الكيميائي المتالق
New Member
Introduction
Hydrogenation of α,β-unsaturated aldehydes into the corresponding unsaturated alcohols is highly desirable from an industrial point of view. On the other hand, the production of the saturated aldehydes from the unsaturated ones has some industrial and biological applications. It is also thermodynamically preferred and can be achieved easily compared to the former one. The selectivity of a hydrogenation catalyst towards the production of unsaturated alcohols rather than the saturated aldehydes may be improved by using additives, such as transition metal salts, promoters, different supports, different solvents, or alloying the base metal of the catalyst with another one. This modifies its electronic and chemisorption properties towards a better adsorption of C=O bond rather than C=C bond, while preserving the hydrogenation activity of the catalyst . For example, Pd catalysts are very active in many hydrogenation reactions, but are poorly selective towards the production of unsaturated alcohols from their unsaturated aldehyde counterparts . In contrast, Sn and Ir have a relatively low activity, but are highly selective in this respect
Unsaturated alcohols of the allyl-type are valuable fine chemicals, widely used in industry to produce perfumes, pharmaceuticals, and flavoring materials. Selective hydrogenation of unsaturated aldehydes with conjugated double bonds is an important route to unsaturated alcohols However, on conventional catalysts of group VIII metals, α, β-unsaturated aldehydes preferentially undergo hydrogenation of the C C bond to yield saturated aldehydes (as initial product) while the selectivity to allylic alcohols is very low. Hydrogenation of the C O bond is disfavored with respect to hydrogenation of the adjacent C C bond, both thermodynamically and kinetically, the selective hydrogenation of the C O bond represents a challenge which is somewhat facilitated for higher aldehydes R1R2C CH CH O, where bulky substituents Ri, interacting with the catalyst, may sterically hinder the activation of the C C bond , The lowest α, β-unsaturated aldehydes – crotonaldehyde (butenal, R1 CH3, R2 H) and especially acrolein (propenal, R1 R2 H) – do not exhibit such steric hindrance and the transformation to unsaturated alcohols is even more difficult. To enable the desired hydrogenation pathway in these cases, the catalyst is assumed to provide a geometry in which the C O bond of the adsorbed aldehyde reactant is notably activated compared to the C C bond There is experimental evidence that enhancing the surface coverage can also facilitate the formation of such desired structures, when the reactant–catalyst interaction is sufficiently strong, as on Pt
1-Reactant structure
Reactant structure can affect the maximum selectivities for desired products obtained over heterogeneous catalysts. The branching in the vicinity of ethylenic double bond favors high selectivities for unsaturated alcohols, which means that the selectivities for unsaturated alcohols decrease in the following reactants: cinnamaldehyde>3methylcrotonaldehyde > crotonaldehyde > acrolein
There are several reasons, which might affect the product selectivity with different substrates, like electronic and inductive effects in the reactant, adsorption mode of the reactant and the geometrical restrictions of the metal surface. Adsorption mode of the reactant can change with changing the initial concentration of the reactant. Moreover the metal surface has geometrical limitations for reactant adsorption, i.e. metal particle size and shape, as well as amount of steps, edges and kinks is essential for high selectivity. The metal surface properties are discussed in Section.
The selectivities to unsaturated alcohols starting from different substrates over conventional monometallic catalysts are reported. Besides unsaturated aldehydes also data on aromatic aldehydes are presented. Additionally, the maximum selectivities to the desired products over heterogeneous catalysts are given. Citral (1) hydrogenation has been investigated in the liquid phase over different catalysts. In citral hydrogenation the desired products can be unsaturated alcohols, nerol (2) and geraniol (3) or citronellol (4) as well as unsaturated aldehyde citronellal (6). Over Ru/C the selectivities to nerol and geraniol are quite low (, but with Pt/SiO2 the selectivity is enhanced to 76%. Recently very high selectivities to nerol and geraniol have been reported by Milone et al, Reyes et al. and Malathi et al. over Au/Fe2O3 Ce/AC Pt-Sn/MgO, Ir/TiO2, Pt/TiO2 In the two latter catalysts, the role of reducible support is important The bimetallic catalysts, like Ir-Ge/TiO2–HTR, were both active and selective. The higher activity of the bimetallic catalyst was attributed to the presence of ionic species of Ge Very high selectivity to nerol and geraniol, 95.2% at 95.6% conversion was obtained over Ru-Fe/C catalyst in citral hydrogenation in methanol under 40 bar hydrogen . The reaction mixture contained, however, a basic additive, triethyl amine, which can increase the chemoselectivity. Similarly bimetallic Pt-Sn/MgO catalyst was active and selective for producing nerol and geraniol in a continuous trickle bed reactor One drawback with this catalyst was the long induction period of 20 h, during which the conversion and selectivity increased. The third desired product from citral hydrogenation, an unsaturated alcohol citronellol can be achieved over Ni/Al2O3 catalyst with the selectivity of 95% in ethanol at 70 °C. The catalyst pretreatment is an important factor for achieving high selectivities to citronellol over this catalyst in alcohol solvents. Citronellal, a partially unsaturated aldehyde, which is used as an intermediate for production of fragrances, can be synthesized over supported Pd catalysts, like Pd supported on polyethylene fibers containing grafted 4-vinylpyridine groups, and Pd/SiO2/AlPO4. In general, Pd as the main metal is very selective for hydrogenating aliphatic unsaturated aldehydes, like citral to citronellal with selectivity of 86% and minor formation of alcohols. Pd/C catalyst was used in citral hydrogenation to citronellal in methanol and in the presence of triethylamine as additive the selectivity to citronellal was 94% at 99.5% conversion.
2_
Type of catalyst
Heterogeneous catalysts used in hydrogenation reactions contain usually a metal supported on a carrier. The metal is able to adsorb hydrogen thus making hydrogenation reaction possible. The carrier is able to disperse the metal to smaller particles as compared to bulk metal enhancing the specific metal surface area. Additionally, smaller metal particles partially behave as non-metals, e.g. have higher electron densities, leading to higher hydrogenation rates compared to larger particles. In addition to electronic effects, which are different for different metals, geometric properties of the metal particles can affect the hydrogenation rate and selectivity if the size of reacting molecule is close to the size of metal crystallites. The electronic properties of the active metal can be modified by a promoter. In this paper the promoter is defined very broadly, i.e. a promoter can be a non-inert support, second metal, oxide or other additive, which can change the geometric and the electronic structure of the metal surface. The metal phase composition and the amount of edges and corners vary with different metal particle sizes. Activity and selectivity depend not only the metal type and dispersion but also on the support. The catalyst carriers should be able to disperse and stabilize the active metal. Support materials can exhibit inert, basic or acidic properties. Furthermore, non-inert carriers can form an alloy with an active metal. In this section, metal selection and the recent catalytic results by using monometallic and bimetallic catalysts from different groups are presented. A separate section is devoted to catalysts containing both one and several metals on reducible supports. Furthermore, different methods for characterization of bimetallic catalysts are shortly discussed. The most important catalyst properties, like metal particle size and shape, metal precursor, support selection, catalyst pretreatment and the effect of modifiers and promoters and their influence on the chemoselective hydrogenation of carbonyl compounds are reviewed below.
2.1. Metal selection
Different geometric and electronic properties of metals can affect the hydrogenation activities and selectivities by influencing not only surface reactions, but also adsorption, the latter being a prerequisite of any heterogeneous catalytic reactions. The adsorption phenomenon itself is limited by many factors, e.g. by the mode of reactant adsorption on a specific metal surface. Different crystal phases of metals have different geometrical
Pt(1 1 1) surface is a flat, dense surface, while number of edges and kinks is increasing with decreasing metal particle size. Transition metals, which are used in catalysts, have partially filled or filled d-orbital. When the d-orbital is filled, the metal is in relative scale catalytically inactive, like in case of silver, gold and copper. The interstitial electron density in metals having filled d-orbital is low and they are not eager to form bonds. However, the electronic structure of the metal can be changed by adding a second metal, leading for instance to alloy formation, by changing the metal particle size or by enhancing interactions with the support.
Generally there has been a correlation between the width of the d bands of metal catalysts and selectivities in the hydrogenation of α,β-unsaturated aldehydes. The d bandwidth increases in the following order Pd < Pt < Ir Os. The repulsive interaction of the metal with the C C bonds is stronger with the metals exhibiting larger d bandwidth. Thus, the chemisorption of C C bond becomes less probable. The hydrogenation activities and selectivities are presented for different groups in the periodic table by starting from group 8. Note that the selectivities to the desired products over metal supported catalysts depend not only on the metal, but also on the reactant structure. In some hydrogenation reactions, like in the hydrogenation of citral (8), the monometallic supported catalysts are very selective, whereas selective hydrogenation of acrolein and crotonaldehyde cannot be achieved over conventional supported monometallic catalysts. Additionally, it could be pointed out that large selectivity changes have been observed in, e.g. citral hydrogenation over a monometallic Pt/SiO2 catalyst just by changing the reaction temperature
One of the most studied reactions of hydrogenation with gold catalysts is the selective hydrogenation of α,β-unsaturated aldehydes Indeed, the selectivity in unsaturated alcohols, the desired products, is difficult to achieve (i) because thermodynamics favors the hydrogenation of the C=C over the C=O bond by about 35 kJ mol−1 and (ii) because for kinetic reasons, the reactivity of the C=C bond is higher than that of the C=O bond. Hence, on conventional supported hydrogenation catalysts (Pt, Ru, Ni, etc.), the main products are the saturated aldehydes. On supported gold catalysts, interesting selectivities to unsaturated alcohol were found, even if the results obtained from one research group to another are not always fully consistent. The selective hydrogenation of crotonaldehyde (but-2-enal) was first studied by Bailie and Hutchings at atmospheric pressure and 250 °C, with Au/ZnO and Au/ZrO2 prepared by coprecipitation, and with Au/SiO2 prepared by impregnation. They found that Au/ZnO was more selective for C=O bond hydrogenation and that the addition of thiophene acted as a promoter for selective hydrogenation. Haruta and colleagues studied the same reaction but at high pressure (1.0 MPa) and 227 °C, with Au/Al2O3 and Au/TiO2 catalysts prepared by deposition-precipitation with NaOH and gold particles sizes of 4.6 and 3.4 nm, respectively, and Au/SiO2 (5.3 nm) prepared by grafting of gold acetyl acetonate. They also found that the selectivity depends on the nature of the support. The highest selectivity to crotyl alcohol was obtained over Au/TiO2 (≈25% instead of <10% for the other samples). Bailie et al. studied the influence of the particle size on the selectivity in the Au/ZnO system. Gold particle sizes were increased by increasing the gold loading in the samples (2 wt% (2–4 nm), 5 wt% (4–5 nm), and 10 wt% (bimodal distribution: 1–3 and 6–10 nm)), or by increasing the reduction temperature of the 5 wt% Au/ZnO catalyst (250 °C (4–6 nm), 300 °C (3–8 nm), 350 °C (10–15 nm), and 400 °C (up to 20 nm)). They reported that the selectivity to crotyl alcohol increased when the gold loading or the reduction temperature increased, and they proposed that the active sites for selective hydrogenation of crotonaldehyde to crotyl alcohol were associated with the presence of large Au particles. Claus et al. also studied the effect of the particle size in a series of three Au/TiO2 catalysts (1.1, 2.0, and 5.3 nm) prepared by three different preparation methods in selective hydrogenation of acrolein under high pressure (2 MPa) and at 240 °C. They observed a drastic decrease in the TOF by two to three orders of magnitude and a decrease of selectivity when the size of gold particles decreased from 2.0 to 1.1 nm, and constant values of TOF and selectivity for average particle sizes of 2.0 and 5.3 nm. They proposed that the structure sensitivity of the reaction in the case of small gold particles was due to quantum size effects, i.e., to alteration of the metallic properties of the smallest gold particles. In a more recent study, Claus and colleagues prepared three Au/ZrO2 catalysts by the same method of deposition-precipitation, using various pH's, to obtain catalysts with various gold particle sizes (4.0, 6.9, and 7.7 nm). They found that the TOF decreased by nearly one order of magnitude, and the selectivity to allyl alcohol increased (from 15 to 35%) when the particle sizes increased. On the other hand, they observed by HRTEM, multiply twinned particles whose number increased with increasing reduction temperature (300 to 500 °C) or reduction time (3 to 18 h). They pointed out the fact that when the number of multiply twinned particles increased, both the selectivity and the TOF decreased. They attributed the higher activity of Au/TiO2 compared to that of Au/ZrO2 (TOF×2) to the fact that the gold particles were round shaped on TiO2 and facetted on ZrO2 (average particle size of 5 nm on both supports). Round-shaped gold particles on TiO2 would contain a higher proportion of low-coordinated surface sites, which would explain the higher TOF.
As shown above, the influence of the gold particle size and of the nature of the support on the activity and selectivity for the hydrogenation of α,β-unsaturated aldehydes is not clear, yet. To attempt to clarify the influence of the first parameter, we decided to investigate the dependence of the catalytic properties with gold particle size in Au/TiO2 catalysts prepared by a single method, the deposition-precipitation with urea The main advantage of this method over the method of deposition-precipitation with NaOH developed by Haruta and colleagues] is that all the gold of the solution is deposited on TiO2, so higher gold loading can be reached with the same gold particle size as with Haruta's method, and the gold loading of the catalysts can be easily controlled. The results obtained are therefore compared with those obtained on Au/TiO2 catalysts prepared by deposition-precipitation with NaOH. The reaction studied is the hydrogenation of crotonaldehyde (CH3---CH=CH---CH=O) at atmospheric pressure, under the same experimental conditions as those used in former studies for Pt/TiO2 catalysts Hence, comparisons of the catalytic behavior between platinum and gold catalysts could be made, especially the influence of the particle size and that of the reduction temperature, since these two parameters are the key parameters which govern the activities and selectivities of Pt/TiO2 catalysts.
2.1.1. Monometallic catalysts
In the conventional monometallic catalysts the metals are selected from the group 10, e.g. Ni, Pd and Pt. In addition, Rh from group 9 and Ru from group 8 have been used in catalysts. Very high selectivities to unsaturated alcohols (83–100%) at high conversions have also been reported over a pyridine poisoned Re catalyst which belongs to group 7. The supports are usually alumina, silica and carbon. The catalytic activity of different metal supported catalysts in hydrogenation is determined by ability to activate C C and C O bonds as well as the activity of hydrogen to react on the metal surface. Hydrogen activity on different metals is relatively well understood on the basis of adsorption studies and fundamental theories. According to Masel, hydrogen is reactive in the surfaces of Co, Ni, Ru, Rh, Pd, Os, Ir, Pt as well as on Sc, Ti, V, Y, Zr, Nb, Mo, La, Hf, Ta, W, Cr, Mn, Fe, Tc and Re. A slower uptake of hydrogen was observed with Cu, whereas Ag and Au were inert. Interestingly to note that ability of hydrogen to react and to dissociate on the surfaces of Au and Ag depends on the size of metal nanoparticles. Exceptional catalytic activities of gold nanoparticles in hydrogenation and oxidation reactions were recently reported Moreover the filling of d-orbital affects very much the adsorption mode. The two latter metals, Au and Ag have completely filled d orbitals, 5d10 and 4d10, and thus they exhibit very low catalytic activities. No adsorption can occur on the metal having a low interstitial electron density Generally there has been a correlation between the width of the d bands of metal catalysts and selectivities in the hydrogenation of α,β-unsaturated aldehydes. The d bandwidth increases in the following order Pd < Pt < Ir Os. The repulsive interaction of the metal with the C C bonds is stronger with the metals exhibiting larger d bandwidth. Thus, the chemisorption of C C bond becomes less probable. The hydrogenation activities and selectivities are presented for different groups in the periodic table by starting from group 8. Note that the selectivities to the desired products over metal supported catalysts depend not only on the metal, but also on the reactant structure. In some hydrogenation reactions, like in the hydrogenation of citral (8), the monometallic supported catalysts are very selective, whereas selective hydrogenation of acrolein and crotonaldehyde cannot be achieved over conventional supported monometallic catalysts. Additionally, it could be pointed out that large selectivity changes have been observed in, e.g. citral hydrogenation over a monometallic Pt/SiO2 catalyst just by changing the reaction temperature
2.1.1.1. Chemoselective hydrogenation over group 8 metals, Ru, Os
The group 8 metals are electropositive, because they have unfilled d orbitals. Supported ruthenium catalysts have been recently used in chemoselective hydrogenations, like in hydrogenation of cinnamaldehyde, crotonaldehyde and citral
Os has been used in very few papers as a catalytic metal supported on SiO2 and on TiO2. In the hydrogenation of oxopromegestone Os was the most selective among tested metals, but the catalytic activity was low. Catalytic activity of Os in hydrogenation reactions is, however, well documented for a long period of time. For instance, F. Haber applied Os catalysts, in ammonia synthesis already a century ago. In citral hydrogenation at 27 °C under 1 bar, hydrogen in hexane the selectivity to nerol and geraniol was close to 100% at 5% conversion. The initial TOF-values were, however, very low over Os/SiO2 catalyst
The performance of Ru/C and Ru/Al2O3 catalysts was compared in citral hydrogenation They exhibited the same TOF in gas phase hydrogenation of crotonaldehyde, but alumina supported Ru catalyst was more selective for crotyl alcohol than Ru/C. The selectivities were however relatively low, i.e. 17 and 3%, respectively. At the same time these two catalysts afforded in citral hydrogenation at the conversion level of 70% selectivities to nerol and geraniol 54 and 38%, respectively The selectivities in citral hydrogenation at 60 °C are additionally lowered by simultaneous side reactions, like cyclisation and acetalization of citronellal, when the solvent was 2-propanol. The Ru/C catalyst exhibited three times higher activity than Ru/Al2O3 calculated per gram of Ru Ru/C was tested also in citral hydrogenation in the work of Mäki-Arvela et al.. This catalyst gave about 22% selectivity to nerol and geraniol in citral hydrogenation in 2-pentanol at 70 °C at 75% conversion. The lower selectivity in compared to might be explained by the origin of the metal precursor, in the former work the precursor was chloride whereas in the latter one it was nitrate. In particular, chloride precursor is known to favor acetalization, which in principle could be avoided in 2-pentanol. However, in hydrophobic 2-pentanol quite large amounts of isopulegol (12%) are formed. High selectivities for nerol and geraniol over Ru/Al2O3 and Ru/TiO2, 47.6 and 73.4% at 100% conversion, respectively, were obtained in citral hydrogenation in n-heptane at 126 °C and 50 barThe selectivity to nerol and geraniol was 73.4% over Ru/TiO2 catalyst reduced at high temperature, when the hydrogenation was carried out at 126 °C and at 50 bar in n-heptane as a solvent. In the selectivity to nerol and geraniol was 58% at 5% conversion of citral in citral hydrogenation at 27 °C under 1 bar hydrogen in hexane. The initial hydrogenation rate, however, over this catalyst was quite low. Relatively low selectivities were obtained over Ru supported catalysts. Only 13% selectivity of cinnamyl alcohol at 20% conversion was obtained over Ru/SiO2 catalyst in 2-propanol at 60 °C and 10 bar, whereas Ru/Al2O3 catalyst gave higher selectivity (about 40%) to cinnamyl alcohol. Cinnamyl alcohol selectivity over Ru/CNF catalyst (CNF stands for carbon nanofiber) was 48% at 60% conversion of cinnamaldehyde at 110 °C and 45 bar in 2-propanol
2.1.1.2. Chemoselective hydrogenation over group 9 metals, Co, Rh, Ir
Monometallic catalysts containing a metal from group 9 have been used in several chemoselective hydrogenations. Applications of Rh and Co and Ir and were reported. Monometallic catalysts including a metal from group 9 can act quite selectively in some cases, like in citral, acetophenone and cinnamaldehyde hydrogenations. The achieved selectivities depend, however, very much on the reactant structure. Ir catalysts have been very selective to unsaturated alcohols, but rather inactive.
Monometallic Rh catalysts have been relatively selective in acetophenone and in cinnamaldehyde hydrogenations, but low selectivities were obtained in crotonaldehyde, citral and cyclohexenone hydrogenations. Acetophenone hydrogenation over Rh/C catalyst resulted in about 70% selectivity to unsaturated alcohol at 25 bar and 80 °C in cyclohexane. Similarly, 70% selectivity to cinnamyl alcohol was achieved over Rh/SiO2 catalyst at 50 °C and 1 bar in ethanol In case of crotonaldehyde the selectivities to crotyl alcohol were low over Rh/SiO2 and over Co/SiO2, i.e. Rh/SiO2 catalyst yielded relatively low selectivities to crotyl alcohol (18% over Rh/SiO2 catalyst prepared by sol–gel method from Rh(acac)) at 100 °C, and 2% over 0.5 wt.% Rh/SiO2 at 35 °C (Rh from RhCl3 prepared by impregnation). The initial selectivity to crotyl alcohol was, however, close to 100% at the conversion of 0.2% in crotonaldehyde hydrogenation at 40 °C over a calcined Rh/Al2O3 catalyst indicating that catalyst pretreatment is important The selectivity to crotyl alcohol decreased linearly with increasing time-on-stream. In citral hydrogenation over Rh/Al2O3 the main product was citronellol (60% selectivity) in 2-pentanol at 70 °C and 1 bar and below 1% nerol and geraniol were obtained Polymer supported Rh catalyst gave as a major product citronellal No unsaturated alcohols were also formed over Rh/SiO2 catalyst in the hydrogenation of 2-cyclohexenone at 25 °C and at 6.8 bar in 2-propanol. The reason for this is the presence of sterically hindered carbonyl bond. In over Rh/SiO2 citronellal was the main product in citral hydrogenation at 27 °C under atmospheric pressure in hexane. Polymer supported Rh catalyst gave as a major product citronellal.
As the d bandwidth of Ir is larger than for Pd or Pt, it could be expected that the hydrogenation of ethylenic bond decrease. In fact Ir has been very selective in citral hydrogenation to nerol and geraniol, but the drawback has been in the low reaction rate Ir/SiO2 was very selective in the hydrogenation of citral at 70 °C and 4.1 bar in ethanol affording 100% selectivity of nerol and geraniol. On the other hand only 7% selectivity to nerol and geraniol was observed in citral hydrogenation at 27 °C and 1 bar hydrogen in hexane at 23% conversion It might be that the solvent also has an effect on selectivityOn Ir/TiO2 selectivity to nerol was also 100% but over both catalysts the total conversion was very low. Ir/graphite catalyst was very selective to cinnamyl alcohol in the hydrogenation of cinnamaldehyde at higher temperatures, i.e. 110 °C and 130 °C in the pressure range of 10–30 bar in 2-propanol Additionally, the selectivity to cinnamyl alcohol increased with increasing conversion.
Co/SiO2 resulted in maximally 20% selectivity of crotyl alcohol at 120 °C and 1 bar in gas phase. In general, however, it can be stated that Co being not a noble metal, is more selective to unsaturated alcohols than Rh, Ru or Pt. The selectivity to unsaturated alcohols can be tuned by changing the surface structure of Co, which exists in four different forms Relatively high selectivities to nerol and geraniol (55%) were obtained at 5% conversion over Co/SiO2 in citral hydrogenation under 1 bar hydrogen at 27 °C in hexane.
2.1.1.3. Chemoselective hydrogenation over group 10 metals, Ni, Pd, Pt
Platinum has been the most intensively used as the active metal in chemoselective hydrogenations, , whereas Ni and Pd supported catalysts are less applied. Some recent results in chemoselective hydrogenations over Pd, Ni and Pt supported catalysts are reviewed below.
Pt is not intrinsically selective to produce unsaturated alcohols, although high selectivities to unsaturated alcohols could be still achieved depending very much on the sterical structure of the reactant. No allyl alcohol was formed in the gas phase hydrogenation of acrolein over Pt/black at 50 °C and the main product was propanal Relatively high selectivities have, however, been achieved in cinnamaldehyde, citral and furfural hydrogenations. In cinnamaldehyde hydrogenation over Pt on graphite and Pt/SiO2 catalysts the selectivities to cinnamyl alcohol were 81.5% in toluene/water mixture at 25 °C and 80% in ethanol. Similarly high selectivity (92%) was obtained over Pt supported on NaY zeolite; the catalyst was prepared by ion-exchange. Even higher selectivity, about 99% towards cinnamylalcohol has been reported over montmorillonite supported Pt
High selectivities to nerol and geraniol can be achieved in citral hydrogenation over Pt/TiO2 at 45 °C and 1 bar in ethanol and Pt/SiO2 catalysts at 150 °C and 20 bar in hexane, i.e. 100% (selectivity to geraniol) and 76%, respectively Furfural can be selectively hydrogenated to furfuryl alcohol (96% at 58% conversion) in 2-propanol–water solvent at 150 °C and 20.6 bar hydrogen over a Pt/C catalyst
Pd differs from the other metals in group 10 from the selectivity point of view, as it is very selective to hydrogenate ethylenic bond. It is believed in that the reason for this is the small d bandwidth of Pd, which affects the adsorption mode of the reactant. Pd favors tetrahepto di-π geometry through both double bonds. Citronellal was the major product in citral hydrogenation over Pd/polymer catalyst at 70 °C and 4.5 bar in 2-propanol as well as over Pd/SiO2 catalyst at 27 °C under 1 bar hydrogen in hexane One exception is hydrogenation of a carbonyl group in aromatic compounds, where selective hydrogenation of carbonyl group can be achieved. For instance 95% selectivity to benzyl alcohol was reported by in benzaldehyde hydrogenation over a cross-linked Pd polymer catalyst at 25 °C and 10 bar hydrogen pressure in methanol. In the liquid phase hydrogenation of cinnamaldehyde over a Pd/SiO2 catalyst, however the maximum selectivity to cinnamyl alcohol was around 32% at 90% conversion at 60 °C and 10 bar in 2-propanol
Ni supported catalysts are very selective in some chemoselective hydrogenations. Ni favors formation of citronellol in citral hydrogenation at 70 °C and 1 bar in ethanol Citronellal was, however, the main product in citral hydrogenation over Ni/SiO2 in hexane at 27 °C under 1 bar hydrogen. The selectivity to citronellal was 77% at 48% conversion. Acetophenone and benzaldehyde hydrogenation was very selective to unsaturated alcohols over Ni/H-Y and Ni/SiO2 ] catalysts, respectively. In acetophenone hydrogenation at 100 °C and 30 bar the highest selectivity to 1-phenylethanol was 100% in n-hexane, but the conversion remained rather low (10%) In alcohol solvents the highest selectivity to 1-phenylethanol was 74% in methanol under the same reaction conditions. It should, however, be pointed out that Ni/Y catalyst was not stable after catalyst recycling, whereas the bimetallic Ni-Pt/Y catalyst exhibited constant activity after recycling. Selectivities close to 100% in benzyl alcohol were achieved in gas phase hydrogenation of benzaldehyde over Ni/SiO2 catalyst at 120 °C. Chemoselective crotonaldehyde hydrogenation is more difficult than previous cases over monometallic supported Ni catalysts. Maximally 59% selectivity to crotyl alcohol was obtained in crotonaldehyde hydrogenation at 60 °C and 1 bar over Ni/TiO2 reduced at 500 °C The relatively high selectivity over Ni/TiO2 was achived due to the existence of interfacial Ni-TiOx sites exhibiting strong metal support interactions The promoted Ni-P/SiO2 catalyst was very active in benzaldehyde hydrogenation at 100 °C and 10 bar in ethanol, but the only reaction product was toluene The reason for this could be the low pH of the reaction mixture (pH 4).
2.1.1.4. Chemoselective hydrogenation over group 11 metals, Cu, Ag, Au
Catalysts having an active metal from group 11 are usually less active than those having a metal from the other groups. This originates from the electronic structure of the d band, which is filled. Recent catalytic results from chemoselective hydrogenations using copper; silver and gold are briefly overviewed here.
Supported monometallic copper catalysts have been used in chemoselective hydrogenations. Cu/SiO2 was very selective towards benzyl alcohol (83%) in benzaldehyde hydrogenation in gas phase at 100 °C at conversion of 68%. No benzyl alcohol was formed over Cu catalysts supported on TiO2, ZrO2, γ-Al2O3 and traces of benzyl alcohol were formed on Cu/CeO2 The highest selectivity to cinnamyl alcohol over 13.7 wt.% Cu/SiO2 was 48% at 60% conversion level of cinnamaldehyde in decalin at 140 °C. The highest selectivity to cinnamyl alcohol was obtained over a Cu/SiO2 prepared by impregnation. The catalyst exhibited large copper crystallites and small amounts of Cu2O at the metal support interface. On the other hand high selectivity to cinnamyl alcohol (78%), but very low activity in cinnamaldehyde hydrogenation was obtained over Cu/ZrO2 catalyst at 140 °C and under atmospheric pressure in propylene carbonate. Very high conversion of furfural and selectivity to furfuryl alcohol (98%) was obtained over Cu/MgO catalyst in the gas phase hydrogenation at 180 °C. The reason for high activity and selectivity was stated to be the presence of defective sites at Cu and MgO in interfacial region.
Silver supported catalysts have been used in the hydrogenation of acetophenone, acrolein, crotonaldehyde and furfural. Unusually high selectivities to unsaturated alcohols have been reported over nanosized silver catalysts. According to Claus, silver is able to adsorb conjugated double bond via bifunctional bonding to both C C and C O bond resulting into formation of 2-propenyloxo intermediate. Possibility of hydrogen adsorption with dissociation was recently confirmed by isotope jumping technique. Selectivity of 94.8% to 1-phenylethanol was reported in acetophenone hydrogenation over Ag/SiO2 catalyst at 140 °C and 2 bar whereas in crotonaldehyde hydrogenation over Ag/SiO2 catalyst at 20 bar and at 140 °C the selectivity to crotyl alcohol was 62.8% at 16% conversion. Ag/SiO2 catalyst was quite selective in the hydrogenation of furfuryl alcohol. The selectivity to furfuryl alcohol was 79% in the gas phase hydrogenation of furfural at 1.9 bar and at 200 °C over 12.1% Ag/SiO2 catalyst.
Gold has been used in the chemoselective hydrogenation of crotonaldehyde and, trans-3-buten-4-phenyl-2-one (benzalaldehyde) acrolein and and citra. Since gold in powder form has been catalytically active in gas phase hydrogenation of acrolein, thus its catalytic activity is not exclusively originated from quantum size effect or from specific metal-support interactions. Quantum size effect alters the electronic properties of a metal particle Moreover the binding energy shifts of core levels have been observed in supported gold nanoparticles. Gold supported on Fe2O3 yielded to very high selectivities of nerol and geraniol in citral hydrogenation at 60 °C in and 1 bar ethanol (95% at 90% conversion of citral) and the selectivity was independent on the metal loading In the gas phase hydrogenation of acrolein at 240 °C and 20 bar the maximum selectivity to allyl alcohol was 37% over Au/ZrO2 catalyst, but the conversion of acrolein was only 5%. The hydrogenation rate was decreasing, while the selectivity to unsaturated alcohol increased with increasing gold particle sizes. The observed structure sensitivity is discussed in detail
2.1.2. Bi- and multimetallic catalysts
Bimetallic catalysts have been very selective in several chemoselective hydrogenations, like Rh-Sn/SiO2 for nerol and geraniol (98%), Pt/ZnO for crotyl alcohol (90%), Rh-Cu/SiO2 for crotyl alcohol (62%) The origin for the high selectivities to unsaturated alcohols has been associated with electron transfer from the less noble metal to more noble metal as well as alloy formation and /or an intimate contact between two metals. This concept was, however, challenged by The second metal can exist as an adatom, in alloy, or in ionic state as well as in partially oxidized form The difference in the electronegativity between two metals can enable the polarization of the carbonyl bond. Additionally, geometrical effects, like a change in metal dispersion decoration of the main metal by the second metal via surface enrichment and organic fragments in the vicinity of an active metal, have been observed in bimetallic catalysts. It is important to note that these effects are often coexistent in bimetallic catalysts and thus one effect, like for instance alloy formation in a model catalyst Sn/Pt(1 1 1), is not alone responsible for improved selectivity because nearly the same selectivities to crotyl alcohol were obtained over non-supported Pt and Pt-Sn model catalysts. Increased selectivities to crotyl alcohol were, however, obtained over supported Pt-Sn catalyst
There exists controversial information of the role of alloys in chemoselective hydrogenation and even bimetallic catalysts non-selective for formation of unsaturated alcohols have been reported Alloys are defined as compounds formed by two or more metals; they can also be formed between a metal and another metal originating from a non-inert support The alloy formation is strongly influenced by the selection of metal precursor and support, catalyst pretreatment as well as the preparation method.
In order to uncover both the selective hydrogenation site and the reaction mechanism it is important to know the chemical nature of the active site. Several physico-chemical methods have been used as tools for characterization of the active sites. The benefits and drawbacks of different characterization methods are discussed in the subsequent section. The following bi- or multimetallic catalysts have been recently investigated in chemoselective hydrogenations from group 8: Ru-Sn, Ru-Fe/C from group 9 Rh-Sn/SiO2, Rh-Cu/SiO2, Rh-Mo/SiO2, Ir-Ge/SiO2, Ir-Fe/SiO2, Mo-Co-B, Sn-Co-B, Ra-Co from group 10 Pt-Sn/Al2O3 and, Pt-Sn/C, Pt-Ge/Al2O3, Pt/ZnO Pt-Sn/MgO, a model catalyst Sn-Pt(1 1 1) without support, Pt-Co/graphite, Ni-P-B, Ni-Cu/SiO2, from group 11 Ag-In/SiO2, Au/ZnO Au-In/ZnO, Cu-Zn-Al, Cu-Ni(Co)-Zn-Al. Additionally, there exist catalysts exhibiting a combined effect of both two metals and reducible oxides, like Pt-Sn/CeO2-SiO2 and Ru-Sn/TiO2. The examples are presented below on a group-by-group basis and the combined effect of bimetallic catalysts on reducible supports is discussed separately
3. Hydrogenation reactions studied using gold catalysts
The literature till the 1960s indicates that gold seems to be very popular for being catalytically inert in hydrogenation. During the following decades most of the work involved mainly the adsorption/activation of hydrogen and hydrogen exchange on gold surfaces, as shown in the previous section, and the hydrogenation of relatively simple molecules, such as CO, CO2, alkenes and alkadienes. Only a very limited number of publications concern the selective hydrogenation of complex organic molecules with more than one unsaturated group to reactions products, which are of interest for the chemical industry. As it was recognized that gold can be prepared as nanoparticles on oxidic supports by wet-chemical techniques, such as deposition–precipitation or coprecipitation methods, a breakthrough came in heterogeneously catalysed hydrogenations using multifunctional organic compounds (e.g. α,β-unsaturated aldehydes, ketones) as educts. In Table 1, a comprehensive overview of the hydrogenation reactions studied together with the gold catalysts used is given considering the literature up to September 2004. It is apparent from the cited reports that supported gold catalysts have in some cases superior prospects for employing them in industrial catalysis. Beside the partial hydrogenation of acetylene to ethylene and 1,3-butadiene to butenes, the main advantage of gold catalysts is their ability to selectively catalyse the hydrogenation of the C=O group of α,β-unsaturated aldehydes producing allyl-type unsaturated alcohols. In this field, attempts of our group to identify the active sites of supported gold catalysts, favoring the adsorption and hydrogenation of the carbonyl group, were successful and the influence of the real structure of the catalyst on the hydrogenation properties was estimated. Both approaches are highlighted in the next section.
Table 1. Hydrogenation of different educts over gold catalysts
Full-size table
View Within Article
4. Selective hydrogenation of α,β-unsaturated aldehydes over supported gold catalysts
The selective hydrogenation of α,β-unsaturated aldehydes to allylic alcohols is of both commercial relevance (e.g. in fine chemical and pharmaceutical intermediate production and specific scientific interest Note that thermodynamics favors hydrogenation of the C=C over the C=O group (stronger negative free reaction enthalpy of 35 kJ mol−1), and for kinetic reasons the C=C bond is more reactive than the C=O group. Thus, in the presence of most of conventional monometallic hydrogenation catalysts, based on Pt, Rh or Pd on silica and alumina as supports, lower α,β-unsaturated aldehydes such as acrolein and crotonaldehyde are hydrogenated predominantly to the saturated aldehydes, and only to a minor extent to unsaturated alcohols Therefore, it is desirable to find catalysts, which are able to control the intramolecular selectivity by hydrogenating preferentially the C=O group, while keeping the olefinic double bond intact.
In our group, the hydrogenation of acrolein) on supported gold catalysts has been used as test reaction to study several aspects of structure sensitivity, i.e., a dependence of activity and selectivity on the size and morphology of gold particles Because of the lack of space-filling substituents (e.g. methyl, phenyl) at the C=C group acrolein is the α,β-unsaturated aldehyde, which is most difficult to hydrogenate to the unsaturated alcohol (allyl alcohol). The latter must be the criterion for the evaluation of the unusual properties of gold as hydrogenation catalyst. Moreover, from an industrial point of view allyl alcohol is much more important for the chemical industry (e.g. used for glycerol synthesis, allyl ester, epoxidation) than, for example, crotyl alcohol which can be obtained by selective hydrogenation of crotonaldehyde over supported gold catalysts. However, it must be kept in mind that in the case of crotonaldehyde (or higher α,β-unsaturated aldehydes), the selectivities reported do not represent the intrinsic selectivity of gold because a methyl (or phenyl) group as substituent on the C=C double bond contributes to the overall selectivity by steric repulsion of the olefinic group changing the adsorption configuration of the aldehyde as shown by a number of experimental data and theoretical calculations,. This is obvious from where the hydrogenation of acrolein and crotonaldehyde over a Au/SiO2 catalyst (dAu = 3.9 ± 2.3 nm, prepared by incipient wetness are compared. Also, in the case of Au/TiO2 catalysts prepared by deposition–precipitation methods the selectivity to the unsaturated alcohol was decreased from a level of 60–70% when using crotonaldehyde to 30% in the case of acrolein. Beside gold on oxide supports, also silver catalysts, are the only candidates, which convert acrolein to allyl alcohol with selectivities one order of magnitude higher than Group VIII hydrogenation catalysts. It is interesting to note that the adsorption of acrolein on both metals (polycrystalline gold, Au (1 1 1), evaporated silver films) was studied by infrared reflection absorption spectroscopy (IRAS) showing a coverage-dependent discrete adsorption geometry. It can be concluded that the adsorbate interacts with the metal surface through the oxygen atom of the C=O group coordinated to a positively charged surface site (kinks, steps).
Fig. 1. Reaction network of acrolein hydrogenation.
View Within Article
Reaction scheme for crotonaldehyde hydrogenation
Table 1.
Comparison of the gas phase hydrogenation of acrolein and crotonaldehyde over 1.6 wt.% Au/SiO2-IW catalysta
Full-size table
a T = 260 °C, ptotal = 2 MPa, molar ratio H2/educt = 20, = 15.3 g h mol−1, = 19 g h mol−1.
b Sum of 2-butanone (via oxygen-π-adsorbed intermediate) and 2-ethylhexanol (via condensation of butanal to 2-ethylhex-1-enal and its subsequent hydrogenation).
View Within Article
However, it has not yet been possible to identify the catalytic phenomena in hydrogenation reactions associated with the size and the morphology of supported gold particles. Therefore, extensive reaction studies coupled with in-depth catalyst characterization were undertaken by our group to focus on the problem of structure-sensitivity and gaining more knowledge of how the unique catalytic behavior of gold can be controlled by the nanostructure of the catalyst. Beside a dependence of the size of the gold particles on the kind of support, the latter has an influence on the degree of rounding of the particles. shows representative gold particles of nearly the same sizes on different supports (SiO2, TiO2, ZrO2, ZnO) after a pretreatment in hydrogen at 573 K. On SiO2 and TiO2, the particles are nearly spherical, whereas on the other supports extended facets of the gold surface are visible, in particular for Au/ZnO. Structure-sensitivity may have its core in the ability of nanostructured matter to occur in metastable non-equilibrium structures, for example as coordinatively unsaturated sites (e.g. edges, kinks) induced by rough surfaces, as matter with lattice strain or multiple-twinned particles (MTPs) [. The question arises how the hydrogenation properties of gold catalysts are influenced by these nanoscale properties of gold. To answer this question, we focussed our work at first on the influence of the geometrical configuration of the supported gold nanoparticles, namely the occurrence of MTPs and the degree of rounding exhibiting the real structure (Significant differences between the Au/ZrO2 samples prepared by deposition–precipitation at different pH values were found. By correlating the HRTEM and catalytic results, it was shown that higher amount of MTPs resulted in a lowering of selectivity to the desired product allyl alcohol (Table 3), and also the turnover frequency (TOF) was decreased. Therefore, the main demand for the catalyst synthesis is that supported gold catalysts with single crystalline gold particles must be prepared to obtain higher selectivities to the unsaturated alcohol. The increase of TOF for Au/TiO2 compared to Au/ZrO2 is attributed to more rounded gold particles, accompanied by a higher relative amount of low-coordinated surface sites.
Fig. 2. Representative gold particles on different supports (DP: deposition–precipitation, I: impregnation)
View Within Article
Fig. 4. HRTEM image of Au–In/ZnO ([35], left) and surface model (right). Indium preferentially decorates the outer faces of the gold particles while the edges remain uncovered. (B) (left
View Within Article
Fig. 3. Representative HRTEM of two Au/ZrO2-DP samples with a more rounded gold particle (A) and a particle with
Table 3.
Influence of the pretreatment of 2.2 wt.% Au/ZrO2-DP catalyst in H2 on nanostructure and catalytic properties during acrolein hydrogenation in the gas phasea
Full-size table
a T = 513 K, ptotal = 2 MPa, molar ratio H2/Ac = 20, = 15.3 g h mol−1
b Reduction temperature.
c Reduction time.
View Within Article
After the finding that the hydrogenation of acrolein preferentially occurs on single crystalline particles rather than on multiply twinned ones, in the next step of our investigation a new approach was applied to distinguish different features of the gold particle morphology. By means of HRTEM, the crystal structure of individual gold particles prepared on ZnO (SBET = 15 m2 g−1) was investigated. Most particles (dAu = 9.0 ± 0.3 nm) were single crystalline and practically no MTPs were found. This is an important result because elucidation of the active site issue cannot be simply done by varying the size of gold particles, since the effects of faceting and multiply twinned particles may interfere in other supported gold nanoparticle systems. In the next step, modification of the gold particle surface by indium has been used to vary the active site characteristics of a suitable catalyst. By detailed HRTEM image analysis, the lattice spacings of these bimetallic Au–In particles perpendicular to the outer surface were measured. Since there were distinct contrast differences between both metals due to the atomic number difference, it was possible to distinguish for each atomic layer between both In- and Au-containing phases. There is clear evidence of the formation of a separate indium phase exhibiting a lattice distance close to the (1 0 1) distance of tetragonal indium. Furthermore, HRTEM imaging revealed clearly a selective decoration of gold particle faces has been observed, leaving edges free. Statistical evaluation of appearance of indium on the cubooctahedral gold surfaces was undertaken which supports the selective face decoration by the second metal. EDX analysis revealed a homogeneous distribution of indium on the gold particles and no monometallic indium deposits on the support were found. With the bimetallic Au–In/ZnO catalyst, the desired allyl alcohol is the main product formed with a selectivity of 63% (at T = 593 K and ptotal = 2 MPa). Note that the Au–In system is the first example showing the important impact of a second metal for tuning the selectivity of a gold catalyst in selective hydrogenation of functional groups. The level of selectivity obtained with Au–In/ZnO is one order of magnitude higher compared with conventional mono- and bimetallic catalysts for acrolein hydrogenation and comparable to our best catalyst for the selective hydrogenation of acrolein to allyl alcohol, Ag–In/SiO2 From the experimentally proven correlation between surface structure and catalytic behavior, the edges of single crystalline gold particles (“edgy catalyst” have been identified as active sites for the preferred C=O hydrogenation.
View Within Article
A further example for the importance of gold as hydrogenation catalyst was illustrated by comparing gold with platinum, both supported on zirconia, in acrolein hydrogenation Preliminary hydrogenation experiments showed that over Pt/ZrO2 the selectivity to allyl alcohol was always very low (ca. 2%). In order to ensure that this was not due to the high catalyst activity (acrolein conversion of 43% at 473 K) and the related secondary reactions of allyl alcohol according to, the modified space time () was varied in such a way that a comparison of selectivities for the catalysts Au/ZrO2 and Pt/ZrO2 was possible with almost the same acrolein conversion. Even under these experimental conditions, propanal was the main product with Pt (S > 90%). Selectivity to allyl alcohol was only 3.9%, i.e., only one-tenth of the selectivity obtained with Au/ZrO2. This finding clearly indicates that gold catalysts are more efficient than platinum catalysts in acrolein hydrogenation. If the α,β-unsaturated aldehyde as educt is changed from acrolein to, for example, crotonaldehyde, then monometallic Pt catalysts (e.g. Pt/ZnO) can also exhibit high selectivity towards the allylic alcohol. However, in that case, as noted above, the increased selectivity is due to the additional steric effect of the methyl group and cannot be ascri3.4. Catalytic behavior
*. Crotonaldehyde hydrogenation
The performance of the catalysts (previously reduced at 473 K) was tested in the vapor phase hydrogenation of crotonaldehyde at 333 K. The evolution of catalytic activity (TOF, calculated in basis of CO uptakes at 300 K) with time on stream for the three catalysts studied is showed in. It can be seen that, after a period of 20 min with a strong deactivation, the activity remained quite stable with time. This fast deactivation of the catalysts is typical of this reaction and has been previously related to the decarbonylation of the reactant molecule yielding carbon monoxide, which is irreversively adsorbed on platinum at the reaction conditions. It can be clearly appreciated that the specific activity strongly depends on the platinum face exposed, which is a clear indicative of the structure-sensitive character of this reaction. Furthermore, it is interesting to observe that polycrystalline Pt/C showed lower TOF values than preferentially oriented catalysts. Englisch et al., in a very complete work on structure sensitivity of crotonaldehyde hydrogenation over Pt/SiO2, also found lower activity for catalysts with smaller metal particles, and this behavior was attributed to a faster deactivation caused by strong adsorption of crotonaldehyde on low coordination sites like edges and corners. However, this fact has not been detected in our case, and deactivation profiles were very similar for the three catalysts under study. In the same way, Singh et al. found a 25-fold increase in TOF for citral hydrogenation as the average platinum crystallite size increases from 1 to 5 nm. Again, the higher participation of low coordination number surface atoms as the metal particle size decreases was invoked to explain this behavior. Note that these highly exposed atoms are assumed to be the most abundant on the surface of Pt/C as indicated by voltammetric and microcalorimetric results).
TOF in crotonaldehyde hydrogenation at 333 K as a function of time on stream for Pt/C, Pt(111)/C, and Pt(100)/C catalysts.
Reaction scheme for crotonaldehyde hydrogenation
View Within Article
Comments on the reaction mechanism of crotonaldehyde hydrogenation on Au/TiO2 catalysts
The selectivity to C=O versus C=C hydrogenation is mainly dictated by the extent of activation of the C=O group Therefore, the adsorption mode of crotonaldehyde is the determining factor, which orientates the selectivity toward the formation of the monohydrogenated products. The possible adsorption modes of α,β-unsaturated aldehydes on monocrystalline metal surfaces were described by Delbecq and Sautet If the adsorption is restricted to that of the C=C moiety of the molecule (η2CC of di-σCC-type) (Scheme 2), hydrogenation would only lead to butanal. Conversely, if the adsorption is restricted to that of the C=O bond (η2CO of di-σCO-type), or to that of the oxygen lone pair (atop), only crotyl alcohol could form. In the case of gold catalysts, both products are formed, and we must then assume either the coexistence of two independent C=C and C=O (or O atop) adsorption modes or a unique adsorption mode involving both C=C and C=O (or O), such as the η3 or η4 modes (Scheme 2).
Scheme 2. Adsorption modes of the molecule of crotonaldehyde on metal, from Ref.
View Within Article
With regards to selectivity, the main product formed in all cases was the saturated aldehyde (butanal). Thus, the selectivity to the unsaturated alcohol was almost negligible for polycrystalline Pt/C and Pt(100)/C catalysts while Pt(111)/C showed certain selectivity to crotyl alcohol (4% molar fraction). It is well known that there is an effect of the metal morfology on the adsorption geometry of unsaturated aldehydes, leading to different types of reaction pathways and finally, different selectivities. The theoretical calculations of Delbecq et al. suggest that, in the case of (100) and open faces or steps, crotonaldehyde adsorption involves both double bonds, in a so-calledη4 planar coordination, this geometry leading to the preferential hydrogenation of the C C bond for kinetic reasons. This fact would explain the low selectivity toward unsaturated alcohol found for Pt(100)/C and Pt/C catalysts, where the fraction of (100) facets and low coordinated atoms, respectively, is very high. On the other hand, the dense (111) metal faces are not favorable for the C C coordination, and crotonaldehyde adsorption via carbonyl bond (di-σC=O) is energetically favored. Thus, a greater participation of these faces in the catalysts surface would improve the selectivity to unsaturated alcohol. The low value of selectivity obtained for Pt(111)/C in our study could be due to the presence of a noticeable fraction of stepped surfaces vicinal to (111) small domains (see prominent peaks at 0.12 and 0.26 V in) since, in these surfaces, the crotonaldehyde coordination is shifted to a πC=C mode. Furthermore, voltammetric results showed that the surface of Pt(111)/C is not exclusively participated by (111) facets and the presence of large (100) domains (could also be responsible of the low value of selectivity to unsaturated alcohol.
Fig. 22. Crotylalcohol selectivity as a function of conversion in consecutive crotonaldehyde hydrogenation over Pt/SnO2 catalyst at 170 °C [69].
Fig. 18. Activity as a function of H2 partial pressure at different temperatures during crotonaldehyde hydrogenation over Cu/AC catalyst.
Fig. 4. Hydrogenation of crotonaldehyde (PH2=752 Torr, Pcrotonaldehyde=8 Torr, Treaction=120 °C). Selectivities versus conversion on DP Urea 16 h/90 reduced under H2 at 120 °C.
*. Cinnamaldehyde hydrogenation
shows the main reaction pathways that can occur during cinnamaldehyde hydrogenation. Under the experimental conditions used in the present work only these four products were detected. Activity and selectivity of the catalysts in the hydrogenation of cinnamaldehyde were evaluated in the liquid phase at 383 K and 5 MPa. plots the conversion of cinnamaldehyde as a function of reaction time for the three catalysts under study. Initial activities were expressed in term of rate constant (min−1), which were obtained from the slope of the cinnamaldehyde concentration-time plot and considering a first rate law equation. The activity per active site (TOF) was calculated using the CO uptakes to determinate the number of surface metal sites and the results are summarized in Table 1. As it can be seen, both the rate constant (Table 1) and the conversion of cinnamaldehyde reached for similar reaction times (followed the trend: Pt/C > Pt(100)/C > Pt(111)/C. However, when the rate is normalized by the number of accessible platinum atoms (TOF), the trend changes to: Pt(100)/C > Pt(111)/C > Pt/C which is consistent with that obtained in crotonaldehyde hydrogenation
High-quality image (43K)Fig. 5. Reaction scheme for cinnamaldehyde hydrogenation.
High-quality image (87K)
Fig. 7. Conversion of cinnamaldehyde (—) and selectivity to cinnamyl alcohol (- - -) for: Au/Al (•), Au/Fe10Al (♦), and Au/HDAE (■).
High-quality image (48K)Fig. 6. Conversion of cinnamaldehyde as a function of time for Pt/C, Pt(111)/C, and Pt(100)/C catalysts.
View Within Article
Table 1.
Rate constant, TOF and selectivity to the saturated aldehyde and the unsaturated alcohol at 10% conversion in cinnamaldehyde hydrogenation for Pt/C, Pt(111)/C, and Pt(100)/C catalysts
Full-size table
a Values at 10% conversion.
View Within Article
Regarding the products distribution, again the main reaction product was the saturated aldehyde (HALD) although, conversely to crotonaldehyde hydrogenation, the three catalysts under study yielded noticeable amounts of cinnamyl alcohol (CALC). This fact is not strange since it is well known that the selective hydrogenation of the carbonyl bond is more difficult to achieve when lower aldehydes are involved and the reaction takes place in gas phase. also shows the selectivity to cinnamyl alcohol (CALC) and hydrocinnamaldehyde (HALD) at equal conversion level (10%) for the three catalysts under study. It is interesting to observe that polycrystalline Pt/C showed lower selectivity to CALC than preferentially oriented samples. This fact can be explained in terms of particle size. It is well known that metal particle size has a strong influence in the selectivity to cinnamyl alcohol in the case of monometallic platinum catalysts. Thus, the selectivity for C O hydrogenation increased with increasing crystallite size. Note that Pt/C showed an average platinum particle size of 3 nm while in preferred oriented samples, Pt(100)/C and Pt(111)/C, platinum is in form of large particles of around 10 nm size (Repulsive interactions between the phenyl ring and the large metal crystallites, which hinders the interaction via the C C bond and enhances interaction via C O bond, is proposed to explain this behavior. This effect is absent in the case of small metal particles because both double bonds, C C and C O, could approach the surface with less repulsive interactions between the phenyl ring and the metal surface.
Furthermore, it can be seen in that Pt(111)/C showed higher selectivity to CALC than Pt(100)/C (48 vs 36%). In this case no size effect can be invoked since both samples showed similar average particle size (10 nm). However, as pointed before, both catalysts differ in the facets distribution of the metal particles, with a higher ratio of (111) domains in the case of Pt(111)/C. Thus, these results clearly indicate that, besides a size effect, the amount of (111) facets composing the metal particle play a crucial role in increasing the selectivity to unsaturated alcohol in these kind of reactions. In this way, previous studies on prenal hydrogenation on well-defined (111) and (110) platinum surfaces, found that the main product (also at 10% conversion) was the unsaturated alcohol and the saturated aldehyde, respectively. Interestingly, the same authors also found that prenal hydrogenation on Pt(553) (which can be considered as a surface with (111) terraces and steps) resulted in a lower selectivity to the unsaturated alcohol compared to (111) flat surface. Again, the simultaneous presence of (111) terraces (where the di-σC=O adsorption mode is preferred) and steps (πC=C adsorption mode preferred) is invoked to explain the lower selectivity to the unsaturated alcohol on this stepped surface. As already seen, it has been proved that the platinum particles composing Pt(111)/C have a noticeable amount of stepped surfaces vicinal to (111) small domains (), which would explain the noticeable yield of saturated aldehyde (obtained for this sample. It is necessary to bear in mind that substituents at the C C bond have a strong influence on the selectivity to the unsaturated alcohol and thus, can influence the intrinsic selectivity determined by the platinum surface. Thus, is not strange that some works, based on the hydrogenation of the simplest unsaturated aldehyde (acrolein), introduced some new points to the factors controlling the selectivity in this kind of reactions. In a recent theoretical work, Loffreda et al concluded that the key point which determines the selectivity to allyl alcohol on Pt(111) is not the adsorption mode of the aldehyde, but the competitive desorption process of the half hydrogenated products (unsaturated alcohol vs saturated aldehyde). Furthermore, in contradiction with earlier assumptions, Mohr et al. claim that the surface sites responsible of the C O hydrogenation are the highly unsaturated surface atoms situated at corners and edges, although these conclusions are obtained for a very special catalytic system (Au/ZnO), and the same authors indicated that these conclusions could not be transferred to other catalytic systems.
scheme
Fig. 1. Conversion of 1% v/v cinnamaldehyde/toluene over 0.50 g of 2% Pd/SiO2 at 25 °C and 47 psi H2 pressure.
for crotonaldehyde hydrogenation
5- Product distributions and activation energy
The product distributions obtained with the Au/TiO2 catalysts are remarkable since they consist of 60–70% of crotyl alcohol, whatever the studied Au/TiO2 catalyst and the temperature of pretreatment (A selectivity-conversion diagram was obtained for catalyst DP Urea 16 h in gradually decreasing flow rates during the time on stream in order to increase the conversion This diagram shows that the selectivity to butanal remains unchanged up to 50% conversion while those in butenes and butanol slightly increase at the expense of that in crotyl alcohol.
Fig. 4. Hydrogenation of crotonaldehyde (PH2=752 Torr, Pcrotonaldehyde=8 Torr, Treaction=120 °C). Selectivities versus conversion on DP Urea 16 h/90 reduced under H2 at 120 °C.
View Within Article
To check if butanal could be formed by isomerization of crotyl alcohol as observed on Pt-based catalysts the hydrogenation of crotyl alcohol was investigated for the DP Urea 16 h catalyst. It was observed that crotyl alcohol reacted at a similar rate as crotonaldehyde, to give rise to 90% of butenes, 3.9% of butanal, and 6.1% of butanol. The butenes formed from crotyl alcohol consisted of 73% of but-1-ene, 7% of trans-but-2-ene, and 20% of cis-but-2-ene. This butene distribution is similar to that obtained from butadiene hydrogenation on the same catalyst (unpublished results). These results suggest that crotyl alcohol transforms by dehydration into butadiene, which further hydrogenates into butenes. This could explain the increased of selectivity to butenes at the expense of that in crotyl alcohol when the conversion of crotonaldehyde increased. This also indicates that the butanal observed in the crotonaldehyde hydrogenation products is directly formed from the reactant since the isomerization of crotyl alcohol into butanal is a minor reaction. Scheme 1 can be proposed.
Scheme 1. Reaction pathway of the hydrogenation of crotonaldehyde with Au/TiO2 catalysts; plain and dotted arrows represent the primary and secondary reactions, respectively, and the lengths of the arrows are roughly proportional to the reaction rates.
In contrast to platinum-based catalysts, the presence of sites, which would hydrogenate the C=C bond only (η2CC structure), is unlikely on gold catalysts. Indeed, gold catalysts are active for the partial hydrogenation of butadieneor acetylene but very poorly active for the hydrogenation of butene or ethylene. We found for the DP Urea 16 h catalyst that butadiene reacted twice as fast as crotonaldehyde under the same conditions of pressure and temperature, but this catalyst was totally inactive for butene hydrogenation (unpublished results). This last result makes unlikely the η2CC adsorption mode of olefin on gold particles, and therefore, that of crotonaldehyde.
On the other hand, the fact that the selectivities to crotyl alcohol and butanal are almost constant whatever the gold particle size and the reduction temperature is not in favor of the coexistence of two types of adsorption modes involving two types of catalytic sites whose proportions depend on the structure/size of the gold particles. Therefore, we rather propose that a single mode of adsorption of crotonaldehyde governs the selectivity, either η3 or η4 adsorption mode and that the addition of two hydrogen atoms gives rise to both butanal and crotyl alcohol. The selectivities to each product would be determined by the relative probabilities of addition of the hydrogen atoms to the different positions in the molecule (O, Cα, Cβ, Cγ).
The dramatic increase of the TOF when the particle size decreases suggests that the gold atoms involved in the hydrogenation process are those located on edges and corners. For instance, the relative number of low-coordinated sites in cubooctahedral-shape particles varies in the same way as the TOF versus the particle size Besides this pure size effect, the evolution of the particle morphology observed on the HRTEM micrographs with the reduction temperature leads to the same conclusion: the gold particles obtained after reduction at 200 °C do not exhibit well-developed facets while those obtained after reduction at 500 °C (are perfectly facetted with low index planes. Therefore, the proportion of low-coordinated sites is much lower after pretreatment at 500 °C than at 200 °C, and the decrease in catalytic activity and TOF when the reduction temperature increases from 200 to 500 °C is attributed to the smoothing of the outer surface of the gold particles, i.e., to the decrease in the number of low-coordinated sites. This is in agreement with the interpretation of Claus and colleagues where they attributed the higher activity of Au/TiO2 compared to that of Au/ZrO2 to the fact that the gold particles were round shaped on TiO2 and facetted on ZrO2. In another work on selective decoration of the faces of gold particles by indium, they proposed that the active sites favoring the adsorption of the C=O group of acrolein are the edges of the gold nanoparticles.
However, these low-coordinated sites are not prone to favor the η4-adsorption mode for which the reactant structure is rather planar or the η3-adsorption mode for which multiple metallic sites are involved but they might be more adapted to dissociate the hydrogen molecule. Therefore, we propose that the activity of gold catalysts in hydrogenation is determined by the dissociation step of molecular hydrogen. The apparent activation energy calculated in the present work, 36.3 kJ mol−1, is the same as that reported by Buchanan and Webb for the selective hydrogenation of butadiene on Au/Al2O3 catalysts, 36.5 kJ mol−1. This could support the hypothesis that butadiene and crotonaldehyde hydrogenation kinetics are governed by a common rate-determining step, the hydrogen dissociation. In the Au/Al2O3 catalysts were prepared by impregnation (1 wt% of Au) and exhibited a broad distribution of particle sizes up to 80 nm with the largest population of size in the 0–10 nm range. Because of the complete absence of activity of the support in butadiene hydrogenation, Buchanan and Webb proposed a mechanism where hydrogen dissociated on alumina and migrated to gold particles to hydrogenate the adsorbed butadiene. However, they did not exclude the fact that the activation of H2 could take place on tiny gold particles, which could not have been detected by electron microscopy (the smallest particle size they could detect with their microscope was not reported in their paper).
The poor catalytic activity in hydrogenation with gold catalysts compared to Pd or Pt catalysts has been generally attributed to the fact that gold hardly dissociates molecular H2, as it has been established experimentally and theoreticallyHowever, several authors have unexpectedly observed the dissociative chemisorption of H2 (D2) on gold films. For instance, Okada et al. who studied the adsorption and reaction of hydrogen (deuterium) on thin gold film grown on Ir (111), noted that surprisingly H2 (D2) can dissociatively adsorb on these Au films. They attributed this property to locally relaxed surface features like defects. On thin gold film deposited on Pyrex glass, Stobinski et al. also found that H2 dissociative chemisorption occurs, providing that the thin Au films were deposited at a low temperature (78 K) and were not sintered afterward. They concluded that H2 chemisorption occurs only on surface atoms of low-coordination number. These conclusions support our proposals for the hydrogenation mechanism.
.
6- Reactor selection
Traditionally batch reactors have been applied in fine chemicals production in the liquid phase, whereas gas phase hydrogenation is usually carried out in fixed bed reactors. It could be, however, attractive from the industrial point of view to produce fine chemicals in a continuous mode. The benefits compared to the conventional batch operation are the higher production capacity, minimization of dead time during the start and end, as well as more equal product quality. According to OECD reference scenario the world chemical demand will increase by 39% from the level in year 2000 to the level in year 2020. Additionally, from the scientific point of view the continuous operation allows an easier way to record and model catalyst stability and deactivation in case of liquid phase reactions.
Liquid phase hydrogenations have been performed in trickle bed, in monolithic and in bubble column reactors as well as in a fixed bed reactor with recirculating liquid phase [ and in continuous flow. One decisive factor for achieving high selectivity to intermediate products in consecutive hydrogenations is the residence time of the reactant in the reactor. Trickle bed reactors are widely used in chemical processes in hydrogenations and oxidations. The drawback in using trickle bed reactors is the poor wetting of the catalyst, which can lead to ineffective use of the catalyst as well as to formation of hot spots. In trickle bed reactors the lower velocities are applied in order to maintain the trickle-flow regime, whereas in monolithic reactor higher velocities should be used in order to avoid maldistribution at lower velocities . Trickle bed reactors have been used in citral, benzaldehyde and in crotonaldehyde hydrogenations. Solvent free citral was hydrogenated in a trickle bed reactor at 100 °C and 20 bar hydrogen. Citral hydrogenation in this reactor exhibited 20 h induction period during which the conversion level of citral increased. After induction period a constant very high conversion (97%) as well as very high selectivity to nerol and geraniol (97%) over Pt-Sn/MgO catalyst was obtained at 100 °C and 20 barInterestingly with a monometallic Pt/MgO lower selectivity was achieved, but also no induction period was observed. In order to avoid the poor catalyst wetting and to enhance the productivity the trickle bed reactor with periodic flow interruption has been applied. As a result an enhanced crotonaldehyde hydrogenation rate was observed over Pd/γ-Al2O3 at 25 °C and 11 bar in such reactor compared to a steady state operation. The performances of a trickle bed and a monolithic reactor were compared in benzaldehyde hydrogenation at 100 °C and 10 bar. The former reactor contained Ni supported on γ-alumina extrudates as a catalyst, whereas a Ni-cordierite catalyst was used in the monolithic reactor. Co-current gas and liquid flows were applied. Monolithic reactor showed higher productivity and selectivity than the trickle bed reactor in the hydrogenation of benzaldehyde. The reason for this might be the sharper residence time distribution in the monolithic reactor.
Monolithic reactors can operate either in single-pass mode or with external liquid recycle. The product selectivity might decrease in a latter case with liquid back-mixing and thus the single pass mode might be preferred. Monolith catalysts were investigated in citral hydrogenation in a screw-impeller tank reactor and in benzaldehyde hydrogenation over a monolith catalyst in a batch mode or in continuous mode. Ni-monolith was used as a catalyst in citral hydrogenation in a screw-impeller stirred tank reactor (SISRCitral hydrogenation was investigated in hexane at 5 and 40 bar hydrogen and at temperatures 40, 80 and 100 °C. Catalyst deactivation was observed after consecutive experiments. The selectivity to citronellal was the highest at 40 °C and 5 bar hydrogen, i.e. 96% at 80% conversion. Higher hydrogen pressures favored formation of 3,7-dimethyloctanal. In benzaldehyde hydrogenation over Ni-monolith catalysts the highest selectivities to benzyl alcohol were obtained in a batch operation without residence time distribution .
A continuous fixed bed reactor operation was demonstrated in citral hydrogenation with a recirculating liquid phase and in a continuous flow. In the former case it was shown in citral hydrogenation at 70 °C in ethanol over knitted Ni-silica fibers that the recirculating reactor operation was equally good as a traditionally batch reactor. In the latter case citral hydrogenation was carried out over 5 wt.% Pd/Al2O3 catalyst at 40° and 190 bar hexane, supercritical CO2 or ethane. Both catalyst deactivation and changes in the product distribution were observed.
A packed bubble column reactor was used in hydrogenation of citral to citronellal. The aim of using recirculating product and circulating hydrogen was to improve the space–time yield, because according to Bröcker et al. the catalyst regeneration is not successful in citral hydrogenation and the larger amounts of catalyst needed to enhance the productivity would decrease the process economy. In a packed bubble column reactor hydrogenation of citral to citronellal over a powdered Pd/C catalyst in the presence of triethyl amine in methanol at 70 °C and 8 bar hydrogen gave 92.9% selectivity to citronellal at 99.7% conversion.
Microstructured carbon-coated reactors have been applied in gas phase hydrogenation of acrolein. The catalyst contained microchannels of 700 μm depth and of 300 μm width with length of 50 mm manufactured from copper wire. The carbon-coating was obtained via carbonization the polymer coating at 550 °C. Thereafter the support was calcined at 350 °C and after fucntionalization of the carbon surface with HNO3 Ru was deposited on the support via ion exchange with [Ru(NH3)Cl]Cl2. Although the comparative specific catalytic activities for microstructured Ru-catalyst and for Ru/C were not measured, it was reported that the selectivity to allyl alcohol was comparable for these catalysts. The short overview of various reactor configurations applied in the recent years for heterogeneous catalytic hydrogenation of carbonyl compounds demonstrates the big although largely unexplored potential of chemical reaction engineering in the field of synthesis of fine chemicals.
. Membrane reactors
Membrane reactors have been applied in liquid phase hydrogenations of cinnamaldehyde Analogously to monolithic systems, when the catalyst is impregnated onto the membrane wall the catalyst separation can be avoided. The technical features of membrane reactors will be discussed Catalytic tubular membrane made from α-Al2O3 was coated with Co-Pt/γ-Al2O3 layer and used for cinnamaldehyde hydrogenation The highest selectivity to cinnamyl alcohol was 75% at conversion of 43.7% at 50 °C in ethanol.
Hydrogen permeating palladium membrane was used as a catalyst in furan hydrogenation at 150 °C The conversion of 16% was achieved with tetrahydrofuran as a main product.
Fig. 27. SISR reactor in citral hydrogenation [164].
Fig. 9. HRTEM image of Au/ZrO2 exhibiting a rounded gold particle (A) and a particle with facets (B). Right: separate size distribution of single crystalline (SC) giold particles, single twinned particles (ST), multiple twinned particles (MTP) together with overall size distribution. Lower row: left: HRTEM images of multiple twinned particles of Au/ZrO2, left: gold particle size distribution [96] (reprinted with permission, copyright (2003) American Chemical Society).
7- Conclusions
Recent literature on the selective hydrogenation of unsaturated aldehydes to unsaturated alcohols over heterogeneous catalysts is overviewed. Both the reactant structure and the nature of metal surface affect the possibilities to achieve high selectivities to unsaturated alcohols. The main challenge in chemoselective hydrogenation is to develop a stable catalyst, which is active and selective, since usually there is an inverse relationship between activity and selectivity.
For supported catalysts both metal and support influence activity and selectivity of the final catalyst. Metals can have different electronic properties for instance due to their different origin, exhibiting various properties. Furthermore, several metals or metal alloys can be used as active components and electronic properties can be tuned by adding promoters. The nature of support material, like porosity, pore size distribution and reducibility is an important parameter in catalyst preparation. Besides conventional oxide and activated carbon supports more advanced support materials have been investigated in the synthesis of fine chemicals. Some examples of tailor-made supports are bifunctional zeolites and mesoporous materials, colloidal metals in polymers as well as reducible and alloy forming supports. It should be pointed out that the final catalysts properties are also a result of specific catalyst pretreatment and a required catalyst could be obtained via fine-tuning the catalyst pretreatment procedures. The selectivity to unsaturated alcohols can in general be increased via increasing the number of active sites activating the carbonyl group.
Design of tailor-made catalysts with specific selectivity is based on the knowledge in reaction mechanisms. Recent studies on molecular modeling and adsorption along with kinetic modeling are discussed. Besides being a way of translating our understanding of the reaction mechanism, kinetic modeling is efficiently used for process development. Three-phase fine chemical syntheses are traditionally carried out in batch reactors over conventional slurry catalysts. There exist, as shown in this review, substantial amount of kinetic data from chemoselective hydrogenations, which can be used in process development. Additionally, the knowledge from conversion–selectivity relationship and a comparison of liquid and gas phase hydrogenations serve as tools for understanding the catalytic phenomena on the surface sites. Due to the increasing future demand of fine chemicals there is, however, a need to develop continuous production technologies for three-phase hydrogenations. These technologies require structured catalysts, which have low-pressure drop and avoiding at the same time maldistribution, thus utilization of several structured supports in chemoselective hydrogenation, like fibers prepared from activated carbon, polymer, silica woven glass fibers has a high potential. Recently, membrane reactors and monolithic catalysts have been also applied in the field of fine chemicals. Besides other advantages continuous technology brings additionally a benefit of understanding the catalyst re-use and deactivation in a well-defined, quantitative way.
Analysis of the state-of-the-art in heterogeneous catalytic hydrogenation strongly emphasizes the necessity of a truly multidisciplinary approach in catalyst development, when the materials chemistry is tightly linked on one hand to surface science and molecular modeling and on the other hand to catalytic reaction engineering, including kinetic modeling and reactor selection
In reactions exhibiting a selectivity problem, e.g. hydrogenation of two C=C bonds or C=C versus C=O groups, the advantage of gold catalysts is the preferred hydrogenation of one of these groups leading to monoenes (also by selective C=C hydrogenation), unsaturated alcohols and unsaturated ketones as reaction products important to the chemical industry. In-depth characterization of gold hydrogenation catalysts clearly shows that elucidation of the active site issue and the observed structure-sensitivity cannot be simply undertaken by varying the size of gold particles because the latter may interfere with the portion of multiply twinned particles (MTPs) and the degree of particle rounding. Because a large portion of MTPs gave low allyl alcohol selectivity (product of C=O group hydrogenation) in acrolein hydrogenation, catalysts with single crystalline gold particles must be prepared by appropriate synthesis methods. However, what are the preparative factors, which control the ratio between gold surface atoms on edges and corners versus those on flat surfaces or the portion of MTPs? The important impact of an appropriate pretreatment in H2 controlling the surface structure, and thus, the catalytic properties was shown by the drastic influence of the prolonged reduction time (which favors the formation of multiple-twinned gold particles non-selective for C=O group hydrogenation. This approach shows that in the case of hydrogenations over gold catalysts, the mechanism of catalysis are more complicated than assuming a simple geometric model (gold particle size) or perimeter-interface model, which is suggested to be the active site in CO oxidation The modification of gold particles by a second metal (indium) was not only a suitable tool to increase the intramolecular selectivity, but was also essential to identify the active sites for this hydrogenation reaction.
No direct evidence has yet been demonstrated whether hydrogen is dissociatively or molecular adsorbed, whether subsurface hydrogen (in analogy to subsurface oxygen in oxidation catalysis) plays a role as indicated at higher coverages, and thus, what the active sites for hydrogen activation are. On carbon (type: Black Pearls, Cabot GmbH) supported gold nanoparticles (dAu = 3.7 ± 1.4 nm), the amount of chemisorbed hydrogen was definitively zero = At present, experimental results indicate, that H2 chemisorption occurs only on Au surface atoms of coordinatively unsaturated sites. Since the adsorption capability of hydrogen on (clean) gold surfaces is limited, the application of hydrogen chemisorption measurements is difficult on gold catalysts, and thus, remains an advanced task. Furthermore, it has to be considered that also surface strain that can be induced in small, thin gold particles by supports =may be responsible for the activation of hydrogen and the hydrogenation properties of supported gold catalysts. In the case of oxygen adsorption and activation on gold nanoparticles, their greater reactivity was attributed to a higher step density and a greater amount of surface strain These factors have to be elucidated in the future with the help of gold model catalysts, in situ techniques =and theory, giving rise to a deeper insight into kinetics and mechanism of gold-catalysed hydrogenations as a key toward future applications.
Thanks to the method of deposition-precipitation with urea, it has been possible to prepare a series of Au/TiO2 samples with the same gold loading ( 8 wt%), and various average particle sizes (1.7 to 8.7 nm), either by varying the duration of the preparation, or by changing the conditions of reduction (nature of the gas, temperature, flow rate). With this series of Au/TiO2 catalysts, it has been possible to study systematically the influence of the gold particle size on their catalytic properties in the hydrogenation of crotonaldehyde, under mild experimental conditions ( 100 °C, 0.1 MPa), already used in former studies on Pt-based catalysts. The selectivity to carbonyl bond hydrogenation (crotyl alcohol formation) was twice as high as for olefinic bond hydrogenation (butanal formation). The selectivity was not sensitive to the reduction temperature, as was the case for platinum catalysts whose selectivity to crotyl alcohol was generally much lower (<50%) than that for gold. The gold catalysts led to very reproducible results. It must be pointed out that the Au/TiO2 catalysts were much less active by about two orders of magnitude than Pt/TiO2 catalysts. The structure sensitivity of the Au/TiO2 catalysts for the hydrogenation of crotonaldehyde was clearly shown by the significant increase in activity and TOF for gold particles of 2 nm. The possible role of the oxygen vacancies of TiO2 at the metal-support interface, whose number is known to increase with the reduction temperature, was not evident as in the case of Pt/TiO2. It is proposed that the dissociation of hydrogen on the low-coordinated sites of gold particles (edges and corners) is the rate-determining step of this reaction, and that crotonaldehyde is adsorbed on gold particles through a η3 or η4 mode of adsorption.
Hydrogenation of α,β-unsaturated aldehydes into the corresponding unsaturated alcohols is highly desirable from an industrial point of view. On the other hand, the production of the saturated aldehydes from the unsaturated ones has some industrial and biological applications. It is also thermodynamically preferred and can be achieved easily compared to the former one. The selectivity of a hydrogenation catalyst towards the production of unsaturated alcohols rather than the saturated aldehydes may be improved by using additives, such as transition metal salts, promoters, different supports, different solvents, or alloying the base metal of the catalyst with another one. This modifies its electronic and chemisorption properties towards a better adsorption of C=O bond rather than C=C bond, while preserving the hydrogenation activity of the catalyst . For example, Pd catalysts are very active in many hydrogenation reactions, but are poorly selective towards the production of unsaturated alcohols from their unsaturated aldehyde counterparts . In contrast, Sn and Ir have a relatively low activity, but are highly selective in this respect
Unsaturated alcohols of the allyl-type are valuable fine chemicals, widely used in industry to produce perfumes, pharmaceuticals, and flavoring materials. Selective hydrogenation of unsaturated aldehydes with conjugated double bonds is an important route to unsaturated alcohols However, on conventional catalysts of group VIII metals, α, β-unsaturated aldehydes preferentially undergo hydrogenation of the C C bond to yield saturated aldehydes (as initial product) while the selectivity to allylic alcohols is very low. Hydrogenation of the C O bond is disfavored with respect to hydrogenation of the adjacent C C bond, both thermodynamically and kinetically, the selective hydrogenation of the C O bond represents a challenge which is somewhat facilitated for higher aldehydes R1R2C CH CH O, where bulky substituents Ri, interacting with the catalyst, may sterically hinder the activation of the C C bond , The lowest α, β-unsaturated aldehydes – crotonaldehyde (butenal, R1 CH3, R2 H) and especially acrolein (propenal, R1 R2 H) – do not exhibit such steric hindrance and the transformation to unsaturated alcohols is even more difficult. To enable the desired hydrogenation pathway in these cases, the catalyst is assumed to provide a geometry in which the C O bond of the adsorbed aldehyde reactant is notably activated compared to the C C bond There is experimental evidence that enhancing the surface coverage can also facilitate the formation of such desired structures, when the reactant–catalyst interaction is sufficiently strong, as on Pt
1-Reactant structure
Reactant structure can affect the maximum selectivities for desired products obtained over heterogeneous catalysts. The branching in the vicinity of ethylenic double bond favors high selectivities for unsaturated alcohols, which means that the selectivities for unsaturated alcohols decrease in the following reactants: cinnamaldehyde>3methylcrotonaldehyde > crotonaldehyde > acrolein
There are several reasons, which might affect the product selectivity with different substrates, like electronic and inductive effects in the reactant, adsorption mode of the reactant and the geometrical restrictions of the metal surface. Adsorption mode of the reactant can change with changing the initial concentration of the reactant. Moreover the metal surface has geometrical limitations for reactant adsorption, i.e. metal particle size and shape, as well as amount of steps, edges and kinks is essential for high selectivity. The metal surface properties are discussed in Section.
The selectivities to unsaturated alcohols starting from different substrates over conventional monometallic catalysts are reported. Besides unsaturated aldehydes also data on aromatic aldehydes are presented. Additionally, the maximum selectivities to the desired products over heterogeneous catalysts are given. Citral (1) hydrogenation has been investigated in the liquid phase over different catalysts. In citral hydrogenation the desired products can be unsaturated alcohols, nerol (2) and geraniol (3) or citronellol (4) as well as unsaturated aldehyde citronellal (6). Over Ru/C the selectivities to nerol and geraniol are quite low (, but with Pt/SiO2 the selectivity is enhanced to 76%. Recently very high selectivities to nerol and geraniol have been reported by Milone et al, Reyes et al. and Malathi et al. over Au/Fe2O3 Ce/AC Pt-Sn/MgO, Ir/TiO2, Pt/TiO2 In the two latter catalysts, the role of reducible support is important The bimetallic catalysts, like Ir-Ge/TiO2–HTR, were both active and selective. The higher activity of the bimetallic catalyst was attributed to the presence of ionic species of Ge Very high selectivity to nerol and geraniol, 95.2% at 95.6% conversion was obtained over Ru-Fe/C catalyst in citral hydrogenation in methanol under 40 bar hydrogen . The reaction mixture contained, however, a basic additive, triethyl amine, which can increase the chemoselectivity. Similarly bimetallic Pt-Sn/MgO catalyst was active and selective for producing nerol and geraniol in a continuous trickle bed reactor One drawback with this catalyst was the long induction period of 20 h, during which the conversion and selectivity increased. The third desired product from citral hydrogenation, an unsaturated alcohol citronellol can be achieved over Ni/Al2O3 catalyst with the selectivity of 95% in ethanol at 70 °C. The catalyst pretreatment is an important factor for achieving high selectivities to citronellol over this catalyst in alcohol solvents. Citronellal, a partially unsaturated aldehyde, which is used as an intermediate for production of fragrances, can be synthesized over supported Pd catalysts, like Pd supported on polyethylene fibers containing grafted 4-vinylpyridine groups, and Pd/SiO2/AlPO4. In general, Pd as the main metal is very selective for hydrogenating aliphatic unsaturated aldehydes, like citral to citronellal with selectivity of 86% and minor formation of alcohols. Pd/C catalyst was used in citral hydrogenation to citronellal in methanol and in the presence of triethylamine as additive the selectivity to citronellal was 94% at 99.5% conversion.
2_
Type of catalyst
Heterogeneous catalysts used in hydrogenation reactions contain usually a metal supported on a carrier. The metal is able to adsorb hydrogen thus making hydrogenation reaction possible. The carrier is able to disperse the metal to smaller particles as compared to bulk metal enhancing the specific metal surface area. Additionally, smaller metal particles partially behave as non-metals, e.g. have higher electron densities, leading to higher hydrogenation rates compared to larger particles. In addition to electronic effects, which are different for different metals, geometric properties of the metal particles can affect the hydrogenation rate and selectivity if the size of reacting molecule is close to the size of metal crystallites. The electronic properties of the active metal can be modified by a promoter. In this paper the promoter is defined very broadly, i.e. a promoter can be a non-inert support, second metal, oxide or other additive, which can change the geometric and the electronic structure of the metal surface. The metal phase composition and the amount of edges and corners vary with different metal particle sizes. Activity and selectivity depend not only the metal type and dispersion but also on the support. The catalyst carriers should be able to disperse and stabilize the active metal. Support materials can exhibit inert, basic or acidic properties. Furthermore, non-inert carriers can form an alloy with an active metal. In this section, metal selection and the recent catalytic results by using monometallic and bimetallic catalysts from different groups are presented. A separate section is devoted to catalysts containing both one and several metals on reducible supports. Furthermore, different methods for characterization of bimetallic catalysts are shortly discussed. The most important catalyst properties, like metal particle size and shape, metal precursor, support selection, catalyst pretreatment and the effect of modifiers and promoters and their influence on the chemoselective hydrogenation of carbonyl compounds are reviewed below.
2.1. Metal selection
Different geometric and electronic properties of metals can affect the hydrogenation activities and selectivities by influencing not only surface reactions, but also adsorption, the latter being a prerequisite of any heterogeneous catalytic reactions. The adsorption phenomenon itself is limited by many factors, e.g. by the mode of reactant adsorption on a specific metal surface. Different crystal phases of metals have different geometrical
Pt(1 1 1) surface is a flat, dense surface, while number of edges and kinks is increasing with decreasing metal particle size. Transition metals, which are used in catalysts, have partially filled or filled d-orbital. When the d-orbital is filled, the metal is in relative scale catalytically inactive, like in case of silver, gold and copper. The interstitial electron density in metals having filled d-orbital is low and they are not eager to form bonds. However, the electronic structure of the metal can be changed by adding a second metal, leading for instance to alloy formation, by changing the metal particle size or by enhancing interactions with the support.
Generally there has been a correlation between the width of the d bands of metal catalysts and selectivities in the hydrogenation of α,β-unsaturated aldehydes. The d bandwidth increases in the following order Pd < Pt < Ir Os. The repulsive interaction of the metal with the C C bonds is stronger with the metals exhibiting larger d bandwidth. Thus, the chemisorption of C C bond becomes less probable. The hydrogenation activities and selectivities are presented for different groups in the periodic table by starting from group 8. Note that the selectivities to the desired products over metal supported catalysts depend not only on the metal, but also on the reactant structure. In some hydrogenation reactions, like in the hydrogenation of citral (8), the monometallic supported catalysts are very selective, whereas selective hydrogenation of acrolein and crotonaldehyde cannot be achieved over conventional supported monometallic catalysts. Additionally, it could be pointed out that large selectivity changes have been observed in, e.g. citral hydrogenation over a monometallic Pt/SiO2 catalyst just by changing the reaction temperature
One of the most studied reactions of hydrogenation with gold catalysts is the selective hydrogenation of α,β-unsaturated aldehydes Indeed, the selectivity in unsaturated alcohols, the desired products, is difficult to achieve (i) because thermodynamics favors the hydrogenation of the C=C over the C=O bond by about 35 kJ mol−1 and (ii) because for kinetic reasons, the reactivity of the C=C bond is higher than that of the C=O bond. Hence, on conventional supported hydrogenation catalysts (Pt, Ru, Ni, etc.), the main products are the saturated aldehydes. On supported gold catalysts, interesting selectivities to unsaturated alcohol were found, even if the results obtained from one research group to another are not always fully consistent. The selective hydrogenation of crotonaldehyde (but-2-enal) was first studied by Bailie and Hutchings at atmospheric pressure and 250 °C, with Au/ZnO and Au/ZrO2 prepared by coprecipitation, and with Au/SiO2 prepared by impregnation. They found that Au/ZnO was more selective for C=O bond hydrogenation and that the addition of thiophene acted as a promoter for selective hydrogenation. Haruta and colleagues studied the same reaction but at high pressure (1.0 MPa) and 227 °C, with Au/Al2O3 and Au/TiO2 catalysts prepared by deposition-precipitation with NaOH and gold particles sizes of 4.6 and 3.4 nm, respectively, and Au/SiO2 (5.3 nm) prepared by grafting of gold acetyl acetonate. They also found that the selectivity depends on the nature of the support. The highest selectivity to crotyl alcohol was obtained over Au/TiO2 (≈25% instead of <10% for the other samples). Bailie et al. studied the influence of the particle size on the selectivity in the Au/ZnO system. Gold particle sizes were increased by increasing the gold loading in the samples (2 wt% (2–4 nm), 5 wt% (4–5 nm), and 10 wt% (bimodal distribution: 1–3 and 6–10 nm)), or by increasing the reduction temperature of the 5 wt% Au/ZnO catalyst (250 °C (4–6 nm), 300 °C (3–8 nm), 350 °C (10–15 nm), and 400 °C (up to 20 nm)). They reported that the selectivity to crotyl alcohol increased when the gold loading or the reduction temperature increased, and they proposed that the active sites for selective hydrogenation of crotonaldehyde to crotyl alcohol were associated with the presence of large Au particles. Claus et al. also studied the effect of the particle size in a series of three Au/TiO2 catalysts (1.1, 2.0, and 5.3 nm) prepared by three different preparation methods in selective hydrogenation of acrolein under high pressure (2 MPa) and at 240 °C. They observed a drastic decrease in the TOF by two to three orders of magnitude and a decrease of selectivity when the size of gold particles decreased from 2.0 to 1.1 nm, and constant values of TOF and selectivity for average particle sizes of 2.0 and 5.3 nm. They proposed that the structure sensitivity of the reaction in the case of small gold particles was due to quantum size effects, i.e., to alteration of the metallic properties of the smallest gold particles. In a more recent study, Claus and colleagues prepared three Au/ZrO2 catalysts by the same method of deposition-precipitation, using various pH's, to obtain catalysts with various gold particle sizes (4.0, 6.9, and 7.7 nm). They found that the TOF decreased by nearly one order of magnitude, and the selectivity to allyl alcohol increased (from 15 to 35%) when the particle sizes increased. On the other hand, they observed by HRTEM, multiply twinned particles whose number increased with increasing reduction temperature (300 to 500 °C) or reduction time (3 to 18 h). They pointed out the fact that when the number of multiply twinned particles increased, both the selectivity and the TOF decreased. They attributed the higher activity of Au/TiO2 compared to that of Au/ZrO2 (TOF×2) to the fact that the gold particles were round shaped on TiO2 and facetted on ZrO2 (average particle size of 5 nm on both supports). Round-shaped gold particles on TiO2 would contain a higher proportion of low-coordinated surface sites, which would explain the higher TOF.
As shown above, the influence of the gold particle size and of the nature of the support on the activity and selectivity for the hydrogenation of α,β-unsaturated aldehydes is not clear, yet. To attempt to clarify the influence of the first parameter, we decided to investigate the dependence of the catalytic properties with gold particle size in Au/TiO2 catalysts prepared by a single method, the deposition-precipitation with urea The main advantage of this method over the method of deposition-precipitation with NaOH developed by Haruta and colleagues] is that all the gold of the solution is deposited on TiO2, so higher gold loading can be reached with the same gold particle size as with Haruta's method, and the gold loading of the catalysts can be easily controlled. The results obtained are therefore compared with those obtained on Au/TiO2 catalysts prepared by deposition-precipitation with NaOH. The reaction studied is the hydrogenation of crotonaldehyde (CH3---CH=CH---CH=O) at atmospheric pressure, under the same experimental conditions as those used in former studies for Pt/TiO2 catalysts Hence, comparisons of the catalytic behavior between platinum and gold catalysts could be made, especially the influence of the particle size and that of the reduction temperature, since these two parameters are the key parameters which govern the activities and selectivities of Pt/TiO2 catalysts.
2.1.1. Monometallic catalysts
In the conventional monometallic catalysts the metals are selected from the group 10, e.g. Ni, Pd and Pt. In addition, Rh from group 9 and Ru from group 8 have been used in catalysts. Very high selectivities to unsaturated alcohols (83–100%) at high conversions have also been reported over a pyridine poisoned Re catalyst which belongs to group 7. The supports are usually alumina, silica and carbon. The catalytic activity of different metal supported catalysts in hydrogenation is determined by ability to activate C C and C O bonds as well as the activity of hydrogen to react on the metal surface. Hydrogen activity on different metals is relatively well understood on the basis of adsorption studies and fundamental theories. According to Masel, hydrogen is reactive in the surfaces of Co, Ni, Ru, Rh, Pd, Os, Ir, Pt as well as on Sc, Ti, V, Y, Zr, Nb, Mo, La, Hf, Ta, W, Cr, Mn, Fe, Tc and Re. A slower uptake of hydrogen was observed with Cu, whereas Ag and Au were inert. Interestingly to note that ability of hydrogen to react and to dissociate on the surfaces of Au and Ag depends on the size of metal nanoparticles. Exceptional catalytic activities of gold nanoparticles in hydrogenation and oxidation reactions were recently reported Moreover the filling of d-orbital affects very much the adsorption mode. The two latter metals, Au and Ag have completely filled d orbitals, 5d10 and 4d10, and thus they exhibit very low catalytic activities. No adsorption can occur on the metal having a low interstitial electron density Generally there has been a correlation between the width of the d bands of metal catalysts and selectivities in the hydrogenation of α,β-unsaturated aldehydes. The d bandwidth increases in the following order Pd < Pt < Ir Os. The repulsive interaction of the metal with the C C bonds is stronger with the metals exhibiting larger d bandwidth. Thus, the chemisorption of C C bond becomes less probable. The hydrogenation activities and selectivities are presented for different groups in the periodic table by starting from group 8. Note that the selectivities to the desired products over metal supported catalysts depend not only on the metal, but also on the reactant structure. In some hydrogenation reactions, like in the hydrogenation of citral (8), the monometallic supported catalysts are very selective, whereas selective hydrogenation of acrolein and crotonaldehyde cannot be achieved over conventional supported monometallic catalysts. Additionally, it could be pointed out that large selectivity changes have been observed in, e.g. citral hydrogenation over a monometallic Pt/SiO2 catalyst just by changing the reaction temperature
2.1.1.1. Chemoselective hydrogenation over group 8 metals, Ru, Os
The group 8 metals are electropositive, because they have unfilled d orbitals. Supported ruthenium catalysts have been recently used in chemoselective hydrogenations, like in hydrogenation of cinnamaldehyde, crotonaldehyde and citral
Os has been used in very few papers as a catalytic metal supported on SiO2 and on TiO2. In the hydrogenation of oxopromegestone Os was the most selective among tested metals, but the catalytic activity was low. Catalytic activity of Os in hydrogenation reactions is, however, well documented for a long period of time. For instance, F. Haber applied Os catalysts, in ammonia synthesis already a century ago. In citral hydrogenation at 27 °C under 1 bar, hydrogen in hexane the selectivity to nerol and geraniol was close to 100% at 5% conversion. The initial TOF-values were, however, very low over Os/SiO2 catalyst
The performance of Ru/C and Ru/Al2O3 catalysts was compared in citral hydrogenation They exhibited the same TOF in gas phase hydrogenation of crotonaldehyde, but alumina supported Ru catalyst was more selective for crotyl alcohol than Ru/C. The selectivities were however relatively low, i.e. 17 and 3%, respectively. At the same time these two catalysts afforded in citral hydrogenation at the conversion level of 70% selectivities to nerol and geraniol 54 and 38%, respectively The selectivities in citral hydrogenation at 60 °C are additionally lowered by simultaneous side reactions, like cyclisation and acetalization of citronellal, when the solvent was 2-propanol. The Ru/C catalyst exhibited three times higher activity than Ru/Al2O3 calculated per gram of Ru Ru/C was tested also in citral hydrogenation in the work of Mäki-Arvela et al.. This catalyst gave about 22% selectivity to nerol and geraniol in citral hydrogenation in 2-pentanol at 70 °C at 75% conversion. The lower selectivity in compared to might be explained by the origin of the metal precursor, in the former work the precursor was chloride whereas in the latter one it was nitrate. In particular, chloride precursor is known to favor acetalization, which in principle could be avoided in 2-pentanol. However, in hydrophobic 2-pentanol quite large amounts of isopulegol (12%) are formed. High selectivities for nerol and geraniol over Ru/Al2O3 and Ru/TiO2, 47.6 and 73.4% at 100% conversion, respectively, were obtained in citral hydrogenation in n-heptane at 126 °C and 50 barThe selectivity to nerol and geraniol was 73.4% over Ru/TiO2 catalyst reduced at high temperature, when the hydrogenation was carried out at 126 °C and at 50 bar in n-heptane as a solvent. In the selectivity to nerol and geraniol was 58% at 5% conversion of citral in citral hydrogenation at 27 °C under 1 bar hydrogen in hexane. The initial hydrogenation rate, however, over this catalyst was quite low. Relatively low selectivities were obtained over Ru supported catalysts. Only 13% selectivity of cinnamyl alcohol at 20% conversion was obtained over Ru/SiO2 catalyst in 2-propanol at 60 °C and 10 bar, whereas Ru/Al2O3 catalyst gave higher selectivity (about 40%) to cinnamyl alcohol. Cinnamyl alcohol selectivity over Ru/CNF catalyst (CNF stands for carbon nanofiber) was 48% at 60% conversion of cinnamaldehyde at 110 °C and 45 bar in 2-propanol
2.1.1.2. Chemoselective hydrogenation over group 9 metals, Co, Rh, Ir
Monometallic catalysts containing a metal from group 9 have been used in several chemoselective hydrogenations. Applications of Rh and Co and Ir and were reported. Monometallic catalysts including a metal from group 9 can act quite selectively in some cases, like in citral, acetophenone and cinnamaldehyde hydrogenations. The achieved selectivities depend, however, very much on the reactant structure. Ir catalysts have been very selective to unsaturated alcohols, but rather inactive.
Monometallic Rh catalysts have been relatively selective in acetophenone and in cinnamaldehyde hydrogenations, but low selectivities were obtained in crotonaldehyde, citral and cyclohexenone hydrogenations. Acetophenone hydrogenation over Rh/C catalyst resulted in about 70% selectivity to unsaturated alcohol at 25 bar and 80 °C in cyclohexane. Similarly, 70% selectivity to cinnamyl alcohol was achieved over Rh/SiO2 catalyst at 50 °C and 1 bar in ethanol In case of crotonaldehyde the selectivities to crotyl alcohol were low over Rh/SiO2 and over Co/SiO2, i.e. Rh/SiO2 catalyst yielded relatively low selectivities to crotyl alcohol (18% over Rh/SiO2 catalyst prepared by sol–gel method from Rh(acac)) at 100 °C, and 2% over 0.5 wt.% Rh/SiO2 at 35 °C (Rh from RhCl3 prepared by impregnation). The initial selectivity to crotyl alcohol was, however, close to 100% at the conversion of 0.2% in crotonaldehyde hydrogenation at 40 °C over a calcined Rh/Al2O3 catalyst indicating that catalyst pretreatment is important The selectivity to crotyl alcohol decreased linearly with increasing time-on-stream. In citral hydrogenation over Rh/Al2O3 the main product was citronellol (60% selectivity) in 2-pentanol at 70 °C and 1 bar and below 1% nerol and geraniol were obtained Polymer supported Rh catalyst gave as a major product citronellal No unsaturated alcohols were also formed over Rh/SiO2 catalyst in the hydrogenation of 2-cyclohexenone at 25 °C and at 6.8 bar in 2-propanol. The reason for this is the presence of sterically hindered carbonyl bond. In over Rh/SiO2 citronellal was the main product in citral hydrogenation at 27 °C under atmospheric pressure in hexane. Polymer supported Rh catalyst gave as a major product citronellal.
As the d bandwidth of Ir is larger than for Pd or Pt, it could be expected that the hydrogenation of ethylenic bond decrease. In fact Ir has been very selective in citral hydrogenation to nerol and geraniol, but the drawback has been in the low reaction rate Ir/SiO2 was very selective in the hydrogenation of citral at 70 °C and 4.1 bar in ethanol affording 100% selectivity of nerol and geraniol. On the other hand only 7% selectivity to nerol and geraniol was observed in citral hydrogenation at 27 °C and 1 bar hydrogen in hexane at 23% conversion It might be that the solvent also has an effect on selectivityOn Ir/TiO2 selectivity to nerol was also 100% but over both catalysts the total conversion was very low. Ir/graphite catalyst was very selective to cinnamyl alcohol in the hydrogenation of cinnamaldehyde at higher temperatures, i.e. 110 °C and 130 °C in the pressure range of 10–30 bar in 2-propanol Additionally, the selectivity to cinnamyl alcohol increased with increasing conversion.
Co/SiO2 resulted in maximally 20% selectivity of crotyl alcohol at 120 °C and 1 bar in gas phase. In general, however, it can be stated that Co being not a noble metal, is more selective to unsaturated alcohols than Rh, Ru or Pt. The selectivity to unsaturated alcohols can be tuned by changing the surface structure of Co, which exists in four different forms Relatively high selectivities to nerol and geraniol (55%) were obtained at 5% conversion over Co/SiO2 in citral hydrogenation under 1 bar hydrogen at 27 °C in hexane.
2.1.1.3. Chemoselective hydrogenation over group 10 metals, Ni, Pd, Pt
Platinum has been the most intensively used as the active metal in chemoselective hydrogenations, , whereas Ni and Pd supported catalysts are less applied. Some recent results in chemoselective hydrogenations over Pd, Ni and Pt supported catalysts are reviewed below.
Pt is not intrinsically selective to produce unsaturated alcohols, although high selectivities to unsaturated alcohols could be still achieved depending very much on the sterical structure of the reactant. No allyl alcohol was formed in the gas phase hydrogenation of acrolein over Pt/black at 50 °C and the main product was propanal Relatively high selectivities have, however, been achieved in cinnamaldehyde, citral and furfural hydrogenations. In cinnamaldehyde hydrogenation over Pt on graphite and Pt/SiO2 catalysts the selectivities to cinnamyl alcohol were 81.5% in toluene/water mixture at 25 °C and 80% in ethanol. Similarly high selectivity (92%) was obtained over Pt supported on NaY zeolite; the catalyst was prepared by ion-exchange. Even higher selectivity, about 99% towards cinnamylalcohol has been reported over montmorillonite supported Pt
High selectivities to nerol and geraniol can be achieved in citral hydrogenation over Pt/TiO2 at 45 °C and 1 bar in ethanol and Pt/SiO2 catalysts at 150 °C and 20 bar in hexane, i.e. 100% (selectivity to geraniol) and 76%, respectively Furfural can be selectively hydrogenated to furfuryl alcohol (96% at 58% conversion) in 2-propanol–water solvent at 150 °C and 20.6 bar hydrogen over a Pt/C catalyst
Pd differs from the other metals in group 10 from the selectivity point of view, as it is very selective to hydrogenate ethylenic bond. It is believed in that the reason for this is the small d bandwidth of Pd, which affects the adsorption mode of the reactant. Pd favors tetrahepto di-π geometry through both double bonds. Citronellal was the major product in citral hydrogenation over Pd/polymer catalyst at 70 °C and 4.5 bar in 2-propanol as well as over Pd/SiO2 catalyst at 27 °C under 1 bar hydrogen in hexane One exception is hydrogenation of a carbonyl group in aromatic compounds, where selective hydrogenation of carbonyl group can be achieved. For instance 95% selectivity to benzyl alcohol was reported by in benzaldehyde hydrogenation over a cross-linked Pd polymer catalyst at 25 °C and 10 bar hydrogen pressure in methanol. In the liquid phase hydrogenation of cinnamaldehyde over a Pd/SiO2 catalyst, however the maximum selectivity to cinnamyl alcohol was around 32% at 90% conversion at 60 °C and 10 bar in 2-propanol
Ni supported catalysts are very selective in some chemoselective hydrogenations. Ni favors formation of citronellol in citral hydrogenation at 70 °C and 1 bar in ethanol Citronellal was, however, the main product in citral hydrogenation over Ni/SiO2 in hexane at 27 °C under 1 bar hydrogen. The selectivity to citronellal was 77% at 48% conversion. Acetophenone and benzaldehyde hydrogenation was very selective to unsaturated alcohols over Ni/H-Y and Ni/SiO2 ] catalysts, respectively. In acetophenone hydrogenation at 100 °C and 30 bar the highest selectivity to 1-phenylethanol was 100% in n-hexane, but the conversion remained rather low (10%) In alcohol solvents the highest selectivity to 1-phenylethanol was 74% in methanol under the same reaction conditions. It should, however, be pointed out that Ni/Y catalyst was not stable after catalyst recycling, whereas the bimetallic Ni-Pt/Y catalyst exhibited constant activity after recycling. Selectivities close to 100% in benzyl alcohol were achieved in gas phase hydrogenation of benzaldehyde over Ni/SiO2 catalyst at 120 °C. Chemoselective crotonaldehyde hydrogenation is more difficult than previous cases over monometallic supported Ni catalysts. Maximally 59% selectivity to crotyl alcohol was obtained in crotonaldehyde hydrogenation at 60 °C and 1 bar over Ni/TiO2 reduced at 500 °C The relatively high selectivity over Ni/TiO2 was achived due to the existence of interfacial Ni-TiOx sites exhibiting strong metal support interactions The promoted Ni-P/SiO2 catalyst was very active in benzaldehyde hydrogenation at 100 °C and 10 bar in ethanol, but the only reaction product was toluene The reason for this could be the low pH of the reaction mixture (pH 4).
2.1.1.4. Chemoselective hydrogenation over group 11 metals, Cu, Ag, Au
Catalysts having an active metal from group 11 are usually less active than those having a metal from the other groups. This originates from the electronic structure of the d band, which is filled. Recent catalytic results from chemoselective hydrogenations using copper; silver and gold are briefly overviewed here.
Supported monometallic copper catalysts have been used in chemoselective hydrogenations. Cu/SiO2 was very selective towards benzyl alcohol (83%) in benzaldehyde hydrogenation in gas phase at 100 °C at conversion of 68%. No benzyl alcohol was formed over Cu catalysts supported on TiO2, ZrO2, γ-Al2O3 and traces of benzyl alcohol were formed on Cu/CeO2 The highest selectivity to cinnamyl alcohol over 13.7 wt.% Cu/SiO2 was 48% at 60% conversion level of cinnamaldehyde in decalin at 140 °C. The highest selectivity to cinnamyl alcohol was obtained over a Cu/SiO2 prepared by impregnation. The catalyst exhibited large copper crystallites and small amounts of Cu2O at the metal support interface. On the other hand high selectivity to cinnamyl alcohol (78%), but very low activity in cinnamaldehyde hydrogenation was obtained over Cu/ZrO2 catalyst at 140 °C and under atmospheric pressure in propylene carbonate. Very high conversion of furfural and selectivity to furfuryl alcohol (98%) was obtained over Cu/MgO catalyst in the gas phase hydrogenation at 180 °C. The reason for high activity and selectivity was stated to be the presence of defective sites at Cu and MgO in interfacial region.
Silver supported catalysts have been used in the hydrogenation of acetophenone, acrolein, crotonaldehyde and furfural. Unusually high selectivities to unsaturated alcohols have been reported over nanosized silver catalysts. According to Claus, silver is able to adsorb conjugated double bond via bifunctional bonding to both C C and C O bond resulting into formation of 2-propenyloxo intermediate. Possibility of hydrogen adsorption with dissociation was recently confirmed by isotope jumping technique. Selectivity of 94.8% to 1-phenylethanol was reported in acetophenone hydrogenation over Ag/SiO2 catalyst at 140 °C and 2 bar whereas in crotonaldehyde hydrogenation over Ag/SiO2 catalyst at 20 bar and at 140 °C the selectivity to crotyl alcohol was 62.8% at 16% conversion. Ag/SiO2 catalyst was quite selective in the hydrogenation of furfuryl alcohol. The selectivity to furfuryl alcohol was 79% in the gas phase hydrogenation of furfural at 1.9 bar and at 200 °C over 12.1% Ag/SiO2 catalyst.
Gold has been used in the chemoselective hydrogenation of crotonaldehyde and, trans-3-buten-4-phenyl-2-one (benzalaldehyde) acrolein and and citra. Since gold in powder form has been catalytically active in gas phase hydrogenation of acrolein, thus its catalytic activity is not exclusively originated from quantum size effect or from specific metal-support interactions. Quantum size effect alters the electronic properties of a metal particle Moreover the binding energy shifts of core levels have been observed in supported gold nanoparticles. Gold supported on Fe2O3 yielded to very high selectivities of nerol and geraniol in citral hydrogenation at 60 °C in and 1 bar ethanol (95% at 90% conversion of citral) and the selectivity was independent on the metal loading In the gas phase hydrogenation of acrolein at 240 °C and 20 bar the maximum selectivity to allyl alcohol was 37% over Au/ZrO2 catalyst, but the conversion of acrolein was only 5%. The hydrogenation rate was decreasing, while the selectivity to unsaturated alcohol increased with increasing gold particle sizes. The observed structure sensitivity is discussed in detail
2.1.2. Bi- and multimetallic catalysts
Bimetallic catalysts have been very selective in several chemoselective hydrogenations, like Rh-Sn/SiO2 for nerol and geraniol (98%), Pt/ZnO for crotyl alcohol (90%), Rh-Cu/SiO2 for crotyl alcohol (62%) The origin for the high selectivities to unsaturated alcohols has been associated with electron transfer from the less noble metal to more noble metal as well as alloy formation and /or an intimate contact between two metals. This concept was, however, challenged by The second metal can exist as an adatom, in alloy, or in ionic state as well as in partially oxidized form The difference in the electronegativity between two metals can enable the polarization of the carbonyl bond. Additionally, geometrical effects, like a change in metal dispersion decoration of the main metal by the second metal via surface enrichment and organic fragments in the vicinity of an active metal, have been observed in bimetallic catalysts. It is important to note that these effects are often coexistent in bimetallic catalysts and thus one effect, like for instance alloy formation in a model catalyst Sn/Pt(1 1 1), is not alone responsible for improved selectivity because nearly the same selectivities to crotyl alcohol were obtained over non-supported Pt and Pt-Sn model catalysts. Increased selectivities to crotyl alcohol were, however, obtained over supported Pt-Sn catalyst
There exists controversial information of the role of alloys in chemoselective hydrogenation and even bimetallic catalysts non-selective for formation of unsaturated alcohols have been reported Alloys are defined as compounds formed by two or more metals; they can also be formed between a metal and another metal originating from a non-inert support The alloy formation is strongly influenced by the selection of metal precursor and support, catalyst pretreatment as well as the preparation method.
In order to uncover both the selective hydrogenation site and the reaction mechanism it is important to know the chemical nature of the active site. Several physico-chemical methods have been used as tools for characterization of the active sites. The benefits and drawbacks of different characterization methods are discussed in the subsequent section. The following bi- or multimetallic catalysts have been recently investigated in chemoselective hydrogenations from group 8: Ru-Sn, Ru-Fe/C from group 9 Rh-Sn/SiO2, Rh-Cu/SiO2, Rh-Mo/SiO2, Ir-Ge/SiO2, Ir-Fe/SiO2, Mo-Co-B, Sn-Co-B, Ra-Co from group 10 Pt-Sn/Al2O3 and, Pt-Sn/C, Pt-Ge/Al2O3, Pt/ZnO Pt-Sn/MgO, a model catalyst Sn-Pt(1 1 1) without support, Pt-Co/graphite, Ni-P-B, Ni-Cu/SiO2, from group 11 Ag-In/SiO2, Au/ZnO Au-In/ZnO, Cu-Zn-Al, Cu-Ni(Co)-Zn-Al. Additionally, there exist catalysts exhibiting a combined effect of both two metals and reducible oxides, like Pt-Sn/CeO2-SiO2 and Ru-Sn/TiO2. The examples are presented below on a group-by-group basis and the combined effect of bimetallic catalysts on reducible supports is discussed separately
3. Hydrogenation reactions studied using gold catalysts
The literature till the 1960s indicates that gold seems to be very popular for being catalytically inert in hydrogenation. During the following decades most of the work involved mainly the adsorption/activation of hydrogen and hydrogen exchange on gold surfaces, as shown in the previous section, and the hydrogenation of relatively simple molecules, such as CO, CO2, alkenes and alkadienes. Only a very limited number of publications concern the selective hydrogenation of complex organic molecules with more than one unsaturated group to reactions products, which are of interest for the chemical industry. As it was recognized that gold can be prepared as nanoparticles on oxidic supports by wet-chemical techniques, such as deposition–precipitation or coprecipitation methods, a breakthrough came in heterogeneously catalysed hydrogenations using multifunctional organic compounds (e.g. α,β-unsaturated aldehydes, ketones) as educts. In Table 1, a comprehensive overview of the hydrogenation reactions studied together with the gold catalysts used is given considering the literature up to September 2004. It is apparent from the cited reports that supported gold catalysts have in some cases superior prospects for employing them in industrial catalysis. Beside the partial hydrogenation of acetylene to ethylene and 1,3-butadiene to butenes, the main advantage of gold catalysts is their ability to selectively catalyse the hydrogenation of the C=O group of α,β-unsaturated aldehydes producing allyl-type unsaturated alcohols. In this field, attempts of our group to identify the active sites of supported gold catalysts, favoring the adsorption and hydrogenation of the carbonyl group, were successful and the influence of the real structure of the catalyst on the hydrogenation properties was estimated. Both approaches are highlighted in the next section.
Table 1. Hydrogenation of different educts over gold catalysts
Educt
Catalyst
CO
Au/TiO2, Au/Fe2O3, Au/ZnO
Au/TiO2, Au/Fe2O3, Au/ZnO
Au/ZrO2
CO2
Au/TiO2, Au/ZnO
CO2
Au/TiO2, Au/ZnO
Au/ZrO2
NO
Au/MgO, Au/Al2O3, Au/SiO2
Ethane
Au/MgO
Propene
Au/SiO2, Au/MgO
NO
Au/MgO, Au/Al2O3, Au/SiO2
Ethane
Au/MgO
Propene
Au/SiO2, Au/MgO
Au/TiO2, Au/ZrO2, Au/ZnO
1-Butene
Au metal
1,3-Butadiene
Au/Al2O3
1-Butene
Au metal
1,3-Butadiene
Au/Al2O3
Au/SiO2
Au/Al2O3, Au/SiO2, Au/TiO2
1-Pentene
Au/Al2O3, Au/SiO2
Acetylene
Au/Al2O3, Au/TiO2
2-Butyne
Au/boehmite
Cyclohexene
Au/SiO2
1-Pentene
Au/Al2O3, Au/SiO2
Acetylene
Au/Al2O3, Au/TiO2
2-Butyne
Au/boehmite
Cyclohexene
Au/SiO2
Au metal, film, powder
Styrene
Au/MCM-41
Acrolein
Au/SiO2, Au/ZrO2, Au/TiO2
Styrene
Au/MCM-41
Acrolein
Au/SiO2, Au/ZrO2, Au/TiO2
Au/ZnO, Au-In/ZnO
Crotonaldehyde
Au/ZrO2
Crotonaldehyde
Au/ZrO2
Au/TiO2
Au/ZnO, Au/ZrO2
Citral
Au/Fe2O3
Benzal acetone
Au/Fe2O3
Citral
Au/Fe2O3
Benzal acetone
Au/Fe2O3
Full-size table
View Within Article
4. Selective hydrogenation of α,β-unsaturated aldehydes over supported gold catalysts
The selective hydrogenation of α,β-unsaturated aldehydes to allylic alcohols is of both commercial relevance (e.g. in fine chemical and pharmaceutical intermediate production and specific scientific interest Note that thermodynamics favors hydrogenation of the C=C over the C=O group (stronger negative free reaction enthalpy of 35 kJ mol−1), and for kinetic reasons the C=C bond is more reactive than the C=O group. Thus, in the presence of most of conventional monometallic hydrogenation catalysts, based on Pt, Rh or Pd on silica and alumina as supports, lower α,β-unsaturated aldehydes such as acrolein and crotonaldehyde are hydrogenated predominantly to the saturated aldehydes, and only to a minor extent to unsaturated alcohols Therefore, it is desirable to find catalysts, which are able to control the intramolecular selectivity by hydrogenating preferentially the C=O group, while keeping the olefinic double bond intact.
In our group, the hydrogenation of acrolein) on supported gold catalysts has been used as test reaction to study several aspects of structure sensitivity, i.e., a dependence of activity and selectivity on the size and morphology of gold particles Because of the lack of space-filling substituents (e.g. methyl, phenyl) at the C=C group acrolein is the α,β-unsaturated aldehyde, which is most difficult to hydrogenate to the unsaturated alcohol (allyl alcohol). The latter must be the criterion for the evaluation of the unusual properties of gold as hydrogenation catalyst. Moreover, from an industrial point of view allyl alcohol is much more important for the chemical industry (e.g. used for glycerol synthesis, allyl ester, epoxidation) than, for example, crotyl alcohol which can be obtained by selective hydrogenation of crotonaldehyde over supported gold catalysts. However, it must be kept in mind that in the case of crotonaldehyde (or higher α,β-unsaturated aldehydes), the selectivities reported do not represent the intrinsic selectivity of gold because a methyl (or phenyl) group as substituent on the C=C double bond contributes to the overall selectivity by steric repulsion of the olefinic group changing the adsorption configuration of the aldehyde as shown by a number of experimental data and theoretical calculations,. This is obvious from where the hydrogenation of acrolein and crotonaldehyde over a Au/SiO2 catalyst (dAu = 3.9 ± 2.3 nm, prepared by incipient wetness are compared. Also, in the case of Au/TiO2 catalysts prepared by deposition–precipitation methods the selectivity to the unsaturated alcohol was decreased from a level of 60–70% when using crotonaldehyde to 30% in the case of acrolein. Beside gold on oxide supports, also silver catalysts, are the only candidates, which convert acrolein to allyl alcohol with selectivities one order of magnitude higher than Group VIII hydrogenation catalysts. It is interesting to note that the adsorption of acrolein on both metals (polycrystalline gold, Au (1 1 1), evaporated silver films) was studied by infrared reflection absorption spectroscopy (IRAS) showing a coverage-dependent discrete adsorption geometry. It can be concluded that the adsorbate interacts with the metal surface through the oxygen atom of the C=O group coordinated to a positively charged surface site (kinks, steps).
Fig. 1. Reaction network of acrolein hydrogenation.
View Within Article
Reaction scheme for crotonaldehyde hydrogenation
Table 1.
Comparison of the gas phase hydrogenation of acrolein and crotonaldehyde over 1.6 wt.% Au/SiO2-IW catalysta
Acrolein
Crotonaldehyde
Conversion (%)
2.7
6.5
2.7
6.5
Selectivities (%) to
Unsaturated alcohol
23.3
38.9
Saturated aldehyde
74.0
32.4
Saturated alcohol
1.4
1.4
Hydrocarbons
1.3
1.0
Other productsb
–
26.3
Unsaturated alcohol
23.3
38.9
Saturated aldehyde
74.0
32.4
Saturated alcohol
1.4
1.4
Hydrocarbons
1.3
1.0
Other productsb
–
26.3
Full-size table
a T = 260 °C, ptotal = 2 MPa, molar ratio H2/educt = 20, = 15.3 g h mol−1, = 19 g h mol−1.
b Sum of 2-butanone (via oxygen-π-adsorbed intermediate) and 2-ethylhexanol (via condensation of butanal to 2-ethylhex-1-enal and its subsequent hydrogenation).
View Within Article
However, it has not yet been possible to identify the catalytic phenomena in hydrogenation reactions associated with the size and the morphology of supported gold particles. Therefore, extensive reaction studies coupled with in-depth catalyst characterization were undertaken by our group to focus on the problem of structure-sensitivity and gaining more knowledge of how the unique catalytic behavior of gold can be controlled by the nanostructure of the catalyst. Beside a dependence of the size of the gold particles on the kind of support, the latter has an influence on the degree of rounding of the particles. shows representative gold particles of nearly the same sizes on different supports (SiO2, TiO2, ZrO2, ZnO) after a pretreatment in hydrogen at 573 K. On SiO2 and TiO2, the particles are nearly spherical, whereas on the other supports extended facets of the gold surface are visible, in particular for Au/ZnO. Structure-sensitivity may have its core in the ability of nanostructured matter to occur in metastable non-equilibrium structures, for example as coordinatively unsaturated sites (e.g. edges, kinks) induced by rough surfaces, as matter with lattice strain or multiple-twinned particles (MTPs) [. The question arises how the hydrogenation properties of gold catalysts are influenced by these nanoscale properties of gold. To answer this question, we focussed our work at first on the influence of the geometrical configuration of the supported gold nanoparticles, namely the occurrence of MTPs and the degree of rounding exhibiting the real structure (Significant differences between the Au/ZrO2 samples prepared by deposition–precipitation at different pH values were found. By correlating the HRTEM and catalytic results, it was shown that higher amount of MTPs resulted in a lowering of selectivity to the desired product allyl alcohol (Table 3), and also the turnover frequency (TOF) was decreased. Therefore, the main demand for the catalyst synthesis is that supported gold catalysts with single crystalline gold particles must be prepared to obtain higher selectivities to the unsaturated alcohol. The increase of TOF for Au/TiO2 compared to Au/ZrO2 is attributed to more rounded gold particles, accompanied by a higher relative amount of low-coordinated surface sites.
Fig. 2. Representative gold particles on different supports (DP: deposition–precipitation, I: impregnation)
View Within Article
Fig. 4. HRTEM image of Au–In/ZnO ([35], left) and surface model (right). Indium preferentially decorates the outer faces of the gold particles while the edges remain uncovered. (B) (left
View Within Article
Fig. 3. Representative HRTEM of two Au/ZrO2-DP samples with a more rounded gold particle (A) and a particle with
Table 3.
Influence of the pretreatment of 2.2 wt.% Au/ZrO2-DP catalyst in H2 on nanostructure and catalytic properties during acrolein hydrogenation in the gas phasea
Tred (K)b
573
573
723
tred (h)c
3
18
3
dAu (nm)
7.7
9.2
8.3
Rel. amount MTPs (%)
25
54
30
Selectivity to allyl alcohol (%)
35
20
37
TOF (s−1)
0.045
0.022
0.045
573
573
723
tred (h)c
3
18
3
dAu (nm)
7.7
9.2
8.3
Rel. amount MTPs (%)
25
54
30
Selectivity to allyl alcohol (%)
35
20
37
TOF (s−1)
0.045
0.022
0.045
Full-size table
a T = 513 K, ptotal = 2 MPa, molar ratio H2/Ac = 20, = 15.3 g h mol−1
b Reduction temperature.
c Reduction time.
View Within Article
After the finding that the hydrogenation of acrolein preferentially occurs on single crystalline particles rather than on multiply twinned ones, in the next step of our investigation a new approach was applied to distinguish different features of the gold particle morphology. By means of HRTEM, the crystal structure of individual gold particles prepared on ZnO (SBET = 15 m2 g−1) was investigated. Most particles (dAu = 9.0 ± 0.3 nm) were single crystalline and practically no MTPs were found. This is an important result because elucidation of the active site issue cannot be simply done by varying the size of gold particles, since the effects of faceting and multiply twinned particles may interfere in other supported gold nanoparticle systems. In the next step, modification of the gold particle surface by indium has been used to vary the active site characteristics of a suitable catalyst. By detailed HRTEM image analysis, the lattice spacings of these bimetallic Au–In particles perpendicular to the outer surface were measured. Since there were distinct contrast differences between both metals due to the atomic number difference, it was possible to distinguish for each atomic layer between both In- and Au-containing phases. There is clear evidence of the formation of a separate indium phase exhibiting a lattice distance close to the (1 0 1) distance of tetragonal indium. Furthermore, HRTEM imaging revealed clearly a selective decoration of gold particle faces has been observed, leaving edges free. Statistical evaluation of appearance of indium on the cubooctahedral gold surfaces was undertaken which supports the selective face decoration by the second metal. EDX analysis revealed a homogeneous distribution of indium on the gold particles and no monometallic indium deposits on the support were found. With the bimetallic Au–In/ZnO catalyst, the desired allyl alcohol is the main product formed with a selectivity of 63% (at T = 593 K and ptotal = 2 MPa). Note that the Au–In system is the first example showing the important impact of a second metal for tuning the selectivity of a gold catalyst in selective hydrogenation of functional groups. The level of selectivity obtained with Au–In/ZnO is one order of magnitude higher compared with conventional mono- and bimetallic catalysts for acrolein hydrogenation and comparable to our best catalyst for the selective hydrogenation of acrolein to allyl alcohol, Ag–In/SiO2 From the experimentally proven correlation between surface structure and catalytic behavior, the edges of single crystalline gold particles (“edgy catalyst” have been identified as active sites for the preferred C=O hydrogenation.
View Within Article
A further example for the importance of gold as hydrogenation catalyst was illustrated by comparing gold with platinum, both supported on zirconia, in acrolein hydrogenation Preliminary hydrogenation experiments showed that over Pt/ZrO2 the selectivity to allyl alcohol was always very low (ca. 2%). In order to ensure that this was not due to the high catalyst activity (acrolein conversion of 43% at 473 K) and the related secondary reactions of allyl alcohol according to, the modified space time () was varied in such a way that a comparison of selectivities for the catalysts Au/ZrO2 and Pt/ZrO2 was possible with almost the same acrolein conversion. Even under these experimental conditions, propanal was the main product with Pt (S > 90%). Selectivity to allyl alcohol was only 3.9%, i.e., only one-tenth of the selectivity obtained with Au/ZrO2. This finding clearly indicates that gold catalysts are more efficient than platinum catalysts in acrolein hydrogenation. If the α,β-unsaturated aldehyde as educt is changed from acrolein to, for example, crotonaldehyde, then monometallic Pt catalysts (e.g. Pt/ZnO) can also exhibit high selectivity towards the allylic alcohol. However, in that case, as noted above, the increased selectivity is due to the additional steric effect of the methyl group and cannot be ascri3.4. Catalytic behavior
*. Crotonaldehyde hydrogenation
The performance of the catalysts (previously reduced at 473 K) was tested in the vapor phase hydrogenation of crotonaldehyde at 333 K. The evolution of catalytic activity (TOF, calculated in basis of CO uptakes at 300 K) with time on stream for the three catalysts studied is showed in. It can be seen that, after a period of 20 min with a strong deactivation, the activity remained quite stable with time. This fast deactivation of the catalysts is typical of this reaction and has been previously related to the decarbonylation of the reactant molecule yielding carbon monoxide, which is irreversively adsorbed on platinum at the reaction conditions. It can be clearly appreciated that the specific activity strongly depends on the platinum face exposed, which is a clear indicative of the structure-sensitive character of this reaction. Furthermore, it is interesting to observe that polycrystalline Pt/C showed lower TOF values than preferentially oriented catalysts. Englisch et al., in a very complete work on structure sensitivity of crotonaldehyde hydrogenation over Pt/SiO2, also found lower activity for catalysts with smaller metal particles, and this behavior was attributed to a faster deactivation caused by strong adsorption of crotonaldehyde on low coordination sites like edges and corners. However, this fact has not been detected in our case, and deactivation profiles were very similar for the three catalysts under study. In the same way, Singh et al. found a 25-fold increase in TOF for citral hydrogenation as the average platinum crystallite size increases from 1 to 5 nm. Again, the higher participation of low coordination number surface atoms as the metal particle size decreases was invoked to explain this behavior. Note that these highly exposed atoms are assumed to be the most abundant on the surface of Pt/C as indicated by voltammetric and microcalorimetric results).
TOF in crotonaldehyde hydrogenation at 333 K as a function of time on stream for Pt/C, Pt(111)/C, and Pt(100)/C catalysts.
Reaction scheme for crotonaldehyde hydrogenation
View Within Article
Comments on the reaction mechanism of crotonaldehyde hydrogenation on Au/TiO2 catalysts
The selectivity to C=O versus C=C hydrogenation is mainly dictated by the extent of activation of the C=O group Therefore, the adsorption mode of crotonaldehyde is the determining factor, which orientates the selectivity toward the formation of the monohydrogenated products. The possible adsorption modes of α,β-unsaturated aldehydes on monocrystalline metal surfaces were described by Delbecq and Sautet If the adsorption is restricted to that of the C=C moiety of the molecule (η2CC of di-σCC-type) (Scheme 2), hydrogenation would only lead to butanal. Conversely, if the adsorption is restricted to that of the C=O bond (η2CO of di-σCO-type), or to that of the oxygen lone pair (atop), only crotyl alcohol could form. In the case of gold catalysts, both products are formed, and we must then assume either the coexistence of two independent C=C and C=O (or O atop) adsorption modes or a unique adsorption mode involving both C=C and C=O (or O), such as the η3 or η4 modes (Scheme 2).
Scheme 2. Adsorption modes of the molecule of crotonaldehyde on metal, from Ref.
View Within Article
With regards to selectivity, the main product formed in all cases was the saturated aldehyde (butanal). Thus, the selectivity to the unsaturated alcohol was almost negligible for polycrystalline Pt/C and Pt(100)/C catalysts while Pt(111)/C showed certain selectivity to crotyl alcohol (4% molar fraction). It is well known that there is an effect of the metal morfology on the adsorption geometry of unsaturated aldehydes, leading to different types of reaction pathways and finally, different selectivities. The theoretical calculations of Delbecq et al. suggest that, in the case of (100) and open faces or steps, crotonaldehyde adsorption involves both double bonds, in a so-calledη4 planar coordination, this geometry leading to the preferential hydrogenation of the C C bond for kinetic reasons. This fact would explain the low selectivity toward unsaturated alcohol found for Pt(100)/C and Pt/C catalysts, where the fraction of (100) facets and low coordinated atoms, respectively, is very high. On the other hand, the dense (111) metal faces are not favorable for the C C coordination, and crotonaldehyde adsorption via carbonyl bond (di-σC=O) is energetically favored. Thus, a greater participation of these faces in the catalysts surface would improve the selectivity to unsaturated alcohol. The low value of selectivity obtained for Pt(111)/C in our study could be due to the presence of a noticeable fraction of stepped surfaces vicinal to (111) small domains (see prominent peaks at 0.12 and 0.26 V in) since, in these surfaces, the crotonaldehyde coordination is shifted to a πC=C mode. Furthermore, voltammetric results showed that the surface of Pt(111)/C is not exclusively participated by (111) facets and the presence of large (100) domains (could also be responsible of the low value of selectivity to unsaturated alcohol.
Fig. 22. Crotylalcohol selectivity as a function of conversion in consecutive crotonaldehyde hydrogenation over Pt/SnO2 catalyst at 170 °C [69].
Fig. 18. Activity as a function of H2 partial pressure at different temperatures during crotonaldehyde hydrogenation over Cu/AC catalyst.
Fig. 4. Hydrogenation of crotonaldehyde (PH2=752 Torr, Pcrotonaldehyde=8 Torr, Treaction=120 °C). Selectivities versus conversion on DP Urea 16 h/90 reduced under H2 at 120 °C.
*. Cinnamaldehyde hydrogenation
shows the main reaction pathways that can occur during cinnamaldehyde hydrogenation. Under the experimental conditions used in the present work only these four products were detected. Activity and selectivity of the catalysts in the hydrogenation of cinnamaldehyde were evaluated in the liquid phase at 383 K and 5 MPa. plots the conversion of cinnamaldehyde as a function of reaction time for the three catalysts under study. Initial activities were expressed in term of rate constant (min−1), which were obtained from the slope of the cinnamaldehyde concentration-time plot and considering a first rate law equation. The activity per active site (TOF) was calculated using the CO uptakes to determinate the number of surface metal sites and the results are summarized in Table 1. As it can be seen, both the rate constant (Table 1) and the conversion of cinnamaldehyde reached for similar reaction times (followed the trend: Pt/C > Pt(100)/C > Pt(111)/C. However, when the rate is normalized by the number of accessible platinum atoms (TOF), the trend changes to: Pt(100)/C > Pt(111)/C > Pt/C which is consistent with that obtained in crotonaldehyde hydrogenation
High-quality image (43K)Fig. 5. Reaction scheme for cinnamaldehyde hydrogenation.
High-quality image (87K)
Fig. 7. Conversion of cinnamaldehyde (—) and selectivity to cinnamyl alcohol (- - -) for: Au/Al (•), Au/Fe10Al (♦), and Au/HDAE (■).
High-quality image (48K)Fig. 6. Conversion of cinnamaldehyde as a function of time for Pt/C, Pt(111)/C, and Pt(100)/C catalysts.
View Within Article
Table 1.
Rate constant, TOF and selectivity to the saturated aldehyde and the unsaturated alcohol at 10% conversion in cinnamaldehyde hydrogenation for Pt/C, Pt(111)/C, and Pt(100)/C catalysts
HALD
CALC
Pt/C
0.016
0.003
82
16
Pt(100)/C
0.008
0.040
61
36
Pt(111)/C
0.005
0.020
49
48
CALC
Pt/C
0.016
0.003
82
16
Pt(100)/C
0.008
0.040
61
36
Pt(111)/C
0.005
0.020
49
48
Full-size table
a Values at 10% conversion.
View Within Article
Regarding the products distribution, again the main reaction product was the saturated aldehyde (HALD) although, conversely to crotonaldehyde hydrogenation, the three catalysts under study yielded noticeable amounts of cinnamyl alcohol (CALC). This fact is not strange since it is well known that the selective hydrogenation of the carbonyl bond is more difficult to achieve when lower aldehydes are involved and the reaction takes place in gas phase. also shows the selectivity to cinnamyl alcohol (CALC) and hydrocinnamaldehyde (HALD) at equal conversion level (10%) for the three catalysts under study. It is interesting to observe that polycrystalline Pt/C showed lower selectivity to CALC than preferentially oriented samples. This fact can be explained in terms of particle size. It is well known that metal particle size has a strong influence in the selectivity to cinnamyl alcohol in the case of monometallic platinum catalysts. Thus, the selectivity for C O hydrogenation increased with increasing crystallite size. Note that Pt/C showed an average platinum particle size of 3 nm while in preferred oriented samples, Pt(100)/C and Pt(111)/C, platinum is in form of large particles of around 10 nm size (Repulsive interactions between the phenyl ring and the large metal crystallites, which hinders the interaction via the C C bond and enhances interaction via C O bond, is proposed to explain this behavior. This effect is absent in the case of small metal particles because both double bonds, C C and C O, could approach the surface with less repulsive interactions between the phenyl ring and the metal surface.
Furthermore, it can be seen in that Pt(111)/C showed higher selectivity to CALC than Pt(100)/C (48 vs 36%). In this case no size effect can be invoked since both samples showed similar average particle size (10 nm). However, as pointed before, both catalysts differ in the facets distribution of the metal particles, with a higher ratio of (111) domains in the case of Pt(111)/C. Thus, these results clearly indicate that, besides a size effect, the amount of (111) facets composing the metal particle play a crucial role in increasing the selectivity to unsaturated alcohol in these kind of reactions. In this way, previous studies on prenal hydrogenation on well-defined (111) and (110) platinum surfaces, found that the main product (also at 10% conversion) was the unsaturated alcohol and the saturated aldehyde, respectively. Interestingly, the same authors also found that prenal hydrogenation on Pt(553) (which can be considered as a surface with (111) terraces and steps) resulted in a lower selectivity to the unsaturated alcohol compared to (111) flat surface. Again, the simultaneous presence of (111) terraces (where the di-σC=O adsorption mode is preferred) and steps (πC=C adsorption mode preferred) is invoked to explain the lower selectivity to the unsaturated alcohol on this stepped surface. As already seen, it has been proved that the platinum particles composing Pt(111)/C have a noticeable amount of stepped surfaces vicinal to (111) small domains (), which would explain the noticeable yield of saturated aldehyde (obtained for this sample. It is necessary to bear in mind that substituents at the C C bond have a strong influence on the selectivity to the unsaturated alcohol and thus, can influence the intrinsic selectivity determined by the platinum surface. Thus, is not strange that some works, based on the hydrogenation of the simplest unsaturated aldehyde (acrolein), introduced some new points to the factors controlling the selectivity in this kind of reactions. In a recent theoretical work, Loffreda et al concluded that the key point which determines the selectivity to allyl alcohol on Pt(111) is not the adsorption mode of the aldehyde, but the competitive desorption process of the half hydrogenated products (unsaturated alcohol vs saturated aldehyde). Furthermore, in contradiction with earlier assumptions, Mohr et al. claim that the surface sites responsible of the C O hydrogenation are the highly unsaturated surface atoms situated at corners and edges, although these conclusions are obtained for a very special catalytic system (Au/ZnO), and the same authors indicated that these conclusions could not be transferred to other catalytic systems.
scheme
Fig. 1. Conversion of 1% v/v cinnamaldehyde/toluene over 0.50 g of 2% Pd/SiO2 at 25 °C and 47 psi H2 pressure.
for crotonaldehyde hydrogenation
5- Product distributions and activation energy
The product distributions obtained with the Au/TiO2 catalysts are remarkable since they consist of 60–70% of crotyl alcohol, whatever the studied Au/TiO2 catalyst and the temperature of pretreatment (A selectivity-conversion diagram was obtained for catalyst DP Urea 16 h in gradually decreasing flow rates during the time on stream in order to increase the conversion This diagram shows that the selectivity to butanal remains unchanged up to 50% conversion while those in butenes and butanol slightly increase at the expense of that in crotyl alcohol.
Fig. 4. Hydrogenation of crotonaldehyde (PH2=752 Torr, Pcrotonaldehyde=8 Torr, Treaction=120 °C). Selectivities versus conversion on DP Urea 16 h/90 reduced under H2 at 120 °C.
View Within Article
To check if butanal could be formed by isomerization of crotyl alcohol as observed on Pt-based catalysts the hydrogenation of crotyl alcohol was investigated for the DP Urea 16 h catalyst. It was observed that crotyl alcohol reacted at a similar rate as crotonaldehyde, to give rise to 90% of butenes, 3.9% of butanal, and 6.1% of butanol. The butenes formed from crotyl alcohol consisted of 73% of but-1-ene, 7% of trans-but-2-ene, and 20% of cis-but-2-ene. This butene distribution is similar to that obtained from butadiene hydrogenation on the same catalyst (unpublished results). These results suggest that crotyl alcohol transforms by dehydration into butadiene, which further hydrogenates into butenes. This could explain the increased of selectivity to butenes at the expense of that in crotyl alcohol when the conversion of crotonaldehyde increased. This also indicates that the butanal observed in the crotonaldehyde hydrogenation products is directly formed from the reactant since the isomerization of crotyl alcohol into butanal is a minor reaction. Scheme 1 can be proposed.
Scheme 1. Reaction pathway of the hydrogenation of crotonaldehyde with Au/TiO2 catalysts; plain and dotted arrows represent the primary and secondary reactions, respectively, and the lengths of the arrows are roughly proportional to the reaction rates.
In contrast to platinum-based catalysts, the presence of sites, which would hydrogenate the C=C bond only (η2CC structure), is unlikely on gold catalysts. Indeed, gold catalysts are active for the partial hydrogenation of butadieneor acetylene but very poorly active for the hydrogenation of butene or ethylene. We found for the DP Urea 16 h catalyst that butadiene reacted twice as fast as crotonaldehyde under the same conditions of pressure and temperature, but this catalyst was totally inactive for butene hydrogenation (unpublished results). This last result makes unlikely the η2CC adsorption mode of olefin on gold particles, and therefore, that of crotonaldehyde.
On the other hand, the fact that the selectivities to crotyl alcohol and butanal are almost constant whatever the gold particle size and the reduction temperature is not in favor of the coexistence of two types of adsorption modes involving two types of catalytic sites whose proportions depend on the structure/size of the gold particles. Therefore, we rather propose that a single mode of adsorption of crotonaldehyde governs the selectivity, either η3 or η4 adsorption mode and that the addition of two hydrogen atoms gives rise to both butanal and crotyl alcohol. The selectivities to each product would be determined by the relative probabilities of addition of the hydrogen atoms to the different positions in the molecule (O, Cα, Cβ, Cγ).
The dramatic increase of the TOF when the particle size decreases suggests that the gold atoms involved in the hydrogenation process are those located on edges and corners. For instance, the relative number of low-coordinated sites in cubooctahedral-shape particles varies in the same way as the TOF versus the particle size Besides this pure size effect, the evolution of the particle morphology observed on the HRTEM micrographs with the reduction temperature leads to the same conclusion: the gold particles obtained after reduction at 200 °C do not exhibit well-developed facets while those obtained after reduction at 500 °C (are perfectly facetted with low index planes. Therefore, the proportion of low-coordinated sites is much lower after pretreatment at 500 °C than at 200 °C, and the decrease in catalytic activity and TOF when the reduction temperature increases from 200 to 500 °C is attributed to the smoothing of the outer surface of the gold particles, i.e., to the decrease in the number of low-coordinated sites. This is in agreement with the interpretation of Claus and colleagues where they attributed the higher activity of Au/TiO2 compared to that of Au/ZrO2 to the fact that the gold particles were round shaped on TiO2 and facetted on ZrO2. In another work on selective decoration of the faces of gold particles by indium, they proposed that the active sites favoring the adsorption of the C=O group of acrolein are the edges of the gold nanoparticles.
However, these low-coordinated sites are not prone to favor the η4-adsorption mode for which the reactant structure is rather planar or the η3-adsorption mode for which multiple metallic sites are involved but they might be more adapted to dissociate the hydrogen molecule. Therefore, we propose that the activity of gold catalysts in hydrogenation is determined by the dissociation step of molecular hydrogen. The apparent activation energy calculated in the present work, 36.3 kJ mol−1, is the same as that reported by Buchanan and Webb for the selective hydrogenation of butadiene on Au/Al2O3 catalysts, 36.5 kJ mol−1. This could support the hypothesis that butadiene and crotonaldehyde hydrogenation kinetics are governed by a common rate-determining step, the hydrogen dissociation. In the Au/Al2O3 catalysts were prepared by impregnation (1 wt% of Au) and exhibited a broad distribution of particle sizes up to 80 nm with the largest population of size in the 0–10 nm range. Because of the complete absence of activity of the support in butadiene hydrogenation, Buchanan and Webb proposed a mechanism where hydrogen dissociated on alumina and migrated to gold particles to hydrogenate the adsorbed butadiene. However, they did not exclude the fact that the activation of H2 could take place on tiny gold particles, which could not have been detected by electron microscopy (the smallest particle size they could detect with their microscope was not reported in their paper).
The poor catalytic activity in hydrogenation with gold catalysts compared to Pd or Pt catalysts has been generally attributed to the fact that gold hardly dissociates molecular H2, as it has been established experimentally and theoreticallyHowever, several authors have unexpectedly observed the dissociative chemisorption of H2 (D2) on gold films. For instance, Okada et al. who studied the adsorption and reaction of hydrogen (deuterium) on thin gold film grown on Ir (111), noted that surprisingly H2 (D2) can dissociatively adsorb on these Au films. They attributed this property to locally relaxed surface features like defects. On thin gold film deposited on Pyrex glass, Stobinski et al. also found that H2 dissociative chemisorption occurs, providing that the thin Au films were deposited at a low temperature (78 K) and were not sintered afterward. They concluded that H2 chemisorption occurs only on surface atoms of low-coordination number. These conclusions support our proposals for the hydrogenation mechanism.
.
6- Reactor selection
Traditionally batch reactors have been applied in fine chemicals production in the liquid phase, whereas gas phase hydrogenation is usually carried out in fixed bed reactors. It could be, however, attractive from the industrial point of view to produce fine chemicals in a continuous mode. The benefits compared to the conventional batch operation are the higher production capacity, minimization of dead time during the start and end, as well as more equal product quality. According to OECD reference scenario the world chemical demand will increase by 39% from the level in year 2000 to the level in year 2020. Additionally, from the scientific point of view the continuous operation allows an easier way to record and model catalyst stability and deactivation in case of liquid phase reactions.
Liquid phase hydrogenations have been performed in trickle bed, in monolithic and in bubble column reactors as well as in a fixed bed reactor with recirculating liquid phase [ and in continuous flow. One decisive factor for achieving high selectivity to intermediate products in consecutive hydrogenations is the residence time of the reactant in the reactor. Trickle bed reactors are widely used in chemical processes in hydrogenations and oxidations. The drawback in using trickle bed reactors is the poor wetting of the catalyst, which can lead to ineffective use of the catalyst as well as to formation of hot spots. In trickle bed reactors the lower velocities are applied in order to maintain the trickle-flow regime, whereas in monolithic reactor higher velocities should be used in order to avoid maldistribution at lower velocities . Trickle bed reactors have been used in citral, benzaldehyde and in crotonaldehyde hydrogenations. Solvent free citral was hydrogenated in a trickle bed reactor at 100 °C and 20 bar hydrogen. Citral hydrogenation in this reactor exhibited 20 h induction period during which the conversion level of citral increased. After induction period a constant very high conversion (97%) as well as very high selectivity to nerol and geraniol (97%) over Pt-Sn/MgO catalyst was obtained at 100 °C and 20 barInterestingly with a monometallic Pt/MgO lower selectivity was achieved, but also no induction period was observed. In order to avoid the poor catalyst wetting and to enhance the productivity the trickle bed reactor with periodic flow interruption has been applied. As a result an enhanced crotonaldehyde hydrogenation rate was observed over Pd/γ-Al2O3 at 25 °C and 11 bar in such reactor compared to a steady state operation. The performances of a trickle bed and a monolithic reactor were compared in benzaldehyde hydrogenation at 100 °C and 10 bar. The former reactor contained Ni supported on γ-alumina extrudates as a catalyst, whereas a Ni-cordierite catalyst was used in the monolithic reactor. Co-current gas and liquid flows were applied. Monolithic reactor showed higher productivity and selectivity than the trickle bed reactor in the hydrogenation of benzaldehyde. The reason for this might be the sharper residence time distribution in the monolithic reactor.
Monolithic reactors can operate either in single-pass mode or with external liquid recycle. The product selectivity might decrease in a latter case with liquid back-mixing and thus the single pass mode might be preferred. Monolith catalysts were investigated in citral hydrogenation in a screw-impeller tank reactor and in benzaldehyde hydrogenation over a monolith catalyst in a batch mode or in continuous mode. Ni-monolith was used as a catalyst in citral hydrogenation in a screw-impeller stirred tank reactor (SISRCitral hydrogenation was investigated in hexane at 5 and 40 bar hydrogen and at temperatures 40, 80 and 100 °C. Catalyst deactivation was observed after consecutive experiments. The selectivity to citronellal was the highest at 40 °C and 5 bar hydrogen, i.e. 96% at 80% conversion. Higher hydrogen pressures favored formation of 3,7-dimethyloctanal. In benzaldehyde hydrogenation over Ni-monolith catalysts the highest selectivities to benzyl alcohol were obtained in a batch operation without residence time distribution .
A continuous fixed bed reactor operation was demonstrated in citral hydrogenation with a recirculating liquid phase and in a continuous flow. In the former case it was shown in citral hydrogenation at 70 °C in ethanol over knitted Ni-silica fibers that the recirculating reactor operation was equally good as a traditionally batch reactor. In the latter case citral hydrogenation was carried out over 5 wt.% Pd/Al2O3 catalyst at 40° and 190 bar hexane, supercritical CO2 or ethane. Both catalyst deactivation and changes in the product distribution were observed.
A packed bubble column reactor was used in hydrogenation of citral to citronellal. The aim of using recirculating product and circulating hydrogen was to improve the space–time yield, because according to Bröcker et al. the catalyst regeneration is not successful in citral hydrogenation and the larger amounts of catalyst needed to enhance the productivity would decrease the process economy. In a packed bubble column reactor hydrogenation of citral to citronellal over a powdered Pd/C catalyst in the presence of triethyl amine in methanol at 70 °C and 8 bar hydrogen gave 92.9% selectivity to citronellal at 99.7% conversion.
Microstructured carbon-coated reactors have been applied in gas phase hydrogenation of acrolein. The catalyst contained microchannels of 700 μm depth and of 300 μm width with length of 50 mm manufactured from copper wire. The carbon-coating was obtained via carbonization the polymer coating at 550 °C. Thereafter the support was calcined at 350 °C and after fucntionalization of the carbon surface with HNO3 Ru was deposited on the support via ion exchange with [Ru(NH3)Cl]Cl2. Although the comparative specific catalytic activities for microstructured Ru-catalyst and for Ru/C were not measured, it was reported that the selectivity to allyl alcohol was comparable for these catalysts. The short overview of various reactor configurations applied in the recent years for heterogeneous catalytic hydrogenation of carbonyl compounds demonstrates the big although largely unexplored potential of chemical reaction engineering in the field of synthesis of fine chemicals.
. Membrane reactors
Membrane reactors have been applied in liquid phase hydrogenations of cinnamaldehyde Analogously to monolithic systems, when the catalyst is impregnated onto the membrane wall the catalyst separation can be avoided. The technical features of membrane reactors will be discussed Catalytic tubular membrane made from α-Al2O3 was coated with Co-Pt/γ-Al2O3 layer and used for cinnamaldehyde hydrogenation The highest selectivity to cinnamyl alcohol was 75% at conversion of 43.7% at 50 °C in ethanol.
Hydrogen permeating palladium membrane was used as a catalyst in furan hydrogenation at 150 °C The conversion of 16% was achieved with tetrahydrofuran as a main product.
Fig. 27. SISR reactor in citral hydrogenation [164].
Fig. 9. HRTEM image of Au/ZrO2 exhibiting a rounded gold particle (A) and a particle with facets (B). Right: separate size distribution of single crystalline (SC) giold particles, single twinned particles (ST), multiple twinned particles (MTP) together with overall size distribution. Lower row: left: HRTEM images of multiple twinned particles of Au/ZrO2, left: gold particle size distribution [96] (reprinted with permission, copyright (2003) American Chemical Society).
7- Conclusions
Recent literature on the selective hydrogenation of unsaturated aldehydes to unsaturated alcohols over heterogeneous catalysts is overviewed. Both the reactant structure and the nature of metal surface affect the possibilities to achieve high selectivities to unsaturated alcohols. The main challenge in chemoselective hydrogenation is to develop a stable catalyst, which is active and selective, since usually there is an inverse relationship between activity and selectivity.
For supported catalysts both metal and support influence activity and selectivity of the final catalyst. Metals can have different electronic properties for instance due to their different origin, exhibiting various properties. Furthermore, several metals or metal alloys can be used as active components and electronic properties can be tuned by adding promoters. The nature of support material, like porosity, pore size distribution and reducibility is an important parameter in catalyst preparation. Besides conventional oxide and activated carbon supports more advanced support materials have been investigated in the synthesis of fine chemicals. Some examples of tailor-made supports are bifunctional zeolites and mesoporous materials, colloidal metals in polymers as well as reducible and alloy forming supports. It should be pointed out that the final catalysts properties are also a result of specific catalyst pretreatment and a required catalyst could be obtained via fine-tuning the catalyst pretreatment procedures. The selectivity to unsaturated alcohols can in general be increased via increasing the number of active sites activating the carbonyl group.
Design of tailor-made catalysts with specific selectivity is based on the knowledge in reaction mechanisms. Recent studies on molecular modeling and adsorption along with kinetic modeling are discussed. Besides being a way of translating our understanding of the reaction mechanism, kinetic modeling is efficiently used for process development. Three-phase fine chemical syntheses are traditionally carried out in batch reactors over conventional slurry catalysts. There exist, as shown in this review, substantial amount of kinetic data from chemoselective hydrogenations, which can be used in process development. Additionally, the knowledge from conversion–selectivity relationship and a comparison of liquid and gas phase hydrogenations serve as tools for understanding the catalytic phenomena on the surface sites. Due to the increasing future demand of fine chemicals there is, however, a need to develop continuous production technologies for three-phase hydrogenations. These technologies require structured catalysts, which have low-pressure drop and avoiding at the same time maldistribution, thus utilization of several structured supports in chemoselective hydrogenation, like fibers prepared from activated carbon, polymer, silica woven glass fibers has a high potential. Recently, membrane reactors and monolithic catalysts have been also applied in the field of fine chemicals. Besides other advantages continuous technology brings additionally a benefit of understanding the catalyst re-use and deactivation in a well-defined, quantitative way.
Analysis of the state-of-the-art in heterogeneous catalytic hydrogenation strongly emphasizes the necessity of a truly multidisciplinary approach in catalyst development, when the materials chemistry is tightly linked on one hand to surface science and molecular modeling and on the other hand to catalytic reaction engineering, including kinetic modeling and reactor selection
In reactions exhibiting a selectivity problem, e.g. hydrogenation of two C=C bonds or C=C versus C=O groups, the advantage of gold catalysts is the preferred hydrogenation of one of these groups leading to monoenes (also by selective C=C hydrogenation), unsaturated alcohols and unsaturated ketones as reaction products important to the chemical industry. In-depth characterization of gold hydrogenation catalysts clearly shows that elucidation of the active site issue and the observed structure-sensitivity cannot be simply undertaken by varying the size of gold particles because the latter may interfere with the portion of multiply twinned particles (MTPs) and the degree of particle rounding. Because a large portion of MTPs gave low allyl alcohol selectivity (product of C=O group hydrogenation) in acrolein hydrogenation, catalysts with single crystalline gold particles must be prepared by appropriate synthesis methods. However, what are the preparative factors, which control the ratio between gold surface atoms on edges and corners versus those on flat surfaces or the portion of MTPs? The important impact of an appropriate pretreatment in H2 controlling the surface structure, and thus, the catalytic properties was shown by the drastic influence of the prolonged reduction time (which favors the formation of multiple-twinned gold particles non-selective for C=O group hydrogenation. This approach shows that in the case of hydrogenations over gold catalysts, the mechanism of catalysis are more complicated than assuming a simple geometric model (gold particle size) or perimeter-interface model, which is suggested to be the active site in CO oxidation The modification of gold particles by a second metal (indium) was not only a suitable tool to increase the intramolecular selectivity, but was also essential to identify the active sites for this hydrogenation reaction.
No direct evidence has yet been demonstrated whether hydrogen is dissociatively or molecular adsorbed, whether subsurface hydrogen (in analogy to subsurface oxygen in oxidation catalysis) plays a role as indicated at higher coverages, and thus, what the active sites for hydrogen activation are. On carbon (type: Black Pearls, Cabot GmbH) supported gold nanoparticles (dAu = 3.7 ± 1.4 nm), the amount of chemisorbed hydrogen was definitively zero = At present, experimental results indicate, that H2 chemisorption occurs only on Au surface atoms of coordinatively unsaturated sites. Since the adsorption capability of hydrogen on (clean) gold surfaces is limited, the application of hydrogen chemisorption measurements is difficult on gold catalysts, and thus, remains an advanced task. Furthermore, it has to be considered that also surface strain that can be induced in small, thin gold particles by supports =may be responsible for the activation of hydrogen and the hydrogenation properties of supported gold catalysts. In the case of oxygen adsorption and activation on gold nanoparticles, their greater reactivity was attributed to a higher step density and a greater amount of surface strain These factors have to be elucidated in the future with the help of gold model catalysts, in situ techniques =and theory, giving rise to a deeper insight into kinetics and mechanism of gold-catalysed hydrogenations as a key toward future applications.
Thanks to the method of deposition-precipitation with urea, it has been possible to prepare a series of Au/TiO2 samples with the same gold loading ( 8 wt%), and various average particle sizes (1.7 to 8.7 nm), either by varying the duration of the preparation, or by changing the conditions of reduction (nature of the gas, temperature, flow rate). With this series of Au/TiO2 catalysts, it has been possible to study systematically the influence of the gold particle size on their catalytic properties in the hydrogenation of crotonaldehyde, under mild experimental conditions ( 100 °C, 0.1 MPa), already used in former studies on Pt-based catalysts. The selectivity to carbonyl bond hydrogenation (crotyl alcohol formation) was twice as high as for olefinic bond hydrogenation (butanal formation). The selectivity was not sensitive to the reduction temperature, as was the case for platinum catalysts whose selectivity to crotyl alcohol was generally much lower (<50%) than that for gold. The gold catalysts led to very reproducible results. It must be pointed out that the Au/TiO2 catalysts were much less active by about two orders of magnitude than Pt/TiO2 catalysts. The structure sensitivity of the Au/TiO2 catalysts for the hydrogenation of crotonaldehyde was clearly shown by the significant increase in activity and TOF for gold particles of 2 nm. The possible role of the oxygen vacancies of TiO2 at the metal-support interface, whose number is known to increase with the reduction temperature, was not evident as in the case of Pt/TiO2. It is proposed that the dissociation of hydrogen on the low-coordinated sites of gold particles (edges and corners) is the rate-determining step of this reaction, and that crotonaldehyde is adsorbed on gold particles through a η3 or η4 mode of adsorption.