Skip to Main Content
Skip Nav Destination

The generation, characterization and catalytic properties of MgO active sites were studied. MgO samples stabilized at different temperatures were used to control the distribution of surface base sites; specifically, MgO was calcined at 673 K, 773 K and 873 K (samples MgO-673, MgO-773 and MgO-873). The nature, density and strength of MgO base sites were characterized by temperature-programmed desorption of CO2 and infrared spectroscopy after CO2 adsorption at 298 K and sequential evacuation at increasing temperatures. MgO samples contained surface sites of strong (low coordination O2− anions), medium (oxygen in Mg2+-O2− pairs) and weak (OH groups) basicity. The density of strong basic sites was predominant on MgO-673. The increase of the calcination temperature drastically decreased the density of strong base sites and to a lesser extent that of weak OH groups, while slightly increased that of medium-strength base sites. The catalytic properties of MgO samples were proved for the aldol condensation of citral with acetone to yieldpseudoionone, the hydrogen transfer reaction of mesityl oxide with 2-propanol to obtain the unsaturated alcohol 4-methyl-3-penten-2ol, and the synthesis of monoglycerides via the transesterification of methyl oleate with glycerol. The effect of calcination temperature on the MgO catalytic properties depended on the basicity requirements for the rate-limiting step of the base-catalyzed reaction. The activity for both the aldol condensation of citral with acetone and the glycerolysis of methyl oleate diminished with the MgO calcination temperature because these reactions were essentially promoted on strongly basic O2− sites. In contrast, the synthesis of 4-methyl-3-penten-2ol by the hydrogen transfer reduction of mesityl oxide with 2-propanol increased with calcination temperature because the reaction intermediate was formed on medium-strength Mg2+-O2− pair basic sites. Additional information on the role played by the MgO active sites on the kinetics of base-catalyzed reactions was obtained by performing molecular modeling studies on our MgO catalysts using Density Functional Theory (DFT) for the glycerolysis of methyl oleate, an unsaturated fatty acid methyl ester (FAME). The molecular modeling of glycerol and FAME adsorptions was carried out using terrace, edge and corner sites for representing the MgO (100) surface. In agreement with catalytic results, calculations predicted that dissociative chemisorption of glycerol with O–H bond breaking occurs only on strong base sites (edge sites) whereas nondissociative adsorption takes place on medium-strength base sites such as those of terrace sites. Results also indicated that glycerol was more strongly adsorbed than FAME. The glycerol/FAME reaction would proceed then through a mechanism in which the most relevant adsorption step is that of glycerol.

Alkaline earth metal oxides catalyze a variety of organic reactions requiring the cleavage of a C–H bond step and the formation of carbanion intermediates. In particular, pure and alkali-promoted MgO has been shown to promote Cannizzaro and Tischenko reactions [1, 2], Michael, Wittig and Knoevenagel condensations [3, 4], transesterification reactions [5–8], double-bond isomerizations [9], self- and cross-condensation reactions [10–13], Henry reaction [14], alcohol coupling [15–17], and H2 transfer reactions [18]. However, the MgO basicity needed for efficiently promoting these reactions depend on the rate-limiting step requirements. MgO can be synthesized in a variety of presentation formats, including nanosheets [19], nanowires [20] and nanoparticles [21], but its catalytic properties depend greatly on the preparation method. Nevertheless, most of reports on the preparation of magnesia deal with the effect of the synthesis method and conditions on the MgO structural and physical properties [22–24]. Very few papers have attempted to tailor the distribution, density, and strength of surface base sites of MgO upon synthesis in order to design the catalyst surface to reaction requirements [25–27]. More insight on the relationship between the synthesis procedure with the generation and control of MgO surface base sites is then required to improve the efficient use of this oxide in catalysis applications.

Detailed characterization of MgO base sites is crucial to establish correlations between the surface basic properties and the catalyst activity and selectivity for a given reaction. The most common methods for characterization of solid basicity are thermal programmed desorption (TPD) and infrared spectroscopy (IR) of preadsorbed probe molecules, and the use of test reactions. TPD studies provide information on the density and strength of base sites while additional insight on the base site nature is often obtained by IR characterization. Carbon dioxide has been largely employed as a probe molecule for evaluating the solid basicity by TPD and IR techniques [28–31] although other acid molecules such as acetic acid have been also used [32]. On the other hand, the test reactions most frequently used for characterizing the catalyst acid-base properties are the decomposition of alcohols, in particular 2-propanol [33–35], 2-butanol [36, 37] and 2-methyl-3-butyn-2-ol [38–40]. In the case of 2-propanol, it is generally accepted that 2-propanol dehydration to propylene occurs on solid acids containing Brønsted acid sites via an E1 mechanism while on amphoteric oxides with acid-base pair sites propylene is obtained through a concerted E2 mechanism [41]. On strong basic catalysts, 2-propanol is dehydrogenated to acetone via an E1cB anionic mechanism [42]. Thus, the catalyst acid-base properties may be related to the propylene/acetone selectivity ratio. In contrast, test reactions have been used only in few cases for characterizing base site strength distributions on solid bases. For example, in a previous work [43], we proposed that on alkali-modified MgO catalysts 2-propanol decomposition to acetone and propylene takes place via an E1cB mechanism in two parallel pathways sharing a common 2-propoxy intermediate; in this mechanism, the intermediate-strength base sites promote acetone formation, whereas high-strength base sites selectively yield propylene. Nevertheless, several studies have shown that the use of test reactions is not sensitive enough to establish a basicity scale of the catalysts [44].

Theoretical calculations of surface sites have been performed for exploration of MgO catalysis. In general, Density Functional Theory (DFT) calculations have shown to be a powerful tool to characterize the thermal stability of hydrated oxide surfaces [45]. Regarding MgO catalysts, DFT studies on the structure of MgO surface defects have been carried out to establish the stability of surface OH groups for water and methanol adsorptions [46, 47]. Recently, combined IR and DFT studies have been performed in an attempt to specify the actual structure of the CO2 species adsorbed on magnesium oxide surface [48]. Unfortunately, theoretical calculations to predict the relationship between the basic site nature and strength and the reaction mechanism have been done only for limited cases.

In this work we study the generation, characterization and catalytic properties of active sites on MgO catalysts. The base properties of MgO samples obtained from Mg(OH)2 decomposition were tuned by modifying the solid calcination temperature. The density and strength of MgO surface base sites were determined by TPD and IR spectroscopy of CO2 adsorbed at 298 K. The activity and selectivity of MgO samples were probed for the liquid-phase cross-aldol condensation of citral with acetone to obtainpseudoionones, the liquid-phase transesterification of methyl oleate with glycerol to yield monoglycerides, and the gas-phase hydrogen transfer reduction of mesityl oxide with 2-propanol toward 4-methyl-3-penten-2ol. Besides, we performed DFT calculations to obtain additional information on the role played by the MgO active sites on the kinetics of base-catalyzed reactions. Specifically, we present molecular modeling studies on our MgO catalysts for the glycerolysis of methyl oleate.

Magnesium oxide samples were prepared by hydration with distilled water of low-surface area commercial MgO (Carlo Erba, 99%, 27 m2/g). 250 ml of distilled water were slowly added to 25 g of commercial MgO and stirred at room temperature. The temperature was then raised to 353 K and stirring was maintained for 4 h. Excess of water was removed by drying the sample in an oven at 358 K overnight. The resulting Mg(OH)2 was decomposed in N2 (30 ml/min STP) to obtain high-surface area MgO which was then treated for 18 h in N2 either at 673, 773 or 873 K to give samples MgO-673, MgO-773 and MgO-873, respectively.

The decomposition of Mg(OH)2 was investigated by differential thermal analysis (DTA) using a Shimadzu DT30 analyzer, by temperature programmed decomposition (TPDe) using a flame ionization detector with a methanation catalyst (Ni/Kieselghur) operating at 673 K and by X-ray diffraction (XRD) in a Shimadzu XD-D1 diffractometer equipped with Cu-Kα radiation source (λ=0.1542nm) and a high temperature chamber. Samples characterized by X-ray diffraction were heated at 5 K/min until 773 K, taking diffractograms at 298, 373, 573, 673 and 773 K.

Surface areas and pore volumes were measured by N2 physisorption at its boiling point using the BET method and Barret-Joyner-Halender (BJH) calculations, respectively, in an Autosorb Quantochrome 1-C sorptometer. The crystalline structure properties of MgO-x samples were determined by X-ray diffraction (XRD) using the instrument described above. Analysis was carried out using a continuous scan mode at 2°/min over a 2θ range of 20°–80°. Scherrer equation was used to calculate the mean crystallite size of the samples.

CO2 adsorption site densities and binding energies were determined from temperature-programmed desorption (TPD) of CO2 preadsorbed at room temperature. MgO-x samples were pretreated in situ in a N2 flow at its corresponding stabilization temperature (673, 773 or 873 K), cooled to room temperature, and then exposed to a mixture of 3% CO2/N2 until surface saturation was achieved (10 min). Weakly adsorbed CO2 was removed by flushing in N2 during 1 h. Finally, the temperature was increased to 773 K at 10 K/min. The desorbed CO2 was converted to methane by means of a methanation catalyst (Ni/Kieselghur) operating at 673 K and monitored using a flame ionization detector.

The chemical nature of adsorbed surface CO2 species was determined by infrared (IR) spectroscopy after CO2 adsorption at 298 K and sequential evacuation at increasing temperatures. Experiments were carried out using an inverted T-shaped cell containing the sample pellet and fitted with CaF2 windows. Data were collected in a Shimadzu FTIR Prestige-21 spectrometer. The absorbance scales were normalized to 20-mg pellets. Each sample was pretreated in vacuum at its corresponding stabilization temperature and cooled to room temperature, after which the spectrum of the pretreated catalyst was obtained. After admission of 5 kPa of CO2 to the cell at room temperature, the samples were evacuated at increased temperatures, and the resulting spectrum was recorded at room temperature. Spectra of the adsorbed species were obtained by subtracting the catalyst spectrum.

The cross-aldol condensation of citral (Millennium Chemicals, 95% geranial+neral) with acetone (Merck, p.a.) was carried out at 353 K under autogenous pressure (≈250 kPa) in a batch Parr reactor, using acetone/citral=49 (molar ratio) and catalyst/(citral+acetone)=1 wt% ratio. The reactor was assumed to be perfectly mixed and interparticle and intraparticle diffusional limitations were verified to be negligible. Reaction products were analyzed by gas chromatography in a Varian Star 3400 CX chromatograph equipped with a FID and a Carbowax Amine 30 M capillary column. Samples of the reaction mixture were extracted every 30 min and analyzed during the 6-h reaction. The main product of the citral/acetone reaction waspseudoionone, PS (cis- and trans-isomers). Moreover, diacetone alcohol and mesityl oxide were simultaneously produced from self-condensation of acetone. Selectivities (Sj, mol of product j/mol of citral reacted) were calculated as Sj (%)=Cj×100/ΣCj, where Cj is the concentration of product j. Yields (ηj, mol of product j/mol of citral fed) were calculated as ηj=SjXCit, where XCit is the citral conversion.

The transesterification of methyl oleate, FAME, (Fluka, >60.0%, with 86% total C18+C16 esters as determined by gas chromatography) with glycerol (Aldrich, 99.0%,) was carried out at 493 K in a seven-necked cylindrical glass reactor that allows: separate loading of the two reactants and the catalyst, stirrer, thermocouple, in-out of inert gas to eliminate methanol of the gas phase, and periodical product sampling.

Glycerol/FAME molar ratio of 4.5 and a catalyst/FAME ratio (Wcat/n0FAME) of 30 g/mol were used. The reactor was operated in a semi-batch regime atatmospheric pressure under N2 (35cm3/min). Liquid reactants were introduced into the reactor and flushed with nitrogen; then the reactor was heated to reaction temperature under stirring (700 rpm). Reaction products were α- and β-glyceryl monooleates (MG), 1,2- and 1,3-glyceryl dioleates (diglycerides) and glyceryl trioleate (triglyceride). Reactant and products were analyzed by gas chromatography in a SRI 8610C gas chromatograph equipped with a flame ionization detector, on-column injector port and a HP-1 Agilent Technologies 15 meter×0.32mm×0.1 µm capillary column after silylation to improve compound detectability, as detailed elsewhere [49]. Twelve samples of the reaction mixture were extracted and analyzed during the 8-h catalytic run.

The gas-phase mesityl oxide/2-propanol reaction was conducted at 573 K andatmospheric pressure in a fixed bed reactor. MgO-x samples sieved at 0.35–0.42mm were pretreated in N2 at the corresponding calcination temperatures for 1 h before reaction in order to remove adsorbed H2O and CO2. The reactants, mesityl oxide (Acros 99%, isomer mixture of mesityl oxide/isomesityl oxide=91/9) and 2-propanol (Merck, ACS, 99.5%), were introduced together with the proper molar composition via a syringe pump and vaporized into flowing N2 to give a N2/IPA/MO=93.4/6.6/1.3, kPa ratio. Reaction products were analyzed by on-line gas chromatography in a Varian Star 3400 CX chromatograph equipped with a flame ionization detector and a 0.2% Carbowax 1500/80–100 Carbopack C column. Main reaction products from mesityl oxide conversion were identified as the two unsaturated alcohol isomers (UOL, 4-methyl-3-penten-2ol and 4-methyl-4-penten-2ol), isomesityl oxide, methyl isobutyl ketone, and methyl isobutyl carbinol.

The base site properties of MgO depend on the preparation method. Usually, MgO is obtained by decomposition of Mg(OH)2 that in turn is produced by different methods such as chemical vapor deposition (CVD), sol-gel, precipitation, and MgO hydration. It has been reported [50] that after Mg(OH)2 decomposition at high temperature (1023 K), the relative distribution of surface low-coordination O2− anions is shifted toward the less coordinated ions along the series MgO-CVD<MgO-hydration≈MgO-precipitation<MgO-sol-gel. The same order was observed for MgO activity to convert 2-methylbut-3-yn-2-ol into acetone and acetylene, a base-catalyzed reaction [50]. The density and strength of base sites on MgO may also be regulated by controlling both the Mg(OH)2 decomposition and MgO activation conditions. For example, Vidruk et al.[51] reported that densification of Mg(OH)2 before its dehydration to obtain MgO generates a significant increase of surface basicity. We have recently investigated [52] the effect of calcination temperature of MgO obtained by Mg(OH)2 decomposition on its base and catalytic properties.

The thermal decomposition of Mg(OH)2 precursor was studied by XRD. The diffractograms in Fig. 1 showed that the Mg(OH)2 brucite structure was stable up to about 573 K, but then, between 573 and 673 K, decomposed to MgO. Figure 1 also shows that the MgO stabilized at 773 K during 18 h is more crystalline than that obtained by dynamic heating up to the same temperature. Consistently, characterization by DTA technique showed that the Mg(OH)2 heating exhibits an endothermic peak between 573 and 673 K arising from the solid decomposition [52]. On the other hand, TPDe experiments revealed the presence of evolved CO2 in the 573–673 K decomposition region, thereby suggesting that the Mg(OH)2 surface is reversibly carbonated by interaction withatmospheric CO2. All these results showed that the thermal treatment of Mg(OH)2 between 575 and 675 K decomposes the solid into crystalline MgO and eliminates adsorbed carbonate species.

Figure 1.1

XRD diffraction patterns of Mg(OH)2 decomposition.

Figure 1.1

XRD diffraction patterns of Mg(OH)2 decomposition.

Close modal

The physical propertied of MgO-x samples are presented in Table 1. The MgO surface area decreased with calcination temperature, from 196 m2/g (MgO-673) to 169 m2/g (MgO-873), while both the mean pore size and the pore volume increased with calcination temperature. The XRD patterns of MgO-x samples exhibited only one crystalline species of MgO periclase. The face-centered cubic unit cell dimensions for MgO-x samples given in Table 1 show that the lattice parameter (a) decreased with calcination temperature. Contraction of the MgO unit cell was accompanied by the increase of crystallite diameter and the sample crystallinity. Data in Table 1 showed that, as expected, the increase of the calcination temperature generated more ordered structures.

Table 1.1

Physical and basic properties of MgO-x samples.

SampleTextural characterizationXRD analysisBase site density (μmol/m2)a
Surface area (m2/g)Pore volume (ml/g)Lattice parameter, a (Å)Crystallite size (Å)Crystallinity (%)Weak nOHMedium nMg-OStrong nOTotal nb
MgO-673 196 0.30 4.243 74.3 85.4 0.71 1.21 2.66 4.58 
MgO-773 189 0.38 4.221 76.5 86.6 0.54 1.26 1.66 3.46 
MgO-873 169 0.44 4.214 143.0 93.4 0.51 1.78 1.19 3.13 
SampleTextural characterizationXRD analysisBase site density (μmol/m2)a
Surface area (m2/g)Pore volume (ml/g)Lattice parameter, a (Å)Crystallite size (Å)Crystallinity (%)Weak nOHMedium nMg-OStrong nOTotal nb
MgO-673 196 0.30 4.243 74.3 85.4 0.71 1.21 2.66 4.58 
MgO-773 189 0.38 4.221 76.5 86.6 0.54 1.26 1.66 3.46 
MgO-873 169 0.44 4.214 143.0 93.4 0.51 1.78 1.19 3.13 
a

By TPD of CO2

The surface basic properties of MgO-x samples were probed by TPD of CO2 and by FTIR of CO2 preadsorbed at room temperature and desorbed at increasing temperatures. Figure 2 shows the IR spectra obtained for MgO-x samples that reveal the presence of at least three different species: unidentate carbonate, bidentate carbonate and bicarbonate [53–57]. Unidentate carbonate formation requires isolated surface O2− ions, i.e., low-coordination anions, such as those present in corners or edges and exhibits a symmetric O-C-O stretching at 1360–1400cm−1 and an asymmetric O-C-O stretching at 1510–1560cm−1. Bidentate carbonate forms on Lewis acid-Brønsted base pairs (Mg2+-O2− pair site), and shows a symmetric O-C-O stretching at 1320–1340cm−1 and an asymmetric O-C-O stretching at 1610–1630cm−1. Bicarbonate species formation involves surface hydroxyl groups and shows a C-OH bending mode at 1220cm−1 as well as symmetric and asymmetric O-C-O stretching bands at 1480cm−1 and 1650cm−1, respectively. Bicarbonate was the most labile species and disappeared on all the samples after evacuation at 373 K. In contrast, both the unidentate and bidentate carbonates remained on the surface after evacuation at 473 K, but only the unidentate carbonate bands were observed upon evacuation at higher temperatures. These results suggest the following strength order for surface basic sites: low-coordination O2− anions>oxygen in Mg2+-O2− pairs>OH groups. On a perfect MgO (1 0 0) surface, Mg2+ and O2− are five coordinated ions (Mg5c and O5c) but on the surface of the high-surface area MgO catalysts used here, both ions are also present with coordination numbers (L) lower than 5 depending on the location in corners or edges. Specifically, L is 5, 4 or 3 for ions in terrace, edge or corner sites, respectively, as shown in Fig. 3[58, 59]. Other authors have confirmed using HRTEM that MgO particles prepared from precipitation of Mg(OH)2 are plenty of surface defects [27]. In Fig. 2, the overlapping adsorption bands giving rise to broad bands for both, unidentate and bidentate carbonate species, suggest the presence of surface sites with different coordination numbers, i.e., in different chemical environment, that bind CO2 with a distribution of basic strength. In contrast, narrower and smoother bands were obtained on more crystalline MgO particles resulting from calcination at higher temperatures [27, 55, 60].

Figure 1.2

Infrared spectra of CO2 adsorbed on MgO-x catalysts upon evacuation at increasing temperatures: (a) 298 K, (b) 373 K, (c) 473 K, (d) 573 K.

Figure 1.2

Infrared spectra of CO2 adsorbed on MgO-x catalysts upon evacuation at increasing temperatures: (a) 298 K, (b) 373 K, (c) 473 K, (d) 573 K.

Close modal
Figure 1.3

Scheme of a stepped MgO (1 0 0) surface with OLc and MgLc ions in different positions (L: coordination number). Terrace sites: O5c, Mg5c; edge sites: O4c, Mg4c; corner sites: O3c, Mg3c.

Figure 1.3

Scheme of a stepped MgO (1 0 0) surface with OLc and MgLc ions in different positions (L: coordination number). Terrace sites: O5c, Mg5c; edge sites: O4c, Mg4c; corner sites: O3c, Mg3c.

Close modal

From the spectra of Fig. 2 we determined the unidentate carbonate/bidentate carbonate band intensity ratios (U.C./B.C.); the obtained values for MgO-673, MgO-773 and MgO-873 samples are plotted in Fig. 4. It is observed that the (U.C./B.C.) intensity ratio on MgO calcined at 673K was 2.5 and then decreased with the calcination temperature. This result may be interpreted by considering that the decomposition of Mg(OH)2 at relatively low temperature, i.e. 673 K, generates hydroxylated MgO containing a high concentration of low-coordination O2− sites located on defects of the crystalline solid surface. Then, the increase of the calcination temperature would remove OH groups and also surface solid defects creating a smoother and thermodynamically more stable structure, as suggested by the XRD data of Table 1. This interpretation is consistent with the results reported by Morterra et al.[54] on MgAl2O4 and by Evans and Whateley [61] on MgO. These authors investigated by IR of CO2 the role of surface hydroxylation on the generation of strong basic sites and concluded that the strong basicity, responsible for unidentate carbonate formation, is promoted by the presence of surface OH groups. Then, the observed loss of unidentate carbonate formation centers (low coordination surface O2− ions) when the calcination temperature is increased can be ascribed to both the elimination of surface defects and the enhancement of surface dehydroxylation.

Figure 1.4

Strong/medium-strength base site and U.C./B.C. ratios as a function of calcination temperature. (B.C.: bidentate carbonate; U.C.: unidentate carbonate).

Figure 1.4

Strong/medium-strength base site and U.C./B.C. ratios as a function of calcination temperature. (B.C.: bidentate carbonate; U.C.: unidentate carbonate).

Close modal

A measure of the number and strength distribution of basic sites on MgO-x samples was obtained by TPD of CO2 preadsorbed at room temperature. The CO2 desorption rate as a function of desorption temperature is presented in Fig. 5. The total base site densities of desorbed CO2 (nb,μmol/m2) were measured by integration of TPD curves in Fig. 5 and are reported in Table 1. It is observed that nb decreased with calcination temperature, from 4.58μmol/m2 (MgO-673) to 3.13μmol/m2 (MgO-873), thereby confirming a solid surface transformation that goes beyond the mere coalescence of the pore structure.

Figure 1.5

TPD profiles of CO2 on MgO-x samples. CO2 adsorption at 298 K, 10 K/min heating rate.

Figure 1.5

TPD profiles of CO2 on MgO-x samples. CO2 adsorption at 298 K, 10 K/min heating rate.

Close modal

Based on the previous IR characterization data, the TPD profiles of Fig. 5 were deconvoluted in three desorption peaks: a low temperature peak at 390 K, assigned to bicarbonates formed on surface OH groups, a middle-temperature peak at 440 K attributed to bidentate carbonates desorbed from Mg2+-O2− pairs, and a high-temperature peak at 550 K resulting from unidentate carbonates released from low-coordination O2− anions. By integrating these three CO2 TPD peaks we determined the density of strongly basic low coordination (O3c and O4c) anions identified as nO in Table 1, medium strength Mg5c-O5c pair sites, nMg-O, and weak OH groups, nOH. Results in Table 1 show that nO and nOH decreased while nMg-O increased with the MgO calcination temperature. In Fig. 4 we plotted the nO/nMg-O ratio as a function of calcination temperature; it is observed that nO/nMg-O values decreased with the calcination temperature following a similar trend that the (U.C./B.C.) ratio determined by IR spectroscopy.

In summary, all these results show that the decomposition of Mg(OH)2 at 673 K generates hydroxylated MgO containing predominantly high-strength low-coordination O2− basic sites located on defects of the crystalline solid surface. The increase of the calcination temperature up to 873 K removes OH groups and also surface solid defects creating more stable structures that contain a higher concentration of medium-strength Mg2+-O2− basic pair sites. Thus, the density, nature and strength of MgO surface basic sites may be regulated by modifying the solid calcination temperature.

Finally, it is significant to note here that in a previous work we have characterized the acid properties of MgO-673 by NH3 TPD and FTIR of adsorbed pyridine [62]. We observed that MgO-673 contained only weak Lewis Mg+2 acid sites; the density of Mg2+ sites as determined by NH3 TPD was 0.14μmol/m2, i.e. about 30 times lower than the density of base sites determined by CO2 TPD (Table 1, nb=4.58μmol/m2).

In order to investigate the effect of MgO calcination temperature on catalyst activity, we carried out two base-catalyzed reactions on our MgO-x samples: the liquid-phase aldol condensation of citral with acetone and the gas-phase hydrogen transfer reduction of mesityl oxide with 2-propanol. For both reactions, catalysts were treated at their calcination temperatures prior to performing the catalytic tests.

The aldol condensation of citral with acetone producespseudoionone (Fig. 6), a valuable acyclic intermediate for the synthesis of ionones which are extensively used as pharmaceuticals and fragrances [63]. The reaction was commercially carried out using diluted bases such as NaOH, Ba(OH)2 or LiOH [64, 65], but it is also efficiently catalyzed on solid bases [13, 66–68].

Figure 1.6

Synthesis ofpseudoionone by citral/acetone aldol condensation.

Figure 1.6

Synthesis ofpseudoionone by citral/acetone aldol condensation.

Close modal

Here, the liquid-phase citral/acetone reaction was performed on the MgO-x samples of Table 1. Figure 7 shows the evolution ofpseudoionone yields (ηPS) as a function of reaction time. At the end of the 6-h catalytic tests, citral conversion was 96%, 80% and 75% for MgO-673, MgO-773 and MgO-873 samples, respectively (Table 2). From the curves of Fig. 7, we determined the initialpseudoionone formation rate (rPS0, mol/h m2) through the initial slopes according to:

graphic
where Wcat is the catalyst weight and nCit0 are the initial moles of citral. The obtained rPS0 values (Table 2) decreased with the calcination temperature, following a trend similar to the density of strong base sites shown in Table 1 (nO values). In all the cases, the initial selectivities topseudoionones were about 100% showing that the conversion of citral via other reactions than its condensation with acetone is negligible. The observed proportionality between rPS0 and nO suggests that under initial conditions the rate-determining step for the citral/acetone reaction towardpseudoionones is promoted by strongly O2− basic sites, which is in agreement with the results reported elsewhere on MgO-based catalysts [13, 62]. The function of surface O2− sites is to abstract the α-proton from acetone, forming a carbanion that consecutively attacks the carbonyl group of the contiguously adsorbed citral molecule, as depicted in Fig. 8. Then a β-hydroxyl ketone intermediate is expected to form; however, this compound was never observed among the reaction products under the reaction conditions of this work. Therefore, this unstable intermediate is assumed to rapidly dehydrate, formingpseudoionone and water and regenerating the active sites on the catalyst surface. The role of surface Mg2+ sites is to provide adsorption sites for acetone through its carbonyl group and to stabilize the reaction intermediates (Fig. 8).

Table 1.2

Catalytic activity data on MgO-x samples.

CatalystReaction
Citral/acetone (liquid phase)aMO/2-propanol (gas phase)bFAME/glycerol (liquid phase)c
XCit, %dInitial PS formation rate, rPS0UOL formation rate, rUOL,XFAME (%)eInitial MG formation rate
mmol/h gmmol/h m2mmol/h gmmol/h m2mmol/h gmmol/h m2
MgO-673 96 93.5 0.477 18.2 0.093 93 28.6 0.146 
MgO-773 80 76.2 0.403 18.3 0.097 84 21.2 0.112 
MgO-873 75 62.9 0.372 18.6 0.110 76 15.2 0.090 
CatalystReaction
Citral/acetone (liquid phase)aMO/2-propanol (gas phase)bFAME/glycerol (liquid phase)c
XCit, %dInitial PS formation rate, rPS0UOL formation rate, rUOL,XFAME (%)eInitial MG formation rate
mmol/h gmmol/h m2mmol/h gmmol/h m2mmol/h gmmol/h m2
MgO-673 96 93.5 0.477 18.2 0.093 93 28.6 0.146 
MgO-773 80 76.2 0.403 18.3 0.097 84 21.2 0.112 
MgO-873 75 62.9 0.372 18.6 0.110 76 15.2 0.090 
a

T=353 K, nDMK0=0.8 moles, nCit0=0.016 moles, WCat.=0.5 g.

b

T=573 K, P=101.3 kPa, N2/IPA/MO=93.4/6.6./1.3, W/FMO0=15 g h/mol.

c

T=493 K; Gly/FAME=4.5; Wcat/n0FAME=30 g/mol K.

d

At the end of the 6-h catalytic runs.

e

After 3 h of reaction

Figure 1.7

Aldol condensation of citral with acetone:pseudoionone yield as a function of time (T=353 K, nAcet0=0.8 moles, nCit0=0.016 moles, WCat.=0.5 g).

Figure 1.7

Aldol condensation of citral with acetone:pseudoionone yield as a function of time (T=353 K, nAcet0=0.8 moles, nCit0=0.016 moles, WCat.=0.5 g).

Close modal
Figure 1.8

Reaction mechanism for citral/acetone aldol condensation.

Figure 1.8

Reaction mechanism for citral/acetone aldol condensation.

Close modal

The selective synthesis of secondary unsaturated (UOL) alcohols from reduction of alkyl vinyl ketones is an important process for pharmaceutical, fragrance and polymer industries. This reaction is hardly achieved on noble metals by conventional hydrogenation that uses high-pressure H2 in multiphase batch reactors because reduction of the CC bond is thermodynamically and kinetically favored over that of the CO group [69, 70]. The substituent at the carbonyl hinders coordination of the CO bond on the surface thereby decreasing the chemoselectivity for the CO bond saturation [71, 72]. In addition, the consecutive UOL isomerization to the corresponding saturated ketone is usually an unavoidable side reaction on metallic catalysts [73]. Hydrogen transfer reduction (HTR) reactions is an alternative route for the catalytic synthesis of UOL by asymmetric reduction of the corresponding a ketone. In the HTR reaction, the carbonyl compound (oxidant) is contacted with a hydrogen donor (reductant) at mild conditions in liquid or gas phase without supply of molecular hydrogen. Heterogeneously catalyzed HTR of unsaturated carbonyl compounds would occur on metal oxides via a Meerwein-Ponndorf-Verley mechanism, which involves the selective reduction of the CO bond preserving the CC bond [74, 75]. In particular, we have studied the gas-phase HTR of 2-cyclohexenone and mesityl oxide (MO) with 2-propanol toward the corresponding unsaturated alcohol on base, acid-base and metal/acid-base catalysts [18, 76, 77]. Here, we present the results obtained for the gas-phase mesityl oxide/2-propanol reaction (Fig. 9) on the MgO-x samples of Table 1.

Figure 1.9

Unsaturated alcohol (UOL) synthesis by hydrogen transfer reduction (HTR) of mesityl oxide (MO) with 2-propanol.

Figure 1.9

Unsaturated alcohol (UOL) synthesis by hydrogen transfer reduction (HTR) of mesityl oxide (MO) with 2-propanol.

Close modal

When both reactants (MO and 2-propanol) are co-fed to the reactor, in addition to the reaction of Fig. 9 several parallel or consecutive reactions can take place, such as:

  1. MO double bond isomerization

    graphic

  2. Selective CC bond reduction of MO or i-MO to MIBK

    graphic

  3. Simultaneous CC and CO bond reduction of MO or i-MO to MIBC:

    graphic

  4. Aldol condensation reactions between C6 carbonyl compounds and acetone toward C9 compounds.

To obtain insight on the reaction pathways of the MO/2-propanol reaction, we investigated the effect of contact time (W/FMO0) on product distribution over MgO-773 at 523 K by varying W/FMO0 between 2.0 and 42.0 gcat h/mol MO [77]. In Fig. 10 we plotted the yields for MO reactions as a function of contact time. UOL formed fast and directly from MO and i-MO, and the UOL yield increased with W/FMO0 up to about 42% without reaching any maximum, thereby suggesting that UOL does not participate significantly in consecutive reactions on MgO-773. MIBK yields of less than 1% were measured regardless of the conversion level thereby confirming that selective reduction of the CC bond of mesityl oxide is unlikely on MgO. The MIBC curve in Fig. 10 is consistent with direct formation from MO at low conversions but also from i-MO or to a lesser extend from UOL at high contact times. Finally, the initial zero slope for C9 product formation reveals that aldol condensation compounds are formed in consecutive pathways. Figure 11 presents a simplified reaction network for the HTR of mesityl oxide with 2-propanol on MgO that is consistent with the catalytic results showed in Fig. 10.

Figure 1.10

HTR of of mesityl oxide with 2-propanol: Yields as a function of contact time (523 K, 100 kPa, N2/2P/MO=93.4/6.6/1.3).

Figure 1.10

HTR of of mesityl oxide with 2-propanol: Yields as a function of contact time (523 K, 100 kPa, N2/2P/MO=93.4/6.6/1.3).

Close modal
Figure 1.11

Scheme of the reaction pathways on MgO for the HTR of mesityl oxide (MO).

Figure 1.11

Scheme of the reaction pathways on MgO for the HTR of mesityl oxide (MO).

Close modal

With the aim of establishing the effect of the calcination temperature of MgO on the catalyst activity for gas-phase HTR reactions we performed the MO/2-propanol reaction on our MgO-x samples using a contact time of 15 g h/mol. The obtained UOL formation rates (rUOL) are shown in Table 2. It is observed that rUOL, expressed either in mass or area basis, slightly increased with the calcination temperature, similarly to the evolution of medium-strength basic sites shown in Table 1 (nMg-O values). This result suggests that the rate limiting step for the formation of UOL is promoted on Mg2+-O2− pair sites. In fact, as it is shown in Fig. 12 the Mg2+-O2− pair sites would promote formation of the six-atom cyclic intermediate needed in Meerwein-Ponndorf-Verley mechanism for preferentially transferring hydrogen from the 2-propanol donor molecule to the CO bond of mesityl oxide. Mesityl oxide adsorbs via the CO bond on a weak Lewis acid Mg2+ cation, whereas 2-propanol adsorbs non-dissociatively on a vicinal Mg2+-O2− pair, giving rise to the required bimolecular six-atom cyclic intermediate [77]. Then, hydride transfer occurs without participation of surface H fragments, selectively forming the unsaturated alcohol.

Figure 1.12

UOL formation by MPV mechanism.

Figure 1.12

UOL formation by MPV mechanism.

Close modal

In summary, our results above show that the density, nature and strength of MgO surface basic sites may be regulated by modifying the solid calcination temperature. But the effect of calcination temperature on the MgO catalytic properties depends on the basicity requirements for the rate-limiting step of the base-catalyzed reaction. For example, the activity for the liquid-phase synthesis ofpseudoionones by condensation of citral with acetone diminishes with MgO calcination temperature because this reaction is predominantly promoted on strongly basic O2− sites. In contrast, the synthesis of 4-methyl-3-penten-2ol by the gas-phase hydrogen transfer reduction of mesityl oxide with 2-propanol is improved by increasing the MgO calcination temperature because the reaction intermediate is formed on medium-strength Mg2+-O2− pair basic sites.

We performed DFT calculations to obtain additional information on the role played by the MgO active sites on the kinetics of base-catalyzed reactions. Specifically, we present molecular modeling studies on our MgO-x catalysts for the synthesis of monoglycerides (MG) from glycerol (Gly) by transesterification (glycerolysis) of methyl oleate (C18:1), an unsaturated fatty acid methyl ester (FAME) (Fig. 13).

Figure 1.13

Synthesis of monoglycerides by transesterification of FAME with glycerol.

Figure 1.13

Synthesis of monoglycerides by transesterification of FAME with glycerol.

Close modal

Monoglycerides present surfactant and emulsifying properties that help hydrophilic and lipophilic substances mix together. Therefore, they can be used in food, detergent, plasticizer, cosmetic and pharmaceutical formulations [78]. The commercial liquid-catalyzed synthesis route to produce MG involves strong mineral bases such as Ca(OH)2 and KOH; this process yields only 40–60% MG, the rest being diglycerides and triglycerides, and entails concerns related to corrosion and disposal of spent base materials. The use of solid catalysts for MG synthesis presents not only the known environmental and practical advantages but also provides the opportunity to increase the MG yield. However, industrial implementation of heterogeneously catalyzed processes for FAME glycerolysis able to efficiently replace the use of liquid bases is still a challenge. Previous works have discussed the different routes for MG synthesis by esterification of fatty acids or by transesterification of triglycerides or fatty acid methyl esters [7, 79]. The base-catalyzed MG synthesis from Gly using FAME instead of fatty acids or triglycerides has several advantages, e.g., FAME is less corrosive than FA, has lower hydrophobic character than triglycerides, and exhibits higher miscibility with glycerol; therefore, the process can be carried out at lower temperatures than TG transesterification. Furthermore, the reaction route from FAME yields MG with a definite acyl group composition (FAME are easier to separate by fractional distillation than fatty acids) whereas in TG glycerolysis the products contain the acyl group distribution of the oil or fat [80]. MgO-based catalysts such as Mg/MCM-41 and Mg-Al mixed oxides have been investigated for the MG synthesis from glycerolysis of FAME [79, 81]. In particular, we have studied the glycerolysis of methyl oleate on MgO-based catalysts [7, 49, 82] and reported the reaction conditions needed to implement this reaction in a four-phase reactor under kinetic control and to reach maximum MG yields. Here, we present the results obtained on our MgO-x samples to get insight into the base site strength requirements for glycerolysis reactions.

Figure 14 shows FAME and Gly conversions and yields on MgO-773 at 493 K and typically illustrates the time-on-stream behavior of the catalyst during the reaction. Results show that monoglycerides can be obtained in high yields (70%) in 8 h using MgO. Only MG was initially formed but as the reaction proceeded, a second transesterification took place forming diglycerides from MG and FAME. MG was obtained in higher selectivity than diglycerides under the present conditions, but glyceride selectivity can be controlled by changing the Gly/FAME ratio so that to modify the availability of both reactants in the reaction zone, as previously discussed [7].

Figure 1.14

FAME conversion and glyceride yields [MgO-773, Gly/FAME=2; T=493 K; Wcat/n0FAME=11 g/mol FAME].

Figure 1.14

FAME conversion and glyceride yields [MgO-773, Gly/FAME=2; T=493 K; Wcat/n0FAME=11 g/mol FAME].

Close modal

Similarly to Fig. 14 for MgO-773, we plotted the evolution of FAME and Gly conversions and yields for MgO-673 and MgO-873 samples (plots not shown here). We observed that the MgO activity decreased with calcination temperature; FAME conversion after 3 h of reaction was, in fact, 93%, 84% and 76% for samples MgO-673, MgO-773 and MgO-873, respectively (Table 2). The initial MG formation rates (rMG0) were determined from the slopes at t=0 of MG yield versus time curves and the results are included in Table 2. The rMG0 values of Table 2 were plotted in Fig. 15 as a function the strong base site density (nO values in Table 1). This result suggested that under initial conditions the rate-determining step for MG formation is essentially promoted on strong base sites, present in corners or edges of the non-uniform surface of MgO catalysts.

Figure 1.15

Initial monoglyceride conversion rate as a function of strong base site density. Reaction conditions as in Fig. 14.

Figure 1.15

Initial monoglyceride conversion rate as a function of strong base site density. Reaction conditions as in Fig. 14.

Close modal

To obtain more insight on the role played by the MgO surface sites in the reaction kinetics we carried out a molecular modeling of Gly and FAME adsorptions on MgO [83] using a cluster model that represents the MgO surface with four different adsorption sites as depicted in Fig. 16: the terrace site contains the Mg5c-O5c pairs (L=5) that model the MgO medium strength base sites; the edge and the O-apical corner sites represent the strongly basic O4c (L=4) and O3c (L=3) sites, respectively, and a Mg-apical corner (Mg3c; L=3) that models a Lewis acid site.

Figure 1.16

Clusters used for modeling the MgO (1 0 0) surface. (a) Perfect terrace site, Mg25O25(Mg-ECP)25; (b) defective edge site, Mg22O22(Mg-ECP)19; (c) defective O-apical corner site, Mg22O22(Mg-ECP)12; (d) defective Mg-apical corner site, Mg23O23(Mg-ECP)14.

Figure 1.16

Clusters used for modeling the MgO (1 0 0) surface. (a) Perfect terrace site, Mg25O25(Mg-ECP)25; (b) defective edge site, Mg22O22(Mg-ECP)19; (c) defective O-apical corner site, Mg22O22(Mg-ECP)12; (d) defective Mg-apical corner site, Mg23O23(Mg-ECP)14.

Close modal

DFT molecular orbital calculations were carried out using the gradient corrected Becke*s three parameters hybrid exchange functional in combination with the correlation functional of Lee, Yang and Parr (B3LYP) [84]. The terrace site at the MgO (100) surface was represented by the Mg25O25 (Mg-ECP)25 cluster consisting of two layers (first layer: Mg9O16; second layer: Mg16O9). For the topological defects at edges and corners similar modeling was used; a Mg22O22(Mg-ECP)19 cluster was used for modeling the edge topological defect of MgO due to the intersection of two [100] and [010] oriented planes; a Mg22O22(Mg-ECP)12 cluster was used for modeling the oxygen corner topological defect of MgO due to the intersection of three [100], [010] and [001] oriented planes whereas the a similarly generated Mg23O23(Mg-ECP)14 cluster was used for modeling a magnesium apical corner.

To take into account the Madelung field due to the rest of the extended surface, the cluster was embedded in an array of ±2 point charges. This embedding technique was used previously for the study of both bulk and surface properties giving results which are in good agreement with those obtained by periodic calculations [85, 86]. Moreover, the positive point charges at the interface were replaced by effective core potentials (ECP) corresponding to Mg2+ to account for the finite size of the cations and to avoid spurious charge polarization.

The O atoms of the MgO surface that interact directly with the glycerol or FAME molecule (all from the first layer) were described with the basis set 6-31+G(d) and Mg atoms with 6-31G(d). For the rest of the oxide atoms in the cluster the basis set 6-31G was used. The 6-31G (d,p) basis set was used for the molecular orbitals of Gly and FAME. The adsorption energy of Gly or FAME (Eads) was evaluated according to the following total energy difference: Eads=E(molecule-MgO cluster)−E(MgO cluster)−E(molecule); where “molecule” is either Gly or FAME. Negative values indicate exothermic adsorption.

On the other hand, the atomic net charges (q) were calculated following the natural bond orbital (NBO) scheme [87], which gives realistic values for the charge partitioning. For all the systems the total charge was zero. Also, Δq(atom) was defined for an atom of Gly or FAME as the atom charge difference between adsorbed and free molecule states.

All the calculations were performed using the Gaussian-03 program package.

First-principles density-functional calculations were performed for the free glycerol molecule and for the adsorption of glycerol on representative terrace, edge, and Mg- and O-corner sites of MgO. The DFT calculations for the optimized geometrical structure of the free Gly molecule resulted in the following intramolecular interatomic distances (d): C-C (d(C–C)≈1.53Å), C-H (d(C–H)≈1.10Å), C-O (d(C-O)≈1.42Å) and O-H (d(O-H)≈0.97Å). For the Gly adsorption on MgO, different initial geometries of the glycerol molecule were evaluated depending on the orientation of the hydroxyl groups toward the MgO surface. Results presented in Table 3 show the optimized geometrical structures obtained for Gly adsorption through one, two or three hydroxyl groups. Species are identified as nOH(m), where n and m represent, respectively, the amount and position of the OH groups participating in the surface species. Thus, m is 1 or 3 for primary hydroxyls and 2 for the secondary one. The Eads values, the intramolecular and the Gly-MgO distances (d) were calculated at equilibrium. The Gly/MgO distances are depicted in the scheme of Fig. 17 as d(O−Mgs) and d(H−Os), considering the closest Gly hydroxyl interacting with the MgO surface.

Table 1.3

Adsorption energies (Eads) and bond distances (d) for Gly adsorption on terrace, edge and O-apical corner sites of MgO (100).a

EntryClusterFinal structure nOH(m)nmEads (eV)d(H−Os) (Å)d(O−Mgs) (Å)d(O–H) (Å)
Terrace (L=5) 1OH(1) −0.65 1.594 3.228 1.008 
Terrace (L=5) 2OH(1,2) 1,2 −0.92 1.588 2.252 1.019 
Terrace (L=5) 3OH 1,2,3 −1.48 1.785 2.293 0.996 
Edge (L=4) 1OH(2) −1.63 1.050 1.988 1.478 
Edge (L=4) 2OH(1,2) 1,2 −1.85 1.035 2.082 1.494 
Edge (L=4) 3OH 1,2,3 −1.62 1.588 2.104 1.013 
O-corner (L=3) 1OH(1) −0.89 1.519 2.802 1.045 
O-corner (L=3) 2OH(1,2) 1,2 −1.55 1.524 2.127 1.037 
O-corner (L=3) 3OH 1,2,3 −0.85 0.985 2.187 1.676 
EntryClusterFinal structure nOH(m)nmEads (eV)d(H−Os) (Å)d(O−Mgs) (Å)d(O–H) (Å)
Terrace (L=5) 1OH(1) −0.65 1.594 3.228 1.008 
Terrace (L=5) 2OH(1,2) 1,2 −0.92 1.588 2.252 1.019 
Terrace (L=5) 3OH 1,2,3 −1.48 1.785 2.293 0.996 
Edge (L=4) 1OH(2) −1.63 1.050 1.988 1.478 
Edge (L=4) 2OH(1,2) 1,2 −1.85 1.035 2.082 1.494 
Edge (L=4) 3OH 1,2,3 −1.62 1.588 2.104 1.013 
O-corner (L=3) 1OH(1) −0.89 1.519 2.802 1.045 
O-corner (L=3) 2OH(1,2) 1,2 −1.55 1.524 2.127 1.037 
O-corner (L=3) 3OH 1,2,3 −0.85 0.985 2.187 1.676 
a

Cluster sites as in Fig. 16; n and m: number and position of OH groups interacting with the surface, respectively

Figure 1.17

Scheme representing the formation of surface glyceroxide species from glycerol and the FAME surface activation.

Figure 1.17

Scheme representing the formation of surface glyceroxide species from glycerol and the FAME surface activation.

Close modal

Table 3 presents the results obtained on a perfect terrace site of MgO (100) where Mg2+ and O2− ions are five-fold coordinated (L=5). Results show that the d(O−Mgs) bond distance diminishes with the number of hydroxyl groups (n) interacting with the surface, probably reflecting a higher electrostatic interaction between Gly and the MgO surface that evidences the hydrophilic properties of MgO. Shortening of d(O−Mgs) suggests the presence of a more stable surface species, as indicated by the larger Eads value.

On the other hand, the values of the intramolecular O-H distance, d(O–H), in Table 3 suggest that regardless of the adsorption species structure, the OH groups maintained their integrity which indicates that glycerol adsorbs non-dissociatively on surface terrace sites. The glycerolysis reaction requires the rupture of an O-H bond at the Gly molecule to proceed, as depicted in Fig. 17. Then, data in Table 3 showing that Gly is non-dissociatively adsorbed on MgO terrace sites strongly suggest that the Mg5c-O5c pairs do not promote the Gly/FAME reaction.

Table 3 also presents the results obtained for the glycerol adsorption on an edge site that models the low coordination (L=4) base sites of MgO. The Eads values obtained for the edge site were higher than those determined for the terrace site, thereby indicating a stronger interaction between Gly and the surface edge site. As a result of the stronger Gly-edge interaction, dissociative adsorption occurred for some of the postulated geometries such as 2OH(1,2)a (Fig. 18(b), entry 5 in Table 3), where the d(O–H) distance was longer than in free glycerol due to O-H bond breaking with formation of both, a new surface OH between the abstracted H and a cluster oxygen, Os, and a surface glyceroxide on a cluster cation, Mgs (Fig. 17). Thus, the d(O−Mgs) and d(H−Os) distances for species 2OH(1,2) considerably shortened compared to 1OH(2) and 3OH surface structures of Table 3 for which non-dissociative adsorption took place.

Figure 1.18

Optimized geometrical structures of species 2OH(1,2) of Gly adsorbed on a MgO (100) surface: (a) non-dissociative adsorption on terrace sites; (b) dissociative adsorption on defective edge sites; (c) non-dissociative adsorption on a defective O-apical corner site.

Figure 1.18

Optimized geometrical structures of species 2OH(1,2) of Gly adsorbed on a MgO (100) surface: (a) non-dissociative adsorption on terrace sites; (b) dissociative adsorption on defective edge sites; (c) non-dissociative adsorption on a defective O-apical corner site.

Close modal

Glycerol adsorption was studied also on an O-apical corner (Table 3). Comparison between the Gly adsorption through the primary OH, structure 1OH(1), on the terrace (entry 1) and on the O-corner (entry 7) sites, indicated that a more stable species was obtained on the O-corner with shortening of the d(O−Mgs) and d(H−Os) distances. However, it seems that for the other two structures of Table 3, the Gly molecule tended to rotate and interacted to a greater extent with the Mg5c-O5c sites of the cluster than with O3c at the cluster corner, probably due to a steric effect that hampered the molecule arrangement on the oxygen corner. In this regard, Fig. 18 illustrates the adsorption of structure 2OH(1,2) of Gly on the terrace, edge and O-corner sites of MgO (100). The increase of the surface oxygen unsaturation from the terrace to the edge site would cause the Gly molecule O-H bond dissociation forming much more stable species. However, a further oxygen unsaturation increase from the edge to the O-corner site, would give rise to a non-dissociated 2OH(1,2) species with an Eads value in between those of the terrace and edge sites. Consistently, the d(O–H) distances of adsorbed Gly structures on an O-apical corner were similar to those of free Gly, suggesting that no O-H bond breaking took place.

DFT calculations were also carried out for FAME adsorption on representative terrace, edge, and Mg- and O-corner sites of MgO. The FAME molecule used in the catalytic experiments was methyl oleate that contains eighteen carbon atoms and one unsaturation (C18:1). For modeling purposes, a shorter molecule containing just five carbon atoms in the acyl chain was used (C5:0). The optimized geometrical structures of free methyl oleate as well as of the FAME used in the calculations are shown in Fig. 19. In the free short C5:0 FAME molecule the calculated intramolecular interatomic distances (d) were: CO (d(CO)=1.212Å), C–O (d(C−OCH3)=1.355Å and d(O−CH3)=1.436Å), C–H (d(C–H)≈1.09Å) and C–C (d(C–H)≈ 1.53Å). All these distances were very similar to those calculated for the free methyl oleate (C18:1) molecule.

Figure 1.19

Optimized geometries for free FAME molecules. (a) FAME used in catalytic experiments, methyl oleate (C18:1); (b) FAME used in theoretical calculations (C5:0).

Figure 1.19

Optimized geometries for free FAME molecules. (a) FAME used in catalytic experiments, methyl oleate (C18:1); (b) FAME used in theoretical calculations (C5:0).

Close modal

Optimized geometrical structures for adsorption of a C5:0 FAME molecule on terrace, edge and corner sites of MgO (100) are shown in Fig. 20 and the results are given in Table 4. Adsorption energy values, intramolecular bond distances and FAME-surface distances to the closest surface atom were calculated at equilibrium. The FAME molecule is expected to adsorb via the CO bond on surface Mg Lewis acid sites as illustrated in Fig. 17. Regardless of the cluster geometry low Eads values for FAME adsorption on MgO were obtained (Table 4), thereby indicating that the FAME-surface interaction is weak. The Eads values in Table 4 were in fact similar to those reported for the adsorption of non-polar low interacting molecules such as methane and benzene on a terrace site of MgO [88].

Table 1.4

Adsorption energies (Eads), carbonyl oxygen charge difference (Δq(OC=O) and bond distances (d) for FAME adsorption on terrace, edge and Mg- and O-apical corner sites of MgO (100).a

EntryCluster siteEads (eV)Δq(OC=O(a.u.)d(OC=O−Mgs) (Å)d(CO) (Å)
Terrace (L=5) −0.05 −0.02 2.474 1.219 
Edge (L=4) −0.02 −0.04 2.298 1.224 
O-corner (L=3) −0.56 −0.07 2.189 1.227 
Mg-corner (L=3) −0.67 −0.07 2.148 1.223 
EntryCluster siteEads (eV)Δq(OC=O(a.u.)d(OC=O−Mgs) (Å)d(CO) (Å)
Terrace (L=5) −0.05 −0.02 2.474 1.219 
Edge (L=4) −0.02 −0.04 2.298 1.224 
O-corner (L=3) −0.56 −0.07 2.189 1.227 
Mg-corner (L=3) −0.67 −0.07 2.148 1.223 
a

Cluster sites as in Fig. 16 

Figure 1.20

Optimized geometrical structures of FAME (C5:0) adsorbed on different MgO surface sites. (a) Terrace site; (b) edge site; (c) O-apical corner site; (d) Mg-apical corner site.

Figure 1.20

Optimized geometrical structures of FAME (C5:0) adsorbed on different MgO surface sites. (a) Terrace site; (b) edge site; (c) O-apical corner site; (d) Mg-apical corner site.

Close modal

The CO bond distance, d(CO), in adsorbed FAME molecules on terrace, edge and corner sites increased as the coordination number of the ions in the cluster decreased from L=5 to L=3, suggesting a stronger FAME-surface interaction. Consistently, the O–Mgs bond distance, d(OC=O−Mgs), decreased when FAME was adsorbed on lower coordination surface ions. Nevertheless, in all the cases the d(CO) distances remained close to that of the free molecule (1.212Å) which shows that the integrity of the adsorbed FAME molecule is preserved. In line with these results, the q(OC=O) values in Table 4 indicated that the oxygen of the CO bond gained some negative charge as a consequence of the adsorption process. However, in all the cases low q(OC=O) values were obtained thereby suggesting that polarization does not proceed to a significant extent on any cluster geometry.

In summary, DFT calculations predict that the proton abstraction from the glycerol hydroxyl groups required in the glycerolysis reaction (Fig. 16) would preferentially occur on low coordination O2− (strong base O4c sites located on edges), in agreement with the catalytic results presented in Fig. 15 and Table 2. FAME adsorption on MgO is weak, even on low coordination Mg3c and O3c sites. Therefore, the Gly/FAME reaction would proceed through a mechanism in which the most relevant adsorption step is that of glycerol.

The density, nature and strength of surface basic sites on MgO obtained from Mg(OH)2 decomposition may be regulated by modifying the solid calcination temperature. Decomposition of Mg(OH)2 at 673 K generates hydroxylated MgO containing mainly strong O2− basic sites located in surface defects such as corners and edges of the crystalline solid surface. The increase of the calcination temperature removes OH groups and also surface solid defects creating more stable structures that contain a higher concentration of medium-strength Mg2+-O2− basic pair sites.

The effect of calcination temperature on the MgO activity and selectivity for a given base-catalyzed reaction depends on the basicity requirements for the rate-limiting step of the reaction mechanism. For example, the activity for the liquid-phase synthesis of monoglycerides by glycerolysis of methyl oleate as well as that ofpseudoionones by condensation of citral with acetone diminish with MgO calcination temperature because both reactions occur predominantly on strong basic O2− sites. In contrast, the gas-phase hydrogen transfer reduction of mesityl oxide with 2-propanol is improved by increasing the calcination temperature because formation of the six-atom cyclic intermediate needed in Meerwein-Ponndorf-Verley mechanism for transferring hydrogen from the 2-propanol donor molecule to the CO bond of mesityl oxide is promoted on medium-strength Mg2+-O2− pair sites.

First-principles density-functional calculations were carried out to obtain more insigth on the role played by the MgO active sites for the glycerolysis of methyl oleate (FAME). DFT calculations were performed for the adsorption of glycerol and FAME on representative terrace, edge, and Mg- and O-corner sites of MgO. In agreement with catalytic results, calculations predict that dissociative chemisorption of glycerol with O–H bond breaking occurs only on low coordination O2− surface sites (edge sites) whereas nondissociative adsorption takes place on medium-strength base sites such as those of terrace sites. The DFT calculations also suggest that FAME adsorption through its CO group on Mg2+ sites is much weaker than glycerol adsorption via an OH group on O2− centers. Therefore, the MgO surface would be mainly covered by glyceroxide anions that would react with weakly adsorbed FAME molecules.

The authors gratefully acknowledge the Universidad Nacional del Litoral (UNL), Consejo Nacional de Investigaciones Científicas y Técnicas (CONICET), and Agencia Nacional de Promoción Científica y Tecnológica (ANPCyT), Argentina, for the financial support of this work. They also thank S. Fuente, R. Ferullo and N. Castellani for their collaboration in DFT calculations and useful discussions.

1.
Tanabe
 
K.
Saito
 
K.
J. Catal.
1974
, vol. 
35
 pg. 
247
 
2.
Tsuji
 
H.
Hattori
 
H.
Catal. Today
2006
, vol. 
116
 pg. 
239
 
3.
Corma
 
A.
Iborra
 
S.
Primo
 
J.
Rey
 
F.
Appl. Catal.
1994
, vol. 
114
 pg. 
215
 
4.
Kabashima
 
H.
Tsuji
 
H.
Hattori
 
H.
Appl. Catal. A: Gen.
1997
, vol. 
165
 pg. 
319
 
5.
Bancquart
 
S.
Vanhove
 
C.
Pouilloux
 
Y.
Barrault
 
J.
Appl. Catal. A: Gen.
2001
, vol. 
218
 pg. 
1
 
6.
Dossin
 
T. F.
Reyniers
 
M. F.
Marin
 
G. B.
Appl. Catal. B: Environ.
2006
, vol. 
62
 pg. 
35
 
7.
Ferretti
 
C. A.
Olcese
 
R. N.
Apesteguía
 
C. R.
Di Cosimo
 
J. I.
Ind. Eng. Chem Res.
2009
, vol. 
48
 pg. 
10387
 
8.
Montero
 
J. M.
Brown
 
D. R.
Gai
 
P. L.
Lee
 
A. F.
Wilson
 
K.
Chem. Eng. J
2010
, vol. 
161
 pg. 
332
 
9.
Gorzawski
 
H.
Hoelderich
 
W. F.
J. Mol. Catal. A: Chem.
1999
, vol. 
144
 pg. 
181
 
10.
Tanabe
 
K.
Zhang
 
G.
Hattori
 
H.
Appl. Catal.
1989
, vol. 
48
 pg. 
63
 
11.
Kurokawa
 
H.
Kato
 
T.
Kuwabara
 
T.
Ueda
 
W.
Morikawa
 
Y.
Moro-Oka
 
Y.
Ikawa
 
T.
J. Catal.
1990
, vol. 
126
 pg. 
208
 
12.
Di Cosimo
 
J. I.
Díez
 
V. K.
Apesteguía
 
C. R.
Appl. Catal.
1996
, vol. 
137
 pg. 
149
 
13.
Díez
 
V. K.
Apesteguía
 
C. R.
Di Cosimo
 
J. I.
J. Catal.
2006
, vol. 
240
 pg. 
235
 
14.
Xi
 
Y.
Davis
 
R. J.
J. Molec. Catal. A: Chem.
2011
, vol. 
341
 pg. 
22
 
15.
Xu
 
M.
Ginés
 
M. J. L.
Hilmen
 
A-M.
Stephens
 
B. L.
Iglesia
 
E.
J. Catal.
1997
, vol. 
171
 pg. 
130
 
16.
Di Cosimo
 
J. I.
Apesteguía
 
C. R.
Ginés
 
M. J. L.
Iglesia
 
E.
J. Catal.
2000
, vol. 
190
 pg. 
261
 
17.
Birky
 
T. W.
Kozlowski
 
J. T.
Davis
 
R. J.
J. Catal.
2013
, vol. 
298
 pg. 
130
 
18.
Ramos
 
J. J.
Díez
 
V. K.
Ferretti
 
C. A.
Torresi
 
P. A.
Apesteguía
 
C. R.
Di Cosimo
 
J. I.
Catal. Today
2011
, vol. 
172
 pg. 
41
 
19.
Zhu
 
K.
Hu
 
J.
Kubel
 
C.
Richards
 
R.
Angew. Chem. Int. Ed.
2006
, vol. 
45
 pg. 
7277
 
20.
Kim
 
H. W
Shim
 
S. H.
Lee
 
J. W.
J. Nanosci. Nanotechnol.
2007
, vol. 
7
 pg. 
4434
 
21.
Utamapanya
 
S.
Klabunde
 
K. J.
Schlup
 
J. R.
Chem. Mater.
1991
, vol. 
3
 pg. 
175
 
22.
Gulková
 
D.
Solcová
 
O.
Zdrazil
 
M.
Microporous Mesoporous Mater.
2004
, vol. 
76
 pg. 
137
 
23.
Ranjit
 
K. T.
Klabunde
 
K. J.
Chem. Mater.
2005
, vol. 
17
 pg. 
65
 
24.
Chen
 
D.
Jordan
 
E. H.
Mater. Lett.
2009
, vol. 
63
 pg. 
783
 
25.
Hattori
 
H.
Appl. Catal. A
2001
, vol. 
222
 pg. 
247
 
26.
Choudary
 
B. M.
Ranganath
 
K. V. S.
Pal
 
U.
Kantam
 
M. L.
Sreedhar
 
B.
J. Am. Chem. Soc.
2005
, vol. 
127
 pg. 
13167
 
27.
Menezes
 
A. O.
Silva
 
P. S.
Hernández
 
E. P.
Borges
 
L. E.
Fraga
 
M. A.
Langmuir
2010
, vol. 
26
 pg. 
3382
 
28.
Hattori
 
H.
Chem. Rev.
1995
, vol. 
95
 pg. 
537
 
29.
Kassner
 
P.
Baerns
 
M.
Appl. Catal. A
1996
, vol. 
139
 pg. 
107
 
30.
León
 
M.
Díaz
 
E.
Bennici
 
S.
Vega
 
A.
Ordóñez
 
S.
Auroux
 
A.
Ind. Eng. Chem. Res.
2010
, vol. 
49
 pg. 
3663
 
31.
Busca
 
G.
Chem. Rev.
2010
, vol. 
110
 pg. 
2217
 
32.
Fung
 
J. L.
Wang
 
I. K.
Appl. Catal. A
1998
, vol. 
166
 pg. 
327
 
33.
Szabó
 
Z. G.
Jóvér
 
B.
Ochmacht
 
R.
J. Catal.
1975
, vol. 
39
 pg. 
225
 
34.
Kurokawa
 
H.
Kato
 
T.
Kuwabara
 
T.
Ueda
 
W.
Morikawa
 
Y.
Moro-Oka
 
Y.
Ikawa
 
T.
J. Catal.
1990
, vol. 
126
 pg. 
208
 
35.
Díez
 
V. K.
Apesteguía
 
C. R.
Di Cosimo
 
J. I.
J. Catal.
2003
, vol. 
215
 pg. 
220
 
36.
Auroux
 
A.
Artizzu
 
P.
Ferino
 
I.
Monaci
 
R.
Rombi
 
E.
Solinas
 
V.
Petrini
 
G.
J. Chem. Soc., Faraday Trans.
1996
, vol. 
92
 pg. 
2619
 
37.
Kús
 
S.
Otremba
 
M.
Tórz
 
A.
Taniewski
 
M.
Appl. Catal. A: Gen.
2002
, vol. 
230
 pg. 
263
 
38.
Lauron-Pernot
 
H.
Luck
 
F.
Popa
 
J. M.
Appl. Catal.
1991
, vol. 
78
 pg. 
213
 
39.
Huang
 
M.
Kaliaguine
 
S.
Catal. Lett.
1993
, vol. 
23
 pg. 
373
 
40.
Thomasson
 
P.
Tyagi
 
O. S.
Knözinger
 
H.
Appl. Catal. A: Gen.
1999
, vol. 
181
 pg. 
181
 
41.
Gervasini
 
A.
Fenyvesin
 
J.
Auroux
 
A.
Catal. Lett.
1997
, vol. 
43
 pg. 
219
 
42.
Bordawekar
 
S. V.
Doskocil
 
E. J.
Davis
 
R.
Catal. Lett.
1997
, vol. 
44
 pg. 
193
 
43.
Díez
 
V. K.
Apesteguía
 
C. R.
Di Cosimo
 
J. I.
Catal. Today
2000
, vol. 
63
 pg. 
53
 
44.
Aramendía
 
M. A.
Borau
 
V.
Jiménez
 
C.
A. Marinas, J. M. Marinas, J. R. Ruiz and F. J. Urbano
J. Molec. Catal. A: Chem.
2004
, vol. 
218
 pg. 
81
 
45.
Arrouvel
 
C.
Digne
 
M.
Breysse
 
M.
Toulhoat
 
H.
Raybaud
 
P.
J. Catal,.
2004
, vol. 
222
 pg. 
152
 
46.
Chizallet
 
C.
Costentin
 
G.
Che
 
M.
Delbecq
 
F.
Sautet
 
P.
J. Phys. Chem. B
2006
, vol. 
110
 pg. 
15878
 
47.
Petitjean
 
H.
Tarasov
 
K.
Delbecq
 
F.
Sautet
 
P.
Krafft
 
J. M.
Bazin
 
P.
Paganini
 
M. C.
Giamello
 
E.
Che
 
M.
Lauron-Pernot
 
H.
Costentin
 
G.
J. Phys. Chem. C
2010
, vol. 
114
 pg. 
3008
 
48.
Cornu
 
D.
Guesmi
 
H.
Krafft
 
J. M.
Lauron-Pernot
 
H.
J. Phys. Chem. C
2012
, vol. 
116
 pg. 
6645
 
49.
Ferretti
 
C. A.
Soldano
 
A.
Apesteguía
 
C. R.
Di Cosimo
 
J. I.
Chem. Eng. J.
2010
, vol. 
161
 pg. 
346
 
50.
Bailly
 
M. L.
Chizallet
 
C.
Costentin
 
G.
Krafft
 
J. M.
Lauron-Pernot
 
H.
Che
 
M.
J. Catal.
2005
, vol. 
235
 pg. 
413
 
51.
Vidruk
 
R.
Landaua
 
M. V.
Herskowitz
 
M.
Talianker
 
M.
Frage
 
N.
Ezersky
 
V.
Froumin
 
N.
J. Catal.
2009
, vol. 
263
 pg. 
196
 
52.
Díez
 
V. K
Ferretti
 
C. A.
Torresi
 
P. A.
Apesteguía
 
C. R.
Di Cosimo
 
J. I.
Catal. Today
2011
, vol. 
173
 pg. 
21
 
53.
Di Cosimo
 
J. I.
Díez
 
V. K.
Xu
 
M.
Iglesia
 
E.
Apesteguía
 
C. R.
J. Catal.
1998
, vol. 
178
 pg. 
499
 
54.
Morterra
 
C.
Ghiotti
 
G.
Boccuzzi
 
F.
Coluccia
 
S.
J. Catal.
1978
, vol. 
51
 pg. 
299
 
55.
Philipp
 
R.
Fujimoto
 
K.
J. Phys. Chem.
1992
, vol. 
96
 pg. 
9035
 
56.
Prinetto
 
F.
Ghiotti
 
G.
Durand
 
R.
Tichit
 
D.
J. Phys. Chem. B
2000
, vol. 
104
 pg. 
11117
 
57.
León
 
M.
Díaz
 
E.
Bennici
 
S.
Vega
 
A.
Ordóñez
 
S.
Auroux
 
A.
Ind. Eng. Chem. Res.
2010
, vol. 
49
 pg. 
3663
 
58.
Knozinger
 
E.
Jacob
 
K.
Singh
 
S.
Hofmann
 
P.
Surf. Sci
1993
, vol. 
290
 pg. 
388
 
59.
Branda
 
M. M.
Ferullo
 
R. M.
Belelli
 
P. G.
Castellani
 
N. J.
Surf. Sci.
2003
, vol. 
527
 pg. 
89
 
60.
Tsuji
 
H.
Shishido
 
T.
Okamura
 
A.
Gao
 
Y.
Hattori
 
H.
Kita
 
H.
J. Chem. Soc. Faraday Trans.
1994
, vol. 
90
 pg. 
803
 
61.
Evans
 
J. V.
Whateley
 
T. L.
Trans. Faraday Soc.
1967
, vol. 
63
 pg. 
2769
 
62.
Díez
 
V. K.
Di Cosimo
 
J. I.
Apesteguía
 
C. R.
Appl. Catal. A: Gen.
2008
, vol. 
345
 pg. 
143
 
63.
Brenna
 
E.
Fuganti
 
C.
Serra
 
S.
Kraft
 
P.
Eur. J. Org. Chem.
2002
pg. 
967
 
64.
Rhodia Inc., P. Gradeff, US Patent, 3 840 601,
1974
65.
Union Camp Corporation, P. Mitchell, US Patent, 4 874 900,
1989
66.
Roelofs
 
J. C.
van Dillen
 
A. J.
de Jong
 
K. P.
Catal. Today
2000
, vol. 
60
 pg. 
297
 
67.
Abelló
 
S.
Dhir
 
S.
Colet
 
G.
Pérez-Ramírez
 
J.
Appl. Catal. A: Gen.
2007
, vol. 
325
 pg. 
121
 
68.
Noda Pérez
 
C.
Henriques
 
C. A.
Antunes
 
O. A. C.
Monteiro
 
J. L. F.
J. Mol. Catal. A: Chem.
2005
, vol. 
223
 pg. 
83
 
69.
Ponec
 
V.
Appl. Catal. A
1997
, vol. 
149
 pg. 
27
 
70.
Claus
 
P.
Appl. Catal. A: Gen.
2005
, vol. 
291
 pg. 
222
 
71.
Augustine
 
R. L.
Catal. Today
1997
, vol. 
37
 pg. 
419
 
72.
de Graauw
 
C. F.
Peters
 
J. A.
van Bekkum
 
H.
Huskens
 
J.
Synthesis
1994
pg. 
1007
 
73.
Musolino
 
M. G.
De Maio
 
P.
Donato
 
A.
Pierpaolo
 
R.
J. Mol. Catal. A: Chem.
2004
, vol. 
208
 pg. 
219
 
74.
Szollosi
 
G.
Bartok
 
M.
J. Mol. Catal. A: Chem.
1999
, vol. 
148
 pg. 
265
 
75.
Corma
 
A.
Domine
 
M. E.
Valencia
 
S.
J. Catal.
2003
, vol. 
215
 pg. 
294
 
76.
Di Cosimo
 
J. I.
Acosta
 
A.
Apesteguía
 
C. R.
J. Mol. Catal. A: Chem.
2004
, vol. 
222
 pg. 
87
 
77.
Di Cosimo
 
J. I.
Acosta
 
A.
Apesteguía
 
C. R.
J. Mol. Catal. A: Chem.
2005
, vol. 
234
 pg. 
111
 
78.
Zheng
 
Y.
Chen
 
X.
Shen
 
Y.
Chem. Rev.
2008
, vol. 
108
 pg. 
5253
 
79.
Corma
 
A.
Hamid
 
S. B. A.
Iborra
 
S.
Velty
 
A.
J. Catal.
2005
, vol. 
234
 pg. 
340
 
80.
Sonntag
 
N. O. V.
J. Amer. Oil Chem. Soc.
1982
, vol. 
59
 pg. 
795A
 
81.
Barrault
 
J.
Bancquart
 
S.
Pouilloux
 
Y.
C. R. Chim.
2004
, vol. 
767
 pg. 
593
 
82.
Ferretti
 
C. A.
Apesteguía
 
C. R.
Di Cosimo
 
J. I.
App. Catal. A: Gen.
2011
, vol. 
399
 pg. 
146
 
83.
Ferretti
 
C. A.
Fuente
 
S.
Ferullo
 
R.
Castellani
 
N.
Apesteguía
 
C. R.
Di Cosimo
 
J. I.
Appl. Catal. A: Gen.
2012
, vol. 
413–414
 pg. 
322
 
84.
Becke
 
A. D.
J. Chem. Phys.
1993
, vol. 
98
 pg. 
5648
 
85.
Calatayud
 
M.
Ruppert
 
A. M.
Weckhuysen
 
B. M.
Chem. Eur. J
2009
, vol. 
15
 pg. 
10864
 
86.
Branda
 
M. M.
Rodríguez
 
A. H.
Belelli
 
P. G.
Castellani
 
N. J.
Surf. Sc.
2009
, vol. 
603
 pg. 
1093
 
87.
Reed
 
A. E.
Curtiss
 
L. A.
Weinhold
 
F.
Chem. Rev.
1988
, vol. 
88
 pg. 
899
 
88.
Trevethan
 
T.
Shluger
 
A. L.
J. Phys. Chem. C
2007
, vol. 
111
 pg. 
15375
 
Close Modal

or Create an Account

Close Modal
Close Modal