structured catalysts based on mg–al hydrotalcite for the synthesis of

8/15/2019 Structured Catalysts Based on Mg–Al Hydrotalcite for the Synthesis Of

1/9

Catalysis Today 216 (2013) 211–219

Contents lists available at SciVerse ScienceDirect

Catalysis Today

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

Structured catalysts based on Mg–Al hydrotalcite for the synthesis of biodiesel

I. Reyeroa, I. Velasco b, O. Sanz b, M. Montes b, G. Arzamendia, L.M. Gandíaa,∗

a Departamento de QuímicaAplicada, Edificio de losAcebos, Universidad Pública de Navarra, Campus de Arrosadía s/n, E-31006 Pamplona, Spainb Departamento de Química Aplicada, UFI 11/65, Facultad de Ciencias Químicas de SanSebastián, UPV/EHU, PaseoManuel de Lardizábal 3, 20018 San

Sebastián, Spain

a r t i c l e i n f o

Article history:Received 13 March 2013Received in revised form 22 April 2013Accepted 25 April 2013Available online 21 May 2013

Keywords:

BiodieselMethanolysisMg–Al hydrotalciteMonolithic stirrer reactorStructured methanolysis catalystTransesterification

a b s t r a c t

The performance of metallic (Fecralloy®) monoliths based on Mg–Al hydrotalcite for the synthesis of biodiesel through the methanolysis of sunflower oilisreported for the first time. A Mg–Al hydrotalcite wassynthesized and the oxides obtained after calcination at 500 ◦C were employed for preparing the slurriesused for washcoating the monoliths. Using methanol as the solvent and sepiolite as binder allowed toreach 96% adherence after sonication. The positive effect of sepiolite on adherence seems to be relatedwith its fibrous nature and ability for crosslinking with the hydrotalcite particles. However, catalytictests have shown a negative effect of sepiolite on the methanolysis activity that can be attributed to thepartial neutralization of the basic sites and/or masking of the hydrotalcite particles. The performance of the structured catalysts has been investigated in a monolithic stirrer reactor under very mild conditions(60 ◦C and 1atm). Using 2 wt.% of catalyst allowed reaching 62–77% oil conversion after 10h. However, thepooradherenceof the coating in the reaction mixture prevented the effective reutilization of the catalysts.Further work is required concerning the formulation of the coating slurries to improve adherence in thechemically aggressive medium involved in the synthesis of biodiesel.

© 2013 Elsevier B.V. All rights reserved.

1. Introduction

The favorable framework established in the European Union bythe Directive on the promotion of the use of energy from renew-able sources continues stimulating the consumption of biofuelsin transport [1]. This consumption has reached 13.6Mtoe in 2011compared to 13.2 Mtoe in 2010,beingbiodiesel themost importantbiofuel with a 78% share [2]. Biodiesel is mainly produced in indus-try through the base-catalyzed transesterification with methanolof the triglycerides composing the vegetable oils and animal fatsto form fatty acid methyl esters (FAMEs or biodiesel) and glycerol.This reaction is also known as methanolysis and is typically con-ducted using homogeneous catalysts such as NaOH, KOH or their

methoxides that are easily dissolved in the alcohol. These catalystsare cheap and very active but pose technological issues such as therequirement of thorough washing steps for removing them fromthe reaction products. This leads to the production of very largevolumes of wastewater and avoids the reutilization of the cata-lysts that also tend to form soaps that complicate the separation of biodiesel and glycerol. It is also important to note that methanoland vegetable oils or melted animal fats are virtually immiscible

∗ Corresponding author. Tel.: +34 948 169605; fax: +34 948 169606.E-mail address: [email protected] (L.M. Gandía).

so reaction rate can become limited by mass transport. For thisreason, mixing and the influence of agitation on drop size distri-butions are critical factors of the performance of the conventionalbatch reactors [3–5].

In the recent years a great research effort is being devotedto the improvement of the biodiesel production technology. Qiuet al. [5], and more recently Santacesaria et al. [6] have reviewedthe several approaches that have been proposed for the intensi-fication of the biodiesel production. Novel reactor concepts havebeen developed for enhancing the physical transportprocesses andthe separation of the reaction products. Most notable advancesinclude the continuous oscillatory flow reactor [7], rotating tubereactor [8], continuous down-flow gas–liquid[9] and liquid–liquid

[10,11] contactors, and the laminar flow reactor–separator [12].The fact that the main reaction products,biodiesel andglycerol, arealso immiscible has been exploited for improved separation usingmembrane reactors [13–18]. Static mixers coupled to stirred tankreactors [19] and static mixers tubular reactors [20,21] have beenused with good results due to enhanced mass transfer. Micropro-cesstechnologyoffersinterestingadvantagesfortheintensificationof multiphase reactions [22]. The short diffusion distances inmicromixers and catalytic-wall microreactors greatly reduce themass-transfer limitations thus significantly improving the reactorperformance for biodiesel synthesis [23]. Sun et al. have tested asbiodiesel synthesis reactors several capillary microreactors with

0920-5861/$ – see front matter © 2013 Elsevier B.V. All rights reserved.

http://dx.doi.org/10.1016/j.cattod.2013.04.022

http://localhost/var/www/apps/conversion/tmp/scratch_3/dx.doi.org/10.1016/j.cattod.2013.04.022http://localhost/var/www/apps/conversion/tmp/scratch_3/dx.doi.org/10.1016/j.cattod.2013.04.022http://www.sciencedirect.com/science/journal/09205861http://www.elsevier.com/locate/cattodmailto:[email protected]://localhost/var/www/apps/conversion/tmp/scratch_3/dx.doi.org/10.1016/j.cattod.2013.04.022http://localhost/var/www/apps/conversion/tmp/scratch_3/dx.doi.org/10.1016/j.cattod.2013.04.022mailto:[email protected]://crossmark.dyndns.org/dialog/?doi=10.1016/j.cattod.2013.04.022&domain=pdfhttp://www.elsevier.com/locate/cattodhttp://www.sciencedirect.com/science/journal/09205861http://localhost/var/www/apps/conversion/tmp/scratch_3/dx.doi.org/10.1016/j.cattod.2013.04.022


2/9

212 I. Reyero et al. / Catalysis Today 216 (2013) 211–219

inner diameters of 0.25 or 0.53mm [24] as well as commer-cial micromixers with characteristic dimensions below 0.4mmconnected to a capillary tube [25,26]. Wen et al. used zigzagmicrochannel reactors of different hydraulic diameters between0.24 and 0.90mm [27]. Da Silva et al. considered a catalyst-freetransesterification process in supercritical ethanol and a stainlesssteel 1/16 in tube with a total volume of 36.5cm3 [28]. Santace-saria et al. used a corrugated brazed plates heat exchanger with1.5–2mm characteristic dimension as microreactor with availablevolume of 200 cm3 [29] and a 10mm internal diameter tube filledwith stainless steel ribbon wool [30].

Except for the supercritical transesterification and some of themembrane reactors, the above-mentioned intensified processesinclude the use of homogeneous catalysts dissolved in the alco-hol. Changing from homogeneous to heterogeneous catalysis is anopportunity for further improvement of the synthesis of biodieseldue to additional economic and environmental benefits [31–33].Kiss et al.have performeda comparative economic assessment thathas revealed the advantages of heterogeneously catalyzed routein terms of higher yield of biodiesel and purity of glycerol, lowercost of catalyst and maintenance, with an estimated productioncosts reduction of 59 US $ per ton of biodiesel compared with theconventional process[31]. Usingheterogeneous catalystsis consid-eredthe mostefficientmanner of producingbiodiesel[32,33]. Maindrawbacks of heterogeneous catalysts for this process are theirlower specific activity and limited chemical stability under reac-tion conditions, specially the influence of leaching as highlightedbyDi Serio et al. [34]. Numerous reviews have appeared in the lastyears on the use of heterogeneous catalysts for the synthesis of biodiesel[6,35–48]. Mostofthestudiesonthesynthesisofbiodieselwith heterogeneous catalysts have been carried out in slurry reac-tors with the solid catalyst in powder form. This configuration isuseful for catalysts screening and formulation development. Forpractical purposes, monolithic catalytic reactors are more inter-esting because they are hydrodynamically superior and avoid thenecessity of catalyst separation thus reducing the production costsand improving the products quality [49,50]. However, few reports

on the use of monolithic methanolysis catalysts have been pub-lished outside of those devoted to immobilized enzymes [51,52].Kolaczkowski et al. used a Zn aminoacid complex deposited ontocordierite monolith as catalyst for rapeseed oil methanolysis. 54%triglycerides conversion was achieved after 3 h of reaction at 195◦Cand 20bar [53]. Tonetto and Marchetti deposited K/Al2O3 catalystonto cordierite honeycombmonolith and used it for the methanol-ysisofsoybeanoil.59%FAMEsyieldwasobtainedafter6hat120 ◦C[54].

In the present work, a Mg–Al hydrotalcite has been depositedonto metallic (Fecralloy®) monoliths and used as catalysts forthe transesterification with methanol of sunflower oil at 60◦Cand atmospheric pressure. Mg–Al hydrotalcites are well-known asactive catalysts for the synthesis of biodiesel. These materials are

among the most studied methanolysis catalysts although in pow-derform[6,35–39,42,45,46] . However, to ourknowledge,this is thefirst report on the use of hydrotalcite-basedstructuredcatalystsforthe synthesis of biodiesel.

2. Experimental

2.1. Catalysts preparation

A Mg–Al hydrotalcite was prepared by the coprecipitationmethod [55,56]. To this end, an aqueous solution of the metal-lic cations was prepared dissolving appropriate amounts of Mg(NO3)2·6H2O and Al(NO3)2·H2O in deionized water to obtain a

Mg/Al molar ratio of 5 and a concentration of aluminum of 0.35M.

Another alkaline solution was prepared containing Na2CO3 0.60 Mand NaOH 2.34M in deionized water (pH = 12.0). Both solutionswere slowly and simultaneously dropped (60cm3/h) under vigor-ous stirring in a glass reactor at room temperature while the pHwas maintained within 10–11. The gel formed was aged for 24h at65 ◦C. Afterwards, the solid obtained was filtered and thoroughlywashed with deionized water until the filtrate conductivity wasbelow 100S. Then, the solid was dried at 120 ◦C for 16h result-ing in the parent Mg/Al hydrotalcite (HT).In order to obtain Mg–Almixed oxides the hydrotalcite was calcined in air at 500◦Cfor6h .

The structured catalysts were prepared by the washcoat-ing method [57]. Several coating suspensions (solids content of 20wt.%) were prepared with the calcined solids, different solvents(water, ethanol, methanol and propanol) and sepiolite which wasadded in different proportions as binder. A natural sepiolite claymineral (Tolsa S.A., Spain) was used whose chemical composition(expressed as oxides, wt.%) is: 60.8 SiO2; 20.3 MgO; 4.6 Al2O3; 1.2Fe2O3; 1.2 CaO; 1.1K2O and 0.4 Na2O. Before suspension prepara-tion,thesolidswereball-milledfor1hat400rpmtosuitablyreducethe particle size (d[3,4]) to 5.1m and 8.5m for the Mg–Al solidsand sepiolite, respectively. The catalytic performance of the solidspresent in the coating slurries has been investigated. After dryingthe slurries, the resulting solids were named as HT Sep X , where X is the percentage of sepiolite used.

Metallic monolithsweremadeof Fecralloy® (FeCr22Al5, 50m,Goodfellow), with 16mm diameter, 30mm of length and 350 cpsi(cells per square inch). After washing with soapy water, they wereimmersedinacetone30minwithultrasounds.Afterthat,theyweredried at 120 ◦C and subsequently calcined at 900◦C for 22h gener-ating -Al2O3 whiskers to favor the adherence between catalystand metallic surface. The coating of monoliths was performed byimmersing them in the suspensions prepared in advance. Theywere kept submerged for 1min and removed at 3cm/min. Theslurry excess was eliminated by centrifugation and then they weredried at 120 ◦C for 30min. This procedure is repeated until achiev-ing a solids load of about 500 mg. After that, monoliths werecalcined at 500 ◦C for 6 h. The monolithic catalysts were named as

M HT Sep X .

2.2. Catalysts activation and adherence tests

As reflected in the literature [58,59] and according to previousresults by our group [60], the parent hydrotalcite and the Mg–Almixed oxides usually display low methanolysis activity under typ-ical reaction conditions. However, the hydrotalcites reconstructedafter rehydration of the calcined solids showed improved catalyticperformance. This fact was attributed to the presence of hydroxideanions in specific points of the rehydrated hydrotalcite plateletsacting as strong basic Brønsted-type sites [60,61]; for this rea-son, rehydration was carried in order to activate the powder andmonolithic catalysts. Rehydration was performed by immersing

the monoliths in boiling deionized water and keeping them understirring for 20min. After that, excess water in the monolith chan-nels was removed by gently blowing with air and finally dried at60 ◦C under vacuum overnight. In the case of the powdered solidstwo rehydration procedures were considered. Rehydration in boil-ing water until complete evaporation of water and immersion inboiling water for 20min and posterior removal of excess water byfiltration. Furthermore, the filtration was performed in two differ-ent ways, by gravity filtration and vacuum filtration. In all cases,the solids were finally dried overnight at 60◦C under vacuum.

Coating adherence tests were performed after the activationof the catalysts. The monoliths were immersed in vials contain-ing petroleum ether (standard medium) or methanol (reactionmedium) andsonicatedfor 30min. After that time, monolithswere

removed, allowed to dry in an oven and then calcined at 500◦

C for


3/9

I. Reyero et al. / Catalysis Today 216 (2013) 211–219 213

6 h. The loss of catalyst during the activation of the monolithic cat-alysts was also checked. To this end, we performedthe rehydrationtreatment whereupon the monolith was calcined (500◦C,6h).Cal-culations were performed comparing the weights of the monolithafter calcination.

2.3. Catalysts characterization

X-ray diffraction (XRD) patterns of the powdered calcined andrehydrated samples were obtained using a D-Max Rigaku diffrac-tometer operating at 40 kV and 80 mA with a Cu anode and agraphitemonochromator( =1.5405 Å),scanningrange3–85◦ (2 ),scan step size of 0.03◦ and 1 s/step. Thermogravimetric analyses(TGA) were performedin a Seiko Exstar6000 thermobalance underair flow and heating rate of 10◦C/min.

The N2 adsorption–desorption isotherms at −196 ◦C weremeasured by the static method in an automatic volumetricMicromeritics ASAP 2020 adsorption analyzer. Specific surfaceareas (SBET) were calculated using the BET method from thenitrogen adsorption isotherms. Specific total pore volumes (V p)were assessed from the amount of nitrogen adsorbed at a rel-ative pressure of 0.99. Particle size distribution of the catalystwas measured with a Laser Particle Size Analyzer MALVERN Mas-terSizer 2000. Electron (SEM) micrographs of the samples weretaken on a HITACHI S-4800 field emission scanning electronicmicroscope.

Thebasicpropertiesofthesolidsweredeterminedthroughtem-perature programmed desorption of CO2 (CO2-TPD) and tests withHammett indicators. CO2-TPD experiments were carried out in aMicromeritics 2900 apparatus equipped with a TCD detector, aU-shaped quartz reactor working at atmospheric pressure, and aprogrammable furnace with temperature control. Typically, about100mg of the solid was placed on a plug of quartz wool; the tem-perature was monitored by a thermocouple in contact with thecatalyst bed. CO2 desorption profiles were recorded by monitor-

ing the TCD signal from about room temperature to 800◦

C at aheating rate of 10◦C/min. In the case of the calcined solids, thesamples were loaded into the reactor and a first treatment upto 800 ◦C in flowing He (50cm3/min) was carried out to removeadsorbedwater andcarbonates. Aftercooling to roomtemperature,25cm3/min of 10% CO2 in He was fed for 1h and then, the flowwas changed 50cm3/min of He. The temperature was increasedup to 100 ◦C to remove physically adsorbed CO2, cooled again toroom temperature, and after 20min for stabilization, the tempera-ture program was run. As for the rehydrated solids, saturation withCO2 was carried out over the fresh material because hydrotalcitewould decompose during the cleaning step. However, as rehy-drated hydrotalcites are very reactive toward atmospheric CO2, ablank experiment was first carried out with each sample by which

the solid was placed into the reactor, and directly heated fromroom temperature to 800 ◦C at a heating rate of 10 ◦C/min under50cm3/min of He while the TCD signal is recorded. After that, afreshsamplewasloadedinthereactorandsubjectedto25cm3/minof10%CO2 in He for 1 h at room temperature. Then the experimentcontinued as mentioned above for the calcined solids.

The basic strength (H ) was determinedby the methodbased onthe color change of Hammett indicators [62]. H is approximatelygiven bythe pK a of the indicator showing, upon adsorption, a colorintermediatebetween those of theacidic andbasic forms.The indi-cators employed in this work were (their pK a values are givenbetween parentheses): bromothymol blue (7.2), phenolphthalein(9.3), thimolphthalein (9.9), alizarine yellow (11.0) and indigocarmine (12.2). The determination of H was carried out placing

about 100 mg of the solid in a test tube, adding about 2cm3

of a

Fig.1. Photographs of themonolithic stirrer reactor usedfor themethanolysisreac-tions with the structured catalysts. Image on the left: emulsion obtained underreaction conditions. Image on the right: two monoliths coated with Mg–Al hydro-talcite attached to thestirrer shaft.

solution of the indicator in methanol and shaking vigorously. Afterequilibration, the suspension was examined for color change.

2.4. Transesterification reactions

Refined sunflower oil (Urzante, Navarra, Spain, acid valueof 0.07mg KOH/g) and HPLC grade methanol (Scharlau) wereselected as the reactants. All catalytic tests were carried out witha methanol/oil molar ratio of 48, catalyst concentration of 2 wt.%referred to the oil mass, atmospheric pressure and 60◦C. Reactionswith thepowder catalystswere conductedin a slurry100 cm3 jack-eted batch reactor equipped with a thermocouple for temperaturemonitoring, a polyamide tube to facilitate sampling and magnetic

stirring. Temperature was controlled with a thermostatic waterbath. Reactions with the structured catalyst were carried out ina monolithic stirrer reactor according to one of the configurationsdescribed by Hoek et al. and shown in Fig. 1 [63]. Two monolithsattached to the stirrer shaft were used in each catalytic run. A250cm3 jacketed glass batch reactor equipped with a thermocou-ple, a polyamide tube for sample extraction and reflux condenserwas used in this case. As can be seen in Fig. 1, good mixing con-ditions were obtained at the working stirring speed of 200rpm,resulting in a homogeneous emulsion.

After activation as described in Section 2.2, the catalysts werequickly transferred to the reactor containing the sunflower oil atthe reaction temperature. Tempered methanol was then addedto start the reaction. The performance of the different catalysts

was monitored by collecting samples at several reaction times.After extraction from the reactor with a syringe, samples werecentrifuged to remove the catalyst and diluted with tetrahydro-furan (THF, Scharlau, HPLC grade). Once filtered with Acrodisc®

filterswith0.2mnylonmembrane,sampleswereanalyzedbysizeexclusion chromatography (SEC) with differential refractive indexdetector at room temperature as described elsewhere [64].

3. Results and discussion

3.1. Catalysts characterization

The XRD patterns of selected samples are shown in Fig. 2. Theas-synthesized parent Mg–Al HT sample exhibits the characteris-

tic peaks of the double-layered hydroxides (LDHs) corresponding


4/9


0 10 20 30 40 50 60 70 80

2

(º)

Sepiolite

HT

HT_Sep10

HT_Sep30

I n t e n s i t y ( a . u . )

Fig. 2. XRD patterns of sepiolite, theparent Mg–Al hydrotalcite(HT) andthe solidsobtained after drying the corresponding slurries for monoliths coating.

to the hydrotalcite structure. Well-defined reflections at 11.7◦ and23.2◦ (2 ) are ascribedto diffraction bythe (00 3)and (00 6)planesof the LDHs structure. These peaks disappear in the patterns of the solids contained in the coating slurries. It should be notedthat the slurries were prepared using the calcined HT materialobtained after decomposing the hydrotalcite, leading to Mg(Al)Omixed oxides as identified by their characteristic diffractions at 43and 63◦ (2 ). It can be clearly seen that the intensity of the peakscorresponding to the sepiolite binder increases with its content inthe solids.

The TGA profiles of the calcined HT used for preparing the coat-ing slurries are shown inFig.3. Ascanbeseen,atotalweightlossof

about 20%takes place that should be due mainly to water adsorbedfrom theatmosphere. About 10% is lost after reaching 200◦C whichcan be attributed to physically adsorbed water[60]. The remaining10wt.% is lost between about 200 and 400 ◦C and can be due tostructural or even interlayer water that could be starting to restorethe hydrotalcite structure. This interpretation is supported by the

50

60

70

80

90

100

0 200 400 600 800 1000

W e i g h t ( % )

Temperature (ºC)

rehydrated HT

calcined HT

Fig. 3. TGA profiles of thecalcined Mg–Al hydrotalcite (HT) and the solid obtained

after rehydration of this sample.

fact that the main weight loss (about 30%) of the sample obtainedafter rehydration of the calcined HT solid (also included in Fig. 3)takes place also in this temperature range. According to Kloproggeet al. the TGA weight losses at these temperatures correspond tostructural hydroxide ions in the brucite-like layers and the decom-position of carbonate anions in the interlayer region [65,66]. Thetotalweightlossisinagreementwiththetypicalvaluesofthewateruptake after rehydration [60].

Regarding the textural properties, the calcined HT sampleexhibits a significantly high specific surface area of 163 m2/g andspecific total pore volume of 0.44cm3/g. However, after rehydra-tion the specific surface area drops to only 1m2/g. This is a typicalbehavior associated to an increased crystallinity of Mg-rich rehy-drated hydrotalcites [60]. The values of the textural properties of the sepiolite used as binder were somewhat similar to that of thecalcined HT, SBET =200m2/g and V p =0.36cm3/g. In the case of thesolids constituting the mixtures present in the coating slurries,their textural properties were in general the corresponding to thecontribution of the different components.

The basic properties of the methanolysis catalysts are of the outmost importance for their catalytic performance inview of the mechanism of the transesterification of triglyceri-des with methanol [67]. Strong basic sites are responsible formethanol activation thus allowing the methanolysis reaction toevolve [61].

TheresultsofthecharacterizationcarriedoutwiththeHammettindicators method are included in Table 1. Due to the limitations of this methodthe results can be considered only in a qualitative way.Regarding the calcined samples, it is interesting to note that thebasic strength decreases with respect to that of the Mg–Al mixedoxides as a consequence of the presence of sepiolite. It should benoted that sepiolite is a cationic clay, in contrast with hydrotalciteswhich are anionic. Therefore, it seems that a sort of neutralizationtakes place among the components of the coating slurries resultingin a lower basic strength. The results with the Hammett indicatorssuggest that little differences exist among the basic strengths of the several mixtures used in the slurries. As expected, the basic

strength increases after rehydration for all the solids. It is usu-ally considered that for rehydrated hydrotalcites, the most activesites seem to be located at the edges and defects of the platelets[68]. The increase of the basic strength is most notable for thesolids with the lowest sepiolite contents. In other words, the basicstrength is higher for the solids with hydrotalcite content above90wt.%.

The CO2-TPD results are included in Fig. 4. The TPD profilesof the calcined solids have two contributions (Fig. 4A). The mainpeak is centered at 155–157 ◦C for the calcined HT and HT Sep5and HT Sep10 samples and shifts to about 150◦C for the sam-ples with highest sepiolite contents. It is reasonable that thissignal corresponds to CO2 adsorbed on very weak basic sites.The CO2-TPD results confirm that sepiolite is a non-basic mate-

rial. The second contribution is associated to a very small peaklocated at about 350 ◦C for all samples that correspond to CO2

Table 1

Basic strength (H ) of the samples indicated according to the Hammett indicatorsmethod.

Sample Basic strength (H )

Calcined Rehydrated

HT 9.3 < H < 9.99 9.9 < H < 11.0Sepiolite – –HT Sep5 7.2 < H < 9.3 9.3 < H < 9.99HT Sep10 7.2 < H < 9.3 9.9 < H < 11.0HT Sep20 7.2 < H < 9.3 7.2 < H < 9.3HT Sep30 7.2 < H < 9.3 7.2 < H < 9.3


5/9


0 100 200 30 0 4 00 50 0 6 00 7 00 8 00

T C D S

i g n a l ( a . u . )

Temperature (°C)

HT

HT_Sep5

HT_Sep10

HT_Sep30

HT_Sep20

Sepiolite

(A)

0 100 200 300 4 00 50 0 60 0 7 00 8 00

Temperature (°C)

T C D s

i g n a l ( a . u . )

without CO2 adsorption

with previous

CO2 adsorption

Rehydrated HT

(B)

Fig.4. CO2-TPD profiles ofthe calcinedsolidsindicated (A), andof therehydratedcalcinedHT hydrotalcitewith andwithout previous adsorption ofCO2 in theTPD apparatus(B).

adsorbed on the strongest basic sites [60]. Integration of the peaksallowed estimating the total amounts of CO2 desorbed, that rangedbetween 0.44mmol/g for HT, 0.39 and 0.37mmol/g for HT Sep5and HT Sep10, respectively, and 0.30mmol/g for both HT Sep20and HT Sep30. These results are in accordance with the previouslydiscussed effect of sepiolite on the mixtures basicity.

The CO2-TPD profiles obtained with the rehydrated HT sampleare shown in Fig. 4B. As can be seen, in this case there is a mainnarrow peak centered at about 360 ◦C that reveals the presence of strong basic sites in this solid. Moreover, as explained in Section2.3, experiments were carried out both without and with previoussaturation of the sample with 10% CO2 in He. Even without previ-ous CO2 adsorption, a peak corresponding to 0.60 mmol CO2/g wasrecorded that should correspond to CO2 adsorbed from the atmo-sphereandevidencesthehighreactivityoftherehydratedmaterial.Furthermore, after previous saturation of the sample in the CO2/Hemixture, theamount of CO2 desorbed increases to 0.91mmol CO2/gthus showing that the stored material still contains reactive basicsites.

3.2. Adherence tests

A first series of adherence tests were carried outwith monolithswashcoated with hydrotalcite slurries prepared using differentsolvents and without any binder. The tests were performedin petroleum ether and the results obtained are included inTable 2.

Water is not suitable for washcoating monoliths with Mg–Alsolids. Rehydration of the calcined solid in the slurry resulted ingel formation and prevented catalyst preparation. The adherencemarkedlyimproved changing thesolvent of theslurriesfrom short-chain alcohols, particularly methanol, that allowed increasing theadherence of the coating to 75%. Despite this improvement, the

Table 2

Characteristics of the slurries indicated and results of the adherence tests inpetroleum ether.

Solvent Solids content (%) Viscosity (cP) Adherence (%)

Water 20 Gel –10 Gel –

Methanol 20 5 75Ethanol 20 6 64Propanol 20 5 23

Table 3

Characteristics of some slurries and results of theadherence tests in methanol.

Sample Solids c ontent (%) Sepiolite content (%) Adherence (%)

HT 20 – 30HT Sep10 20 10 87HT Sep30 20 30 96

monoliths prepared using methanolic suspensions of the calcinedHT showed poor adherence in methanol (30%), as indicated inTable 3. It should be noted that methanol is one of the reactantsof the biodiesel synthesis so the adherence in this medium is animportant factor to take into account. This is the reason why wedecided to include in thecoating slurryformulation a typical bindersuch as sepiolite fibers. As can be seen, the presence of sepioliteremarkably improved the adherence in methanol medium. More-over, the adherence increased with the sepiolite content, reaching96% for the HT Sep30 sample. The reason for this improvementappears to be the ability of the sepiolite fibers to crosslink the par-ticles forming aggregates which increase the adherence. This canbeseenin Fig. 5, particularly for the sample HT Sep30 (Fig. 5C) thatshows a high density of crosslinking.

Fig. 5. SEM imagesof the calcinedHT (A),HT Sep10 (B),HT Sep30 (C) and sepiolite

(D).


6/9


3.3. Catalytic performance

The catalytic performance of the parent Mg–Al hydrotalcite willbe considered first. In accordance with previous literature results[60,69], this sample resulted almost inactive for the methanoly-sis of sunflower oil. Rehydration after calcination results in theformation of meixnerite (Mg6Al2(OH)18·4H2O), a hydrotalcite inwhich hydroxides are the only compensating anions present in theinterlayer space. This feature is called as retro-topotactical trans-formation and also known as “memory effect” that is characteristicof hydrotalcites provided that the decomposition temperature islow enough as to avoid the formation of the MgAl2O4 spinel thatmakes the thermal decomposition irreversible [70]. Xi and Davisfound that the methanolysis activity of Mg–Al hydrotalcites recon-structed after rehydration was much higher than that of the parenthydrotalcite or the oxides resulting after calcination [71,72]. It wasconcluded that the interlayer hydroxide anions of meixnerite actas Brønsted sites active for the methanolysis reaction whereas thehydroxide anions in the brucite-like layers are considered to havenegligible catalytic activity.

In this work, rehydration has been carried out by immersionin boiling water and three distinct procedures for removing theexcess water have been considered as explained in Section 2.2:until total evaporation of the water contained in the rehydrata-tion recipient (HT RH ev), and removing excess water by gravityfiltration (HT RH gf) orvacuumfiltration (HT RH vf).The evolutionof the sunflower oil conversion with reaction time obtained usingthe rehydrated HT hydrotalcite in powder form obtained accord-ing to these procedures is shown in Fig. 6. It can be seen that themost active sample is the one obtained when water is removedin the rehydration flask by evaporation to dryness but avoidingoverheating. In this case, about 96% oil conversion is reached after24h of reaction. In contrast, if water is removed by filtration theoil conversion after 24h decreases to 70–72% with few differencesbetween gravity and vacuum filtration. As discussed in the pre-vious section, rehydrated hydrotalcites are very reactive towardatmospheric CO2. Water removal by filtration takes more time in

contact with the atmosphere which can explain the lower activ-ity due to deactivation of part of the active sites by carbonation.For this reason, the results shown hereafter correspond to samples

0

0,2

0,4

0,6

0,8

1

0 5 10 15 20 25

O i l c o n

v e r s i o n

Time (h)

HT_RH_ev

HT_EtOH

HT_THF

HT_THF/EtOH

HT_RH_gf

HT_RH_vf

Fig. 6. Evolution of the sunflower oil conversion with reaction time over the rehy-drated HT hydrotalcite obtained after applying different procedures for removingthe water in excess as well as for recovered, washed with different solvents, and

reactivated hydrotalcites.

rehydrated according to the procedure by which water in excess isremoved by evaporation.

Another important issue is the possibility of catalyst reuti-lization. It is well-know that although hydrotalcites are verystable against leaching during methanolysis, they suffer fromsignificantdeactivationbystrongadsorptionofglyceridesandglyc-erol [59,73]. In order to investigate the catalysts reusability, theHT RH ev sample was recovered after reaction by centrifugation of the reaction mixture, and thoroughly washed using different sol-vents. Four cycles of washing/centrifugation (6000rpm for 20min)were applied in all cases. The solvents used were THF (HT THF),ethanol (HT EtOH), and in one case the first cycle was conductedwith THF and the other 3 with ethanol (HT THF/EtOH). The recov-ered and washed solids behaved very similarly in a subsequentmethanolysisreaction, reaching about 40% oilconversion after 24hof reaction.The activity of thesamples recoveredand washed couldbe improved after subjecting them to a reactivation through cal-cination at 500◦C for 6 h and rehydration. This can be seen inFig. 6 which shows that the activity of the reactivated sample ishigher when only THF was used as cleaning solvent. So whereasthe HT THF rehydrated solid gave 69% oil conversion after 24h,this value was 49%when thepreviouswashing wasperformedonlywith ethanol.

The catalytic performance after rehydration of the solidsobtained by drying the slurries used for washcoating the metal-lic monoliths is shown inFig.7. It can be seen that the sunflower oilconversion after 24h of reaction ranges from 96% for the bare HThydrotalcite to 41%for the sample with thehighest content of sepi-olite (HT Sep30). The activity results correlate with that of thebasicstrength of these materials, and then, with their sepiolite content.Indeed, both HT and HT Sep10 give very similar results, and theywere also the most active catalysts. As indicated in Table 1, afterrehydration, these solids exhibited the strongest basic sites, capa-ble of transforming alizarine yellow (pK a = 11.0) to its basic form.Theconversion obtainedwith HT Sep5 wasslightlylower,86%after24h ofreaction. Forsolidswithsepiolite contentsabove10 wt.%,theoil conversion clearly decreases as the sepiolite content increases.

Conversion values after 24h of reaction decrease from81 to 64andfinally41%for the HT Sep17.5,HT Sep20 and HT Sep30 samples.Asdiscussed in Section 3.1, the detrimental effect of sepiolite on the

0

0,2

0,4

0,6

0,8

1

0 5 10 15 20 25

Time (h)

HT_Sep10

HT

O i l c o n v e r s i o n

HT_Sep17.5

HT_Sep20

HT_Sep30

HT_Sep5

Fig. 7. Evolutionof thesunflower oil conversion with reaction time over thesolidsobtained after drying the slurries used to washcoat the metallic monoliths. All the

catalysts were activated through rehydration.


7/9


0

0,2

0,4

0,6

0,8

1

0 2 4 6 8 10 12

O i l c o n

v e r s i o n

Time (h)

M_HT_Sep10

M_HT

M_HT_Sep17.5

M_HT_Sep20

M_HT_Sep30

M_HT_Sep5

Fig. 8. Evolutionof thesunfloweroil conversion with reaction time over themono-lithic catalysts. Filled symbols: performance during the first use. Open symbols:performance of the recovered, washed and reactivated monoliths.

catalytic activity can be interpreted in terms of acid–base interac-tions that lead to a partial neutralization of the hydrotalcite sitesactive for methanolysis. However, other effects such as maskingcannot be ruled out. Indeed, Rasmussen et al. [74] and Ávila et al.[75] studied the influence of the sepiolite binder on the catalyticproperties of mixed oxides. The presence of increasing quantitiesof sepiolite produced a progressive worsening of the catalytic per-formance because it tended to cover the metal oxide particles. Itwas found that high sepiolite contents strongly masked the cata-lyst, changing its morphology and increasing the average particlesize. In our case, both HT and HT Sep10 have very similar particlesize (about5 m); however, the particle size of HT Sep30 increases

to 9m. Moreover,the FESEM image shows thesepiolite fibers cov-ering the Mg–Al particles (see Fig. 5C) which can explain the lowercatalytic activity.

The catalytic performance of the monolithic catalysts is shownin Fig. 8. When comparing these results with that of the pow-der catalysts in Fig. 7 it should be realized that the maximumduration of the catalytic tests with the monoliths was 10–11h.As can be seen there are no big differences between the per-formances of the monoliths with coatings containing less than17.5wt.% of sepiolite. The oil conversion after 10h of reactionranged between 62% for M HT Sep10 and 77% for M HT Sep5.Similar values were obtained with these catalysts in the powderform which after 10h of reaction achieved between 65% oil con-version for HT Sep17.5 and 79% for both HT and HT Sep10 (see

Fig. 7). The monoliths with coatings containing 20 and 30wt.%of sepiolite exhibited poor catalytic performance. They roughlyachieved 30 and 10% oil conversion, respectively, after 6h of reaction.

It should be noted that monoglycerides and diglycerides areintermediate products of the methanolysis reaction that involves 3reversiblestepsinseries [76,77]. Theevolutionoftheyieldsoftheseproducts asa function of the oil conversion is shown inFig.9 whichincludes the results obtained with all the powder and monolithiccatalysts included in this study. As can be seen, it can be hardlydistinguished between the performance of the powders and thatof the monoliths which indicates that the chemical nature of thecoatings andthe powders is similar. Maximum yield of diglyceridesis about 8% at 20% oil conversion, whereas the maximum yield of

monoglycerides, about 6%, is reached when the oil conversion is

0,00

0,05

0,10

0,15

0,20

0 0,2 0,4

0,6 0,8

1Oil conversion

M o n o g l y c e r i d e s y i e l d

0,05

0,10

0,15

0,20

D i g l y c e r i d e s Y i e l d

Fig. 9. Evolution of the monoglycerides and diglycerides yields as function of thesunflower oil conversion during the methanolysis reaction. Filled symbols: powdercatalysts. Open symbols: monolithic catalysts.

almost complete, in accordance with the in series scheme of thetriglycerides methanolysis reaction.

Our best results, obtained with the M HT Sep5 monoliths, arevery modest. Nevertheless, it should be emphasized the very mildreaction conditions considered in this work. Indeed, working at60 ◦C and atmospheric pressure are typical conditions of the con-ventional synthesis of biodiesel withhomogeneous catalystswhich

intrinsically exhibit much higher specific activity that the hetero-geneous catalysts. There are very few reports on the methanolysisof vegetable oils with structured catalysts. Kolaczkowski et al.obtained 54% rapeseed oil conversion after 2h of reaction work-ing at 195◦C, 20bar, and methanol to oil molar ratio of 12 using acordierite monolithcontaining 0.3g of a Zn aminoacidcomplexcat-alyst [53]. Tonetto and Marchetti used a monolithic stirrer reactorloadedwithcordieritemonolithswashcoatedwithK/Al2O3 catalyst[54]. 59% FAMEs yield was obtained after 6h of reaction at 120 ◦Cwith 0.5 wt.% catalyst concentration and methanol to soybean oilmolar ratio of 32. It can be expected that the methanolysis activityof the hydrotalcite-based monolithic catalysts would significantlyincrease under the reaction conditions, particularly the high reac-tion temperatures, used in these works.The main problem with the

monoliths washcoated with hydrotalcite is, however, that of stabil-ity under reaction conditions. In spite of the good results obtainedduring the adherence tests in methanol (see Section 3.2), the lackof adherence in the reaction mixture leads to a poor performanceof the reutilized monoliths (Fig. 8, open symbols). The reactionmedium is a complex mixture containing, in addition to methanol,oil, biodiesel, glycerol and monoglycerides and diglycerides. Expe-rience with biodiesel in the transportsectorhas demonstrated thatthis biofuel has extremely good lubricating and solvent properties.So, achieving good adherence in this aggressive medium consti-tutes a challenge from the point of view of the preparation of structured methanolysis catalysts. Further work is under progressinour laboratoriesin orderto improve theadherenceof thecoatingsbased on Mg–Al hydrotalcite which is a promising methanolysiscatalyst.


8/9

http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0325http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0325http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0325http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0325http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0325http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0325http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0325http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0325http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0325http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0325http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0325http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0325http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0325http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0320http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0320http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0320http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0320http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0320http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0320http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0320http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0320http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0320http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0320http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0320http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0320http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0320http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0320http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0320http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0320http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0320http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0320http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0315http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0315http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0315http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0315http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0315http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0315http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0315http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0315http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0315http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0315http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0315http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0315http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0315http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0315http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0315http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0310http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0310http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0310http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0310http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0310http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0310http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0310http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0310http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0305http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0305http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0305http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0305http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0305http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0305http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0305http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0305http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0305http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0305http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0305http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0305http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0305http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0305http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0305http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0305http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0305http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0305http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0300http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0300http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0300http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0300http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0300http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0300http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0300http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0300http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0300http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0300http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0300http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0300http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0300http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0300http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0300http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0300http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0300http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0300http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0300http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0300http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0300http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0300http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0300http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0300http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0300http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0300http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0295http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0295http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0295http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0295http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0295http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0295http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0295http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0295http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0295http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0295http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0295http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0295http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0295http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0295http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0295http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0295http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0295http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0295http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0290http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0290http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0290http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0290http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0290http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0290http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0290http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0290http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0290http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0290http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0290http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0290http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0290http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0290http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0290http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0290http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0290http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0285http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0285http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0285http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0285http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0285http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0285http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0285http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0285http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0285http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0285http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0285http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0285http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0285http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0285http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0285http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0285http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0285http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0285http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0285http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0280http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0280http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0280http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0280http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0280http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0280http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0280http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0280http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0275http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0275http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0275http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0275http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0275http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0275http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0275http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0275http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0275http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0275http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0275http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0275http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0270http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0270http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0270http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0270http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0270http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0270http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0270http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0270http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0270http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0270http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0270http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0265http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0265http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0265http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0265http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0265http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0265http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0265http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0265http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0265http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0265http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0265http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0265http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0260http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0260http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0260http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0260http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0260http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0260http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0260http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0260http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0260http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0260http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0260http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0260http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0260http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0260http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0260http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0260http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0260http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0260http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0260http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0255http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0255http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0255http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0255http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0255http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0255http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0255http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0255http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0255http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0255http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0255http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0255http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0255http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0255http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0255http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0255http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0255http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0255http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0255http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0250http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0250http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0250http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0250http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0250http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0250http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0250http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0250http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0250http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0250http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0250http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0250http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0250http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0250http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0245http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0245http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0245http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0245http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0245http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0245http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0245http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0245http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0245http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0245http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0245http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0245http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0245http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0245http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0245http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0245http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0245http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0240http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0240http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0240http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0240http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0240http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0240http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0240http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0240http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0240http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0240http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0240http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0240http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0240http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0240http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0240http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0240http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0240http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0240http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0235http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0235http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0235http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0235http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0235http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0235http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0235http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0235http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0235http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0235http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0235http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0235http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0235http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0235http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0235http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0235http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0235http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0230http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0230http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0230http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0230http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0230http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0230http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0230http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0230http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0230http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0230http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0230http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0230http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0230http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0225http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0225http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0225http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0225http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0225http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0225http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0225http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0225http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0225http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0225http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0225http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0225http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0225http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0225http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0220http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0220http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0220http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0220http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0220http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0220http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0220http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0220http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0220http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0220http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0220http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0220http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0220http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0220http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0220http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0215http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0215http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0215http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0215http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0215http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0215http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0215http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0215http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0215http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0215http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0215http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0215http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0215http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0215http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0215http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0210http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0210http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0210http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0210http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0210http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0210http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0210http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0210http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0210http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0210http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0210http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0210http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0205http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0205http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0205http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0205http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0205http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0205http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0205http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0205http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0205http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0205http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0205http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0205http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0205http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0200http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0200http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0200http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0200http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0200http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0200http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0200http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0200http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0200http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0200http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0200http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0200http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0200http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0200http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0200http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0200http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0200http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0200http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0200http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0200http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0195http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0195http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0195http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0195http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0195http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0195http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0195http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0195http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0195http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0195http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0195http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0195http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0195http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0195http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0195http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0195http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0195http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0190http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0190http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0190http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0190http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0190http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0190http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0190http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0190http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0190http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0190http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0190http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0190http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0190http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0190http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0185http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0185http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0185http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0185http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0185http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0185http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0185http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0185http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0185http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0185http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0185http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0185http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0185http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0185http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0180http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0180http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0180http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0180http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0180http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0180http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0180http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0180http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0180http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0180http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0180http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0180http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0180http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0180http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0180http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0180http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0180http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0180http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0180http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0175http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0175http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0175http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0175http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0175http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0175http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0175http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0175http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0175http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0175http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0175http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0175http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0175http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0175http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0175http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0175http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0175http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0170http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0170http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0170http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0170http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0170http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0170http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0170http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0170http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0170http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0170http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0170http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0170http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0170http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0170http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0170http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0170http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0170http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0170http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0170http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0170http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0170http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0165http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0165http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0165http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0165http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0165http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0165http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0165http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0165http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0165http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0165http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0165http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0165http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0160http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0160http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0160http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0160http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0160http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0160http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0160http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0160http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0160http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0160http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0160http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0160http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0160http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0160http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0160http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0155http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0155http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0155http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0155http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0155http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0155http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0155http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0155http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0155http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0155http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0155http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0155http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0155http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0155http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0155http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0155http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0155http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0155http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0155http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0150http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0150http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0150http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0150http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0150http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0150http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0150http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0150http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0150http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0150http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0150http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0150http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0150http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0150http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0150http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0150http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0150http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0150http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0150http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0150http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0150http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0150http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0145http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0145http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0145http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0145http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0145http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0145http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0145http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0145http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0145http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0145http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0145http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0145http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0145http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0145http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0145http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0145http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0145http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0145http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0145http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0145http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0145http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0145http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0145http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0140http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0140http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0140http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0140http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0140http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0140http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0140http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0140http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0140http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0140http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0140http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0140http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0140http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0140http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0140http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0140http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0140http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0140http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0140http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0140http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0135http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0135http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0135http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0135http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0135http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0135http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0135http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0135http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0135http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0135http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0135http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0135http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0135http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0135http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0135http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0135http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0135http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0130http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0130http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0130http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0130http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0130http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0130http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0130http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0130http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0130http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0130http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0130http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0130http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0130http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0130http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0130http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0130http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0130http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0130http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0130http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0130http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0125http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0125http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0125http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0125http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0125http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0125http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0125http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0125http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0125http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0125http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0125http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0125http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0125http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0125http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0125http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0125http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0125http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0120http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0120http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0120http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0120http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0120http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0120http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0120http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0120http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0120http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0120http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0120http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0120http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0120http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0120http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0120http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0120http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0120http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0120http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0120http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0120http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0115http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0115http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0115http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0115http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0115http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0115http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0115http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0115http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0115http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0115http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0115http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0110http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0110http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0110http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0110http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0110http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0110http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0110http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0110http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0110http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0110http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0110http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0110http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0110http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0110http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0105http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0105http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0105http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0105http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0105http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0105http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0105http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0105http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0105http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0105http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0105http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0105http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0105http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0105http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0105http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0105http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0105http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0105http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0105http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0105http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0105http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0105http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0105http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0100http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0100http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0100http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0100http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0100http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0100http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0100http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0100http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0100http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0100http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0100http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0100http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0100http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0100http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0100http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0100http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0100http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0100http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0100http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0100http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0100http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0100http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0095http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0095http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0095http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0095http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0095http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0095http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0095http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0095http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0095http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0095http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0095http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0095http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0095http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0095http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0095http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0095http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0090http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0090http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0090http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0090http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0090http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0090http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0090http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0090http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0090http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0090http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0090http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0090http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0090http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0085http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0085http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0085http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0085http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0085http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0085http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0085http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0085http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0085http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0085http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0085http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0085http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0085http://refhub.elsevier.com/S0920-5861(13)00211-3/sbref0085http://refhub

structured catalysts based on mg–al hydrotalcite for the synthesis of

Documents