A new peroxo-route for the synthesis of Mg–Zr mixed oxides catalysts: Application in the gas phase acetone self-condensation

Applied Ca alysis A: Gene al 477 (2014) 26–33 Con en s lis s a ailable a ScienceDi ec Applied Ca alysis A: Gene al jou n al hom epage: www.else ie .com/loca e/apca a A new pe oxo- ou e o he syn hesis o Mg–Z mixed oxides ca alys s: Applica ion in he gas phase ace one sel -condensa ion Igo K i so a, Lau a Fabab, E a Díazb, Sal ado O dó˜ nezb,∗, Viachesla A dina, Se gei Khainako c, Jose R. Ga ciac aNano echnology Resea ch and Educa ion Cen e , Depa men o Chemis y, Sou h U al S a e Uni e si y, Lenina a . 76, Chelyabinsk, Russia bDepa men o Chemical and En i onmen al Enginee ing, Uni e si y o O iedo, JuliánCla e ía s/n, O iedo 33006, Spain cDepa men o O ganic and Ino ganic Chemis y, Uni e si y o O iedo, JuliánCla e ía s/n, O iedo 33006, Spain a i c l e i n o A icle his o y: Recei ed 20 Decembe 2013 Recei ed in e ised o m 3 Ma ch 2014 Accep ed 4 Ma ch 2014 A ailable online 13 Ma ch 2014 Keywo ds: Magnesia–zi conia Ace one condensa ion Aldol condensa ion Zi conium pe oxocomplex a b s a c We p opose in his manusc ip a new pe oxo-media ed p ocedu e o p epa ing magnesia–zi conia mixed oxides, wi h Mg/Z mola a io be ween 1 and 3, wi h enhanced dis ibu ion o basic si es. The mixed magnesia–zi conia oxides ha e been p epa ed om he gelled complex by Pechini- ype me hod. The MgO–Z O2ma e ials ha e been cha ac e ized and used as ca alys s o ace one aldol condensa ion. The p oposed p epa a ion me hod p o ides a high deg ee o molecula homogenei y and a ou s he o ma- ion o magnesia-s abilized zi conia phase. Ace one gas-phase sel -condensa ion was ca ied ou o e hese ca alys s as model eac ion equi ing he p esence o basic si es. The condensa ion yields diace one alcohol and mesi yl oxide as mean C6 p oduc s, and pho ones, isopho ones and mesi ylene as C9 p od- uc s. In compa ison o Mg–Z oxide p epa ed by co-p ecipi a ion, hese new ma e ials p esen be e con e sions and highe selec i i y o linea dime s and ime s (as mesi ylene), whe eas he selec i i y o isopho ones is significan ly lowe . © 2014 Else ie B.V. All igh s ese ed. 1. In oduc ion Complex oxides o IV g oup o elemen s o he pe iodic able and alkaline ea h me als ha e a ac ed g ea a en ion o scien is s and indus y o a long ime, since hey can be used as conduc i e ce amics [1], biocompa ible ma e ials [2], adso ben s [3], and ca a- lys s [4]. This las applica ion is g owing up in impo ance due o he a ising demand o ca alys s wi h bo h basic and acid si es [5–10]. In he case o magnesia–zi conia ca alys s, he mos common p epa a- ion p ocedu e is he co-p ecipi a ion o he ino ganic sal s, ni a es o chlo ides [11], and mo e a ely ul a-dilu ion me hods [12], and ci a e- o oxala e p ocessing [13,14]. The p ocesses o phase o ma- ion, polymo phic ansi ions (in p ecipi a ed magnesia–zi conia o p epa ed ia con en ional sol–gel Pechini-me hod) as well as hei su ace p ope ies, a e ho oughly s udied [15,16]. Howe e , i is known ha many p ope ies o sol–gel o p ecipi a ed ma e- ials based on zi conia a e de e mined by ac o s such as he sal anion used in he syn hesis [17], he pH o he eac ion mix u e and e en he o de o mixing wi h p ecipi a ing agen [18]. These ∗Co esponding au ho . Tel.: +34985103437; ax: +34985103434. E-mail add ess: [email p o ec ed] (S. O dó˜ nez). pa ame e s a ec he p ope ies and he s uc u e o he final p od- uc as hey de e mine he hyd olysis a e, and he oligome iza ion and polyme iza ion o aqueous complexes o zi conium, no exis - ing in monome ic o m e en in e y acidic medium. Applica ion o alkoxides sol es some p oblems by con olling he hyd olysis and polyme iza ion o zi conia species and, as esul , high su - ace a ea mesopo ous ma e ials a e ob ained [19]. Alkoxide sol–gel echnique has p o ed o be he bes me hod o p epa e high pe - o mance me al oxide ca alys s, bu i has se e al disad an ages s anding on he way o i s applica ion on indus ial scale. Alkoxides a e much mo e expensi e han ino ganic sal s, uns able and eas- ily hyd olyzed unde he influence o a mosphe ic wa e apo s, hus special p ecau ions a e necessa y when handling hem [20]. Fu he mo e, alkoxides a e highly oxic compounds and a g ea amoun o o ganic sol en s is indispensable du ing alkoxide-based sol–gel syn hesis, hus, hey do no mee equi emen s o sus ain- able chemical p ocesses. The pe oxocomplexes o some ansi ion me als could be con- side ed as a “g een” al e na i e o me al alkoxide p ecu so s [21]. Al hough, di e en s udies on he syn hesis o pe oxocomplexes [21,22] ha e been epo ed, s udies on zi conium pe oxocomplexes as he p ecu so s o Z O2o zi conium-based complex oxides a e sca ce. Gao e al. [23] epo ed he pe oxo-me hod o zi conia h p://dx.doi.o g/10.1016/j.apca a.2014.03.008 0926-860X/© 2014 Else ie B.V. All igh s ese ed. I. K i so e al. / Applied Ca alysis A: Gene al 477 (2014) 26–33 27 hin-film p epa a ion, whe eas Pa k e al. [24] ound some ad an- ages o pe oxo-media ed p ocedu e o Z O2films syn hesis, as such as he homogeneous dis ibu ion on he subs a e and high densi y o he film. High deg ee o homogenei y o he complex oxide is also ound o be one o he ad an ages o pe oxo- ou es, as i allows lowe ing he empe a u es o hea ea men while ob aining he desi able phase [25,26]. A deep knowledge abou he cha ac e is ics o mixed oxides o magnesia and zi conia is o key impo ance in basic he e ogeneous ca alysis. Mixed MgO/Z O2has shown high ca aly ic ac i i y in u - u al aldol condensa ion wi h ace one [27] and, as i has ecen ly been epo ed, in gaseous ace one sel -condensa ion [28]. Ace one sel -condensa ion is a highly impo an indus ial and scien ifi- cally in e es ing eac ion, since eme ging chemical and biological p ocesses ha e become he ace one in o a bio-based pla o m molecule. Upg ading o ace one elies on he o ma ion o new C–C bonds, yielding mo e complex molecules, being mesi yl oxide, isopho ones, and mesi ylene he mos in e es ing p oduc s [29]. Despi e he g ea in e es o his eac ion, he easons o selec- i i y owa ds ce ain p oduc s a e s ill unclea o he esea che s, since many p ope ies o he ca alys s ha e influence on he way o he eac ion. I is likely ha no only he dis ibu ion o acid-basic si es de e mines he ac i i y and selec i i y o he ca alys , bu also he deg ee o molecula mixing in he oxide may play a significan ole, as i has been poin ed ou by Sádaba e al. [30]. A new p ocedu e o he p epa a ion o MgO/Z O2mixed oxides (using wa e -soluble pe oxocomplex as p ecu so s) is epo ed, desc ibing hei s uc u al, ex u al and physicochemi- cal p ope ies, and co ela ing hese p ope ies wi h he ca aly ic pe o mance in he gas phase ace one sel -condensa ion. 2. Expe imen al 2.1. Chemicals Mixed magnesia–zi conia oxides we e p epa ed using mag- nesium sul a e (MgSO4) ob ained om P olabo (98% pu e) and zi conium oxychlo ide aqueous solu ion (Z OCl2), con aining 19–21 w % o Z O2, pu chased om MEL Chemicals as MgO and Z O2sou ces, espec i ely. Hyd ogen pe oxide 30 w % wa e solu- ion, anhyd ous ci ic acid (97% pu e) and ace one (99.8% pu e) we e supplied by Ald ich; whe eas sodium hyd oxide (99%pu e) was supplied by P olabo. All chemicals we e used as ecei ed wi h- ou u he pu ifica ion. As zi conium sal s usually o m polyme ic hyd oxo-species in aqueous solu ion, leading o p ecipi a ion and changing o concen a ion, he con en o zi conium was de e - mined g a ime ically p io o i s use o syn hesis. 2.2. Syn hesis o MgO/Z O2mixed oxides A new pe oxocomplex-media ed ou e o p epa e mixed magnesia–zi conia oxides was accomplished in se e al s ages. Se - e al s eps a e coinciden wi h a p ocedu e desc ibed ea lie o he p epa a ion o SiO2/TiO2mixed oxides [31]. In he fi s s age, he con en ional co-p ecipi a ion p ocedu e desc ibed by A amen- dia e al. [11] was applied. Ini ially, he s a ing aqueous solu ions o MgSO4(0.2, 0.4 and 0.6 M concen a ions we e used in o de o achie e he desi ed Mg/Z a io) and 50-mL o Z OCl2(0.2 M) we e mixed and p ecipi a ed while s i ing wi h 1.5 M NaOH solu- ion a pH alue equal o 10.7. The a e o s i ing o a magne ic s i e a ied om 200 pm in he beginning o he p ecipi a ion o 600 pm when he suspension became iscous. The p epa ed p ecipi a es we e cen i uged a 3000 pm and washed 8 imes wi h deionized wa e o emo e sodium sul a es and chlo ides. The pe oxocomplexes we e p epa ed in he ollowing way: he mix u e o med by 20 mL o H2O2and 20 mmol o anhyd ous ci ic acid was added o he p ecipi a es, he ob ained suspensions we e hea ed a 373 K unde s i ing (200 pm) o 15 min. The p ecip- i a e wi h Mg/Z = 3 was o ally dissol ed and clea solu ion was ob ained, whe eas in o he cases he p ecipi a es dissolu ion was incomple e and he solu ion was isola ed om he solid phase by cen i uga ion a 3000 pm. The final s age o he syn hesis was accomplished ia Pechini- ype sol–gel p ocess. Wa e om he solu ions was e apo a ed on a wa e ba h ha caused gela ion, hen gels we e d ied in an o en a 333 K o 24 h, powde ed in an aga e mo a , hea ed a 473 K o 3 h ( o decomposi ion o he com- plexes), and finally ea ed a 873 K o 4 h in ai o ob ain mixed oxides wi h Mg/Z mola a ios equaling 1.0, 2.0 and 3.0 ( u he designa ed as 1MZ, 2MZ and 3MZ). Solid wa e -soluble pe oxo- complexes we e also subjec ed o in es iga ion. Fi s , hey we e isola ed om wa e solu ions by p ecipi a ion wi h e hanol. The p ecipi a es we e washed wi h 25 mL o e hanol o 6 imes and d ied a 333 K o 24 h p io o in es iga ion. The 4MZ sample was p epa ed by co-p ecipi a ion, acco ding o he p ocedu e desc ibed in a p e ious wo k [27]. 2.3. Ca alys cha ac e iza ion X- ay di ac ion pa e ns we e egis e ed om he powde ed samples by a Phillips X’Pe di ac ome e , ope a ing a Cu K␣ line ( = 0.154 nm) in he ange o 2 be ween 20 and 70◦. The - moanaly ical in es iga ion o he complex was ca ied ou using simul aneous TG/DSC he mal analyze Ne zsch 449F1. The pow- de ed sample was placed in he pla inum c ucible and hea ed in ai om oom empe a u e o 1273 K wi h a hea ing a e o 5 K/min. FTIR spec a we e egis e ed by a B uke Tenso 27 spec ome e in he ange o 400–4000 cm−1 om he pelle s o he complex and mixed oxides powde ed wi h KB wi h a esolu ion o 4 cm−1. The ex u al cha ac e iza ion was ca ied ou by ni ogen physiso p ion a 77 K in he Mic ome i ics ASAP 2020 su ace a ea and po osi y analyze . Be o e analysis, he samples we e ou gassed a acuum condi ions (<10−3kPa) a 473 K o 4 h. Su ace a ea was ob ained om he BET me hod, whe eas he po e size dis ibu ion and o al po e olume we e ob ained om he BJH app oach. The s eng h and dis ibu ion o he basic/acid si es we e de e mined by em- pe a u e p og ammed deso p ion o p eadso bed CO2and NH3, espec i ely, in a Mic ome i icsTPD/TPR 2900 appa a us. Samples (10 mg) we e p e ea ed in He a 723 K o 2.5 h and exposed o CO2 o NH3(2.5% NH3in He) s eam a 298 K empe a u e un il sa u- a ion co e age was eached. Weakly adso bed CO2o NH3was emo ed by flushing wi h He a he same empe a u e o abou 1.3 h. The empe a u e was hen inc eased a a linea a e o 5 K/min om 298 K o 723 K and he CO2o NH3deso p ion was moni o ed by mass spec ome y. An Ox o d Ins umen EDS a ached o a Jeol 6460LV scanning elec on mic oscope was used o de e mining he elemen al composi ion o he oxides. The su ace composi- ion o he mixed oxides was measu ed by X- ay Pho oelec on Spec oscopy (XPS), using a SPECS sys em equipped wi h a Hemi- sphe ical Phoibos de ec o ope a ing in a cons an pass ene gy, using Mg-K␣ adia ion (h· = 1253.6 eV). 2.4. Ace one sel -condensa ion A 150 mg sample o each ca alys was placed in a 0.4 cm i.d. U- shaped qua z ube loca ed in a PID- con olled u nace. This eac o was connec ed o he eac ion ini ial gas flow. The ace one was injec ed by a sy inge pump as a liquid in a He flow (0.05 L/min) and apo ized in si u, ob aining a olume concen a ion o 3.2%. Ou - going gases om he eac o we e analyzed online by a Shimadzu GC-2010 gas ch oma og aph wi h a FID de ec o . Be o e any ba ch, he ca alys was p e ea ed in He a 723 K o 1 h. The ace one gas 28 I. K i so e al. / Applied Ca alysis A: Gene al 477 (2014) 26–33 75010001250150017502000 500 Wa enu mbe s (c m –1 ) T ansmi ance (a.u.) 910 ν (O–O) Fig. 1. FTIR spec a o he p ecu so o 1MZ sample. phase condensa ion was s udied in a empe a u e ange om 473 o 723 K wi h s eps o 50 K. Mo e de ails abou he eac ion p oce- du e, he eac o configu a ion and equipmen we e desc ibed in p e ious s udies [32]. 3. Resul s and discussion 3.1. Ca alys cha ac e iza ion 3.1.1. In es iga ion o he pe oxocomplexes P elimina y expe imen s es ablished ha dissolu ion o he p e- cipi a es occu s only o Mg/Z a ios be ween 1 and 3. An inc ease in he magnesium con en leads o he o ma ion o he p ecipi- a e a he fi s s age o syn hesis, which canno be dissol ed in he hyd ogen pe oxide solu ion. Con a y, lowe amoun s o mag- nesium in he s a ing solu ion cause he o ma ion o complexes con aining undesi able sodium ions in la ge excess. The chemical composi ion o he p epa ed oxides was ana- lyzed by EDS (Table 1), concluding ha he Mg con en is equal o sligh ly lowe han p edic ed. A simila a ia ion was al eady epo ed in he li e a u e, being a ibu ed o he elimina ion o highly dispe sed Mg(OH)2species du ing he washing s ep [30]. A esidual amoun o sodium (<1 w %) was also de ec ed. Su ace composi ion was analyzed by XPS and he esul s a e also lis ed in Table 1. In all samples i is obse ed an en ichmen o Mg on he su ace. The p ecu so o 1MZ sample has been analyzed by FTIR and TG/DSC echniques, in o de o show he o ma ion o he pe oxocomplex and e eal some ea u es o i s he mal decompo- si ion. The FTIR spec um o he p epa ed complex (Fig. 1) shows he main abso bance peaks co esponding o symme ic s e ching o O=C–O a 1630 cm−1and asymme ic ones a 1395 cm−1, which a e a ibu ed o a ci a e ligand, and also a cha ac e is ic band a 910 cm−1. This las abso p ion can be a ibu ed o he p esence o pe oxo-g oups in he complex. In his way, in pu e H2O2, O–O s e ching ib a ion co esponds o a band a 877 cm−1, bu in com- plexes i migh appea a lowe o g ea e wa enumbe s [33]. The da a on he posi ions o O–O s e ching ib a ions in IR spec a o me al pe oxocomplexes a e no unequi ocal. Ta a de e al. [34] obse ed O–O band ib a ions a 840 cm−1. O he esea che s [35] assigned he band a 850 cm−1 o ␯(O–O). In his way, he band a 860 cm−1was epo ed o be a cha ac e is ic band o (O–O) in zi conium pe oxocomplexes [36], which is simila o he esul s ob ained o i anium pe oxoci a e [37]. Howe e , i is known ha O–O bands can be obse ed a wa enumbe s as high as 935 cm−1 [33]. Thus, we can sugges ha he dissolu ion o he p ecipi a e p epa ed om zi conium and magnesium sal s in hyd ogen pe ox- ide leads o he o ma ion o he zi conium pe oxocomplex wi h magnesium in i s ca ionic o m, he s abili y o he complex being p o ided by he addi ion o ci ic acid. I is highly p obable ha he syn hesized complexes ha e composi ions and s uc u es simila o hose desc ibed p e iously o ba ium– i anium pe oxoci a es [38]. The mog a ime ic analyses e eal ha he isola ed complex decomposes in wo main s eps. A ypical p ofile (MZ1) is shown in he Supplemen a y In o ma ion, all he ma e ials p epa ed by his p ocedu e p esen ing simila p ofiles. The fi s one, occu ing a empe a u es be o e 473 K and ha ing mass loss o app oxi- ma ely 15%, co esponds o dehyd a ion and emo al o he excess o he ligand. The second s age o decomposi ion in he ange o 673–773 K is accompanied by an exo he mic e ec on he DSC cu e; asc ibed o he oxida ion o he o ganic pa o he com- plex. Fu he mo e, a maximum a 503 K is obse ed which could be a ibu ed o he up u e o he O–O b idges and ansi ion o he single oxygen o O2- s a e, in ag eemen wi h he esul s ob ained by Ichinose e al. [39], in hei s udies abou pe oxo i ana es. A hi d b oad exo he mic peak is also obse ed in he ange o 928–1028 K, a ibu ed o he polymo phic ansi ions in zi conia and magnesia sys ems (Supplemen a y In o ma ion). 3.1.2. XRD and FTIR s udies o MgO/Z O2 X- ay di ac ion analysis o he p epa ed samples de ec s he p esence o poo ly c ys alline phases, which can be a ibu ed o: cubic MgO phase (ICDD PDF2 99-200-4113), cubic o e agonal zi conia (ICDD PDF2 99-101-0922, ICDD PDF2 00-042- 1167) undis inguishable due o b oad and insepa able peaks, o he magnesia-s abilized zi conia phase wi h a gene al o mula MgxZ 1-xO2-x(ICDD PDF2 00-080-0967 and 00-080-0964) (Fig. 2a). In he 1MZXRD pa e n, he p esence o cubic magnesia is ha dly no iceable, bu a highe Mg/Z a ios he sepa a ion o c ys alline MgO om he mixed oxide becomes mo e ob ious. The c ys alli e size o magnesia eaches 7 nm in he 3MZ mixed oxide. The (1 1 1) eflec ion o cubic/ e agonal zi conia phase is shi ed o highe 2 deg ees on he XRD o all magnesia–zi conia samples in compa - ison wi h pu e zi conia p epa ed ia simila p ocedu e (Fig. 2a), ha indica es he o ma ion o cubic magnesia-s abilized zi conia phase. The c ys alli e size o MgxZ 1-xO2-x, es ima ed by Sche e equa ion is ound o be in he ange o 3–4 nm o all MgO/Z O2 a ios. This shi ing (Fig. 2b), caused by inco po a ion o magnesium ions in c ys alline la ice o zi conia, was al eady no iced. Con a y o his case, Sádaba e . al [30] ound his peak con inuously shi ed o he highe 2 deg ees wi h he inc easing MgO/Z O2 a ios up o 4.88, indica ing highe deg ee o inco po a ion o magnesium ions in o he zi conia ma ix. Howe e , he abili y o zi conia ma ix o accommoda e magnesium ions is limi ed, as he highes Mg/Z a io c ys alline MgxZ 1-xO2-xpossesses, acco ding o ICDD PDF2 da abase, is 0.4/1. The p esence o MgxZ 1-xO2-xphase was no de ec ed in 4MZ p epa ed by co-p ecipi a ion and only peaks co - esponding o pe iclase MgO and e agonal Z O2we e iden ified [27]. The di e ence in he o ma ion o he MgxZ 1-xO2-xphase in he ma e ials p epa ed ia co-p ecipi a ion and pe oxo-me hod can be explained in he ollowing way. Du ing he co-p ecipi a ion, he inco po a ion o magnesium in o amo phous zi conia could be accompanied by he o ma ion o magnesium hyd oxide. The inc ease o magnesium ions concen a ion p omo es hei inclu- sion in Z O2la ice o he highe ex en , when he o he pa o magnesium ions is p ecipi a ed in he o m o hyd oxide. The p o- cess o s uc u e o ma ion o oxide phase om he pe oxocomplex is di e en . In his case, zi conia pe oxo-species and magnesium ions a e pa s o he same s uc u al uni ; his ea u e ensu es high molecula homogenei y o he ma e ial, as he oxide phase o ms di ec ly by complex decomposi ion, a oiding he s age o p ecipi- a ion o me al hyd oxides. This jus ifies he highes inco po a ion o magnesium ions in o Z O2la ice o he 1MZ sample. I. K i so e al. / Applied Ca alysis A: Gene al 477 (2014) 26–33 29 Table 1 Mo phological p ope ies o magnesia–zi conia samples. Samples Chemical composi ion Mg/Z su ace a omic a io (XPS) Su ace a ea, SBET, m2/g Po e olume Vp, cm3/g Mean po e diame e Dp, Å C ys alli e sizes o MgxZ 1-xO2-x; MgO, nm 1MZ Mg1Z 1Na0.05 1.18 87 0.08 31 3; – 2MZ Mg1.8Z 1Na0.01 4.30 48 0.08 136 3; – 3MZ Mg2.6Z 1Na0.1 5.08 30 0.07 102 3; 7 4MZaMg4.5Z 1Na0.02 11.7 78 0.8 342 –; 11 aDa a aken om he s udy [27] o he bulk magnesia–zi conia p epa ed ia p ecipi a ion me hod. Chemical composi ion was ob ained by EDS o he pe oxocomplexes. Fo sample 4MZ, by ICP-OES. 20 30 40 50 60 70 Z O2 MZ1 MZ2 MZ3 2The a (deg ees) Coun s (a.u.) * * 25 26 27 28 29 30 31 32 33 34 35 2The a (deg ees) Coun s (a.u.) MZ1 MZ2 MZ3 b a Fig. 2. XRD pa e ns o he zi conia–magnesia samples (a), and enla ged egion o 2 alues showing he eflec ion a ibu ed o MgxZ 1-xO2-xcubic phase (b). Symbols: cubic MgO (); e agonal zi conia ();monoclinic zi conia (*);cubic MgxZ 1-xO2-x(). The mos b oad and in ense doubling peak in he IR spec a o all samples (Fig. 3a) is obse ed in he ange o 1600–1400 cm−1co - esponding o Mg–O s e ching and biden a e ca bona e bonded o he su ace o oxides. Abso p ion a 1150 cm−1, p esen in all spec a, can be a ibu ed o CO32- s e ching, and i is ound o ha e he highes in ensi y in he spec um o 1MZ sample. P ob- ably, due o he adso p ion o ca bona e species on he s ong basic Mg–O–Z cen e s o he MgxZ x-1O2-xmixed oxide phase. The bands a 550–630 and 430–450 cm−1can be a ibu ed o he Mg–O ib a ions in he c ys alline pe iclase. I is wo h men ioning ha 2000180 0160 0 b a 1400120 0100 080 060 040 0 Wa enu mbe s (c m–1) Wa enu mbe s (c m–1) T ansmi ance (a.u.) 1150 MZ1 MZ2 MZ3 14471537640465 3500 3600 3700 3800 3900 T ansmi ance (a.u.) MZ1 MZ2 MZ3 Fig. 3. FTIR spec a o he magnesia–zi conia samples (a) and he egion o OH- g oups s e ching ib a ions (b). he band nea 3700 cm−1(Fig. 3b) does no appea o 1MZ sam- ple. The abso p ion a his ange o wa e numbe s unambiguously co esponds o he ib a ion o OH g oups bonded o magnesium a oms. I is mo e likely ha Mg–OH g oups ha e been o med in he esul o e-hyd oxyla ion o MgO su ace wi h a mosphe ic wa e apo s. The success in he inco po a ion o magnesium a oms in he zi conia la ice ia b idged oxygen bonded o zi conium, can sa - is ac o ily explain he absence o his band in he 1MZ spec um. This obse a ion also suppo s he o ma ion o mo e homogeneous mixed magnesia–zi conia phase in hese samples, made on he basis o XRD analysis. 30 I. K i so e al. / Applied Ca alysis A: Gene al 477 (2014) 26–33 0.0 0.2 0.4 0.6 0.8 1.0 10 20 30 40 50 60 Rela i e P essu e (P/P0) Volume N2 adso bed (cm3g–1) 0.0 0.2 0. 40. 60. 81. 0 0 20 40 60 80 100 Rela i e P essu e ( P/P0) Volume N2 adso bed (cm3g–1) 0.0 0.2 0. 4 0.6 0. 81. 0 0 10 20 30 40 50 60 70 Rela i e P essu e (P/P0) Volume N2 adso bed (cm3g – 1) (a) (b) (c) Fig. 4. Adso p ion–deso p ion iso he ms o he magnesia–zi conia samples: (a) 1MZ, (b) 2MZ and (c) 3MZ. 3.1.3. Tex u al and mo phological p ope ies Tex u al p ope ies o he samples a e summa ized in Table 1. Ni ogen adso p ion–deso p ion iso he ms de e mined a 77 K (Fig. 4) co espond in all he cases o ype IV (mesopo ous solids), acco ding o IUPAC classifica ion. The hys e esis loops a e a he na ow and e ically o ien ed in he case o 2MZ and 3MZ, being associa ed o ype H1 – po ous ma e ials exhibi ing a na ow dis i- bu ion o ela i ely uni o m (cylind ical po es). Same ype o loop was p e iously obse ed in 4MZ p epa ed by co-p ecipi a ion. In he case o 1MZ, he hys e esis loop shows he ypical shape o H2 ype, ypical o mo e complex po e ne wo ks consis ing o po es wi h ill-defined shape and wide po e size dis ibu ion. Compa - ing he samples ob ained by he pe oxocomplex ou e wi h he bulk magnesia–zi conia p epa ed ia p ecipi a ion me hod, a sligh inc ease in su ace a ea and po e olume is obse ed o he 1MZ sample— he sample wi h he highes inco po a ion o magnesium ions in o Z O2la ice sample, acco ding o XRD pa e ns—, whe eas highe magnesium con en dec eases he specific su ace a ea. Con- a y, Sádaba e al. [30] no iced ha he su ace a ea inc eased wi h magnesium con en . The su ace a ea is ela ed o he ac- ion o c ys alline MgO phase. In ha wo k, he c ys alliza ion o pe iclase goes h ough he o ma ion o he in e media e highly dis- pe sed magnesium hyd oxide, whose u he dehyd a ion esul ed in he MgO phase [30]. This me hod could allow he o ma ion o po ous magnesia wi h mo e de eloped su ace in compa ison o 250 350 45 055 065 0 750 CO 2 deso bed (a.u.) Tempe a u e (K ) 1MZ 2MZ 3MZ Fig. 5. CO2-TPD cu es o he magnesia–zi conia samples. 900800700600500400300 NH 3 deso bed (a.u.) Tempe a u e (K) 1MZ 2MZ 3MZ Fig. 6. NH3-TPD cu es o he magnesia–zi conia samples. he me hod p oposed in he p esen s udy, whe e MgO is o med a oiding hyd olysis and p ecipi a ion s age. 3.1.4. Basic and acid p ope ies CO2-TPD p ofiles o he magnesia–zi conia oxides a e depic ed in Fig. 5, whe eas he concen a ions and s eng h-dis ibu ions o he basic si es a e summa ized in Table 2. Compa ing all he esul s wi h he basici y dis ibu ion o he bulk 4MZ p epa ed by co-p ecipi a ion me hod [27], a no iceable dec ease in bo h, con- cen a ion and s eng h, can be obse ed. Ma e ials p epa ed by he pe oxo- ou e p esen mainly weak and medium-s eng h basic si es (bica bona es and biden a es) and he concen a ion o hese si es dec eases as Mg/Z a io inc eases: om 66 ␮mol/g (1MZ) o 25 ␮mol/g (3MZ). The e is no an impo an di e ence in he deso p ion empe a u es among he h ee ma e ials. These esul s a e in ag eemen wi h FTIR expe imen s. The medium s eng h basic si es a e he main, co esponding o he mos in ense peak (1600–1400 cm−1) a ibu ed o biden a e ca bona es bonded o he oxide su ace. These biden a e ca bona es equi e he pa ici- pa ion o an adjacen ca ionic si e (M+–O2−pai s, medium s eng h base si es), ha is, he p esence o an associa ed acid si e, as demon- s a ed using mic ocalo ime ic s udied wi h hyd o alci e-de i ed Mg–Al mixed oxides [40]. Likewise, 1MZ sample exhibi s also he s onges basic si es, in ag eemen wi h he band a 1150 cm−1 obse ed o his ma e ial. The analysis o he acidi y was ca ied ou by NH3-TPD. P o- files ob ained a e shown in Fig. 6 whe eas he dis ibu ion o acid si es was summa ized in Table 2. The ma e ial p epa ed by co- p ecipi a ion (4MZ) showed highe acidi y han he pe oxocomplex ou e p epa ed ma e ials [27]. Likewise, he acidi y o he 4MZ I. K i so e al. / Applied Ca alysis A: Gene al 477 (2014) 26–33 31 Table 2 Basic and acid p ope ies o magnesia–zi conia samples. Samples Basic si es, CO2-TPD, ␮mol/g (K) Acid si es, NH3-TPD, ␮mol/g (K) Bica bona e Biden a e Monoden a e Weak (<503) Medium (503 < T < 673) S ong (>673) 1MZ 21 (330) 45 (458) 6 (666) 34 (378) – 45 (808) 2MZ 11 (317) 34 (455) – 59 (399) 38 (505,569) – 3MZ 12 (329) 13 (423) – 49 (428) 61 (574) – 4MZa– 121 (444) 13 (585) 62 (353) 74 (566) – aDa a aken om he s udy [27] o he bulk magnesia–zi conia p epa ed ia p ecipi a ion me hod. 0 20 40 60 450 550 65 0 750 Ace one con e sion (%) Tempe a u e (K) 1MZ 2MZ 3MZ 4MZ Fig. 7. Ac i i y o he ace one condensa ion o e he di e en Mg–Z ca alys s. sample showed simila ends, wi h di e en concen a ions bu simila s eng h dis ibu ions, o 2MZ and 3MZ samples. In all cases, he acidic si es a e dis ibu ed in o weak and medium s eng h acid si es. In he case o 1MZ ca alys , medium-s eng h acid si es disappea ed and a deso p ion peak a empe a u es abo e 800 K appea ed. F om a p e ious wo k [27], i was concluded ha s ong acid si es do no pa icipa e in he eac ion, hus his peak will be no conside ed. 3.2. Reac ion s udies The ace one gas-phase condensa ion was ca ied ou a empe - a u es be ween 373 and 723 K using he di e en Mg/Z ma e ials as ca alys s. The di e en compounds iden ified we e cong u- en wi h he gene al mechanism p oposed o his eac ion. This mechanism is summa ized in Scheme 1. The condensa ion eac- ion o ace one o ms diace one alcohol. A subsequen dehyd a ion o his compound leads o ei he isomesi yl oxide o mesi yl oxide. Condensa ion o hese C6 compounds wi h ace one o ms pho ones (C9), which ea ange o yield cyclic C9 compounds called isopho ones. I was p e iously s udied how he dis ibu ion o acid and basic si es in he ca alys condi ion he selec i i y o each p oduc ob ained in his eac ion [28]. The absence o di us- ional limi a ions a wo king condi ions was co obo a ed using he Thiele modulus modified by Weisz [41] and he negligible ace one con e sion a hese condi ions wi hou any ca alys s was p e i- ously confi med [28]. I should be also no ed ha he densi y o he s udied ma e ials is e y simila , leading o simila ca aly ic bed leng hs, and disca ding any a ia ion o he e ec i e esidence ime when using di e en ca alys s. The e olu ion o he ace one con e sion wi h he empe - a u e using he di e en ca alys s is plo ed in Fig. 7. Resul s ob ained wi h 4MZ we e also plo ed o easily discuss he di e - ences be ween bo h p epa a ion me hods. The ace one con e sion inc eases wi h he empe a u e, eaching alues o he pe oxo- me hod p epa ed MgZ oxides highe in all cases han o he bulk co-p ecipi a ed MgZ oxide. The con e sion a 723 K eaches he 44% o 1MZ, 2MZ and 3MZ, whe eas he bulk oxide did no exceed 35%. Selec i i y o he main eac ion p oduc s–diace one alcohol (C6), mesi yl oxide (C6), pho ones (C9), isopho ones (C9) and mesi ylene (C9) a e summa ized in Table 3 o wo empe a- u es, as well as he ca bon balance. Selec i i ies below 573 K we e mainly owa ds C6 p oduc s, whe eas a 723 K, he C9 compounds become mo e impo an . Selec i i y owa ds ime s a he highes empe a u e, eaches he alue o 30.1, 20.2 and 4.9% owa ds pho ones (P), mesi ylene (M) and isopho ones (IP), espec i ely. Fo compa a i e pu poses, he amoun o IP ob ained is lowe han o e Mg-Al oxides (13.9%) [42] and Mg–Z oxides (18%) [28]. Fu he , i is in e es ing o poin ou ha he selec i i ies owa ds C9 compounds a e sligh ly highe (55.2 o 1MZ e sus 42% o Mg–Z suppo ed on high su ace a ea g aphi e), al hough he con e sion is lowe (44.8 e sus 54%) [28]. The alues he e ob ained o C9 compounds a e simila o ha ob ained o e a high su ace a ea MgO (53.8%), ob ained by an hyd o he mal ea men [32]. Focusing on he pe oxo-me hod p epa ed MgZ oxides, i is obse ed ha he con e sion eached, especially a he highes empe a u e, is e y simila in he h ee cases, being mo e no o- ious he di e ences in selec i i ies. A empe a u es lowe han 523 K, only C6 p oduc s a e o med in ele an concen a ions. Only he 1MZ ca alys , he ma e ial wi h he lowes con e sion a his empe a u e, eaches a 17.6% o selec i i y owa ds pho ones. In all cases, he diace one alcohol ollows he ypical p ofile o a p i- ma y p oduc , whose selec i i y eaches a maximum and dec eases by he decomposi ion in o o he p oduc s. Among MO and IMO, mesi yl oxide is he main p oduc a hese condi ions, ollowing IMO an unclea end. The o ma ion o DAA is ca alyzed by medium s eng h basic si es, and i s decomposi ion in o MO o IMO needs acid-basic pai s. These icinal M+–O2−si es a e o med by he oxygen anion, a medium-s eng h base, and he M+, weakly acid. Likewise, i was shown in p e ious wo ks ha add o medium, he p esence o weak basic si es enhances his eac ion [28]. 1MZ sam- ple exhibi s he highes amoun o biden a e and also bica bona e cen e s, hus he highes eac ion ex ension. Compa ing he esul s o he Mg–Z oxides p epa ed by he pe oxo-me hod wi h hose p e iously published on cop ecipi a ed Mg–Z oxides, i is e iden ha he ma e ials p esen ed in his wo k ha e a mo e homoge- neous dis ibu ion o basic and acid si es, in ag eemen wi h XRD analysis, whe e a mo e homogeneous mixed magnesia–zi conia phase in hese samples was obse ed. A 723 K, he selec i i y owa ds C6 dec eases and C9 com- pounds become mo e impo an , eaching o 1MZ selec i i ies o C9 o 55.2%. The fi s ime o med is he pho one, which eac s o o m ei he isopho ones o mesi ylene. Isopho ones, he mos aluable p oduc , a e ob ained in lowe ex ension han o e co- p ecipi a ed Mg–Z oxides o Mg–Al oxides. This s ep equi es he p esence o s ong basic si es, only p esen ed on 1MZ ca alys , su - p isingly, wi h he lowes selec i i y owa ds hese compounds. Howe e , i is no o ious he high selec i i y owa ds mesi ylene, p oduc ha can be ob ained by wo di e en ou es. I can be o med om linea pho ones, in ol ing basic and acid si es [28], bu also om isopho ones, in ol ing only acid si es. Thus, he high 32 I. K i so e al. / Applied Ca alysis A: Gene al 477 (2014) 26–33 Scheme 1. Reac ion mechanism o he ace one sel -condensa ion. Table 3 E olu ion o selec i i y o ace one condensa ion owa ds he eac ions p oduc s a h ee empe a u es: diace one alcohol (DAA), isomesi yl oxide (IMO), mesi yl oxide (MO), pho one (P), mesi ylene (M) and isopho one (IP). Ca alys Con e sion (%)/ca bon balance (%) Selec i i y (%) DAA IMO MO P M IP 523 K 723 K 523 K 723 K 523 K 723 K 523 K 723 K 523 K 723 K 523 K 723 K 523 K 723 K 1 MZ 13.3/95.0 44.8/91.5 11.1 7.6 3.9 1.0 64.2 36.3 17.6 30.1 3.2 20.2 0.0 4.9 2 MZ 16.7/96.0 44.6/82.2 13.3 7.7 0.0 0.1 79.4 71.3 4.5 3.9 2.1 9.8 0.8 7.3 3 MZ 18.1/85.8 42.5/80.5 13.1 24.8 0 0.1 82.1 53.6 4.7 6.8 0.0 6.2 0.0 8.6 4 MZa14.5/94.6 31.5/91.3 1.3 8.0 2.1 0.5 76.7 69.6 1.6 0.5 2.2 0.1 16.1 25.5 aDa a aken om he s udy [28] o he bulk magnesia–zi conia p epa ed ia p ecipi a ion me hod and used as he ca alys o ace one sel -condensa ion. selec i i y he e obse ed owa ds mesi ylene a 723 K is in good ag eemen wi h he isopho ones selec i i y p ofile and wi h he ele an concen a ion o s ong acid si es obse ed in 1MZ sam- ple. By con as s, on 2MZ and 3MZ, mesi ylene could be o med mainly om pho ones, ou e which equi es he p esence o basic and weak acid si es, p esen in hese ma e ials (Table 2) Compa ing he esul s ob ained wi h hese ma e ials and wi h he bulk oxide (4MZ), i should be no ed ha al hough he con- e sions a e only sligh ly lowe wi h he 4MZ ma e ial, selec i i y pa e ns a e comple ely di e en . In gene al e ms, 4MZ shows highe selec i i y owa ds isopho ones, and lowe o he o ma- ion o pho ones and mesi ylene. In he mos accep ed mechanisms o ace one condensa ion, i is assumed ha in amolecula aldol condensa ions (needed o he o ma ion o isopho ones) equi e highe s eng h o basic si es han he ace one aldol condensa- ion o MO and IMO o ma ion [28]. Ob ained esul s sugges ha he s eng h o he basic si es in he ma e ials p epa ed by he pe oxo-media ed p ocedu e is enough o aldol condensa ion, bu no enough o he in amolecula condensa ion. F om he poin o iew o he acid si es dis ibu ion, bo h kinds o MgZ oxides p esen a e y simila acidi y pa e n. Howe e , he a io be ween he amoun o acid and basic si es is highe o he mixed oxides p epa ed by he pe oxo-media ed p ocedu e. In good ag eemen wi h his ac , mesi ylene selec i i y inc eases o hese ma e ials. I should be no ed ha mesi ylene o ma ion om pho one needs he p esence o acid si es [28]. 4. Conclusions We p opose in his wo k a new pe oxo-media ed p ocedu e o he p epa a ion o mixed magnesia–zi conia ca alys s. The dis- solu ion o he magnesia-zi conia hyd ogel in hyd ogen pe oxide in p esence o ci ic acid leads o he o ma ion o he amo - phous pe oxocomplex, whose s abili y and solubili y a e ensu ed by he addi ion o ca boxylic acid. This me hod p o ides a high deg ee o inco po a ion o magnesium ions in o he c ys alline I. K i so e al. / Applied Ca alysis A: Gene al 477 (2014) 26–33 33 la ice o zi conia, hus o ming cubic magnesia-s abilized zi co- nia phase wi h a gene al o mulaMgxZ x-1O2-x. Wi h he inc easing o magnesia con en in he samples he sepa a ion o he pe iclase om he mixed oxide phase becomes mo e no iceable, whe eas no significan changes in he MgxZ x-1O2-xla ice pa ame e s can be obse ed. The dec ease o su ace a ea in he samples wi h Mg/Z mola a io highe han 1 is a ibu ed o highe c ys allini y o mag- nesium oxide. To al basici y and acidi y o he p epa ed ma e ials a e ound o be lowe han i is o p ecipi a ed mixed oxide. Only he 1MZ sample exhibi s s ong basic and acid si es. The ace one gas-phase sel condensa ion esul s e idence he homogeneous dis ibu ion o acid and basic si es on he Mg–Z oxides p epa ed by he pe oxo-complex ou e. Likewise, i was obse ed ha he selec i i y owa ds mesi ylene is especially impo an o he 1MZ sample, ca alys wi h he highes amoun o bo h medium and high s eng h basic si es and also s ong acid si es. Acknowledgemen This wo k was suppo ed by he Spanish Go e nmen (con ac CTQ2011-29272-C04-02). Appendix A. Supplemen a y da a Supplemen a y ma e ial ela ed o his a icle can be ound, in he online e sion, a h p://dx.doi.o g/10.1016/j.apca a.2014. 03.008. Re e ences [1] S. Yoon, T. Noh, W. Kim, J. Choi, H. Lee, Ce am. In . 39 (2013) 9247–9251. [2] L. Hao, J. Law ence, Ma e . Sci. Eng. A364 (2004) 171–181. [3] X. Jiao, L. Li, N. Zhao, F. Xiao, W. Wei, Ene g. Fuels 27 (2013) 5407–5415. [4] G. Pa ameswa am, M. S ini as, B. Ha iBabu, P.S. Sai P asada, N. Lingaiah, Ca al. Sci. Technol. (2013), h p://dx.doi.o g/10.1039/C3CY00532A. [5] V. E a, P. Mäki-A ela, J. Wä nå, T. Salmi, J.-P. Mikkola, D.Y. Mu zin, Appl. Ca al. A 404 (2011) 39–46. [6] M.B. Gawande, A.K. Ra hi, P.S. B anco, T.M. Po ewa , A. Velhinho, I.D. Noguei a, A. Tols ogouzo , C.A.A. Ghummane, O.M.N.D. Teodo o, RSC Ad . 3 (2013) 3611–3617. [7] C.J. Ba e , J.N. Chheda, G.W. Hube 1, J.A. Dumesic, Appl. Ca al. B 66 (2006) 111–118. [8] S. Liu, X. Zhang, J. Li, N. Zhao, W. Wei, Y. Sun, Ca al. Commun. 9 (2008) 1527–1532. [9] D. Jiang, G. Pan, B. Zhao, G. Ran, Y. Xie, E. Min, Appl. Ca al. A 201 (2000) 169–176. [10] V.V. O domsky, V.L. Sushke ich, I.I. I ano a, J. Mol. Ca al. A 333 (2010) 85–93. [11] M.A. A amendia, V. Bo áu, C. Jiménez, A. Ma inas, J.M. Ma inas, J.A. Na io, J.R. Ruiz, F.J. U bano, Colloid. Su . A 234 (2004) 17–25. [12] M.B. Gawande, P.S. B anco, K. Pa ghi, J.J. Sh ikhande, R.K. Pandey, C.A.A. Ghum- man, N. Bundaleski, O.M.N.D. Teodo od, R.V. Jaya amb, Ca al. Sci. Technol. 1 (2011) 1653–1664. [13] H. Gocmeza, H. Fujimo i, Ma e . Sci. Eng. B 148 (2008) 226–229. [14] T. Se u, Ce am. In . 26 (2000) 517–521. [15] M.A. A amendia, V. Bo au, C. Jiménez, A. Ma inas, J.M. Ma inas, J.R. Ruiz, F.J. U bano, J. Mol. Ca al. A 218 (2004) 81–90. [16] R.H.R. Cas o, P.J.B. Ma cos, A. Lo iaux, M.C. S eil, L. Gengemb e, P. Roussel, D. Gou ea, Chem. Ma e . 20 (2008) 3505–3511. [17] M. Tahmasebpou , A.A. Babaluo, M.K.R. Aghjeh, J. Eu . Ce am. Soc. 28 (2008) 773–778. [18] I.V. K i so , M.V. Ilkae a, V.V. A din, D.A. Zhe eb so , J. Non C ys . Solids 362 (2013) 95–100. [19] I. Sádaba, M. Ojeda, R. Ma iscal, R. Richa ds, M. Lopez G anados, Chem. Phys. Chem. 13 (2012) 3282–3292. [20] Lisa. C. Klein (Ed.), Sol–Gel Op ics: P ocessing and Applica ions, Kluwe Aca- demic Publishe , 1994. [21] M. Kakihana, M. Kobayashi, K. Tomi a, V. Pe ykin, Bull. Chem. Soc. Jpn. 83 (2010) 1285–1308. [22] J.-Y. Piquemal, E. B io , J.-M. B égeaul c, Dal on T ans. 42 (2013) 29–45. [23] Y. Gao, Y. Masuda, H. Oh a, K. Koumo o, Chem. Ma e . 16 (2004) 2615–2622. [24] J.H. Pa k, Y.B. Yoo, K.H. Lee, W.S. Jang, J.Y. Oh, S.S. Chae, H.K. Baik, ACS Appl. Ma e . In e aces 5 (2013) 410–417. [25] L. an Rij, L. Winnubs , L. Jun, J. Schoonman, J. Ma e . Chem. 10 (2000) 2515–2521. [26] M.D. Gonc¸ al es, R. Muccillon, Ce am. In . 40 (2014) 911–917. [27] L. Faba, E. Díaz, S. O dó˜ nez, Appl. Ca al. B 113-114 (2012) 201–211. [28] L. Faba, E. Díaz, S. O dó˜ nez, Appl. Ca al. B 142–143 (2013) 387–395. [29] P.J. Da da, V.V. Ranade, Chem. Eng. J. 207–208 (2012) 349–367. [30] I. Sádaba, M. Ojeda, R. Ma iscal, J.L.G. Fie o, M. Lopez G anados, Appl. Ca al. B101 (2011) 638–648. [31] I.V. K i so , M.V. Ilkae a, V.D. Samokhina, V.V. A din, S.A. Khainako , D.A. Uchae , J.R. Ga cia, J. Sol–Gel Sci. 63 (2013) 665–669. [32] M. León, L. Faba, E. Díaz, S. Bennici, A. Vega, S. O dó˜ nez, A. Au oux, Appl. Ca al. B 147 (2014) 796–804. [33] J. Muhlebach, K. Mulle , G. Shwa zenbach, Ino g. Chem. 9 (1970) 2381–2390. [34] M.T.H. Ta a de , P. Bha acha jee, A.K. Sa ka , Polyhed on 1 (1992) 795–798. [35] D.D. Aga wal, R. Jain, R.P. Bha naga , S. S i as a a, Polyhed on 9 (1990) 1405–1409. [36] A.C. Dengel, W.P. G i fi h, Polyhed on 8 (1989) 1371–1377. [37] I. T uijen, A. Ha dy, M.K. Van Bael, H. Van den Rul, J. Mullens, The mochim. Ac a 456 (2007) 38–47. [38] M. Tada, K. Tomi a, V. Pe ykin, M. Kakihana, Solid S a e Ionics 151 (2002) 293–297. [39] H. Ichinose, M. Te asaki, H. Ka suki, J. Ce am. Soc. Jpn. 107 (1996) 715–718. [40] M. León, E. Díaz, S. Bennici, A. Vega, S. O dó˜ nez, A. Au oux, Ing. Eng. Chem. Res. 49 (2010), 3663–3671. [41] P.B. Weisz, Science 179 (1973) 433. [42] S. O dó˜ nez, E. Díaz, M. León, L. Faba, Ca al. Today 167 (2011) 71–76.