5 Application of calcined Mg-Al hydrotalcites for Michael additions

5.1 Introduction

↓66

Hydrotalcite (HT)-like compounds, also known as layered double hydroxides (LDH) and anionic clays, are natural or synthetic crystalline materials consisting of positively charged two-dimensional sheets with water and exchangeable charge-compensating anions in the interlayer region. They may be represented by the general formula [M2+ 1-xM3+ x (OH)2]x+[An x/n·mH2O]x . M2+ are divalent anions (e.g., Mg2+, Zn2+, Mn2+, Ni2+, Co2+, Fe2+), M3+ are trivalent metal ions (e.g., Al3+, Cr3+, Fe3+, Co3+, Ga3+) and An- is the interlayer anion with charge n. The structure of hydrotalcite (HT) itself, Mg6Al2(OH)16CO3⋅4H2O, is similar to that of brucite, Mg(OH)2, in which Mg2+ is octahedrally coordinated by hydroxyl groups. These octahedra share adjacent edges to form sheets or layers. In HT, part of the Mg2+ ions is replaced by Al3+ ions, resulting in positively charged layers. The space between the stacked brucite-like cation layers is filled with charge compensating anions (e.g., CO3 2 , Cl, NO3 , SO4 2 , OH and many others) and water molecules. In naturally occurring HT, mostly carbonate is the interlayer anion. The synthesis, the textural and acid-base properties, and the catalytic application of hydrotalcites and hydrotalcite-related catalysts have been reviewed in [ 1, 2, 3, 4]. Hydrotalcites and hydrotalcite-related materials are used in different forms as follows:

As-synthesized hydrotalcite

As-synthesized hydrotalcite is used as flame retardant, ion exchanger, adsorbent for waste water in industry. As-synthesized hydrotalcite can also be used as catalyst and showed a high catalytic activity for decomposition of 2-methyl-3-butyn-2-ol into acetone and acetylene [5]. Halide exchange reactions between alkyl chloride and bromide/iodide are catalyzed by the catalysts having Cl, Br, or I in the interlayers [6]. Moreover, after anion exchange with inorganic heteropolyacids, hydrotalcite may also exhibit acid properties, which mainly are associated with the anions in interlayers [206].

↓67

Calcined hydrotalcite

The calcined hydrotalcites are normally referred to as mixed oxides and most widely used as catalysts. The most interesting and important properties of the mixed oxides obtained by calcination are the following: a) high surface area; b) basic properties; c) homogeneous mixtures of oxides with very small crystal size; d) memory effect, which allows the reconstruction of the original hydrotalcite-related structure under mild conditions [206]. In the field of heterogeneous catalysis (hydrogenation, reforming, basic catalysis and catalyst support), properties a), b) and c) have found various applications [206,208]. Recently, in base-catalyzed reactions, calcined hydrotalcites have attracted more attention. They are attractive alternatives to the use of dissolved alkali hydroxides or alkoxides.

Reconstructed hydrotalcite

↓68

When calcined hydrotalcites are rehydrated in water or in flowing nitrogen saturated with water, hydrotalcites structure are reconstructed. This is the “memory effect” of calcined hydrotalcites. The reconstructed material contains OH ions in the interlayers. It is also possible to introduce OH ions by direct ion exchange. The base strength of the reconstructed hydrotalcite containing OH ions is stronger than the original hydrotalcite containing CO3 2 ions. Figueras and co-workers found that the reconstructed materials were very useful catalysts for aldol condensation [123,7] and Michael addition [141]. The OH ions in the interlayers are believed to be active sites for these reactions.

The hydrotalcite-type structure and the formation process of the hydrotalcite-related materials are presented in Scheme 5.1 [207,209].

Scheme 5.1 The hydrotalcite-type structure (up) and formation process of the hydrotalcite-related materials (down)

↓69

Hydrotalcites are commercially available, cheap solid bases. As mentioned above, hydrotalcites as well as calcined hydrotalcites are highly active, selective catalysts and play an important role in many base-catalyzed reactions, e.g., Claisen-Schmidt condensations [131] and Knoevenagel condensations [ 8]. Their surface base sites were characterized by temperature programmed desorption (TPD) after adsorption of CO2 [ 9, 10], FTIR spectroscopy with CO2 [ 11, 12, 13], and gas-phase microcalorimetry (with CO2 and SO2) [216, 14,15,16]. In addition to the base sites, acid sites or acid-base pairs on these materials also influence catalytic performance. Acid-base sites on mixed oxides are highly active sites for many reactions: Meerwein-Ponndorf-Verley reactions [17], cycloadditions of carbon dioxide to epoxides [ 18], and aldol condensations to form 2-nonenal [19]. Acid-base properties of Mg-Al mixed oxides are governed by the Mg/Al molar ratio [215,216,217,218], calcination temperature [219], and preparation condition [214,218, 20, 21]. While the acid-base properties are important for understanding catalytic activity and selectivity, the relation between catalytic behavior in Michael additions, the acid-base properties, and the composition of hydrotalcite and calcined hydrotalcites is still poorly understood.

Thus, the aim of this chapter was to a) find efficient, selective catalyst that is easily acquired or prepared for further Michael additions and b) study the influence of its acid-base properties and chemical composition on the catalytic performance of the calcined hydrotalcites [22]. The Michael additions of 1,3-diones with different pK a values to methyl vinyl ketone were examined on calcined commercial Mg-Al hydrotalcites including an Al-rich (Mg/Al = 0.6) sample. MgO and Al2O3 were also involved for comparison. Acid-base properties of the catalysts were investigated by gaseous probe molecules with FTIR spectroscopy and microcalorimetry. Microcalorimetric measurements with benzoic acid were carried out for the study of catalyst basicity under similar liquid-phase reaction conditions.

5.2 Preparation and characterization of calcined Mg-Al hydrotalcites

5.2.1 Catalyst preparation

Hydrotalcites (PURAL MG 30, 50, 61, 70 from SASOL Germany GmbH) [ 23] were calcined at 550 °C in air for 3 h. Al2O3 was obtained by calcination of the AlO(OH) (Pural SB, SASOL Germany GmbH) [228] under the same conditions. Calcination of dried Mg(OH)2 at 600 °C in air for 4 h produced MgO. The hydroxide was precipitated from Mg(NO3)2 with KOH and dried at 110 °C (see chapter 3). In this chapter, the following sample codes are used: HT0.6 or CHT0.6 for the sample with a bulk Mg/Al molar ratio of 0.6; the C in the sample code indicates that the sample has been calcined (Table 5.1).

5.2.2 Characterization of calcined hydrotalcites

↓70

Physical properties

The physical properties of hydrotalcites (HT), calcined hydrotalcites (CHT), MgO and Al2O3 are given in Table 5.1. N2 adsorption experiments at 77 K show that the HT samples have low specific surface areas, S BET (13–20 m2/g) and high average pore diameter, d P (167–205 Å) with the exception of Al-rich HT0.6, which has a much higher specific surface area of 163 m2/g and pore diameter of 76 Å, which may be due to the existence of boehmit (Fig. 5.1A: d). Sample calcination increases S BET independent of composition; the CHT samples have significantly higher surface areas (>200 m2/g) with exception of CHT2.2 (114 m2/g). The average pore diameter (d P) are similar (45–52 Å) with the exception of 81 Å for CHT0.6 (Table 5.1). It is interesting to note that the surface areas of the CHT samples decreased dramatically when the fresh calcined samples were kept for several days even in closed containers prior to adsorption experiments.The reference compounds, MgO and Al2O3, have surface areas of 75 and 234 cm2/g and pore diameters of 258 and 45 Å, respectively. The pore volumes (V P) are the highest for these two samples with values of 0.96 (MgO) and 0.54 cm3/g (Al2O3).

Table 5.1Textural properties of hydrotalcites, calcined hydrotalcites, MgO and Al2O3

Sample

Mg/Al molar ratio

S BET

(m2/g)

V p a

(cm3/g)

d p b

(Å)

Cc

(wt%)

Weight

Lossd(wt%)

Theor.

Bulk

(ICP)

Surface

(XPS)

  

HT0.6

0.5

0.6

n.d.

163

0.32

76

1.42

31

HT1.4

1.25

1.4

n.d.

13

0.05

167

1.19

36

HT2.2

2.0

2.2

n.d.

15

0.05

128

2.42

42

HT3.0

3.0

3.0

n.d.

20

0.10

205

2.27

44

CHT0.6

0.5

0.6

0.8

257

0.52

81

-

-

CHT1.4

1.25

1.4

1.4

201

0.23

45

-

-

CHT2.2

2.0

2.2

2.0

114

0.13

46

-

-

CHT3.0

3.0

3.0

2.0

203

0.22

52

-

-

MgO

-

-

-

75

0.96

258

-

-

Al2O3

-

-

-

234

0.54

45

-

-

a: BJH desorption cumulative pore volume of pores between 17.0 and 3000.0 Å diameter
b: Average pore diameter by BET
c: Carbon content
d: Based on the weights before and after calcination in air at 550 °C

↓71

ICP-OES results show that the Mg/Al molar ratios of the corresponding HT and CHT samples are identical and, thus, independent of the samples’ thermal treatment. In comparison, the surface Mg/Al molar ratios (XPS measurements) deviate from the bulk ratios depending on the composition (Mg/Al molar ratio) of the CHT samples. The surface Mg/Al ratio for CHT0.6 of 0.8 is slightly higher than the bulk ratio (Mg-enriched surface, Alsurface < Albulk), whereas the identical surface Mg/Al molar ratios of CHT2.2 and CHT3.0 are lower in value (both 2.0) than that of the bulks (Al-enriched surfaces, Alsurface > Albulk). The bulk and surface Mg/Al ratios of CHT1.4 are identical (Alsurface = Albulk). If XPS is considered a pure surface method independent of surface area [214,215], the ICP and XPS results confirm the homogeneity of CHT1.4. This homogeneity suggests that smaller Al3+ cations are more easily incorporated into the main Mg oxygen network (Mg/Al > 1) than the larger Mg2+ cations into the main Al oxygen network (Mg/Al < 1: CHT0.6). The “nonincorporation” effect becomes more significant (Al-enriched surfaces) the higher the Mg/Al ratio (>> 1); the degree of Al enrichment of the CHT3.0 surface is higher than that of CHT2.2. Al-enriched surfaces were also found with XPS (Mg/Al = 1.5 and 2.4) for calcined Mg-Al hydrotalcites with Mg/Al molar ratios of 1.87 and 4.57, respectively [214].

XRD

The XRD patterns of the HT samples show that crystallinity of the HT samples increases with Mg content (Fig. 5.1A: a–d) and that the HT samples have a hydrotalcite-like structure similar to those presented in [225,226, 24,25]. Additional weak peaks at 28 and 49° in the HT0.6 pattern (Fig. 5.1A: d) can be assigned to boehmit (cf. Fig. 5.1A: e for Pural SB, AlO(OH), PDF No. 49-133) . Calcination of the HT samples (CHT3.0–1.4) results in the destruction of the hydrotalcite-like structure and the formation of a periclase MgO (PDF No. 45-946, Fig. 5.1B: a, b, c) also found for the pure MgO sample (Fig. 5.1B: f). The MgO peaks (at about 43 and 63°) of the CHT samples including CHT0.6 (Fig. 5.1B: a, b, c and d) have higher 2 values than those of pure MgO; this is caused by the incorporation of smaller Al3+ cations in the bulk lattice of all four CHT samples [14,18,29,30]. These peaks shift from higher to lower 2values as the Mg/Al ratio increases, which shows that the amount of Al3+ cations incorporated decreases accordingly. These results agree with the XPS results: CHT0.6 (Alsurface < Albulk) and CHT3.0 (Alsurface > Albulk). The AlO(OH) reference material forms -Al2O3 after calcination (PDF No. 29-63, Fig. 5.1B: e). In the pattern of the homogeneous CHT1.4 sample (Alsurface = Albulk, Fig. 5.1B: c), weak broad peaks for MgAl2O4 (35°, PDF No. 21-1152) and -Al2O3 (66°) are

↓72

Fig. 5.1 XRD patterns of hydrotalcites (A) and hydrotalcites calcined at 550 °C (B) with Mg/Al molar ratios of 3.0 (a), 2.2 (b), 1.4 (c), and 0.6 (d). The sample in A: d was further calcined at 700 °C to give the XRD pattern B: d1. The MgO reference material is shown after calcination in B: f. The Al2O3 reference oxide is shown before (boehmit in A: e) and after calcination (-Al2O3 in B: e).

observed. The intensity decrease of the MgAl2O4 and -Al2O3 peaks from CHT0.6 to CHT3.0 also indicates finely dispersed Al3+ cations in the bulk MgO network [219]. In the pattern of the very amorphous CHT0.6 sample, the broad peak at ca. 43° (Fig. 5.1B: d) may be attributed to regions of MgO, -Al2O3, and an additional spinel phase of MgAl2O4 in this sample [215]. Its maximum narrows, sharpens, and shifts to a slightly higher 2 value (45°) as a separate spinel phase forms during further calcination at 700 °C (Fig. 5.1B: d1); weak, broad peaks for MgO at ca. 43 and 63° (Fig. 5.1B: b) decrease in intensity after calcination at 700 °C, whereas the peaks for MgAl2O4 at 37, 45, 59, and 65° become stronger (Fig. 5.1B: d1). Thus, the increased calcination temperature of 700 °C causes a separate phase of MgAl2O4 to form [219, 26]. With higher Mg contents (higher Mg/Al ratios), MgO is the main phase (strong peak at ca. 63° in Fig. 5.1A: a), and the spinel peak disappears (2 = 37°).

27Al MAS NMR

↓73

The 27Al MAS NMR spectra of -Al2O3 and calcined hydrotalcites are shown in Fig. 5.2. Two broad signals at about 5 and 64 ppm are observed in the spectra of both pure Al2O3 and the CHT samples for octahedrally and tetrahedrally coordinated aluminum cations (AlO and AlT), respectively [214,217,219,229,231]. The identical spectra of -Al2O3 and CHT0.6 (highest Al content) regarding signal shapes, in particular that of the 64-ppm signal, indicate a similar Al coordination in these two samples and explains the weak, broad peaks in the XRD patterns (Fig. 5.1B: b and d). The maximum of the signal at 5 ppm shifts slightly to lower fields (from ca. 2 to 6.5 ppm) with increasing Mg/Al ratio, which indicates increased bonding of the AlO6 octahedra in the bulk to a less electronegative element (Mg): Mg–O–AlO [217,219]. A lack of signal splitting [219] suggests that distinct Al coordination sites do not exist in the samples. CHT1.4 has a relatively narrow signal (at 5 ppm) compared to other samples, which indicates a low degree of disorder of the AlO6 octahedra and may be related to the homogeneity of this sample confirmed by ICP-OES and XPS (Albulk = Alsurface).

The AlT/AlO ratio calculated from the integrated signals for Al2O3 is 0.38 (or 28% AlT) and is close to the value 0.4 of alumina given in [214]. The 28% AlT observed here is somewhat lower than that found in cubic aluminas (31–33%), but higher than the amount of AlT found in hexagonal alumina (15–25%) [27]. The samples, CHT0.6, CHT1.4, and CHT2.2, all have a AlT/AlO ratio identical to that of Al2O3 (0.38); however, CHT3.0 has a much higher ratio of 0.59 (37% AlT). This indicates that the octahedral coordination AlO is predominant in the structure of all the samples, yet tetrahedral AlT cations are more common in CHT3.0 possibly by forming inverse spinel-like domains: Mg–O–AlT [217]. AlO4 tetrahedra are generally formed during calcination [231,28] (pre-calcination: exclusively AlO6 octahedra); the preparation method [214] and Al content [214,217] can also influence the AlT/AlO ratio. The CHT samples were all prepared and calcined under the same conditions; thus, the higher AlT/AlO ratio in CHT3.0 (Alsurface >> Albulk) can only be the result of the strong difference in Al content between the bulk and surface not observed in the other samples. The influence of surface Al on the AlT/AlO ratio has been suggested before [214].

Fig. 5.2 27Al MAS NMR spectra of Al2O3 and calcined hydrotalcites

5.2.3 TG-DTA

↓74

The TG-DTA curves of HT0.6 and HT3.0 are shown in Fig. 5.3. In the TG-DTA curves of HT0.6 (Fig. 5.3a) and HT3.0 (Fig. 5.3b), gradual weight loss is observed from about 60 to approximately 600 °C with two main endothermic effects at about 250 and 416 (432) °C; a broad, weak endothermic peak is found at 90 °C for HT0.6. The first large endothermic effect at about 250 °C may be attributable to the loss of inter-layer water [225, 29]; the second endothermic effect of HT3.0 signifying the loss of OH groups and the decomposition of the CO3 2– anion in the brucite-like layers of hydrotalcites [225,234] occurs at 432 °C. The total weight loss observed for HT0.6 and HT3.0 (ca. 33.5 and 42.5%, respectively) is similar to that given in Table 5.1 measured pre- and post-calcination at 550 °C (31 and 44%, respectively).

Fig. 5.3 TG-DTA profiles of hydrotalcites with Mg/Al molar ratios of 0.6 (a) and 3.0 (b)

The monitored m/z peak at 18 (H2O+) for H2O or OH shows an immediate, continuous increase in intensity upon heating up to 416 °C with slight maxima at 247 and 416 °C (Fig. 5.3a) and 250 and 432 °C (Fig. 5.3b) for HT0.6 and HT3.0, respectively. The shape of these curves confirms the release of water and OH groups as the main cause of weight loss. Intensity increases of the mass peak, m/z = 44 (CO2 +), are only observed during the second main step of weight loss from 300 (HT0.6) or 400 (HT3.0) to 600 °C (both) with maxima at around 460 °C [231]. This increase is also gradual for HT0.6 at temperatures higher than 300 °C, but immediate and more defined for HT3.0 at 410 °C. The intensity differences of the m/z = 44 maxima are a direct result of the varying carbon contents (1.42 for HT0.6 < 2.27 for HT3.0,Table 5.1) and are less significant to the samples’ overall weight loss (1 mole CO3 2– vs. 16 moles OH and 4 moles crystal water). Due to the removal of water and dehydroxylation, as well as the decomposition of carbonate anions observed by the thermal analysis, surface areas and pore volumes increased after calcination (Table 5.1).

5.3 Catalytic behavior of calcined Mg-Al hydrotalcites

↓75

Michael addition of 2-methylcyclohexane-1 ,3 -dione to methyl vinyl ketone – comparison of hydrotalcites and calcined hydrotalcites

A comparative catalytic study of the HT and CHT samples is shown in Fig. 5.4A and 5.4B, respectively. Both types of samples (HT and CHT) are very active and produce the target product with 100% selectivity. In contrast, the calcined hydrotalcite studied in [141] exhibited no catalytic activity in 1,4-Michael additions.This could not be confirmed here. The catalytic activity of the sample was evaluated from the yield of the Michael adduct as a function of reaction time. After a reaction time of 24 h, the final catalytic yields decreased in the following order regardless of sample calcination: (C)HT0.6 > (C)HT3.0 > (C)HT1.4 > (C)HT2.2 (> Al2O3 > MgO≈ 0 in Fig. 5.4B). A linear correlation between Mg/Al molar ratio and catalytic activity was not found; samples with the highest and lowest Mg content, (C)HT0.6 and (C)HT3.0, respectively, achieved the highest yields. HT0.6 and CHT0.6 gave respective target product yields of 98 and 100% after 24 h; the use of HT3.0 and CHT3.0 resulted in yields of 82 and 94%, respectively. The sample with a Mg/Al ratio of 2.2 (HT2.2 and CHT2.2) was the least active of these samples. The catalytic yield a) remained more or less the same (Mg/Al = 1.4 and 0.6), b) decreased from 69 to 49% (Mg/Al = 2.2), or c) increased significantly from 82 to 94% (Mg/Al = 3.0) after calcination of the corresponding hydrotalcite at 550 °C. Due to weight loss during calcination, i.e., amount of catalyst (0.225 g) used was weighed out prior to calcination, CHT0.6 and CHT3.0 showed enhanced yields. This weight loss could also explain the drastic decrease in the catalytic activity of CHT2.2; however, HT2.2 also gave the lowest catalytic yield suggesting that other factors may be involved in the low yield of CHT2.2.

In the case of the pure oxides (Al2O3 and MgO), the catalytic activity of Al2O3 was lower than that of the CHT samples (final yield of 23% after 24 h, Fig. 5.4B), whereas MgO was found to have a considerably different, less selective behavior. In the case of MgO, 65% of the target product was formed within 30 min, but the amount of the Michael adduct in the reaction mixture dropped to 6% in a reaction time of 24 h. In the 1H NMR spectra of the reactionmixtures, additional CH3 signals (0.97 and 1.23 ppm) were found for 6-hydroxy-1,6-dimethyl-2,9-dioxobicyclo[3.3.1]nonane formed by the aldol cyclization of 2-methyl-2-(3-oxo-butyl)-cyclohexane-1,3-dione. In our experiments, when Mg(OH)2 or different MgO catalysts were used, cyclization under the same conditions also took place (see chapter 3).

↓76

Fig. 5.4 Results of the Michael addition of 2-methylcyclohexane-1,3-dione (15 mmol) to methyl vinyl ketone (22.5 mmol) in 10 mL methanol over the hydrotalcites (A) and over the calcined hydrotalcites, Al2O3 and MgO (B). Catalyst amounts were 0.225 g for the hydrotalcites, Al2O3, and MgO and the amount remaining after calcination of 0.225 g hydrotalcite at 550 °C

Michael addition of 2-acetylcyclopentanone and 2-acetylcyclohexanone to methyl vinyl ketone over calcined hydrotalcites

The Michael addition of 2-acetylcyclopentanone (pK a = 7.8) and 2-acetylcyclohexanone (pK a = 10.1) to methyl vinyl ketone was performed (Fig. 5.5A and 5.5B, respectively) to investigate possible further applications and compare the basic site strengths and numbers of CHT samples.

↓77

The catalytic results (Fig. 5.5) using these liquid diones can not be compared directly to those in Fig. 5.4B (2-methylcyclohexane-1,3-dione, a solid, is only slightly soluble in methanol at room temperature). A similar decreasing order of catalytic activity (CHT0.6 > CHT3.0 > CHT2.2 > CHT1.4 > Al2O3 > MgO) was found in these two Michael additions after 24 h. Note, CHT1.4 was the least active of the CHT samples for these reactions in comparison to the results of the reaction with the1,3-dione.

Fig. 5.5 Results of the Michael additions of 2-acetylcyclopentanone (15 mmol) (A) and 2-acetylcyclohexanone (15 mmol) (B) to methyl vinyl ketone (22.5 mmol) in 10 mL methanol over calcined hydrotalcites (0.225 g)

↓78

Fig. 5.6 Dependence of Michael adduct yield on Mg/Al ratio of catalyst after 24 h

The correlation between catalytic activity and the Mg/Al molar ratio was again nonlinear (Fig. 5.6) and identical for both reactions. Independent of the sample used (CHT, MgO, or Al2O3), the addition of 2-acetylcyclopentanone to methyl vinyl ketone proceeds faster than that of 2-acetylcyclohexanone (i.e., slope of curves is higher in Fig. 5.5A than in Fig. 5.5B up to a reaction time of 8 h). The higher slopes (Fig. 5.5A) and final yields reached after 24 h (Fig. 5.6) are both easily explained by the higher acidity of 2-acetylcyclopentanone (lower pK a, deprotonation ease). In the case of the less acidic 2-acetylcyclohexanone, CHT0.6 exhibited a much higher activity than that of the other catalysts (Fig. 5.5B and Fig. 5.6).It has been suggested that a decrease in Al contents results in a decrease of the total basic sites, but the fraction of stronger ones increases [213]. The high activity observed for CHT0.6 could also be caused by strong Lewis acid sites discussed below.It was also suggested that the catalytic performance of Mg y AlO x catalysts is related not only to the surface density of active sites but also to the nature of the bulk structure and transformations in the cation (Mg2 + or Al3 +) environment induced by a change in the chemical composition [217]. The higher initial yield after 1 h for MgO in the cyclohexanone addition reaction (cf. Fig. 5.5A and 5.5B) indicates a slowed (consecutive) reaction.

Fig. 5.7 Results of the solvent-free Michael additions of 2-acetylcyclopentanone (15 mmol) (A) and 2-acetylcyclohexanone (15 mmol) (B) to methyl vinyl ketone (22.5 mmol) on CHT0.6 and CHT3.0 (0.225 g)

↓79

Solvent-free Michael additions were carried out to examine the possibility of this type of environmentally friendly implementation of the reaction. This was checked with the most active samples, CHT0.6 and CHT3.0, in the Michael additions with 2-acetylcyclopentanone (Penta) and 2-acetylcyclohexanone (Hex) (Fig. 5.7). The reactions were catalyzed without the presence of a solvent, and CHT0.6 is again the more active catalyst; however, yields are decreased for both catalysts: from 99 (Penta) and 90% (Hex) to 55 (Penta) and 64% (Hex) for CHT0.6 and from 93 (Penta) and 40% (Hex) to 28 (Penta) and 18% (Hex) for CHT3.0 after 24 h. Moreover, in the case of CHT0.6, the solvent-free addition of 2-acetylcyclopentanone (pK a = 7.8) to methyl vinyl ketone proceeded slightly slower than that of 2-acetylcyclohexanone (pK a = 10.1). This result of CHT0.6 was unexpected since the pK a value of the former 1,3-dione is lower; the addition of former 1,3-dione should proceed faster. However, for CHT0.6, the reactions seem to be independent of the pK a values of the reactants that are normally determined in water or methanol. The similar result was also observed in base-catalyzed Knoevenagel condensation before [ 30] and could be explained by the role of the level of stabilization of the carbanionic intermediate throughout the particular interactions between reactant and the catalyst [235]. In contrast, the reaction with cyclopentanone proceeded slightly faster than the latter in the case of CHT3.0. Meanwhile, yields of the target products were lower than those found when a solvent was used for both catalysts after 24 h. The low (delayed) product formation may be related to the absence of solvation necessary to stabilize the carbanion generated from the CH-acid compound for the addition to methyl vinyl ketone. Nonetheless, the investigated heterogeneously catalyzed Michael-additions can be performed under solvent-free conditions with only slightly longer reaction times.

In summary, although the reaction rates and product yields of solvent-free Michael additions were, respectively, slower and lower than those observed in methanol, the solvent-free results are promising regarding their environmental impact.

5.4 Acid-base properties of calcined hydrotalcites

To understand the catalytic behavior of the catalysts, the acid-base properties were characterized by FTIR studies after pyridine and CO2 adsorption. Temperature programmed desorption (TPD) of the adsorbed probe molecules was used to study the strength of the acid and base sites. Microcalorimetric measurements were also involved in the investigation.

5.4.1 FTIR study of pyridine adsorption

↓80

The spectra measured after pyridine adsorption at 40 °C on Al2O3, MgO and CHT samples are shown in Fig. 5.8. The spectra of all the samples show bands for stretching vibrations of pyridine molecules coordinatively bonded to Lewis acid sites between 1630 and 1600 cm–18a mode) and between 1450 and 1440 cm–119b mode). Protonation of pyridine on Brønsted acid sites (bands at 1640 and 1540 cm-1) [31, 32] was not observed on any of the samples studied here.

Fig. 5.8 Pyridine adsorption on Al2O3 (a), calcined hydrotalcites with Mg/Al molar ratios of 0.6 (b), 1.4 (c), 2.2 (d) and 3.0 (e) and MgO (f). Spectra were measured after adsorption and evacuation at 40 °C

The negligible Lewis acidity of MgO (Fig. 5.8f) indicated by weak bands at 1600 and 1440 cm–1 corresponds to coordinatively unsaturated Mg2+ cations [237]. The shift of the ν8a band distinguishes the strength of the Lewis acid sites, whereas the intensity of the ν19b band characterizes their number [33]. With increasing aluminum content, the ν8a band shifts to higher wave numbers from 1601 (CHT3.0, Fig. 5.8e) to 1613 cm–1 Al2O3 (Fig. 5.8a). In general, bands at 1618–1612 and 1626–1620 cm–1 on Al2O3 are attributed to pyridine bonded to a pair of coordinatively unsaturated Al3+ ions in octahedral and tetrahedral positions and coordinatively unsaturated tetrahedral Al3+, respectively [34,35,36,37]. In the spectrum of Al2O3 (Fig. 5.8a), an additional band at 1622 cm–1 can be ascribed to pyridine strongly adsorbed on Al3+ ions in tetrahedral positions, AlT. A band similar in wave number, but of much lower intensity was observed for the CHT0.6 sample (Fig. 5.8b). These results indicate the presence of stronger Lewis acid sites (involving AlT) on the surface of Al2O3 and CHT0.6, but that CHT0.6 has a much lower number of these sites. The XPS results show that the Mg/Al ratio on the surface of CHT0.6 is slightly higher than in the bulk. The presence of MgO domains probably leads to the decrease in the number of strong acid sites on the surface of CHT0.6. Although 27Al MAS NMR showed that the bulk AlT/AlO ratio was identical for Al2O3 and CHT0.6–2.2 (0.38), no band was observed at 1622 cm–1 in the IR spectra of CHT1.4 and 2.2 after pyridine adsorption. The AlT/AlO ratio of CHT3.0 was higher (0.59, i.e., octahedral coordination predominates, but the number of AlT is higher for this sample than for other CHT samples), and the Mg/Al molar ratio was much lower on the surface of CHT3.0 than in the bulk. This would suggest that more Al tetrahedra are on the surface of CHT3.0 and can easily adsorb pyridine. Although this may be the case, it does not lead to the formation of strongly bonded pyridine adsorbate complexes (absence of band at 1622 cm–1 in Fig. 5.8e). The presence of the band in Fig. 5.8a and b may have a structural origin. When comparing the CHT patterns, that of CHT0.6 (Fig. 5.1B: d) is the only one with peaks for -Al2O3 (e.g., at 66°). The presence of the -Al2O3 structure in CHT0.6 could be of significance in the formation of strong acid sites. No indication of -Al2O3 is seen in the pattern of CHT3.0 (Fig. 5.1B: a). The influence of increasing Al contents and consequent structural changes on the catalytic activity of Mg-Al mixed oxides has been mentioned before [215,217].

↓81

Fig. 5.9 Temperature programmed desorption of pyridine. Effect of the evacuation temperature on the intensity of the band of adsorbed pyridine around 1440 cm–1

The total amount of absorbed pyridine molecules was estimated by the integration ofthe band (ν19b mode) around 1440 cm–1. The value calculated after evacuation at different temperatures is plotted against the evacuation temperature in Fig. 5.9. The total number of Lewis acid sites is highest for Al2O3 and CHT0.6 (practically identical band intensities at 40 °C) and decreases with Al content in the following order: CHT1.4 > CHT3.0 > CHT2.2. Few Lewis acid sites were detected on the surface of MgO, which agrees with weak bands found in Fig. 5.8f. As the evacuation (desorption) temperature was increased, the integrated intensity, i.e., number of pyridine molecules still adsorbed, decreased. At an evacuation temperature of 200 °C, most of the pyridine on CHT1.4–3.0 had desorbed; higher intensities (> 0.5 a.u.) are only observed for Al2O3 and CHT0.6. After evacuation at 300 °C, Al2O3 is the only sample with a band intensity over 0.5 a.u. These experiments show that while the CHT1.4–3.0 have varying amounts of acid sites (different band intensities after evacuation at 40 °C), they have a similar number of stronger acid sites (similar intensities after evacuation at temperatures higher than 150 °C). Thus, the amount of weaker acid sites increases from CHT2.2 to CHT3.0 to CHT1.4, which may explain the slightly lower activity of this sample (CHT1.4) in the Michael addition with 2-methylcyclohexane-1,3-dione in comparison to that of the other CHT samples.

In summary, CHT0.6 has the highest acidity among the CHT samples with respect to the number and strength of acid sites. The total amount of acid sites on the other CHT samples (CHT1.4–3.0) is much lower. The Al-rich sample (CHT0.6) and Al2O3 have the same total amount of acid sites (identical intensities at 40 °C), yet the stronger the sites are (the higher the evacuation temperature) the lower their numbers (amount of pyridine still adsorbed) for CHT0.6 compared to that for Al2O3. This may begin to explain the drastic difference in the activity of a) CHT0.6 and Al2O3 and b) CHT0.6 and CHT1.4–3.0.

5.4.2 FTIR study of CO2 adsorption

↓82

Because CO2 is acidic, it adsorbs specifically on basic sites. However, a lot of species can be formed. CO2reacts with basic hydroxyl groups forming hydrogen carbonate (bicarbonate) species (Scheme 5.1) and with basic oxygen ions forming different kinds of carbonate species (Scheme 5.2) [ 38, 39].

Scheme 5.1 Bicarbonate species

Scheme 5.2 Carbonate species

↓83

In the infrared spectra of bulk alkaline bicarbonates, which are present in dimer form in solid state, bands are observed at 2620-2450 cm-1OH…O), 1655-15 (asym νC=O), 1430-1370 (sym νC=O) and 1300 cm-1OH…O).

The free carbonate ion (D3h symmetry) has three IR active vibrations:

ν3 (E)

asymmetric νCO vibration:

1415 cm-1

ν2 (A2’)

out of plane π (CO3) deformation:

879 cm-1

ν4 (E)

in plane δ (CO3) deformation:

680 cm-1

and one Raman active vibration:

ν1 (A1’)

symmetric νCO vibration:

1063 cm-1.

↓84

In the adsorbed state, the symmetry is lowered (Scheme 5.2) and the doubly degenerate ν3 and ν4 vibrations are splitted and the ν1 vibration becomes IR active. The splitting of the ν3 mode is related to the structure of the carbonate (the angular distortion from D3h symmetry within the carbonate ligand) [244]:

Δν3 ca. 100 cm-1

monodentate

Δν3 ca. 300 cm-1

bidentate

Δν3 ca. 400 cm-1

bridged

However, much smaller values (50–150 cm-1) also have been observed for bridging carbonates. The thermal stability of surface carbonates should also be taken into account [5]. For instance, monodentate carbonates are usually less stable than the corresponding bidentate carbonates.

↓85

The FTIR spectra (1100–1800 cm–1) of Al2O3, MgO and CHT samples after CO2 adsorption at 40 °C are shown in Fig. 5.10. The thermal stability of the surface species was studied by desorption experiments at temperatures up to 300 °C (Fig. 5.11, 5.12 and 5.13). Analysis of the spectra allows us to distinguish two different types of surface basic sites on all catalysts, Brønsted base sites (basic OH groups) and Lewis base sites (basic O2– ions), and to estimate the respective amount and strength of the basic sites.

Fig. 5.10 CO2 adsorption on Al2O3 (a), calcined hydrotalcites with Mg/Al ratios of 0.6 (b), 1.4 (c), 2.2 (d) and 3.0 (e), and MgO (f). Spectra were measured after adsorption and evacuation at 40 °C

FTIR study of CO2 adsorption on Al2O3

↓86

The predominant species observed on Al2O3 (Fig.5.10a and Fig.5.11) is bicarbonate (HCO3 ) with bands at 3609 (νOH), 1647 (asym νC=O), 1476 (sym νC=O) and 1233 cm–1OH), which forms on surface basic OH groups. Bands at1756, 1704, 1537, 1265, and 1204 cm–1 (Fig. 5.11)and the one at 1450 cm–1 can be attributed to surface carbonate species, CO3 2–. The adsorption of CO2 on surface oxygen atoms can form different types of carbonates, monodentate, bidentate, bridged or polydentate, as mentioned above, depending on the environment of the oxygen atom [243,244]. One of the three IR active vibrations of the free carbonate ion, which possesses D 3 h symmetry, is the doubly degenerate ν3 (E) mode at 1415 cm–1. In the adsorbed state, the symmetry of the CO3 2– ion is lowered, and the ν3 vibration splits into two components (Δν3) [243,244] dependent on the structure of the surface carbonates. In the case of Al2O3, this splitting leads to bands at 1756 and 1204 cm–1 (Δν3 splitting = 552 cm–1), which can be assigned to the νC=O vibrations of bridged “organic-like” complexes; the shoulders at 1704 and 1265 cm–1 (Δν3 splitting = 439 cm–1) can be assigned to bidentate carbonate species. The assignment of the bands at 1537 cm–1 and 1450 cm–1 is less clear, which could indicate the presence of non-coordinated symmetrical carbonate or monodentate carbonate.

Fig. 5.11 TPD of CO2 adsorbed on Al2O3, spectra taken in vacuum at 40, 100, 150, 200 and 250 °C

FTIR study of CO2 adsorption on MgO

↓87

The spectrum of MgO (Fig. 5.10f and Fig. 5.12) after CO2 adsorption is complex with numerous bands between 1800 and 1100 cm–1. The major bicarbonate species is indicated by the bands at 1630, 1452 and 1217 cm–1. Weaker bands (shoulders) at 1600, 1448 and 1200 cm–1 (Fig. 5.12) could belong to another type of bicarbonate species. Pairs of bands at 1686/1365 and 1658/1304 cm–1 can be assigned to the ν3 vibrations of two different bidentate carbonate structures. Based on desorption experiments (spectra not shown),the most thermally stable species on MgO have band pairs at 1578/1444 and 1525/1381 cm–1.Based on their low Δν3 splitting (134 and 144 cm–1, respectively), these bands could be assigned to monodentate carbonate species [244, 40, 41]. However, since they are observed even at higher evacuation temperatures of 200 °C, i.e., high thermal stability of these corresponding species, they are probably caused by polydentate carbonates [243].

Fig. 5.12 TPD of CO2 adsorbed on MgO, spectra taken in vacuum at 40, 100, 150, 200, 250 and 300 °C.

FTIR study of CO2 adsorption on CHT samples

↓88

Fig. 5.13 TPD of CO2 adsorbed on CHT0.6 (left) and CHT3.0 (right), spectra taken in vacuum at 40, 100, 150, 200, 250 and 300 °C

Compared with MgO, a lower structural variety of the carbonates was formed on the surface of CHT samples (Fig. 5.10b–e and Fig. 5.13). Very broad bands are seen in two regions, 1750 to 1500 and 1500 to 1250 cm–1, with a single band between 1233 and 1217 cm–1; the general band shape in these regions closely resembles that found in the spectrum of MgO without the distinct maxima. The band for bicarbonates formed on the surface broadens and shifts to lower wave numbers (from 1233 cm–1 for CHT0.6 in Fig. 10b to 1220 cm–1 for CHT3.0 in Fig. 5.10e) as the Mg/Al ratio increases; its intensity also gives information about the number of OH species and compares the basicities of surface hydroxyl groups [243244]. Based on the Δν3 splittings between 188 and 226 cm–1, the maxima at 1591–74 and 1386–65 cm–1 could be assigned to monodentate surface carbonates. On the other hand, the thermal stability of these species is rather high. The bands are observed even after evacuation at 300 °C. Thus, as in the case of MgO, the formation of polydentate structures is likely, similar to that of bulk species [244].

The strength and the amount of base sites

↓89

According to Davydov et al. [246], the formation of different types of carbonates is related to the basicity of the surface oxygen atoms. It was postulated that the structural nonuniformity of oxygen ions on the catalyst surface leads to differences in the effective negative charge of the oxygen ions, i.e., their basicity, and that this causes the formation of different carbonate structures. The spectral parameter, Δν3, was proposed as a measure of the relative strength of surface basic sites [246]: the smaller the Δν3 value, the stronger the surface basic site is involved in the interaction with CO2.

Based on the ν3 splittings, the following conclusions can be drawn. Al2O3 exhibits the weakest Lewis basicity with Δν3 values of 552 and 439 cm–1 for the bridged and bidentate carbonate species, respectively. The strongest basicity was observed for MgO,which formed monodentate species with CO2 and exhibits the smallest Δν3 of 134 cm–1. The calcined hydrotalcites are less basic than MgO, but more basic than Al2O3. Here, the strength of basic sites (decrease in Δν3) of the CHT samples decreases with increasing magnesium content in the following order: CHT0.6 (188 cm–1) > CHT1.4 (209 cm–1) > CHT2.2 (225 cm–1) ≈ CHT3.0 (226 cm–1). The relative amount of surface, basic oxygen atoms (Lewis base sites) was estimated by the integration of the carbonate bands (1572, 1444, 1517 and 1388 cm–1 for MgO, and bands at 1574–1581 and 1368–1386 cm–1 for CHT samples)after evacuation at 150 °C, i.e., after removal of bicarbonate species (spectra not shown). When the amount of Lewis base sites is normalized by weight, intensities decreased for the samples in the following order: CHT3.0 ~ CHT1.4 ~ MgO > CHT2.2 > CHT0.6 >> Al2O3. CHT3.0 has the highest number of weakest Lewis base sites (higher splitting), whose strengths are similar to those found on CHT2.2; Lewis base sites on CHT0.6, the most catalytically active sample, are the strongest, but few in number These findings agree with results of Corma et al. who observed that the total amount of basic sites increases, but the proportion of stronger basic sites decreases with increasing Mg/Al molar ratio of calcined hydrotalcites [229].

Finally, the intensity of the OH deformation band (around 1230 cm–1) for surface bicarbonates after evacuation at 40 °C has been used to quantitatively determine the number of basic OH species [245]. The amount of basic OH groups (Brønsted base sites) on the surface of the calcined hydrotalcites does not correlate linearly with an increase in the Mg/Al ratio (Fig. 5.14). Interestingly enough, however, the decrease in the band intensities (amount of OH species), CHT0.6 > CHT3.0 > CHT1.4 > CHT2.2 (Fig. 5.14), does correlate with catalytic activities found for the CHT samples in the Michael addition of 2-methyl-cyclohexane-1,3-dione. This is not true of the Michael addition with the cyclopentanone and cyclohexanone, in which CHT1.4 was less active than CHT2.2. The shift of the 1230-cm–1 band to lower wave number indicates that the strengths increase with Mg content from CHT0.6 to MgO. Thus, CHT0.6 has a high number of Brønsted base sites; the Brønsted base sites on CHT3.0 are lower in number, but stronger. This may explain why both of these samples achieved the highest catalytic yields.

↓90

Fig. 5.14 Intensity of band around 1220 cm–1 after evacuation at40 °C

5.4.3 Microcalorimetric measurements

The base sites were further characterized by a series of microcalorimetric measurements using a gaseous probe molecule, CO2, and a liquid probe molecule, benzoic acid, to gain new insights into and supplementary information on the basic sites and differences in their behavior in gas atmosphere and in solution.

Fig. 5.15 Differential heat of CO2 adsorption at 40 °C as a function of CO2 uptake

↓91

The curves of the differential heat of CO2 adsorption as function of CO2 uptake on calcined hydrotalcites, Al2O3 and MgO at 40 °C are shown in Fig. 5.15. The curves for CHT samples are similar in shape; however, MgO and Al2O3 do not adsorb CO2 at high uptakes in comparison to the CHT samples. All of the initial heats are between 130 and 155 kJ/mol. The absence of a plateau of the differential heats at lower uptakes, i.e., the heat of adsorption decreases immediately and continuously with the amount of adsorbed CO2, indicates basic sites with different strengths exist on all the samples. Based on the differential heats in the entire range of CO2 uptake (Fig. 5.15), it seems that the strength of basic sites on CHT samples increase with Mg content. In contrast, the FTIR studies of CO2 adsorption at 150 °C showed that the strength of the Lewis basic sites was highest for MgO, increased with decreasing Mg content, and was the lowest for Al2O3: MgO > CHT0.6 > CHT 1.4 > CHT 2.2 ~ CHT3.0 > Al2O3 using the spectral splitting (Δν3). Since the microcalorimetric measurements were performed at 40 °C [close to room (reaction) temperature] instead of 150 °C (FTIR measurements), nonspecific adsorption of CO2 on basic sites of varying strength can not be excluded from the differential heats measured [42]. Therefore, the differential heat of adsorption of calcined hydrotalcites measured at 40 °C does not accurately determine the strength of the base sites.The measured differential heats may be average values for the sites that the molecules adsorb on; differences among these sites may not be detectable.

To further understand the specific catalytic sites, the amounts of totally and irreversibly adsorbed CO2 were determined from the adsorption and re-adsorption isotherms at 0.2 Torr; this pressure was chosen because the level of irreversible adsorption is practically constant above it. The quantitative results are shown in Table 5.2.

Table 5.2Total and irreversible adsorption of CO2 at p = 0.2 Torr at 40°C

Catalyst

V tot (μmol/g)

V irr (μmol/g)

V irr/V tot

CHT0.6

155.6

110.3

0.71

CHT1.4

189.4

154.6

0.82

CHT2.2

184.9

140.5

0.76

CHT3.0

201.4

166.5

0.83

Al2O3

70.2

44.0

0.63

MgO

97.0

75.1

0.77

↓92

The total amount of adsorbed CO2 including reversible and irreversible adsorption decreases in the following order: CHT3.0 > CHT 2.2 ~ CHT 1.4 > CHT 0.6 > MgO > Al2O3. The CO2 amount decreases with a decrease in Mg content except for MgO. Reversible adsorption involves weak interactions and is ascribed mainly to physisorption. The irreversibly adsorbed CO2 corresponds to the chemisorption of CO2; the amount of chemisorbed CO2 (V irr) decreases in the following way: CHT3.0 > CHT1.4 > CHT2.2 > CHT0.6 > MgO > Al2O3. The portion of CO2 chemisorbed on Al2O3 of 0.63 is slightly lower than that of the other samples: 0.71–0.83. MgO has a lower total CO2 uptake (97.0 μmol/g) and a lower amount of chemisorbed CO2 (75.1 μmol/g) than those of the calcined hydrotalcites. This may result from the formation of bulk carbonates on strong basic sites on MgO, which are then blocked and inaccessible to CO2. Species (CO2) with high thermal stability that exhibit a rather low Δν3 splitting (134 cm–1 for MgO is smallest observed here), can have a polydentate structure similar to that of bulk species. The intensity of the bands, which corresponds to the amount of Lewis basic sites, decreases similarly to V irr values observed: CHT3.0 ~ CHT1.4 ~ MgO > CHT2.2 > CHT0.6 > Al2O3 with exception of MgO (FTIR). This suggests the results of microcalorimetric gas-phase CO2 adsorption under our conditions do give some information about the amount of basic sites, but are not accurate in regard to strength.

Fig. 5.16 Differential heat (up) and integral heats (down) of adsorption of benzoic acid in toluene measured by microcalorimetry at 70 °C, except for that of MgO (at 40 °C)

Benzoic acid is used for the measurement of the strength and amount of basic sites by indicator titration [11]. Microcalorimetric measurements with benzoic acid in toluene were also performed here to measure basicity under liquid phase conditions identical to that of the catalytic test reactions studied here. The curves of the differential heat and integral heat of adsorption are shown in Fig. 5.16 for MgO and CHT0.6–2.2. The number of base sites (amount of benzoic acid uptake) differs from that found using CO2 and decreases in the following order: CHT0.6 > MgO > CHT3.0 > CHT2.2 ~ CHT1.4; in this case, not CHT3.0, but CHT0.6 has the most base sites and MgO has more sites using the benzoic acid probe molecule than the rest of the CHT samples.Based on the integral heats measured, the following order of decreasing strength was observed: CHT3.0 > CHT0.6 > MgO > CHT1.4 ≈ CHT2.2. The change in the order of decreasing amounts of acid sites, i.e., MgO after CHT0.6 has the second highest amount of base sites found using benzoic acid, could be a direct effect of the absence of site blocking by the probe molecule.

↓93

Since chemisorption of CO2 occurs on the catalytically active base sites, the amount of chemisorbed CO2 should correlate with the catalytic results. However, here, the gas phase microcalorimetric results do not agree with the activities found for the liquid phase Michael reactions. In contrast, the activities of CHT samples in the Michael additions directly relate to the amount of the base sites determined by liquid phase microcalorimetric measurements. This reflects the difference in behavior of base sites in the gas and liquid phase. Differences in the base site properties of MgO with the probe molecules, however, also suggest that the chemistry of the probe molecule may influence their number and strength. Interestingly enough, the base sites of CHT1.4 and 2.2 were similar in strength and number using the probe molecule, benzoic acid; these sites were also similar in number based on gas phase microcalorimetry experiments, but CHT1.4 chemisorbed more CO2 than CHT2.2. These gas phase result indicate some differences between CHT1.4 and 2.2 and may begin to explain the reactant-dependent catalytic behavior of these two samples.

5.5 Correlation of catalytic behavior and the acid-base properties

Information regarding the nature, strength and number of the acid-base sites obtained from the FTIR and microcalorimetric measurements is crucial to understanding the relationship between acid-base properties and catalytic activities. To clarify this relationship the results of these measurements and findings are discussed here by answering several questions.

Is the acidity responsible for the activity?

↓94

Michael additions can be catalyzed by not only base catalysts, but also acid catalysts. Acid-catalyzed Michael additions have been reported for such catalysts as acetic acid [159] (homogeneous catalyst) and metal complexes [146] or TS-1 molecular sieve [147] (both heterogeneous catalysts). However, among the oxide catalysts studied here, Al2O3, which has the strongest and highest amount of Lewis acid sites (IR, pyridine adsorption), as well as the weakest and lowest amount of base sites compared with CHT samples (IR, CO2 adsorption) was the least active in the Michael additions. Moreover, when HS-AlF3 [ 43], a very strong Lewis acid, was used as a catalyst for the Michael addition with 1,3-dione, no obvious activity was observed. These results suggest that the acid sites are not the main ones responsible for catalytic activity and that they cannot catalyze the Michael additions independently of base sites. Interestingly enough, Al2O3 does show a considerable activity (50% catalytic yield) in the Michael addition of 2-acetylcyclopentanone (pK a = 7.8) (Fig. 5.5A). The presence of weak base sites combined with mild acid sites on Al2O3 can catalyze the Michael addition to some extent. In other words, the important active sites determining the catalytic activity for Michael additions investigated here are base sites. However, the cooperation of the acid sites with the base sites can not be excluded. This is indicated by the very high activity of CHT0.6, the only sample to exhibit acid sites similar to those found in Al2O3, but much lower in number (IR, pyridine adsorption). These acid sites are much stronger than those in the other CHT samples.

Which base sites contribute to the Michael additions, Brønsted base or Lewis base sites or both?

For hydrotalcites, Brønsted base sites (OH groups) are the main base sites. However, Lewis and Brønsted base sites (O2– anions) are present on the surface of MgO and the calcined hydrotalcites (CHT samples). As demonstrated in the Michael additions of 1,3-diones, both kinds of sites catalyze the reaction. This result is very different from that of Choudary et al.[141], who observed that Brønsted base sites exclusively catalyzed the Michael additions. Moreover, it is interesting that the amount of Brønsted base sites (Fig. 5.14) on the CHT samples correlates with the catalytic yields; CHT0.6 and CHT3.0, which have the highest number of OH groups (Brønsted base sites, Fig. 5.14), achieved the highest catalytic yields. However, this does not prove that the Brønsted base sites are the only active sites for the reactions. Lewis base sites on the CHT samples predominate with respect to their amount and the strength in comparison to those of the Brønsted base sites; this suggest that OH groups enhance catalytic activity, but by no means exclusively cause it.

↓95

Does the base strength influence the selectivity?

MgO did not selectively produce the target Michael adducts. This could be attributed to the very strong Lewis basic character of MgO, which is demonstrated by FTIR; the reaction pathway of the resulting consecutive reaction is shown in chapter 3. Mg(OH)2 exhibits the same poor selectivity as MgO in the reaction with the 1,3-dione because of its expected strong Brønsted basicity, yet the reaction was much faster with Mg(OH)2 (4 h). This indicates that the reaction time needed decreases with an increasing number of Brønsted base sites. It seems clear that Michael additions investigated here are selectively catalyzed by moderate strength base sites, Brønsted base or Lewis base sites (Δν3), in methanol. The strength of the base sites that is lower than those on MgO prevents consecutive reactions and yields high selectivities.

What is more important: the strength or the amount of moderate basic sites?

↓96

And what is the possible mechanism of reaction?

For the Michael additions, the performance of the calcined hydrotalcites in the reactions depended on the Mg/Al molar ratios and the acidity of the Michael donors. The general activity of the catalysts has following order: CHT0.6 > CHT3.0 > CHT2.2 ~ CHT1.4. A direct (linear) correlation between the Mg/Al molar ratios and the catalytic behavior was not observed. However, based on the catalytic behaviors, the four CHT catalysts can be divided into two groups. One is the Al-rich sample CHT0.6, which showed the highest activity in all three reactions. High yields (90–99%) were obtained after 24 h for all three Michael additions. The second group includes the Mg-rich samples (Mg/Al > 1.0). Why CHT0.6 is the most efficient catalyst should be related to its special acid-base properties. Regarding strength, CHT0.6 (Δν3 = 188 cm–1) has the strongest Lewis base sites after those on MgO (Δν3 = 134 cm–1) and medium-strength Lewis acid sites comparable to strength of those on Al2O3. Regarding amount of base sites, CHT0.6 has the most Brønsted base sites and highest total number of base sites determined by IR and liquid phase microcalorimetric measurement, respectively. It is interesting to point out that the activity directly correlates with the amount of the base sites determined by the liquid phase microcalorimetric measurements. However, as mentioned above, the acid sites may participate in and enhance the catalytic process, because the difference in the amount of the Lewis base sites between CHT0.6 and the other three samples does not completely explain its outstanding behavior in the Michael addition of 1,3-dione, especially with 2-acetylcyclohexanone with a high pK a value.

Therefore, a plausible mechanism based on the cooperation of the acid sites with the basic sites (synergy) is presented in Scheme 5.3. Initially, the 1,3-dione is exclusively activated by deprotonation of the α-C atom between the two ketone groups by base site on the surface of the catalyst to produce a carbanion. At the same time, methyl vinyl ketone via its O atom is bonded to the surface of the adjacent Lewis acid site. Then, the carbanion attacks the carbon atom at β-position of methyl vinyl ketone to form the product anion. The product anion picks up a proton from the catalyst surface or from solvent to yield product. Basic and acidic sites located close to each other cooperate to yield high activities. A similar synergy on Mg-Al oxide has also been suggested by Yamaguchi et al. for the cycloaddition of carbon dioxide to epoxides [223]. Climent et al. also observed a cooperative effect between weak acid and base sites in aldol condensations on solid catalysts [44]. Indeed, if the appropriate strength of the active site is one of the keys in the catalytic cycle, an important aspect is the stabilization of the reaction intermediates [90]. When the methyl vinyl ketone is bonded to the adjacent Lewis acid sites, which exist on CHT samples and Al2O3, the intermediates are stabilized and the methyl vinyl ketone becomes more reactive. In fact, for CHT0.6, the acid sites are stronger and higher in number than those of the Mg-rich samples; these acid sites make the cooperative effect between base and acid sites more apparent and distinct in the Michael addition of 1,3-dione, especially with 2-acetylcyclohexanone with a high pK a value.

↓97

Scheme 5.3 Plausible mechanism of Michael addition of 1,3-dione to methyl vinyl ketone over acid-base pairs on calcined hydrotalcites

In the absence of moderate-strength acid sites for the first step of the reaction, although there is slight difference in the strength of base sites on CHT samples depending on the Mg/Al molar ratio, the strengths of the base sites of the Mg-rich CHT samples are all strong enough to activate the 1,3-dione. Thus, the amount of base sites, normalized to the same weight of the catalyst, may determine the rate of the reactions and the yields of Michael adduct in the same reaction time. In this case, the behavior CHT3.0 > CHT2.2 ~ CHT1.4 also directly correlates the results of the FTIR and microcalorimetric measurements; the amount of base sites decreases in the following order: CHT3.0 > CHT2.2 ~ CHT1.4. Both the intermediate stabilization by acid sites and the strength of the base sites decreases with increasing Mg/Al ratio, whereas the amount of basic sites increases with higher Mg/Al ratios; this “compensation” explains the V-like (first decrease then increase, Fig. 5.6) catalytic behavior of the CHT samples with increasing Mg/Al ratio.

In conclusion, the optimal balance of acid-base properties compared to that of Al2O3 and MgO makes the Al-rich calcined hydrotalcite an excellent catalyst in the Michael addition of numerous 1,3-diones independent of their pK a values.

5.6 Conclusions

↓98

Calcined hydrotalcites exhibit Lewis acid, Brønsted base, and Lewis base sites. The Al-rich sample (Mg/Al molar ratio of 0.6) possesses Lewis acid sites similar in strength to those found on Al2O3, but stronger than those found on the Mg-rich hydrotalcites.

Hydrotalcites and calcined hydrotalcites catalyze the Michael additions of 2-methylcyclohexane-1,3-dione, 2-acetylcyclopentanone, and 2-acetylcyclohexanone to methyl vinyl ketone with high selectivity (no side product formation). These heterogeneously catalyzed Michael additions in methanol proceeded faster than the solvent-free reactions.

Pure Al2O3 was the least active among the investigated catalysts. Both the Al-rich hydrotalcite and calcined hydrotalcite (Mg/Al molar ratio of 0.6) gave product yields above 95% within 24 h.

↓99

A nonlinear correlation between the Mg content and catalytic activity suggested the activity of pure MgO surpasses that of the calcined hydrotalcites, but causes consecutive reactions of the Michael addition products, which reduce the product selectivities and yields.

Catalytic activity correlates with the amount of the base sites determined by benzoic acid microcalorimetry dependent on the Mg/Al molar ratio.

An optimal balance of acid-base properties may make the Al-rich calcined hydrotalcite an excellent catalyst in the Michael addition of numerous 1,3-diones independent of their pK a values.


Footnotes and Endnotes

1 [] F. Cavani, F. Trifiro, A. Vaccari, Catal. Today 11 (1991) 173–301.

2 [] A. Vaccari, Catal. Today 41 (1998) 53–71.

3 [] B.F. Sels, D.E. De Vos, P.A. Jacobs, Catal. Rev. -Sci. Eng. 43 (2001) 443–488.

4 [] D. Tichit, B. Coq, Cattech 7 (2003) 206–217.

5 [] V.R.L. Constantino, T.J. Pinnavaia, Catal. Lett. 23 (1994) 361–367.

6 [] E. Suzuki, M. Okamoto, Y. Ono, J. Mol. Catal. 61 (1990) 283–294.

7 []K.K. Rao, M. Gravelle, J. Sanchez-Valente, F. Figueras, J. Catal. 173 (1998) 115–121.

8 [] A. Corma, V. Fornés, R.M. Martin-Aranda, F. Rey, J. Catal. 134 (1992) 58–65.

9 [] A.L. McKenzie, C.T. Fishel, R.J. Davis, J. Catal. 138 (1992) 547–561.

10 [] J.I. Di Cosimo, V.K. Díez, M. Xu, E. Iglesia, C.R. Apesteguía, J. Catal. 178 (1998) 499–510.

11 [] J. Shen, J.M. Kobe, Y. Chen, J.A. Dumesic, Langmuir 10 (1994) 3902–3908.

12 [] V.K. Díez, C.R. Apesteguía, J.I. Di Cosimo, J. Catal. 215 (2003) 220–233.

13 [] F. Prinetto, G. Ghiotti, R. Durand, D. Tichit, J. Phys. Chem. B 104 (2000) 11117–11126.

14 [] J. Shen, M. Tu, C. Hu, J. Solid State Chem. 137 (1998) 295–301.

15 [] S. Casenave, H. Martinez, C. Guimon, A. Auroux, V. Hulea, A. Cordoneanu, E. Dumitriu, Thermochim. Acta 379 (2001) 85–93.

16 [] S. Casenave, H. Martinez, C. Guimon, A. Auroux, V. Hulea, E. Dumitriu, J. Therm. Anal. Cal. 72 (2003) 191–198.

17 [] P.S. Kumbhar, J.S. Valente, J. Lopez, F. Figueras, Chem. Commun. (1998) 535–536.

18 [] K. Yamaguchi, K. Ebitani, T. Yoshida, H. Yoshida, K. Kaneda, J. Am. Chem. Soc. 121 (1999) 4526–4527.

19 [] D. Tichit, D. Lutic, B. Coq, R. Durand, R. Teissier, J. Catal. 219 (2003) 167–175.

20 [] D. Tichit, M.H. Lhouty, A. Guida, B.H. Chiche, F. Figueras, A. Auroux, D. Bartalini, E. Garrone, J. Catal. 151 (1995) 50–59.

21 [] M.J. Climent, A. Corma, S. Iborra, K. Epping, A. Velty, J. Catal. 225 (2004) 316–326.

22 [] H.A. Prescott, Z.-J. Li, E. Kemnitz, A. Trunschke, J. Deutsch, H. Lieske, A. Auroux, J. Catal. 234 (2005) 119–130.

23 [] www.sasol.com, product information.

24 [] A. Corma, V. Fornés, F. Rey, J. Catal. 148 (1994) 205–212.

25 [] G. Fornasari, M. Gazzano, D. Matteuzi, F. Trifiro, A. Vaccari, Appl. Clay Sci. 10 (1995) 69–82.

26 [] K.J.D. MacKenzie, R.H. Meinhold, B.L. Sherriff, Z. Xu, J. Mater. Chem. 3 (1993) 1263–1269.

27 [] R.H. Meinhold, R.C.T. Slade, R.H. Newman, Appl. Magn. Reson. 4 (1993) 121–140.

28 [] W.T. Reichle, S.Y. Kang, D.S. Everhardt, J. Catal. 101 (1986) 352–359.

29 [] S. Miyata, Clays Clay Miner. 28 (1980) 50–56.

30 [] M.J. Climent, A. Corma, V. Fornés, A. Frau, R. Guil-Lόpez, S. Iborra, J. Primo, J. Catal. 163 (1996) 392–398.

31 [] H. Knözinger, Adv. Catal. 25 (1976) 184–271.

32 [] G. Busca, Phys. Chem. Chem. Phys. 1 (1999) 723–736.

33 [] N. Fripiat, R. Conanec, A. Auroux, Y. Laurent, P. Grange, J. Catal. 167 (1997) 543–549.

34 [] C. Morterra, A. Chiorino, G. Ghiotti, E. Garrone, J. Chem. Soc. Faraday Trans. 1, 75 (1979) 271–288.

35 [] C. Morterra, S. Coluccia, A. Chiorino, F. Boccuzzi, J. Catal. 54 (1978) 348–364.

36 []F. Abbattista, S. Delmastro, G. Gozzelino, D. Mazza, M. Vallino, G. Busca, V. Lorenzelli, G. Ramis, J. Catal. 117 (1989) 42–51.

37 [] P. Nortier, P. Fourre, A.B. Mohammed Saad, O. Saur, J.C. Lavalley, Appl. Catal. 61 (1990) 141–160.

38 []G. Busca, V. Lorenzelli, Mater. Chem. 7 (1982) 89–126.

39 [] J.C. Lavalley, Cata. Today 27 (1996) 377–401.

40 [] J.A. Lercher, C. Colombier, H. Noller, J. Chem. Soc., Faraday Trans. 1, 80 (1984) 949–959.

41 []A.A. Davydov, M.L. Shepot’ko, A.A. Budneva, Kinet. Catal. 35 (1994) 299–306.

42 [] V. Solinas, I. Ferino, Catal. Today 41 (1998) 179–189.

43 [] E. Kemnitz, U. Groß, S. Rüdiger, C.S. Shekar, Angew. Chem. Int. Ed. 42 (2003) 4251–4254.

44 [] M.J. Climent, A. Corma, V. Fornés, R. Guil-Lopez, S. Iborra, Adv. Synth. Catal. 344 (2002) 1090–1096.



© Die inhaltliche Zusammenstellung und Aufmachung dieser Publikation sowie die elektronische Verarbeitung sind urheberrechtlich geschützt. Jede Verwertung, die nicht ausdrücklich vom Urheberrechtsgesetz zugelassen ist, bedarf der vorherigen Zustimmung. Das gilt insbesondere für die Vervielfältigung, die Bearbeitung und Einspeicherung und Verarbeitung in elektronische Systeme.
DiML DTD Version 4.0Zertifizierter Dokumentenserver
der Humboldt-Universität zu Berlin
HTML generated:
13.09.2005