TiO₂ (Degussa P-25 S_BET= 50 m²g^-1) was impregnated with H₂IrCl₆ to obtain Ir/TiO₂ catalysts with Ir loading of 1 wt %. The impregnated solid was dried at 343 K for 6 h, calcined in air at 673 K for 4 h and reduced at 773 K (HTR: High temperature reduction) for 2 h.

The BET-surface area was evaluated from nitrogen adsorption isotherms at 77 K and the number of active sites of each catalyst were obtained by hydrogen chemisorption at 298 K, both performed in a Micromeritic ASAP 2010. The surface acidity was determined by temperature programmed desorption, TPD, of NH₃. Transmission electronic microscopy (TEM) studies were carried out in a JEOL Model JEM-1200 EXII microscope.

Furfuraldehyde and furfuryl alcohol used in all experiments, both of analytical reagent grade were supplied by Aldrich. Catalytic reactions were carried out in a stainless steel batch reactor at a constant stirring rate (1000 rpm). To carry out the kinetic study over the catalysts only one variable was modified in each experiment, keeping constant all the others. The effect of furfural concentration was studied in the concentration range 0.025 to 0.1 M. The hydrogen partial pressure was studied in the range 0.48 MPa to 0.84 MPa. The temperature was varied in the range 323 to 363 K and the catalyst weight, ranged from 0.1 to 0.3 g.

Prior the experiment, all catalysts were treated in situ under hydrogen flow of 20 cm³min^-1 at atmospheric pressure and temperature of 363 K to remove possible surface oxide species generated during handing. The avoid the presence of oxygen, once the reactor was loaded with the reactant mixture and catalyst, the system was flushed with He at atmospheric pressure during 30 min.

Analysis of the reaction mixture and products was carried out using a gas chromatograph Varian 3400 furnished with a HP Wax column of 30 m length and 0.53 mm ID. The GC analysis was performed using a flame ionization detector, using He as carrier. The products of reaction were analyzed in a Varian 3800-Saturn 2000 GC-MS provided with a ionic trap and using the same separation conditions.

3.1 Catalyst characterization
The results of catalysts characterization have been provided previously [11], however some relevant aspects are considered here. These solids when they are reduced at high temperature in H₂, lead to a partial reduction of the support, TiO_2-x which covers the metal surface, by the so called SMSI effect (strong metal support interaction). This behavior is associated with a decrease of H/Ir ratio in chemisorption measurements (table 1) but without any effect in the metal particle size because there are no changes in the metal particle size in catalysts of Ir/TiO₂ reduced at different temperatures of reduction. This phenomenon has been widely studied by TEM [1,11]. Surface acidity measurements were carried out by DTP of ammonia, previously adsorbed at 393 K. Ir/TiO₂ exhibits a very low acidity compared with others supports [12].

Table 1. Characterization results of Ir/TiO₂ catalyst

3.2 Mass transfer
In the liquid phase hydrogenation of furfuraldehyde, mass-transfer processes (gas-liquid, liquid-solid, and intraparticle diffusion) can influence the rates of reaction [13]. For the kinetic study these diffusional resistances should be absent. The liquid-side mass-transfer coefficient and liquid-solid mass-transfer coefficient depend on the intensity of turbulence in the liquid phase. For this reason highest stirring speed were used (1000 rpm). For prevent intraparticle mass-transfer resistance, it was used smallest catalyst particles sizes (100 mm). An experimental approach following the guidelines of Satterfield and Sherwood (eq. 1) [14] for corroborate the absence of gas-liquid or liquid-solid mass transfer resistance was applied. The equation 1 describes that the reciprocal of the conversion rate as a function of the reciprocal of catalyst mass should give a straight line indicating absence of external mass-transfer limitation.

The figure 1 displays the results obtained applying this equation. It can be observed the absence of external mass-transfer limitation. Therefore, the mass-transfer rates thus obtained were considerably higher than the maximum reaction rates observed in the kinetic experiments, consequently, it was concluded that external mass transfer did not have an effect on the observed rates in the hydrogenation experiments.

Figure 1. Effect of catalysts weight on the initial rates in the furfural hydrogenation at 363 K, 0.1 M and 0.62 Mpa

3.3 Effect of solvent
Some solvents were used for furfural hydrogenation, such as, ethanol, heptane, ethanol-heptane (1:1). Table 2 summarizes the solvents used, the conversion level at 60 min of reaction time as well as the initial activity expressed as TOF (turnover frequency) for Ir/TiO₂ catalysts.

Table 2. Catalytic activity in furfural hydrogenation at 363 K and 0.62 MPa, expressed as conversion, TOF and selectivity at 60 min, S_FOL=Selectivity to furfuryl alcohol, S_FDA to 2-furaldehyde-diethylacetal

As polarity increases also conversion as well as in the TOF is observed at higher solvent, but the selectivity to the unsaturated alcohol decreases at higher polarity. Similar trends have been reported for other hydrogenation reactions [15]. It is interesting note the synergestic effect of mixed solvents which increases both activity and selectivity.

The results in table 2 indicate clearly that the main product is furfuryl alcohol produced by hydrogenation of C=O bond in the furfural.

The acetalization takes place over acid sites, specially Brönsted acid sites on the catalyst surface, these sites are placed near to iridium crystals. However, it has been suggested that the acid sites in the catalysts did not modify the active metal properties [16] and favors the electron transfer from acid sites in the support to the active metal sites, by suppressing the hydrogenation of C=C bond. The rate of formation of acetals is faster during the initial period of reaction but the selectivity of acetal decreased with the consumption of reactant and became lower than 1 % at high conversion (figure 2).

Figure 2. Conversion vs selectivity, effect of solvent in furfural hydrogenation for Ir/TiO₂ at 363 K and 0.62 MPa with different solvents used; (Δ) heptane, (○) ethanol, (□) ethanol-heptane 1:1

The Figure 3 shows the hydrogenation products of furfural over Ir/TiO₂, arising from the reduction of the C=O group and the formation of 2-furaldehyde-diethylacetal. In this catalyst was not observed the hydrogenation of the furan ring, or other compounds derived from secondary reactions, such as hydrogenolysis of the C-O bond, decarbonylation, hydrogenation and furan ring opening that may appear [17].

Figure 3. Reaction pathway for furfural hydrogenation over Ir/TiO2

3.4 Analysis of initial rate data
The effect of initial concentration of furfural and H₂ pressure on the initial rate data of furfural hydrogenation was studied. Experiments were conducted at various initial concentrations of furfural in the range of 0.025 M to 0.1 M, at 363 K and 0.62 MPa hydrogen partial pressure. The initial rates were expressed as TOF. The order calculated taking out the plot of -log initial concentration vs. -log TOF for Ir/TiO₂ was of 0.5 (figure 4).

]]> Figure 4. Effect of initial concentration on hydrogenation of furfural over Ir/TiO₂ as a plot of -log C₀ furfural vs -log TOF.

An explanation for this result can be the adsorption mode of furfural on the catalysts surface. Furfural possibly adsorbs preferentially towards C=C with plane geometry competitive with atop geometry similar in cinnamaldehyde [18]. This could be attributed to the fact that C=C on the ring of furfural are conjugated. At first time reactions the coverage of surface of molecules of furfural with atop geometry is larger, but whiles the reaction progress the surface is totally coverage of furfural in plane geometry.

The effect of the hydrogen partial pressure on the initial rates was studied in the range of 0.48 MPa to 0.84 MPa at constant initial concentration (0.1 M) and temperature (363 K). The results are displayed in figure 5. The order calculated for initial rate was -1,0 using the initial rates expressed as TOF.

Figure 5. Effect of hydrogen partial pressure in furfural hydrogenation over Ir/TiO₂ catalysts as a plot of -log PH₂ vs -log TOF

3.5 Kinetic model
Two types of kinetic equations were employed in the quantitative description of the experimental results: first, empirical power-law equations based on initial rates and a second, equations based on the Langmuir-Hinshelwood mechanism. Power-law equations for the hydrogenation of furfural were written as

where i can be Ir/TiO₂, p and q are the order obtained at initial rate with respect to furfural concentration and partial pressure of H₂. Thus for Ir/TiO₂ the global order is -0,5. Because the order with respect to concentration of furfural is fractional, it is desirable to obtain insight into the kinetics via Langmuir-Hinshelwood models.

Considering that the hydrogenation of aldehydes a,b-unsaturated is potentially a complex combination of series and parallel reactions, can be expected a parallel reactions that involve the hydrogenation of both functional groups of the furfural (C=C, C=O), however the results showed for Ir/TiO₂ in furfural hydrogenation produces a highly selectivity towards furfuryl alcohol, being the result of the presence of SMSI effect on this type of support. For this catalyst studied the selectivity towards unsaturated alcohol is greater of 98% which can be interpreted as the presence of a single site present (i.e. Ir^d+ species are preferential).

Considering these observations and assuming on basis of the experimental observation the sequence of reaction is expressed as:

Where q is the active site, FAL is furfural, q_H, q_FAL, q_FOL, q_* are respectively the adsorbed hydrogen, furfural adsorbed, furfuryl alcohol adsorbed and vacant sites. Based on the basis asumptions mentioned above, the reaction rate can be written as:

Where V/W_c is the ratio of he reaction volume to the mass of catalyst; C_FAL is the concentration of furfural, k_r1 is the rate constant for the reaction.

K_FAL, K_FOL and K_H2 are equilibrium adsorption constants. These expressions can be substituted into the RDS step.

Considering that the expression for the fractional surface coverage of sites is:

In agreement of the results of hydrogenation reactions at 363 K with addition of the unsaturated alcohol (furfuryl alcohol) to the reaction mixture in the beginning had no effect on the rate of reaction the product term q_FOL was omitted.

Finally the reaction rate for model 1 is expressed as:

The reaction rate can be written as:

With the same procedure describe above, the rate expression for model II can be expressed as:

3.6 Confirmation of reaction mechanism
In order to select a suitable rate equation, a nonlinear least-squares regression analysis was used for each rate equation to obtain the best values of the parameters using the values of the rate constants were fitted of the experimental concentration versus time profiles of the chemical species involved in the reaction, using a non-linear regression program with an iterative Gauss-Newton method. The objective function to be minimized is residual sum of squares, RSS=∑(r*-r^calc)² where r* is the experimental relative rate, r^calc is the relative concentration calculated with the model. A comparison of the RSS between the models described above, indicates that model II gives the best fit to the experimental data (table 3).

Table 3. Optimized rate parameters for the hydrogenation of furfural at different temperatures, RSS between experimental data and results simulated by mechanism II

Figure 5. Parity plot of calculated concentrations (Eq. 12) with experimental concentrations at various temperatures

The reactions were carried out at different temperatures in the range of 323 to 363 K. Table 3 displays the optimized rate parameters at different temperatures for model II. From the temperature dependence of rate parameters the activation energy was calculated. The surface activation energy near of 12 kcal mol^-1 is similar that the reported for Vaidya et al, for furfural hydrogenation over Pt/C [13]. The enthalpy of adsorption and the entropy of adsorption of furfural and hydrogen were calculated from K_FAL and K_H2 using the van't Hoff equation, these values are listed in Table 4. Both enthalpies and entropies of adsorption are negative, as they should be, and quantitatively satisfies guidelines established, according to their thermodynamic consistency [19].

Table 4. Thermodynamic parameters for adsorption K_R from expressions encountered for furfural hydrogenation over Ir/TiO₂