form of CuO insuch cases. Besides, H2/Cu ratio was calculated for each catalyst, as-suming that Cu was present as CuO in each catalyst and completelyreduced to Cu. In addition, based on stoichiometry of CuO reduction(CuO+H2→ Cu+H2O), a value of H2/Cu near 1 would mean that Cuis entirely present in Cu2+ state and had been completely reduced toCu, while H2/Cu ratio near 0.5 would show that Cu+ is the dominantstate of Cu or it is not completely reduced. Although no proper trend inthe extent of Cu reduction was observed in the study, the H2/Cu ratiowas improved from 0.90 to 1.11 by increasing the Cu content from 5 to10wt%. Meanwhile, in the Cu/ZrO2 system, it had been reported thatsome quantity of Cu2+ was incorporated in Zr4+ lattices, which wasdifficult to be reduced [27]. In this regard, improvement of Cu reduc-tion can be justified with the increment of Cu fraction outside theselattices by increasing Cu content from 5 to 10wt%. Approximately,100% Cu reduction was obtained for catalysts up to 15 wt% Cu loading,which not only confirmed our assumption in calculating the extent ofCu reduction, but it also revealed that Cu2+ was completely reduced toCu instead of Cu+. Similar observations were also recorded by Lo’pez-Sua’rez et al. in Cu/Al2O3 catalyst [52]. Nevertheless, H2/Cu ratio wasdeclined when Cu exceeded beyond 15wt%. The decline in H2/Cu ratioclearly indicated the agglomeration of copper.

Furthermore, the distribution of reducible copper was also affectedby Cu loading. Peak α remained the major reduction peak in H2-TPRprofile of catalysts with Cu loading up to 15wt%. Nevertheless, theintensity of peak α was reduced, and consequently, peak β was en-hanced as the Cu content was increased beyond 15 wt%. This resultindicated that Cu experiences phase transformation from Cu2+ to Cu+

by further increase of Cu loading. The high dispersion of CuO remainedconstant in the range of 68–72% until 15 wt% of Cu loading. Above15wt%, the CuO dispersion dropped to 23% for Cu loading of 25 wt%.This behavior further supports the results of N2O chemisorption

Table 2H2-TPR data of CZC catalysts with different Cu content.

Sample H2Consumed(µmoles/g)

H2/Cu Red. Temp.(°C) H2 Consumed(µmoles/g)

A (α)/(A(α)+A(β))a (%)

Peak α Peak β Peakα

Peakβ

CZC5 709 0.90 268 385 503 205 71CZC10 1752 1.11 254 333 1260 491 72CZC15 2145 0.90 258 337 1455 689 68CZC20 2715 0.86 256 474 1038 1676 38CZC25 3162 0.80 283 400 734 2427 23

a fraction of highly dispersed Cu.

Fig. 5. XPS of (A) Cu 2p of (a) CZC5, (b) CZC10, (c) CZC15, (d) CZC20 and (e) CZC25 catalysts and (B) Zr 3d of (a) CZC5, (b) CZC10, (c) CZC15, (d) CZC20 and (e) CZC25 catalysts.

I.U. Din et al. Chemical Engineering Journal 334 (2018) 619–629

624

reported earlier.In the current case TPR findings are summarized as

1. CuO is completely reduced as indicated by H2/Cu values at Table 2.2. Stepwise reduction of copper (Cu2+→ Cu+→ Cu) was revealed by

TPR investigations, as confirmed by XPS studies.3. Highly dispersed Cu was dominant with catalyst up to 15 wt% Cu

loading, however it was adversely effected by further Cu promotion.

3.1.7. Chemical state of catalysts componentsXPS spectra of Cu 2p and Zr 3d for catalysts having different Cu

content are shown in Fig. 5A and B.Each catalyst exhibited Cu 2p3/2 parental peak at 934.4 eV accom-

panied by a broad satellite peak around 943 eV and another parentalpeak for Cu 2p1/2 at 954 eV accompanying with a shakeup peak ataround 962 eV. The existence of satellite peaks along with parentalpeaks is a clear indicator that Cu predominantly existed as Cu2+ in allcalcined samples. More importantly, Cu 2p3/2 was observed at 934.4 eVfor each sample, which suggested the highly dispersed state of copperwhereas it is generally observed at about 933 eV in the case of bulkcopper [53,54]. Similarly, XPS spectra of Zr 3d is shown in Fig. 4B. Twoprominent XPS peaks were identified as Zr 3d3/2 and Zr 3d5/2 at 184.6and 182.2 eV, respectively. The presence of two XPS peaks with an2.4 eV energy gap indicated the presence of of Zr4+ species [55].

As shown in tabulated data (Table 3), the binding energies of Cu2p3/2 slightly shifted from 934.45 eV to 934.23 eV with increasing Cuconcentration from 5 to 15wt%. This shift of binding energies fromhigher to lower magnitude with the increase of Cu loading indicatedintensification in interaction between Cu and ZrO2 on one hand, and onthe other hand, it was also a clear manifestation of increasing copperdispersion [56]. However, both interaction and dispersion of copperwere adversely affected when the Cu loadings reached beyond 15wt%.These findings further validated N2O chemisorption and TEM results.Moreover, such variation in binding energies, as a consequent of in-creasing Cu loadings, was also reported by Liu et al., where lowerbinding energies were related to dispersed CuO and higher bindingenergies were declared for crystalline CuO [57]. In contrast, bindingenergies of Zr 3d5/2 were less affected by increment in Cu concentrationup to 15wt%. However, further increase in Cu content shifted the po-sition of Zr 3d5/2 to slightly higher binding energies, suggesting di-minishing the mutual interaction and reducing Zr dispersion with theincrease in Cu content.

Besides, CuO has a nearly square-planar symmetry, where Cu2+ ionis surrounded by four oxygen anions. So, in such a system when thesymmetry of Cu2+ ion is distorted, other charge-transfer excitationsmay take place due to electron correlation effects [58].

The broad shake-up satellite peak observed in the current case couldbe due to the slight distortion of CuO symmetry. This could be furtherelaborated by examining the change in FWHM value of parent Cu 2p3/2peaks as a function of Cu addition. The magnitude of FWHM slightlyincreased by increasing Cu loadings and maximum values were re-corded for catalysts with 15 and 20wt% of Cu content. Therefore, itcould be infirmed from the XPS data that addition of Cu content led tocreation of additional Cu2+–Cu2− weak bonding with neighboring O2−

ions, which consequently distorted coordination symmetry of Cu2+ ionto more distorted octahedral symmetry [59]. Similarly, the FWHMvalues of parent Zr 3d5/2 were found to be invariant throughout therange of Cu concentration. Therefore, no particular information per-taining to bonding and the nature of Zr4+ ions could be obtained.Furthermore, relative atomic ratio of Cu/Zr is also presented in Table 3,whereby the values in parentheses are nominal ratio of the two metal inparent catalysts. This atomic ratio increased with the increase of coppercontent, but as for the bimetallic catalyst with higher loadings, it wassignificantly lower than the nominal ratio. This showed enrichment ofsurface Zr concentration as a consequent of surface Cu depletion. Thiswas due to agglomeration at higher loadings and had been consistentwith the agglomeration of the copper particles observed by XRD athigher content. Recently, similar results were also reported by Martinet al. for Cu–ZrO2 with increased Cu content [60].

To identify the different Cu species in the catalysts, Cu2p3/2 peak ofeach catalyst was resolved in different peaks and is displayed in Fig. 6.

Cupric ion can easily be recognized owing to its characteristiccoupling phenomenon between unpaired electrons. However, bindingenergies of Cu0 and Cu+ are so similar that they could not be dis-tinguished. In this work, higher energy Cu2p3/2 peak observed at933.7 eV and 935.5 were assigned to Cu2+ ion whereas peak of lowerenergy at 932.7 eV was ascribed to Cu+ ion [61]. The XPS analysissupported the TPR findings by showing different CuO species as wassuggested by H2-TPR studies for step-wise reduction of CuO.

3.1.8. Basicity profile of CZC catalystsIn order to get insight about the basic sites of the catalysts CO2-TPD

was performed. TPD-CO2 profile of catalysts with different copperloading is presented in Fig. 7.

Similarly, magnitude of basic sites and their respective densities aredocumented in Table 4. A wide distribution of basic sites was observedranging from 100 to 700 °C. The sites below 450 °C were categorized asmedium and those above this temperature range as strong basic sites.

In addition, the total amount of desorbed CO2 increased with theincrease of Cu content in the CZC catalysts. This linear relationship wasmaintained with the addition of Cu up to 20wt%. However, it wasreduced on further Cu loading of 25 wt%. This increase in intensity ofbasic sites with increasing Cu loading is in line with the literature [62].Since CuO has been reported to be the adsorption site of CO2 [63],hence it is quite understandable that CO2 adsorption is improved byincreasing Cu loading in CZC catalysts. Besides, the basic sites dis-tribution was remarkably altered by variation in Cu loadings. A smallshoulder at 300 °C confirmed the existence of medium basic sites, butbasic sites were predominantly distributed in strong and very strongadsorption regions, as indicated by desorption peaks at 480, 560, and700 °C for CZC5 catalyst. Meanwhile, the incorporation of Cu content to10 wt% created a new desorption peak in the weak basic sites regionwith desorption maximum at around 421 °C. However, the majority ofbasic sites were still distributed at higher temperature regions in CZC10catalyst. A remarkable shift of basic sites from higher temperature re-gion to lower temperature region was observed when Cu content wasincreased to 15 wt%. Nevertheless, on further increasing of Cu content,the distribution of basic sites was shifted to the right side of the TPDspectra. A similar trend was observed for density of basic sites variationas a function of Cu content.

A comparative study of the current catalyst with the recently re-ported data for CO2 hydrogenation to methanol has been presented inTable 5. As depicted from the tabulated data the current catalystshowed good activity for methanol synthesis and excellent CO2 con-version rate as compared to the catalysts supported on other traditionalsupports like Al2O3, SiO2, alumina, MCM-41, SBA-15 and MSU-F sup-ports.

On the industrial scale, methanol is synthesized by using natural gasor coal gasification. By this way the resulting syngas is generallycomprised of CO, CO2 and H2 reactant gases to produce methanol.

Table 3XPS data of CZC catalysts with different Cu content.

Sample Binding energy (eV) FWHM (eV) Atomic Cu/Zr ratio

Cu 2p3/2 Zr 3d5/2 Cu 2p3/2 Zr 3d5/2

CZC5 934.45 182.18 4.12 4.41 1 (0.6)CZC10 934.20 182.20 4.15 4.42 0.9 (1.35)CZC15 934.23 182.24 4.22 4.43 1.2 (2.02)CZC20 934.32 182.60 4.22 4.40 1.1 (2.7)CZC25 934.43 182.62 4.17 4.42 0.9 (3.37)

I.U. Din et al. Chemical Engineering Journal 334 (2018) 619–629

625

Syngas-based methanol synthesis not only consume large amount offossil resources but also produce about 0.6–1.5 tons of CO2 for each tonof methanol produced. Methanol synthesis by pure CO2 hydrogenationis preferred over conventional syngas route. Similarly, CO2 hydro-genation is preferred because of its high abundance, nontoxicity, lowcost and greater potential as renewable energy source [64]. Likewise, interms of kinetics CO2 hydrogenation is faster than CO hydrogenation[65]. In addition, as evident from tabulated data of Table 5, CNFs basedcatalyst (current study) showed excellent activity for CO2 hydrogena-tion to methanol. Keeping in view the high mechanical strength ofCNFs, it has great potential to be used in bulk quantity at industrial

Fig. 6. XPS Cu 2p peak fitting curves of (a) CZC10, (b) CZC15, (c) CZC20 and (d) CZC25 catalysts.

Temperature ( oC)

CO

2de

sorb

ed(a

.u.)

200 300 400 500 600 700

setiscisabgnortSsetiscisabmuideM

Fig. 7. TPD-CO2 profile of (a) CZC5, (b) CZC10, (c) CZC15, (d) CZC20 and (e) CZC25catalysts.

Table 4CO2 TPD and activity data of CZC catalysts.

Catalyst No. of basicsites (mmol/g.cat)

Density ofbasic sites(mmol/m2)

Meth.Activity (g/kg.cat.h)

CO2

Conv.(%)

TOFMeOH X10−4 (s−1)

CZC5 9.40 0.08 08 4 5.0CZC10 11.2 0.09 13 7 4.7CZC15 15.83 0.12 20 11 5.4CZC20 17.04 0.18 18 12 5.2CZC25 15.83 0.17 16 09 5.1

I.U. Din et al. Chemical Engineering Journal 334 (2018) 619–629

626

level.

3.2. Structure-activity relationship as a function of Cu content

Table 4 presents activity data of catalysts with various Cu loadings.Methanol synthesis rate was significantly increased by increasing Cucontent from 5 to 10wt%. The trend was continued with further ad-dition of Cu and optimum activity was recorded for catalysts with 15 wt%. However, when the Cu content was increased further, the additionalamount of Cu led to a decline in the methanol activity. The trend inmethanol synthesis with respect to Cu content had been closely asso-ciated with the variations in Cu surface area and dispersion. Similarly,Bell et al. also reported such observations [68]. In terms of methanolsynthesis, production rate of almost similar magnitude was reportedover Ag/ZnO/ZrO2 and Au/ZnO/ZrO2 catalysts by Sloczynski et al.[66]. Similarly, CO2 conversion gradually increased with the increase ofCu content and maximum conversion was achieved with CZC20 cata-lyst. In case of TOF, although maximum value was recorded for CZC15catalyst, no proper variation was observed with varying Cu content.

Due to the acidic nature of CO2, the number of basic sites play animportant role in CO2 conversion reactions. As shown in Fig. 8, the

performance of catalysts was significantly altered by the number ofbasic sites.

The rate of methanol productivity was increased by increasing thenumber of basic site with an exception of increase in basic sites from15.8 to 17.04mmol/g.cat, which led to a decrease in methanol pro-ductivity from 20 to 18 g/kg.cat.h, respectively. This indicates thatphysical parameters like BET surface area, SCu and agglomeration havesuperseded the role of basic sites in terms of methanol productivity athigher Cu loadings. However, CO2 conversion was progressively in-creased by increasing the number of basic sites. This signified the vitalrole of magnitude of basic sites in CO2 conversion. Similar correlationof improved CO2 conversion as a function of the number of basic siteshad also been previously reported [69].

On top of that, the BET surface area is a vital factor in discussing thecatalyst activity. In the current work, a correlation of BET surface areawith methanol productivity, as a consequent of varying Cu content, wasinvestigated.

As shown in Fig. 9, an increase in Cu content increased the BETsurface area of the catalyst, which subsequently improved the methanolproductivity rate. The trend continued with further Cu addition andoptimum values of both BET surface area and methanol productivityrate were obtained for catalyst with 15 wt% of Cu content. However,further increase of Cu loading reduced the BET surface area due toagglomeration, and consequently, methanol productivity rate was de-pressed. This pointed out that BET surface area had been a predominantcontribution factor in activity profiles of CZC catalyst to methanolformation. Similar observations were also documented for methanolproduction over Cu/ZnO catalyst by Pan et al. [70].

As suggested by reported mechanistic studies, magnitude of SCu hasvital role in CO2 hydrogenation to methanol. In order to assess the roleof copper surface area (SCu) on the activity pattern of CZC catalysts, agraph portraying SCu, methanol production, and turnover frequency(TOF) is shown in Fig. 10.

As evident from the graph, a linear dependence of methanol yield toSCu was found, which is in accordance with the literature [71]. Ob-viously, promotion in SCu provided more atomic hydrogen. Conse-quently, more atomic hydrogens were supplied to the ZrO2 sites forreduction of CO2 adsorption, which led to higher production rate.However, TOF variation was not totally aligned with variant SCu values.Based on Boudart’s theory, straight line should be obtained between SCuand TOF in case where catalyst activity is only dependent on SCu [72].Based on this observation, it could be inferred that in the current case,although the rate of methanol synthesis was accelerated by increasingSCu, however it was not the sole factor that affect the activity profile.

Table 5Activity comparison of this novel catalyst with the recent reported data.

Catalyst Reaction conditions Meth. activity(g/kg.h)

CO2 conversion(%)

Refs.

CZC15 (180 °C & 30 bar) 20 11 Thiswork

Pd/ZnO/Al2O3

(250 °C & 30 bar) 24 4.4 [39]

Ag/ZrO2 (230 °C & 80 bar) 27 – [55]Au/ZnO/

ZrO2

(220 °C & 80 bar) 19 2 [66]

Ag/ZnO/ZrO2

(220 °C & 80 bar) 15 2 [66]

Ag/Zn (200 °C & 80 bar) 15 – [55]Ag/Zn/Zr (200 °C & 80 bar) 23 – [55]Cu/Zr (200 °C & 80 bar) 7 – [55]Pd/SiO2 (250 °C & 41 bar) 10 3 [67]Cu/SiO2 (250 °C & 41 bar) 6 2.8 [67]Cu-Pd/SiO2 (250 °C & 41 bar) 35 6.6 [67]Cu-Pd/

MCM-41

(250 °C & 41 bar) 21 6.2 [67]

Cu-Pd/SBA-15

(250 °C & 41 bar) 23 6.5 [67]

Cu-Pd/MSU-F

(250 °C & 41 bar) 13 5.3 [67]

Fig. 8. Catalysts performance versus number of basic site.

Fig. 9. Correlation of Cu content with BET surface area and methanol productivity rate.

I.U. Din et al. Chemical Engineering Journal 334 (2018) 619–629

627

4. Conclusion

This study reports the effect of Cu loading on the physicochemicalproperties and activity pattern of carbon nanofibers supported Cu-ZrO2

catalyst in liquid phase CO2 hydrogenation. CuO/ZrO2/CNF catalystswith 5, 10, 15, 20 and 25wt% of Cu loading were synthesized usingdeposition-precipitation method. Both physicochemical and activityprofiles of catalysts were significantly affected by variation of Cu con-tent. Physicochemical investigation suggested 15wt% of Cu as an op-timum amount for the CNF supported Cu/ZrO2 catalysts. Activity dataconfirmed 15CZC catalyst as the most efficient catalyst in terms ofmethanol productivity (20 g/kg.cat.h).

Acknowledgments

Financial assistant from Universiti Teknologi PETRONAS (UTP) andthe Ministry of Higher Education Malaysia, FRGS No: FRGS/1/2011/SG/UTP/02/13 is gratefully acknowledged. The authors acknowledgeuse of facilities within the UTP Centralized Analytical Laboratory(CAL).

References

[1] F. Meng, Q. Zhang, G. Yang, R. Yang, Y. Yoneyama, N. Tsubaki, Structural andkinetical studies on the supercritical CO2 dried Cu/ZnO catalyst for low-tempera-ture methanol synthesis, Chem. Eng. J. 295 (2016) 160–166.

[2] A.A. Kiss, J.J. Pragt, H.J. Vos, G. Bargeman, M.T. de Groot, Novel efficient processfor methanol synthesis by CO2 hydrogenation, Chem. Eng. J. 284 (2016) 260–269.

[3] T. Witoon, J. Chalorngtham, P. Dumrongbunditkul, M. Chareonpanich,J. Limtrakul, CO2 hydrogenation to methanol over Cu/ZrO2 catalysts: effects ofzirconia phases, Chem. Eng. J. 293 (2016) 327–336.

[4] T. Phongamwong, U. Chantaprasertporn, T. Witoon, T. Numpilai, Y. Poo-arporn,W. Limphirat, W. Donphai, P. Dittanet, M. Chareonpanich, J. Limtrakul, CO2 hy-drogenation to methanol over CuO–ZnO–ZrO2–SiO2 catalysts: effects of SiO2contents, Chem. Eng. J. 316 (2017) 692–703.

[5] H. Zhang, X. Li, F. Zhu, K. Cen, C. Du, X. Tu, Plasma assisted dry reforming ofmethanol for clean syngas production and high-efficiency CO2 conversion, Chem.Eng. J. 310 (Part 1) (2017) 114–119.

[6] G.C. Chinchen, P.J. Denny, J.R. Jennings, M.S. Spencer, K.C. Waugh, Synthesis ofmethanol: part 1. catalysts and kinetics, Appl. Catal. 36 (1988) 1–65.

[7] J.C.J. Bart, R.P.A. Sneeden, Copper-zinc oxide-alumina methanol catalysts re-visited, Catal. Today 2 (1987) 1–124.

[8] G. Jia, Y. Tan, Y. Han, A comparative study on the thermodynamics of dimethylether synthesis from CO hydrogenation and CO2 hydrogenation, Ind. Eng. Chem.Res. 45 (2006) 1152–1159.

[9] F. Arena, K. Barbera, G. Italiano, G. Bonura, L. Spadaro, F. Frusteri, Synthesis,characterization and activity pattern of Cu–ZnO/ZrO2 catalysts in the hydrogena-tion of carbon dioxide to methanol, J. Catal. 249 (2007) 185–194.

[10] H.-D. Zhuang, S.-F. Bai, X.-M. Liu, Z.-F. Yan, Structure and performance of Cu/ZrO2

catalyst for the synthesis of methanol from CO2 hydrogenation, J. Fuel Chem.Technol. 38 (2010) 462–467.

[11] W.H. Hg Guo, J.L. Shen, H.Z. Liu, Studies on deactivation of copper-based mathanol

synthesis catalysts, Indus. Catal. 11 (2003) 39–42.[12] I.A. Fisher, A.T. Bell, In-situinfrared study of methanol synthesis from H2/CO2 over

Cu/SiO2 and Cu/ZrO2/SiO2, J. Catal. 172 (1997) 222–237.[13] D. Bianchi, T. Chafik, M. Khalfallah, S.J. Teichner, Intermediate species on zirconia

supported methanol aerogel catalysts: III. Adsorption of carbon monoxide on coppercontaining solids, Appl. Catal. A 112 (1994) 57–73.

[14] J. Wambach, A. Baiker, A. Wokaun, CO2 hydrogenation over metal/zirconia cata-lysts, PCCP 1 (1999) 5071–5080.

[15] J. Ma, N. Sun, X. Zhang, N. Zhao, F. Xiao, W. Wei, Y. Sun, A short review of catalysisfor CO2 conversion, Catal. Today 148 (2009) 221–231.

[16] Z. Li, S. Yan, H. Fan, Enhancement of stability and activity of Cu/ZnO/Al2O3 cat-alysts by microwave irradiation for liquid phase methanol synthesis, Fuel 106(2013) 178–186.

[17] J. Toyir, P.R.r. de la Piscina, J.L.G. Fierro, N.s. Homs, Highly effective conversion ofCO2 to methanol over supported and promoted copper-based catalysts: influence ofsupport and promoter, Appl. Catal. B 29 (2001) 207–215.

[18] V. Deerattrakul, P. Dittanet, M. Sawangphruk, P. Kongkachuichay, CO2 hydro-genation to methanol using Cu-Zn catalyst supported on reduced graphene oxidenanosheets, J. CO2 Util. 16 (2016) 104–113.

[19] Y. Sun, L. Chen, Y. Bao, G. Wang, Y. Zhang, M. Fu, J. Wu, D. Ye, Roles of nitrogenspecies on nitrogen-doped CNTs supported Cu-ZrO2 system for carbon dioxidehydrogenation to methanol, Catal. Today (2017).

[20] I. Kang, Y.Y. Heung, J.H. Kim, J.W. Lee, R. Gollapudi, S. Subramaniam,S. Narasimhadevara, D. Hurd, G.R. Kirikera, V. Shanov, M.J. Schulz, D. Shi,J. Boerio, S. Mall, M. Ruggles-Wren, Introduction to carbon nanotube and nanofibersmart materials, Compos. B Eng. 37 (2006) 382–394.

[21] M. Peer, A. Qajar, B.-P.M. Holbrook, R. Rajagopalan, H.C. Foley, Platinum em-bedded within carbon nanospheres for shape selective liquid phase hydrogenation,Carbon 57 (2013) 485–497.

[22] J.H. Bitter, J.K.P. De, D.L.M.K. Van, D.A.J. Van, Carbon nanofibre composites,preparation and use, Google Patents, 2005.

[23] J.K. Chinthaginjala, K. Seshan, L. Lefferts, Preparation and application of carbon-nanofiber based microstructured materials as catalyst supports, Ind. Eng. Chem.Res. 46 (2007) 3968–3978.

[24] S.K.K.A.M. Zainoodin, M.S. Masdar, W.R.W. Daud, A.B. Mohamad, J. Sahari Appl.Energy 113 (2014) 946–954.

[25] I. Ud Din, M.S. Shaharun, D. Subbarao, A. Naeem, Synthesis, characterization andactivity pattern of carbon nanofibers based copper/zirconia catalysts for carbondioxide hydrogenation to methanol: influence of calcination temperature, J. PowerSources 274 (2015) 619–628.

[26] I.A. Fisher, H.C. Woo, A.T. Bell, Effects of zirconia promotion on the activity of Cu/SiO2 for methanol synthesis from CO/H2 and CO2/H2, Catal. Lett. 44 (1997) 11–17.

[27] J.B. Ko, C.M. Bae, Y.S. Jung, D.H. Kim, Cu–ZrO2 Catalysts for water-gas-shift re-action at low temperatures, Catal. Lett. 105 (2005) 157–161.

[28] A. Karelovic, P. Ruiz, The role of copper particle size in low pressure methanolsynthesis via CO2 hydrogenation over Cu/ZnO catalysts, Catal. Sci. Technol. (2015).

[29] A.T. Bell, Molecular design of highly active methanol synthesis catalysts, in:J.J.S.E. Iglesia, T.H. Fleisch (Eds.), Stud. Surf. Sci. Catal. Elsevier, 2001, pp. 13–19.

[30] J. Liu, J. Shi, D. He, Q. Zhang, X. Wu, Y. Liang, Q. Zhu, Surface active structure ofultra-fine Cu/ZrO2 catalysts used for the CO2+H2 to methanol reaction, Appl.Catal. A 218 (2001) 113–119.

[31] M.K. van der Lee, J. van Dillen, J.H. Bitter, K.P. de Jong, Deposition precipitationfor the preparation of carbon nanofiber supported nickel catalysts, JACS 127 (2005)13573–13582.

[32] J.-S. Kim, S.-B. Lee, M.-C. Kang, K.-W. Lee, M.-J. Choi, Y. Kang, Promotion of CO2hydrogénation to hydrocarbons in three-phase catalytic (Fe-Cu-K-Al) slurry re-actors, Korean J. Chem. Eng. 20 (2003) 967–972.

[33] X.-L. Du, Q.-Y. Bi, Y.-M. Liu, Y. Cao, H.-Y. He, K.-N. Fan, Tunable copper-catalyzedchemoselective hydrogenolysis of biomass-derived [gamma]-valerolactone into 1,4-pentanediol or 2-methyltetrahydrofuran, Green Chem. 14 (2012) 935–939.

[34] J.W. Evans, M.S. Wainwright, A.J. Bridgewater, D.J. Young, On the determinationof copper surface area by reaction with nitrous oxide, Appl. Catal. 7 (1983) 75–83.

[35] J. Agrell, M. Boutonnet, I. Melián-Cabrera, J.L.G. Fierro, Production of hydrogenfrom methanol over binary Cu/ZnO catalysts: Part I. Catalyst preparation andcharacterisation, Appl. Catal. A 253 (2003) 201–211.

[36] H. Ahouari, A. Soualah, A. Le Valant, L. Pinard, P. Magnoux, Y. Pouilloux, Methanolsynthesis from CO2 hydrogenation over copper based catalysts, React. Kinet. Catal.Lett. 110 (2013) 131–145.

[37] Q. Sun, Y.-L. Zhang, H.-Y. Chen, J.-F. Deng, D. Wu, S.-Y. Chen, A novel process forthe preparation of Cu/ZnO and Cu/ZnO/Al2O3 ultrafine catalyst: structure, surfaceproperties, and activity for methanol synthesis from CO2+H2, J. Catal. 167 (1997)92–105.

[38] Y. Nie, Surface silanization of carbon nanofibers and nanotubes for altering theproperties of epoxy composites Ilmenau University of Technology, 2012.

[39] X.-L. Liang, X. Dong, G.-D. Lin, H.-B. Zhang, Carbon nanotube-supported Pd–ZnOcatalyst for hydrogenation of CO2 to methanol, Appl. Catal., B 88 (2009) 315–322.

[40] C.-M. Chen, Y.-M. Dai, J.G. Huang, J.-M. Jehng, Intermetallic catalyst for carbonnanotubes (CNTs) growth by thermal chemical vapor deposition method, Carbon 44(2006) 1808–1820.

[41] Q.-Y.L. Ya-Hao Li, Hong-Qiang Wang, You-Guo Huang, Xiao-Hui Zhang, Qiang Wu,Hong-Quan Gao, Jian-Hong Yang, Applied Energy (2014).

[42] B. Neppolian, Q. Wang, H. Yamashita, H. Choi, Synthesis and characterization ofZrO2–TiO2 binary oxide semiconductor nanoparticles: application and interparticleelectron transfer process, Appl. Catal. A 333 (2007) 264–271.

[43] K. Rajaram, E. Savarimuthu, S. Arumugam, Structural, optical and morphologicalcharacteristics of in doped CuO nano particles prepared by precipitation method,

Fig. 10. Correlation of SCu with methanol production and TOF.

I.U. Din et al. Chemical Engineering Journal 334 (2018) 619–629

628

Chem. Sci. Rev. Lett. 2 (2014) 626–629.[44] I.C. Freitas, S. Damyanova, D.C. Oliveira, C.M.P. Marques, J.M.C. Bueno, Effect of

Cu content on the surface and catalytic properties of Cu/ZrO2 catalyst for ethanoldehydrogenation, J. Mol. Catal. A: Chem. 381 (2014) 26–37.

[45] B.L. Kirsch, S.H. Tolbert, Stabilization of isolated hydrous amorphous and tetra-gonal zirconia nanoparticles through the formation of a passivating alumina shell,Adv. Funct. Mater. 13 (2003) 281–288.

[46] O.A. Yeshchenko, I.M. Dmitruk, A.A. Alexeenko, A.M. Dmytruk, Size-dependentmelting of spherical copper nanoparticles embedded in a silica matrix, Phys. Rev. B75 (2007) 085434.

[47] F. Kazemi, A. Saberi, S. Malek-Ahmadi, S. Sohrabi, H. Rezaie, M. Tahriri, NovelMethod of synthesis of metasatable tetragonal zirconia nanopowder at low tem-peratures, Ceramics-Silikáty 55 (2011) 26–30.

[48] C.J.G. Van Der Grift, P.A. Elberse, A. Mulder, J.W. Geus, Preparation of silica-supported copper catalysts by means of deposition-precipitation, Appl. Catal. 59(1990) 275–289.

[49] Z. Xu, L. Chen, Y. Shao, D. Yin, S. Zheng, Catalytic hydrogenation of aqueous nitrateover Pd−Cu/ZrO2 catalysts, Ind. Eng. Chem. Res. 48 (2009) 8356–8363.

[50] J.Y. Yan, W.M.H. Sachtler, H.H. Kung, Effect of Cu loading and addition of modi-fiers on the stability of Cu/ZSM-5 in lean NOx reduction catalysis, Catal. Today 33(1997) 279–290.

[51] P. Da Costa, B. Modén, G.D. Meitzner, D. Ki Lee, E. Iglesia, Spectroscopic andchemical characterization of active and inactive Cu species in NO decompositioncatalysts based on Cu-ZSM5, PCCP 4 (2002) 4590–4601.

[52] F.E. López-Suárez, A. Bueno-López, M.J. Illán-Gómez, Cu/Al2O3 catalysts for sootoxidation: copper loading effect, Appl. Catal., B 84 (2008) 651–658.

[53] J.P. Espinós, J. Morales, A. Barranco, A. Caballero, J.P. Holgado, A.R. González-Elipe, Interface effects for Cu, CuO, and Cu2O deposited on SiO2 and ZrO2. XPSdetermination of the valence state of copper in Cu/SiO2 and Cu/ZrO2 catalysts, J.Phys. Chem. B 106 (2002) 6921–6929.

[54] Y. Matsumura, Stabilization of Cu/ZnO/ZrO2 catalyst for methanol steam reformingto hydrogen by coprecipitation on zirconia support, J. Power Sources 238 (2013)109–116.

[55] R. Grabowski, J. Słoczyński, M. Śliwa, D. Mucha, R.P. Socha, M. Lachowska,J. Skrzypek, Influence of polymorphic ZrO2 phases and the silver electronic state onthe activity of Ag/ZrO2 catalysts in the hydrogenation of CO2 to methanol, ACSCatal. 1 (2011) 266–278.

[56] X. Guo, D. Mao, G. Lu, S. Wang, G. Wu, The influence of La doping on the catalyticbehavior of Cu/ZrO2 for methanol synthesis from CO2 hydrogenation, J. Mol. Catal.A: Chem. 345 (2011) 60–68.

[57] Z. Liu, M.D. Amiridis, Y. Chen, Characterization of CuO supported on tetragonalZrO2 catalysts for N2O decomposition to N2, J. Phys. Chem. B 109 (2005)

1251–1255.[58] Y. Okamoto, K. Fukino, T. Imanaka, S. Teranishi, Surface characterization of copper

(II) oxide-zinc oxide methanol-synthesis catalysts by x-ray photoelectron spectro-scopy. 1. Precursor and calcined catalysts, J. Phys. Chem. 87 (1983) 3740–3747.

[59] Y.-W. Suh, S.-H. Moon, H.-K. Rhee, Active sites in Cu/ZnO/ZrO2 catalysts for me-thanol synthesis from CO/H2, Catal. Today 63 (2000) 447–452.

[60] D. Durán-Martín, M. Ojeda, M.L. Granados, J.L.G. Fierro, R. Mariscal, Stability andregeneration of Cu–ZrO2 catalysts used in glycerol hydrogenolysis to 1,2-propane-diol, Catal. Today 210 (2013) 98–105.

[61] I.U. Din, M.S. Shaharun, D. Subbarao, A. Naeem, F. Hussain, Influence of niobiumon carbon nanofibres based Cu/ZrO2 catalysts for liquid phase hydrogenation ofCO2 to methanol, Catal. Today 259 (Part 2) (2016) 303–311.

[62] G. Avgouropoulos, T. Ioannides, TPD and TPSR study of CO interaction withCuO–CeO2 catalysts, J. Mol. Catal. A Chem. 296 (2008) 47–53.

[63] C.S. Chen, J.H. Wu, T.W. Lai, Carbon dioxide hydrogenation on Cu nanoparticles, J.Phys. Chem. C 114 (2010) 15021–15028.

[64] Z. Huo, M. Hu, X. Zeng, J. Yun, F. Jin, Catalytic reduction of carbon dioxide intomethanol over copper under hydrothermal conditions, Catal. Today 194 (2012)25–29.

[65] G. Centi, S. Perathoner, Advances in Catalysts and Processes for Methanol Synthesisfrom CO2, in: M.D. Falco, G. Iaquaniello, G. Centi (Eds.), CO2: A Valuable Source ofCarbon, Springer, London, 2013, pp. 147–169.

[66] J. Słoczyński, R. Grabowski, A. Kozłowska, P. Olszewski, J. Stoch, J. Skrzypek,M. Lachowska, Catalytic activity of the M/(3ZnO·ZrO2) system (M=Cu, Ag, Au) inthe hydrogenation of CO2 to methanol, Appl. Catal. A 278 (2004) 11–23.

[67] X. Jiang, N. Koizumi, X. Guo, C. Song, Bimetallic Pd–Cu catalysts for selective CO2hydrogenation to methanol, Appl. Catal. B 170 (2015) 173–185.

[68] M.D. Rhodes, A.T. Bell, The effects of zirconia morphology on methanol synthesisfrom CO and over catalysts: Part I. Steady-state studies, J. Catal. 233 (2005)198–209.

[69] Q. Pan, J. Peng, T. Sun, S. Wang, S. Wang, Insight into the reaction route of CO2

methanation: promotion effect of medium basic sites, Catal. Commun. 45 (2014)74–78.

[70] W.X. Pan, R. Cao, D.L. Roberts, G.L. Griffin, Methanol synthesis activity of CuZnOcatalysts, J. Catal. 114 (1988) 440–446.

[71] C. Baltes, S. Vukojević, F. Schüth, Correlations between synthesis, precursor, andcatalyst structure and activity of a large set of CuO/ZnO/Al2O3 catalysts for me-thanol synthesis, J. Catal. 258 (2008) 334–344.

[72] X. Guo, D. Mao, G. Lu, S. Wang, G. Wu, Glycine–nitrate combustion synthesis ofCuO–ZnO–ZrO2 catalysts for methanol synthesis from CO2 hydrogenation, J. Catal.271 (2010) 178–185.

I.U. Din et al. Chemical Engineering Journal 334 (2018) 619–629

629