Latin American Applied Research 33:79-86 (2003)

EFFECT OF THE ACID-BASE PROPERTIES OF Mg-Al MIXED OXIDES ON THE CATALYST DEACTIVATION DURING ALDOL CONDENSATION REACTIONS

V.K. DÍEZ, C.R. APESTEGUÍA and J.I. DI COSIMO*

Instituto de Investigaciones en Catálisis y Petroquímica -INCAPE- (UNL-CONICET). Santiago del Estero 2654, (3000) Santa Fe, Argentina, e-mail: [email protected]

Abstract−− The effect of chemical composition of have reported that Mg-Al mixed oxides efficiently Mg-Al mixed oxides on both the acid-base properties catalyze the gas-phase self-condensation of acetone to and the deactivation process during the gas phase α,β-unsaturated ketones such as mesityl oxides and self-condensation of acetone was studied. The isophorone (Di Cosimo et al., 1998a). Unfortunately, in activity and selectivity for acetone oligomerization coupling reactions like aldol condensations, basic depended on the catalyst acid-base properties. Mg- catalysts are often deactivated either by the presence of rich catalysts selectively yielded mesityl oxides byproducts such as water in the gas phase or by coke whereas Al-rich MgyAlOx oxides produced mainly build up through secondary side reactions. Deactivation isophorone. The initial deactivation rate, increased has traditionally limited the potential of solid basic linearly with the density of surface basic sites, catalysts to replace environmentally problematic and thereby suggesting that although MgyAlOx oxides corrosive liquid bases. However, few works in the promote the self-condensation of acetone by both literature deal with the deactivation of solid bases under acid- and base-catalyzed mechanisms, the reaction conditions. Studies relating the concerted and deactivation rate would be closely related to the sequential pathways required in the deactivation surface basic properties. The MgyAlOx activity mechanism with the acid-base properties of the catalyst declines in the acetone oligomerization reaction due surface are specially lacking. In this work, we studied to a blockage of both base and acid active sites by a the deactivation of Mg-Al mixed oxides in the gas phase carbonaceous residue formed by secondary oligomerization of acetone. We prepared and reactions. The amount and the nature of the carbon characterized calcined Mg-Al hydrotalcites with Mg/Al deposits were characterized by temperature- atomic ratios of 1-9. The effect of composition on both programmed oxidation technique. MgyAlOx and the surface and catalytic properties as well as the Al2O3 formed more and heavier coke than pure MgO catalyst deactivation were investigated by combining but the latter deactivates faster. The deactivation several characterization methods with catalytic data. rate and coke composition are defined by the nature of the active site involved in the coke-forming II. METHODS reactions at different catalyst compositions rather than by the carbon amount or polymerization A. Synthesis Procedures degree. Mg/Al hydrotalcite with Mg/Al atomic ratios (y) of 1, 3, Keywords−− Catalyst Deactivation, Acetone 5 and 9 were prepared by coprecipitation of Mg(NO3)2.6 Oligomerization, Acid-Base Catalysis H2O and Al(NO3)3.9 H2O at 333 K and pH = 10. After drying overnight at 348 K, precursors were decomposed in N2 at 673 K in order to obtain the MgyAlOx mixed I. INTRODUCTION oxides. More experimental details are given elsewhere Mg-Al mixed oxides obtained by thermal (Di Cosimo et al., 2000). Reference samples of Al2O3 decomposition of anionic clays of hydrotalcite structure, and MgO were prepared following the same procedure. present acidic or basic surface properties depending on their chemical composition
  (Di Cosimo et al., 2000). B. Methods for the Sample Characterization These materials contain the metal components in close The crystalline structure of the samples was determined interaction thereby promoting bifunctional reactions that by powder X-ray diffraction (XRD) methods. Acid site are catalyzed by Brönsted base-Lewis acid pairs. densities were measured by temperature-programmed Among others, hydrotalcite-derived mixed oxides desorption (TPD) of NH3 preadsorbed at room promote aldol condensations (Reichle, 1985), temperature. The structure of CO2 chemisorbed on alkylations (Velu and Swamy, 1996) and alcohol hydrotalcite-derived samples was determined by eliminations (Di Cosimo et al., 2000). In particular, we infrared spectroscopy (IR). CO2 adsorption site

* To whom correspondence should be addressed

79 Latin American Applied Research 33:79-86 (2003)

densities were obtained from TPD of CO2 preadsorbed gaseous evolution develops a significant porous at 298 K. Total surface areas (Sg) were measured by N2 structure and gives rise to relatively high surface area physisorption at 77 K. mixed oxides as shown in Table 1. In the compositional range of this study, the analysis by XRD of the C. Acetone Self-Condensation Procedures MgyAlOx oxides showed only one crystalline phase, Self-condensation of acetone was carried out at 473 K MgO periclase (ASTM 4-0829). No traces of and 100 kPa in a flow system with a differential fixed- hydrotalcite phases were detected after the thermal bed reactor. Acetone was vaporized in H2 (H2/acetone = decomposition at 673 K. 12) before entering the reaction zone. The standard Table 1 presents the acid and base site densities 0 contact time (W/F Acetone) was 49 g catalyst h/mole measured for all the samples by TPD of preadsorbed acetone. Main reaction products were mesityl oxides NH3 and CO2, respectively. The base site density (nb) (MO's), isophorone (IP) and mesitylene (MES). Traces of the MgyAlOx oxides was always between the values of phorone and light hydrocarbons were also identified. measured for MgO and Al2O3. Basic sites of different Coke formed on the catalysts was characterized after chemical nature were observed by FTIR of CO2 and reaction, ex-situ, in a temperature-programmed deconvolution of the CO2 TPD traces; the strength of oxidation (TPO) unit. The TPO experiments were the basic sites follows the order: isolated low- 2- 2- carried out in a microreactor loaded with 50 mg of coordination O ions > O in metal-oxygen pairs > OH catalyst and using a 3 % O2/N2 carrier gas. Sample groups (Di Cosimo et al., 1998b). The density of temperature was increased linearly from room metal–oxygen pairs measured by CO2 TPD decreased temperature to 973 K at 10 K/min. The reactor exit with increasing the Al content at expenses of the surface gases were fed into a methanator operating at 673 K to hydroxylation. convert COx in methane and then analyzed by flame The acid site density (na), on the other hand, ionization detector. increases with increasing Al content. By deconvolution of the NH3 TPD traces it was found that the MgyAlOx III. RESULTS AND DISCUSSION oxides contain both Brönsted OH groups (low temperature peak) and Lewis acid sites provided by A. Catalyst Characterization metal cations (high temperature peak) (Prinetto et al., 2000; Berteau et al., 1991). Since NH adsorption was The hydroxycarbonate precursors showed diffraction 3 performed at room temperature, the measured Brönsted patterns consistent with the layered double hydroxide acidity may be attributed on all the samples to the structure proposed for hydrotalcite-like compounds, i.e., reversible H-bonded NH adsorption on surface brucite-like sheets formed by OH groups and cations 3 hydroxyls. Lewis acidity of Mg AlO oxides (Mg2+ and Al3+) in octahedral coordination. Water y x corresponds to the irreversible coordinated NH molecules and charge-compensating CO 2- anions are 3 3 adsorption on surface defects with accessible Al3+ located in the interlayers.The hydrotalcite-like precursor + − cations whereas on alumina, the Lewis acidity was more r 2 3+ composition is [Mg1−rAlr (OH)2 ] (CO3 )r/2 .mH 2 O likely assigned to Al cations in the abundant surface 3+ 2- with r = Al/(Al+Mg); the stoichiometric hydrotalcite Al -O pairs. On MgO, on the other hand, the high compound, Mg6Al2(OH)16CO3.4H2O, is obtained for r = temperature peak may result from the irreversible 2+ 0.25. Upon thermal decomposition at 673 K, the coordinated NH3 adsorption on Mg cations combined hydroxycarbonate precursors lose water and CO2. This with heterolytic dissociative adsorption leading to

Table 1. Chemical Composition, Crystalline Structure, Surface Area and Acid-Base Properties of MgyAlOx, MgO and Al2O3. 10 Acetone Catalyst r = Al/(Al+Mg) Crystalline Sg Site Density

b 2 ) µ 2 (molar)a Phase (m2/g) ( mol/m )

c d MO's nb na

MgO 0.00 MgO 143 3.13 0.51

mol/hm 1

Mg AlO 0.11 MgO 114 1.17 0.81 µ 9 x ( j

  r IP Mg5AlOx 0.18 MgO 184 0.46 0.85

Mg3AlOx 0.24 MgO 238 0.59 1.02 Mg AlO 0.47 MgO 231 0.83 1.57 0.1 1 x 0246810 Al2O3 1.00 Amorphous 230 0.34 1.34 a b c d Time (h) by AAS; by XRD; by TPD of CO2; by TPD of NH3 Figure 1: Time dependence of the acetone conversion and product formation rates on Mg9AlOx (T = 473 K, PAcetone = 7.7 kPa)

80 V. K. DÍEZ, C. R. APESTEGUÍA, J. I. DI COSIMO

45 3.6 80 IP

40 r 0 ) 3.2 Acetone 2 35 2.8 60 ( µ mol/m 30

2.4 m mol/h µ

( 40 a 2.0 25

1.6 20 + n + MO's 2

b 20 ) (carbon atom %) (carbon atom n j 1.2 15 0

S MES 0.8 10 0 0.0 0.2 0.4 0.6 0.8 1.0 0.0 0.2 0.4 0.6 0.8 1.0 r = Al /(Al + Mg) r = Al/(Al+Mg) Figure 2: Initial acetone conversion rate and total Figure 3: Initial product selectivity as a function of number of active sites as a function of composition of composition on MgyAlOx catalysts (T = 473 K, PAcetone = 7.7 kPa, X0 = 12-14 %) MgyAlOx oxides (T = 473 K, PAcetone = 7.7 kPa, Acetone 0 W/F Acetone = 49 g h/mol)

- + 2+ 2- NH2 - H species on Mg - O pairs (Garrone and Mg9AlOx sample which lost about 60 % of its initial 0 Stone, 1985). activity after 10 h-run. Initial reaction rates (r j) and 0 product selectivities (S j) were calculated by B. Sample Composition, Surface Acid-Base extrapolating the reaction rates vs. time curves to zero. Properties and Catalytic Activity 0 In Figure 2 we have represented both the r Acetone Acetone aldol condensation proceeds on either acidic or values and the total measured site density (nT = nb + na) basic catalysts. On basic catalysts, the reaction products as a function of catalyst composition. Qualitatively, the 0 are mainly α,β-unsaturated ketones (Di Cosimo et al., variation of r Acetone with increasing Al content is similar 1996) whereas on acidic materials formation of to that followed by nT thereby suggesting that acetone aromatics and olefins is favored (Chang and Silvestri, conversion depends on both acid and base sites. Pure 1977). In our catalytic tests, the main reaction products MgO was the most active catalyst whereas Al2O3 were mesityl oxides (MO's) and isophorone (IP). Over showed the lowest activity. This is because Al-O pairs all the samples the reaction rate diminished as a are much less active than Mg-O pairs for promoting the function of time-on-stream as shown in Figure 1 for the proton abstraction and carbanion stabilization steps

O OH O O C CH + H O 22CHCH3 C CH3 (CH3)2 C CH2 C CH3 (CH3)2 C CH 3 2 Acetone Aldol Diacetone Alcohol Mesityl Oxide O

Aldol + CH3 C CH3

Aldol -H2O Michael -H2O Aldol

CH3 O O CH3 O O

(CH3)2 C CH C CH C CH3 CH3 C CH2 C CH2 C CH3 (CH3)2 C CH C CH C (CH3)2 4,6- dimethyl hepta-3,5-dien-2-one Phorone CH3 4,4-dimethyl’ hepta-2,6-dione (2,6-dimethyl-2,5-heptadien-4-one) 1,6-Aldol 1,6 - -H2O Michael 1,6-Aldol -H O
  2 -Michael 1,6 O CH3

CH3

CH CH3 CH3 CH3 3 Mesitylene Isophorone (3,5,5-trimethyl-2-ciclohexen-1-one)

Figure 4: Acid- and base-catalyzed pathways for acetone conversion by aldol-like reactions on MgyAlOx catalysts.

81 Latin American Applied Research 33:79-86 (2003) involved in aldol condensation reactions. We have TPO TPO(N) shown (Di Cosimo et al., 2000) that the aldolization rate

0.1 is controlled on basic catalysts by the number of metal r = 0 cation-oxygen anion surface pairs. Mg-rich MgyAlOx oxides are less active than MgO because they exhibit a lower base site density and also poor acidic properties.

In contrast, Al-rich MgyAlOx oxides are more active 0.11 than Al2O3 due to a proper combination of acid and base sites. In Figure 3 the initial product selectivities measured 0.18 at isoconversion are represented as a function of catalyst composition. It is observed that the initial selectivities toward IP and MES increase with increasing the Al 0.24

content whereas high selectivity toward MO's was FID Signal (V) obtained on Mg-rich catalysts. The presence of MES in 0.47 low concentrations on Al-rich samples is indicative of an acid-catalyzed reaction pathway (Chang and

Silvestri, 1977). Brönsted OH groups present on the surface of Al-rich catalysts, probably account for MES 1 formation in accord with previous work (Di Cosimo et 400 500 600 700 800 900 al., 1998b) by X-ray photoelectron spectroscopy in Temperature (K) which we demonstrated that in Al-rich MgyAlOx mixed oxides, the binding energy of O 1s signal approached Figure 5: Coke burning experiments after the values of surface OH groups thereby confirming a acetone aldol condensation at 473 K on higher surface hydroxylation degree compared to Mg- MgyAlOx catalysts. r = Al/(Al + Mg) rich catalysts. Fig. 4 shows the main reaction scheme for the aldol results are summarized in Table 2. The amount of condensation sequence leading to IP, where the dotted carbon measured on the samples was between 12 and 22 lines represent the reaction pathways to MES and IP wt. % and, as a general trend, Mg AlO and Al O formation that take place exclusively on acidic catalysts. y x 2 3 formed more coke than pure MgO. The dimer, MO is the detected primary product formed by the initial self-condensation of acetone after fast The TPO curves of Fig. 5 show the existence of two dehydration of the unstable diacetone alcohol whereas different carbon deposits; the main one gives a broad IP is the final product arising from the consecutive aldol oxidation peak centered at 650-700 K (low-temperature condensation between MO and acetone. peak); the other one arises as a shoulder at temperatures above 720 K (high-temperature peak). The low- C. Coke Formation and Characterization temperature peak maximum shifts to higher In a previous work (Di Cosimo and Apesteguía, 1998), temperatures with increasing Al content in the samples we found that MgO-based catalysts are deactivated (Table 2). In order to obtain more insight on the nature during acetone oligomerization because of the formation of the carbon deposits, stripping experiments in pure N2 of carbon deposits that block pores and surface active were carried out. In these experiments, coked samples sites. Here, the coke formed on the samples after the were first treated in N2 from room temperature up to 10-h catalytic runs was characterized by TPO technique. 973 K at 10 K/min before performing standard TPO The TPO curves are shown in Figure 5 and quantitative runs. The TPO curves obtained after the N2 treatment are identified here as TPO(N) and are Table 2. Coke Characterization displayed in dotted lines in Fig. 5. The amount of carbon oxidized in TPO(N) Catalyst TPO TPO(N) experiments was between 4 and 8 wt % Carbon Low-Temperature Carbon Peak (Table 2). The shape and temperature Amount Peak Amount Temperature maximum of the TPO(N) curves were (wt
  %) (K) (wt %) (K) similar to those of the high-temperature peaks in the corresponding TPO curves. MgO 12.5 623 4.7 730 These results show that the nitrogen treatment removes almost completely Mg9AlOx 12.1 625 5.2 724 the carbon deposits that account for Mg5AlOx 13.6 650 5.5 736 low-temperature peaks in TPO traces of Mg3AlOx 13.6 642 5.0 750 Fig. 5. Both kind of carbon deposits Mg AlO 22.4 697 7.6 760 seem to be related to the catalyst acid- 1 x base properties since both peak maxima Al2O3 15.8 709 6.0 787 gradually shifted to higher temperatures

82 V. K. DÍEZ, C. R. APESTEGUÍA, J. I. DI COSIMO

Heavy α,β-unsaturated ketones Mg-O pair sites + n Acetone Tetramers (isoxylitones) Michael Addition + Aldol Cond. H C CH3 Mg-O pair sites 3 + Acetone O C CH Michael Addition 3 + Acetone + Acetone LinearTrimeric CH or Aldol Cond. H3C C = O Intermediates H C CH3 -HO -H2O M-O pair sites or 3 CH 2 C = O (C9 ketones) IP 3 Aldol Cond. Aldol Cond. Brönsted acid sites Acetone CH M-O pair sites 3 M-O pair sites . MO Aldol Cond. Brönsted acid sites CH3

H3C CH3 MES Brönsted acid sites Hydrocarbons Brönsted acid sites

M: Mg or Al Aromatic coke

Figure 6: Coke-forming reactions during acetone oligomerization on MgyAlOx

at increasing Al contents as presented in Table 2. rearrangements (Reichle, 1980). Selective formation of In previous work (Di Cosimo and Apesteguía, 1998) one trimer in preference to the others depends on the we postulated that the coke precursor species on MgO- catalyst acid-base properties. Phorone is the preferred based catalysts are highly unsaturated linear intermediate on basic catalysts whereas formation of compounds, such as 2,6-dimethylhepta-2,5-dien-4-one MES on acidic catalysts (Fig. 3) takes place only from (phorone), 4,6-dimethylhepta-3,5-dien-2-one and 4,4'- 4,6-dimethylhepta-3,5-dien-2-one (Reichle, 1980). To dimethylhepta-2,6-dione which are trimeric compounds obtain more information on the heavy compounds formed from MO and acetone by secondary aldol present in trace amounts in the gaseous products, we condensation reactions as shown in Figure 4. These performed a GC-Mass Spectroscopy analysis of the linear trimers (C9 ketones) are non-detectable in the gas reactor effluent collected in a condenser during acetone phase during reaction because they remain bound to the aldolization on Mg1AlOx catalyst. Besides acetone, surface due to the strong interaction between the MO's, IP and MES, we detected the trimer 4,6- unsaturated chain and the basic centers. All of them are dimethylhepta-3,5-dien-2-one (m/z = 138), tetramers able to give IP by consecutive internal aldol or Michael like isoxylitones (m/z = 178) and also

1.0 0.5 75

W/F°Acetone = 73.4 gh/mol ) -1 MO 70 S 0.8 (h 0.4 0 0 65 j d (carbon atom %) atom (carbon a 60 0.6 0.3 55 49.0 IP 20 0.4 0.2 24.4 15 Activity, Activity,

  12.2 10 0.2 0.1 7.3 5

0.0 deactivation, Initial 0.0 0 0246810 0 20406080 0 Time (h) W/F Acetone (g h/mol) Figure 7: Acetone condensation activity as a function Figure 8: Initial catalyst deactivation and product 0 of time at different contact times (W/F Acetone) over selectivities as a function of contact time over MgO MgO (T = 573 K, PAcetone = 7.7 kPa) (T = 573 K, PAcetone = 7.7 kPa)

83 Latin American Applied Research 33:79-86 (2003)

) 1.0 -1 0.4 (h 0

d 0.8 Al2O3 0.3 a 0.6 MgyAlOx 0.2 0.4 Activity, 0.1 0.2

Initial deactivation, deactivation, Initial MgO 0.0 0.0 468101214 0246810 Carbon Amount (wt%) Time (h) Figure 9: Initial catalyst deactivation as a function of Figure 10: Acetone condensation activity as a function the carbon amount deposited on MgO after 10-h of time over MgyAlOx, MgO and Al2O3 catalysts (T = catalytic runs at different contact times (T = 573 K, 0 473 K, PAcetone = 7.7 kPa, W/F Acetone = 49 g h/mol) PAcetone = 7.7 kPa) ▲ r = 0.11; • r = 0.18; > r = 0.24;♦ r = 0.47 tetramethyltetralones (m/z = 202) and other heavy that the activity decline of MgO is not related to the polymers of m/z = 218 and 240. formation of any terminal product in the consecutive Results can be explained by considering that the reaction pathways. In Fig. 8 we also plotted the initial catalyst surface acid-base properties determine the selectivities toward MO and the final product (IP) as a preferential formation of a given trimeric intermediate function of contact time. The fact that d0 parallels the which in turn defines both the formation of the product MO selectivity dependence confirms that IP is not released to the gas phase (MO, IP or MES) and the responsible for MgO deactivation. We can also discard nature of the carbon deposit. This interpretation is water as responsible for deactivation of MgO since IP depicted in Figure 6. Coke formed from the 4,6- synthesis involves the concomitant formation of two dimethylhepta-3,5-dien-2-one intermediate on acidic water molecules; water partial pressure on the catalyst catalysts (dotted lines in Fig. 6) will be fundamentally surface should therefore increase with increasing aromatic in composition since MES can further react to contact time, which is the opposite trend observed for d0 xylenes and other C9+ aromatic hydrocarbons (Chang in Fig. 8. Then, the effect of water as a surface poison and Silvestri, 1977; Salvapati et al., 1989). On the other able to convert the Lewis acid sites (Mg2+) of MgO into hand, coke formed on more basic materials from inactive Mg-OH surface species can be neglected phorone, will be mostly heavy α,β-unsaturated ketones because although generation of these species in situ is such as isoxylitones, tetralones and other C12+ possible, the labile nature of the surface bond probably oxygenates (Salvapati et al., 1989; Lippert et al., 1991). makes them unstable under reaction conditions. In fact, Then, the gradual shift from one intermediate to another we have shown previously by IR of CO2 the weak causes a concomitant change in coke composition and it character of bicarbonate species formed on surface could explain the higher temperatures needed to burn hydroxyls of MgO (Di Cosimo et al., 1998b). the carbon deposits at increasing Al content in the Since MO itself is not a strong coking agent samples, as observed in Fig. 5. (Reichle, 1980), it is reasonable to assume that on MgO coke may be formed from the unsaturated trimers as D. Relationship between catalyst deactivation and postulated in Fig. 6. Furthermore, in Fig 9 we present a acid-base properties linear correlation between the d0 values and the carbon The relative catalytic activity is defined as a = rt /r0 amount measured by TPO after the catalytic runs of where rt and r0 are the reaction rates at time t and zero Fig. 7, indicating that the carbon deposit formed on the time respectively.
  Fig. 7 shows the activity decay of very active Mg-O pair sites causes the deactivation of MgO as a function of time at different contact times. To the MgO catalyst. quantify the effect of contact time on catalyst activity, Plots of the relative catalytic activity as a function of the experimental points of Fig. 7 were fitted with first time on stream at 473 K (Fig. 10) showed that the order deactivation functions and from the slopes of the activity decay on Mg-Al mixed oxides is faster than on curves, we calculated the initial deactivation rate as Al2O3 but slower than on MgO. Pure MgO presented a d0 = [- da/dt ]t=0. A plot of the resulting d0 values as a rapid deactivation and its residual activity after the 10-h function of contact time is presented in Fig. 8; it is run was very low. From the slope of the plots of Fig. observed that the initial deactivation rate decreases 10, the initial deactivation rate was once more rapidly with increasing contact time. This result shows calculated. In Fig. 11, the d0 values obtained for all the

84 V. K. DÍEZ, C. R. APESTEGUÍA, J. I. DI COSIMO

Carbon Amount (wt %) On MgO, the fast deactivation is due to the fact that 12 16 20 24 aldolization and coke formation reactions compete for 0.6 the same active site, the Mg-O pairs, thereby showing a ) -1 linear correlation between d0 and the carbon amount (h

0 0.5 (Fig. 9). The aldol reaction stops almost completely d even when the carbon deposition is less intense than on 0.4 Al-containing samples. On MgyAlOx oxides, coke will mainly form on Mg-O pairs (Mg-rich samples) or Brönsted acid sites (Al-rich samples) following the Al 0.3 content in the sample as shown in Fig. 6.

In Fig. 11, the d0 values obtained for MgO, Al2O3 0.2 and MgyAlOx oxides were plotted as a function of the

Initial deactivation, density of basic sites, nb (Table 1). It is observed that 0.1 the initial deactivation correlates linearly with the 0123 density of basic sites. No correlation was found 2 Base site density, n (µmol/m ) between the NH3 site density (na, Table 1) and d0. b These results show that although MgyAlOx oxides Figure 11: Initial catalyst deactivation of MgyAlOx promote the self-condensation of acetone by both acid- oxides, MgO and Al2O3 as a function of the base site and base-catalyzed pathways (Fig. 2), the initial density and carbon amount deposited after 10-h deactivation rate is essentially related to the surface catalytic runs (T = 473 K, P = 7.7 kPa, W/F0 Acetone Acetone basic properties. This dependence of d0 with nb for = 49 g h/mol) MgyAlOx oxides probably reflects the fact that the acetone oligomerization rate decline is significantly samples are represented in open symbols as a function higher when coke poisons very active Mg-O pair sites of the carbon content measured after the catalytic runs than when eliminates moderately active Al-O pair site (Table 2). Clearly, it does not exist any correlation because of the more basic features of the former. between d0 and the amount of coke for MgyAlOx catalysts, in contrast to what was obtained on pure MgO IV. CONCLUSIONS (Fig. 9). Initial deactivation is lower on Al2O3 (d0 = -1 -1 0.14 h ) than on MgO (d0 = 0.53 h ), in spite that The activity and selectivity for acetone oligomerization alumina forms more coke during reaction and that the on MgyAlOx oxides depend on the catalyst acid-base coke is more difficult to oxidize as compared to MgO properties, which in turn are determined by the (Table 2 and Fig. 5). aluminum content in the sample. Mg-rich catalysts These results show that neither the coke amount nor selectively yield mesityl oxides whereas Al-rich its polymerization degree account for the catalyst MgyAlOx oxides produce mainly isophorone. MgyAlOx deactivation order observed in Fig. 10. A better samples are more active than alumina. Intermediate explanation is obtained by considering the nature of the Mg/Al compositions show unique catalytic properties surface sites that are responsible for the formation of because fulfill the reaction requirements for the density coke precursors on pure Al2O3 or MgO. Alumina and strength of basic and acid sites. contains Brönsted (OH groups) and Lewis (metal The MgyAlOx activity declines in the acetone cations) acidic sites and probably coke forms oligomerization reaction due to a blockage of both basic preferentially on Brönsted sites rather than on Al-O and acidic active sites by a carbonaceous residue formed acid-base pair sites. Contrarily to the known catalytic by secondary aldol condensation or hydrocarbon- properties of Brönsted OH sites in liquid-phase forming reactions, respectively. The key intermediate heterogeneously catalyzed acetone aldolization at low species for coke formation are highly unsaturated linear temperatures, these sites are not active for the gas-phase trimers formed by aldol condensation of mesityl oxide reaction to unsaturated ketones and the reaction would with acetone that remain strongly bound to the catalyst proceed, therefore, at an appreciable rate on Al-O pairs surface. The catalyst surface acid-base properties even in the presence of a high carbon content deposited determine the preferential formation of a given trimeric on the Brönsted sites.
  intermediate, which in turn defines the chemical nature Contrarily to what has been discussed above about of the carbon deposit. Aromatic hydrocarbons are the disregarding the poisoning effect of water on MgO, we main component of coke formed on acidic Al-rich have no evidence on the role of water in poisoning MgyAlOx samples whereas heavy α,β-unsaturated Lewis acid sites of alumina. In fact, surface ketones preferentially form on basic Mg-rich catalysts. transformations caused by water leading to Al-H and Alumina forms more and heavier coke than MgO Al-OH species with the consequent diminution of active but the latter deactivates faster. This is explained by the surface Al-O pairs should be also taken into account in fact that on MgO primary aldol condensation reactions addition to coke formation when dealing with and formation of coke precursors species take place on deactivation of alumina or Al-rich mixed oxides. the same sites, i.e., the Mg-O pairs in which oxygen

85 Latin American Applied Research 33:79-86 (2003) presents moderate basic properties. In contrast, main 177-185 (1998). reaction to ketones takes place on the Al-O pairs of Di Cosimo, J.I., V.K. Díez and C.R. Apesteguía, alumina whereas CHx coke precursors form on Brönsted “Syntesis of α,β-unsaturated ketones over OH groups that are inactive for aldol condensation thermally activated Mg-Al hydrotalcites”, Appl. reactions to unsaturated ketones. Clay Science 13, 433-449 (1998a). Deactivation of MgyAlOx oxides is caused by carbon Di Cosimo, J. I. , V. K. Díez, M. Xu, E. Iglesia and C. deposition but neither the carbon amount nor its R. Apesteguía, “Structure and Surface and polymerization degree are key factors in terms of the Catalytic Properties of Mg-Al basic oxides”, J. deactivation rate. The nature of surface active sites for Catal. 178, 499-510 (1998b). coke formation, which depend on the catalyst chemical Di Cosimo, J.I., C.R. Apesteguía, M.J.L Ginés and E. composition, seems to define the activity decay rate. Iglesia, “Structural requirements and reaction

Deactivation of MgyAlOx oxides during acetone pathways in condensation reactions of alcohols on oligomerization reactions is essentially related to the MgyAlOx catalysts”, J. Catal. 190, 261-275 surface basic properties. This is because the activity (2000). decline is significantly higher when coke poisons very Garrone, E. and F. Stone, “UV diffuse reflectance active basic Mg-O pairs than when eliminates spectra of ammonia adsorbed on alkaline earth moderately active acidic Al-O pairs. oxides and analogy with charge-transfer-to-solvent spectra”, in M. Che and G.C. Bond (Eds.), Acknowledgements Adsorption and Catalysis on Oxide Surfaces, Elsevier, Amsterdam (1985). Support of this work by the Consejo Nacional de Lippert, S., W. Baumann and K. Thomke, “Secondary Investigaciones Científicas y Técnicas (CONICET), reactions of the base-catalyzed aldol condensation Argentina and by the Universidad Nacional del Litoral, of acetone”, J. Molec. Catal. 69, 199-214 (1991). grant CAI+D 94-0858-007-049, Santa Fe, Argentina, is Prinetto, F., G. Ghiotti, R. Durand and D. Tichit, gratefully acknowledged. “Investigation of acid-base properties of catalysts obtained from layered double hydroxides”, J. REFERENCES Phys. Chem. B 104, 11117-11126 (2000). Berteau, P., B. Delmon, J.L. Dallons and A. Van Gysel, Reichle, W.T., “Pulse microreactor examination of the “Acid-base pproperties of silica-aluminas: use of vapor-phase aldol condensation of acetone,” J. 1-butanol dehydration as a test reaction”, Appl. Catal. 63, 295-306 (1980). Catal. 70, 307-323 (1991). Reichle, W.T., “Catalytic reactions by thermally Chang, C.D. and A.J. Silvestri, “The conversion of activated, synthetic, anionic clay minerals”, J. methanol and other O-compounds to hydrocarbons Catal. 94, 547-557 (1985). over zeolite catalysts”, J. Catal. 47, 249-259 Salvapati, G.S., K.V. Ramanamurty and M. (1977). Janardanarao, “Selective catalytic self- Di Cosimo, J.I., V.K. Díez and C.R. Apesteguía, “Base condensation of acetone”, J. Molec. Catal. 54, 9- catalysis for the síntesis of α,β−unsaturated 30 (1989). ketones from the vapor-phase aldol condensation Velu, S. and C.S. Swamy, “Kinetics of the alkylation of of acetone”, Appl. Catal. 137, 149-166 (1996). phenol with methanol over magnesium-aluminium Di Cosimo, J.I. and C.R. Apesteguía, “Study of the calcined hydrotalcites”, Appl. Catal. 145, 225-230 catalyst deactivation in the base catalyzed (1996). oligomerization of acetone”, J. Molec. Catal. 130,

 

Received: September 16, 2001. Accepted for publication: July, 26, 2002. Recommended by Guest Editors J. Cerdá, S. Díaz and A. Bandoni.

86