The true catalyst in hydrogen transfer reactions with alcohol donors in the presence of RuCl₂(PPh₃)₃ is ruthenium(0) nanoparticles

Abstract

The widespread soluble complex RuCl₂(PPh₃)₃ is evidently not the true catalyst in numerous hydrogen transfer reactions where it has been utilized. In the presence of alcohol donors, particularly under boiling conditions, the complex is swiftly reduced to Ru(0) and forms nanosized clusters which are the genuine catalysts. The so formed nanoparticles are not stable and they slowly agglomerate into larger non-active assemblies which are observable by the naked eye. Both the formation and the agglomeration of these nanoparticles are enhanced in the presence of a base such as NaOH. Conversely, addition of stabilizers, such as surface active agents or active carbon, inhibits the agglomeration process and elongates the life time of the catalyst although the observed activity is reduced. The presence of nanoparticles and their unique role in the catalysis of hydrogen transfer reactions between alcohols and ketones were corroborated by TEM imaging, NMR diffusion measurements, XPS and UV-vis spectroscopy and by kinetic studies.

---

Introduction

Catalytic hydrogen transfer (CTH) reactions were developed over the last 40 years as an alternative to traditional catalytic hydrogenation processes where hydrogen gas is used under pressure.¹ In a typical CTH reaction a hydrogen donor and an acceptor are reacted, in the presence of a catalyst, to yield a reduced AH₂ and an oxidized D, according to the general equation shown in Scheme 1.^2,3

Scheme 1 General formula of a catalytic hydrogen transfer reaction.

Prevalent donors are alcohols and formate salts while typical acceptors are ketones, olefins, nitroaromatic derivatives, aryl halides, imines and other unsaturated molecules. Frequently the reactions are reversible in nature and the equilibrium conversion is dependent on the character of the reactants and the reaction conditions.

Although heterogeneous CTH reactions are known,⁴ the extensive methodology is based on liquid phase homogeneous catalysis.⁵ Ruthenium complexes are recognized as the most active catalysts in these transformations. A well studied CTH reaction is the reversible process shown in Scheme 2 where hydrogen is exchanged between isopropanol and cyclohexanone to form acetone and cyclohexanol.

Scheme 2 Typical CTH reaction.

Several generations of ruthenium catalysts were developed through the years for the above reaction and for other similar transformations. The theme was set off with the long-established, RuCl₂(PPh₃)₃⁶ continued with RuH₂(PPh₃)₃,⁷ [RuCl₂(p-cymene)]₂⁸ and the chiral Ru-TsDPEN developed by Noyori et al.⁹ in the 1990's which was remarkably applied in enantioselective CTH. More recently reported catalysts are carbene based ruthenium complexes¹⁰ and different types of pincer complexes¹¹ followed by the Shvo's catalyst¹² and complexes carrying peptide ligands.¹³ A very active family of ruthenium(II) NNN CTH catalysts was recently reported by Yu and coworkers.¹⁴

A variety of molecular mechanisms were proposed for the alcohol to ketone CTH reactions. These were reviewed by Backvall et al.¹⁵ and include monohydride and dihydride mechanisms, inner and outer sphere mechanisms, metal ligand bifunctional catalysis, concerted vs. stepwise mechanisms and ionic mechanisms. Some of the latter are supported also by theoretical studies.

It was early on demonstrated that the above catalysis is far more effective under basic conditions^16,17 and that RuH₂(PPh₃)₃ is an intermediate and a more active catalyst than the original dihalide derivative for transfer hydrogenation.¹⁸ Indeed, bases often served as activators of hydrogen transfer catalysts and as promoters for hydrogen transfer reactions.¹⁹ In fact, some reports depict systems where the base alone catalyzes the reaction portrayed in Scheme 2. Nonetheless, several authors advocated a “base free” hydrogen transfer catalysis although a base is reportedly present in the preparation of the catalyst.²⁰

Concurrent with the above developments, a number of authors have introduced neat alcohols²¹ and combined, alcohol/base²² systems as effective reducing agents for the fabrication of metallic nanoparticles via reduction of various transition metal salts and complexes.

Chemical reduction of transition metal salts is the most widely used method for the generation of colloidal suspensions of metallic particles. Typical reducing agents are hydrogen, borohydrides, formic acid, hydrazine, ascorbic acid, citric acid, etc.²³ Some transition metal salts can be straightforwardly reduced in refluxing alcohol.²⁴ In this process, the alcohol functions both as a solvent and as a reducing agent. During the reduction, the alcohol is oxidized to the corresponding carbonyl compounds as shown in Scheme 3.

Scheme 3 Reduction of a metallic salt to a free metal by alcohol.

Hirai et al.,²⁵ and more recently, Delmas et al.,²⁶ have extensively used aqueous alcohols as reducing agents in the syntheses of colloidal transition metals such as Rh, Pt, Pd, Ag, Os, or Ir, normally in the presence of stabilizing agents to prevent the nanoparticles from agglomeration and precipitation. Several groups explored the impact of various experimental parameters on the characteristics of the generated nanoparticles. Thus, for example, with alteration of the nature and amount of the alcohol and base,²⁷ structure and concentration of the stabilizing agent and the properties of the metallic precursor,²⁸ the particle size distribution and the morphology of the produced clusters could be modified and occasionally controlled.²⁹ Recently, it was disclosed that ruthenium nanoparticles can be effectively produced by the reduction of RuCl₃ in low boiling point alcohols, such as n-butanol, n-propanol and ethanol, upon prolonged refluxing.³⁰ Although the common precursors in the above reductions are metal salts, examples where the starting materials are metal complexes or chelates are also known. Thus Zhu and coworkers have prepared Ru nanoparticles by reduction of [CpRuCp*RuCp*]PF₆ (Cp* = C₅Me₅) in ethylene glycol with hydrogen at 180°. [IrCl(cod)]₂ and Ru(cod)(cot) were also used as precursors for generation of Ir(0) and Ru(0) nanoparticles.³¹

It was envisaged by Widegren and Finke³² that in numerous, allegedly homogeneous, catalytic hydrogenation reactions, the true catalysts are the in situ formed metallic nanoparticles and not the original soluble metal complexes. Several methods were exercised to corroborate the role of nanocluster heterogeneous catalysis in the above reactions. These include sigmoidal kinetic profiles, TEM analysis for detection of nanoparticles, mercury poisoning that affects only heterogeneous catalysts, dibenzocyclooctatetraene (DCT) which is a specific inhibitor to certain homogeneous catalysts, light scattering, centrifugation and filtering of the in situ formed particles. Putative homogeneous catalysts such as Pt(1,5-COD)Cl₂, Pt(1,5-COD)(CH₃)₂, Ru(η⁶-C₆Me₆)O₂CMe)₂ and RuCl₂(μ-H)₂(μ-Cl)(η⁶-C₆Me₆) were all demonstrated to operate only as catalyst precursors and that nanoparticles are the true catalysts in all these cases. Of particular interest is the role of the original ligands in the formation, properties and fate of the metallic nanoparticles. Finke has provided a comprehensive list of catalytic reactions (mainly hydrogenations) assumed to be homogeneous which warrant a detailed analysis of the true nature of the catalyst. These suspected reactions engage either easily reduced transition metal complexes, forcing reaction conditions, the presence of nanocluster stabilizers,³³ formation of dark reaction solution or metallic precipitates or occurrence of an induction period before the reaction starts. His list did not incorporate catalytic hydrogen transfer reactions which are the theme of this study.

Distinguishing metal–complex homogeneous catalysis from the in situ-formed metal–particle heterogeneous catalysis is not trivial.³⁴ It is a task which has caused considerable confusion in the literature. The challenge is most demanding when a soluble colloidal/nanocluster catalyst is involved, partly because colloidal solutions often appear homogeneous to the naked eye. Additionally, nanoparticles/colloids can be as small as 0.5–1 nm in diameter, which makes them difficult to detect even by some instrumental methods.

Several criteria and tests were put forward by Finke to distinguish nanocluster catalysts from traditional homogeneous catalysts.³⁵ In addition, based on comprehensive kinetic studies, a novel autocatalytic mechanism for nanoparticles formation has been proposed. The latter consists of a slow initial nucleation followed by fast autocatalytic surface growth leading to near mono-dispersed nanoclusters.³⁶

The nature and abundance of active sites present on metal–particle may change with particle size, the method of preparation and reaction conditions. Thus, the catalyst activity, selectivity, stability, and lifetime strongly depend on the experimental conditions and on the presence of stabilizing agents.²⁹

In the present study, we have unequivocally established that the first generation ruthenium complexes, RuCl₂(PPh₃)₃⁶ and RuH₂(PPh₃)₃,¹⁸ are not the true catalysts in the CTH reactions reported. In fact, in the presence of alcohols, a rapid reduction of the latter to metallic Ru(0) is taking place. We presume that the nanosized Ru clusters formed are modified and partially stabilized by the presence of the original ligands such as PPh₃ that adsorb to the metallic surface. We confirmed that these clusters are very active catalysts for CTH reactions although, if not stabilized properly, they tend to lose activity via agglomeration, particularly under basic conditions.

Results and discussion

The classical reduction of cyclohexanone to cyclohexanol (Scheme 2) with 2-propanol, both as a solvent and a H donor and RuCl₂(PPh₃)₃ (1) as catalyst, was selected as the model reaction following Backvall's account.¹⁶ By carefully following the procedure reported we clearly noticed, after one hour of reaction, the formation of a solid dark slurry in the reaction mixture. SEM analysis of the latter, after standing for 48 hours, revealed the presence of particles with size of 20–100 nm which were totally inactive in catalysis shown in Scheme 2. This is shown in Fig. 1.

Fig. 1 Reaction mixture (Scheme 2) after one hour (left); agglomeration of nanoparticles detected by SEM after reaction and standing for 48 h (right).

Reaction kinetics

The kinetics in Scheme 2 was studied, under the exact Backvall's conditions.¹⁶ Four consecutive runs were performed with addition of a fresh substrate at the end of each batch.

Samples were taken from the reaction mixture and analyzed by gas chromatography. The reaction profiles of the four runs are shown in Fig. 2.

Fig. 2 Reaction profiles of four consecutive runs. Reagents and conditions: 1 (0.01 mmol), cyclohexanone (5 mmol), isopropanol (15 ml), NaOH (0.24 mmol), 82 °C under N₂. Addition of fresh 5 mmol cyclohexanone after 3 hours to each batch.

We calculated the TOF in these four runs at the point of maximum velocity for each: 16.6 min⁻¹ for the first batch; 2.5 min⁻¹ for the second and 0.7 min⁻¹ for the third. In the fourth batch no reaction was observed. The catalyst decay under these conditions is evident. We attribute the deactivation of the catalyst to agglomeration of the metallic nanoparticles which is noticeable by the naked eye as early as after the first batch.

Very similar behavior was observed when other ketones such as acetophenone or cyclopentanone were reacted with 2-propanol under identical conditions. Conversely, similar transfer hydrogenation of benzaldehyde proceeded very slowly.

An important stage in the original procedure¹⁶ is a pretreatment step where the catalyst is dissolved in isopropanol and warmed to reflux for several minutes followed by addition of cyclohexanone and another period of heating prior to the addition of NaOH that initiates the process. In attempt to figure out what occurs in the course of the pretreatment advocated by Backvall, we examined the impact of heating 1 in isopropanol for periods of 0 minute to 5 hours prior to the addition of the substrate and the base. The kinetic profiles of these five runs, after addition of the substrate and the base, are shown in Fig. 3a and b.

Fig. 3 Influence of the pretreatment on the rate of the catalytic reaction (time needed for 100% conversion as a function of pretreatment time). Reagents and conditions: 1 (0.01 mmol), cyclohexanone (10 mmol), temperature (82 °C), isopropanol (15 ml), NaOH (0.24 mmol), under N₂.

It is evident that the rate and the general behavior of the reaction shown in Scheme 2 strongly depend on the duration of the pretreatment. The highest rate is measured after pretreatment for 30 minutes, while shorter or longer treatment results in inferior performance of the catalyst. Note that after a pretreatment for 6 hours the reaction is slow and the catalyst is deactivated before 100% conversion is attained. An induction time of 8 minutes is observed only in the run without pretreatment.

Note that in all these runs the catalyst was completely dissolved in the solvent at 25 °C before the start of the experiment.

Similar kinetic behaviour was realized when two other first generation CTH catalysts were tested, namely RuCl₃ and RuH₂(PPh₃)₃. The latter exhibited a shorter induction period but decayed faster than 1 while RuCl₃ has a longer induction but retained its activity for longer time. This is shown in Fig. 4.

Fig. 4 Catalytic performance of RuCl₃, RuH₂(PPh₃)₃ and 1 shown in Scheme 2 without pretreatment. Reagents and conditions: 1 or RuH₂(PPH₃)₃ (0.01 mmol) and RuCl₃ (0.1 mmol), cyclohexanone (10 mmol), temperature (82 °C), isopropanol (15 ml), NaOH (0.24 mmol).

No induction period and relatively poor activity were observed when a commercial heterogeneous 5%Ru/C catalyst was used in reaction shown in Scheme 2. Same was realized when we used ruthenium nanoparticles prepared in our hands as catalysts. The latter was fabricated by reduction of RuCl₃ with sodium borohydride. Kinetic profiles are shown in Fig. 5.

Fig. 5 Catalysis by 1, Ru/C and Ru nanoparticles in Scheme 2 (no pretreatment). Conditions as in Fig. 3.

Color changes

Unique color changes take place in the course of reaction shown in Scheme 2 (Fig. 6). The initial color of the solution of 1 in isopropanol is green (2i) which turns into grey upon heating to boiling point (2a). After addition of the base, the color turns red (2b) and in the course of the reaction the solution becomes pale yellow with the formation of a black precipitate which is easily noticeable after standing for several hours at room temperature (4f). Some of these color changes were also observed by Backvall.¹⁶

Fig. 6 Color changes during the reaction (Scheme 2). Reagents and conditions: 1 (0.01 mmol), cyclohexanone (10 mmol), temperature (82 °C), isopropanol (15 ml), NaOH (0.24 mmol).

Interestingly some authors did notice the formation of a black precipitate in the course of alcohol dehydrogenations and CTH reactions but no attempt was made to figure out the role of this phenomenon in the catalytic process.^1,16,37

Catalyst isolation and characterization

TEM study. The consequence of the pretreatment in isopropanol on the catalyst was studied by TEM analysis.

A sample of 1 in isopropanol was maintained at 82 °C (reflux) for 30 minutes and after drying it was analyzed by transmission electron microscopy (TEM). The result is shown in Fig. 7. This experiment clearly confirms the presence of nanoparticles, in the size range of 0.5–1.5 nm. The particle size distribution at this stage of the process is shown in Fig. 8.

Fig. 7 Ruthenium nanoparticles after 30 min pretreatment of 1 in isopropanol at 82 °C visualized by TEM.

Fig. 8 Particle size histograms of ruthenium nanoparticles after pretreatment of 1 in isopropanol.

Moreover, analysis of the latter sample by Electron Diffraction and High Resolution Imaging showed a 1D fringe with a d spacing of about 0.20–0.21 nm (Fig. 9). This result corroborates the effective reduction of the initial ruthenium(II) to ruthenium(0). The regular structure of the Ru metal is hexagonal close-packed with the unit cell parameter equal to 0.27 nm and 0.43 nm and d spacing equal to 2.06 Å and 2.14 Å corresponding to the crystal plane.³⁸

Fig. 9 TEM image through Electron Diffraction (right) and High Resolution (left) reveals the fundamental state of ruthenium(0).

The appearance of the above nanoparticles is changing upon addition of NaOH to the mixture. We observed that with NaOH the nanoparticles swiftly agglomerate to generate large amorphous assemblies that are essentially inactive in catalysis (Fig. 10).

Fig. 10 Agglomeration of Ru nanoparticles after addition of NaOH following pretreatment with 2-propanol at 82 °C for 30 min.

Similar TEM analyses following identical pretreatment were carried out with RuCl₃ (Fig. 11, left) and RuH₂(PPh₃)₃ (Fig. 11, right). Ru(0) nanoparticles with diameters of 0.5–2 nm were detected and identified by electron diffraction. Evidently RuH₂(PPh₃)₃ is not the true catalyst (Scheme 2) as previously asserted.¹⁵

Fig. 11 Nanoparticles of ruthenium detected by TEM formed from RuCl₃ (left) and from RuH₂(PPh₃)₃ (right).

Diffusion NMR. Diffusion ordered spectroscopy (DOSY)³⁹ obtained through pulsed field gradient (PFG) NMR experiments⁴⁰ has become a multi-purpose tool in the chemical laboratory due to its ability to gain insight into the inner behavior of solution systems through the measurement of diffusion coefficients (D) and their correlation to molecular properties such as size, weight, shape, etc. or the nature of interaction between different molecules.⁴¹ The DOSY technique has been applied to study the dynamics of formation of nanoparticles, having lower diffusion rates than the starting molecules.⁴² Although the metallic nanoparticles themselves cannot be detected by NMR, measurements of diffusion coefficients of species adsorbed on the catalyst particle surface allow us to perform the test and draw critical and indisputable conclusions concerning the formation and behavior of nanoparticles. In our case these adsorbed species probably comprise hydrides and free triphenylphosphine molecules. The self-diffusion constant is measured in m² s⁻¹ and for metallic nanoparticles it is expected to be larger by several orders of magnitude than that of typical organic or organometallic molecules.

We initially measured the self-diffusion coefficient of RuCl₂(PPh₃)₃ and of triphenylphosphine in benzene-d₆ at 25 °C. These were 0.11–6.3 × 10⁻¹⁰ m² s⁻¹ for 1 (Fig. 12, left) and 2.5–1.3 × 10⁻¹⁰ m² s⁻¹ for the ligand (Fig. 12, right). Using the Stokes–Einstein equation⁴³ we estimated the molecular diameter of RuCl₂(PPh₃)₃ at 6–11 Å and of PPh₃ at 3.5.–5.5 Å (note that the viscosity of benzene d₆ is 0.6076 mPa s).

Fig. 12 DOSY spectrum of RuCl₂(PPh₃)₃ (left) and of triphenylphosphine (right) in benzene-d₆.

Then the reaction mixture of 1 with 2-propanol in the presence of NaOH was studied by diffusion 1H NMR spectroscopy with the aim of detecting the formation of nanoparticles. The spectrum of 1 heated in isopropanol to 82 °C in the presence of NaOH under nitrogen for 30 minutes is shown in Fig. 13. It is apparent that the original peaks of coordinated triphenylphosphine at δ = 6.9–7.7 ppm were replaced by different peaks at δ = 6, 4 and 2 ppm (marked b) which are attributed to species adsorbed on the metallic surface. The latter were measured to have a low diffusion coefficient of 2.5–6.31 × 10⁻¹¹ m² s⁻¹. Traces of fragments absorbing at δ = 7–8 ppm with a diffusion coefficient of 3.98 × 10⁻⁹ m² s⁻¹ are also observable. Using the Stokes–Einstein equation we estimated the van der Waals diameter of the main product to be 10.8–28 nm and of the smaller fragment to be 1.8 Å. The latter is most likely a fragment of triphenylphosphine removed from the original complex in the course of formation of nanoparticles.

Fig. 13 (Left) H-NMR and (right) DOSY spectrum of ruthenium in benzene-d₆ separating Ru nanoparticles (b) and fragments of ligand molecules (a).

Interestingly we did not detect any peaks with δ < 0 that could be assigned to the presence of an intermediate hydride complex such as RuH₂(PPh₃)₃.⁴⁴

XPS measurement. To further confirm the formation of metallic ruthenium nanoclusters in the course of Scheme 2, XPS was employed to determine the oxidation state of surface ruthenium atoms obtained in the above pretreatment after evaporation of the solvent. This is shown in Fig. 14. The observed binding energy of Ru 3d_5/2 (280.0 eV) and Ru 3d_3/2 (284.2 eV) in the clusters are in accordance with those of the bulk ruthenium metal.⁴⁵ Note the significant fraction of oxide species detected on the particles' surface. These could be attributed to 2-propanol molecules adsorbed on the surface or to ruthenium oxide species formed by air oxidation of the nanoparticles.

Fig. 14 XPS spectrograms for ruthenium nanoclusters formed in a typical pretreatment.

UV measurement. UV-visible absorption spectrophotometry has been considered as an effective method to monitor the evolution of metal species in the preparation of colloidal metal clusters.⁴⁶

We have examined the visible spectra of a 10⁻³ M solution of 1 in 2-propanol at 82 °C as a function of time. Results are shown in Fig. 15A. It is apparent that the original complex solution shows λ_max at 664 nm. The latter gradually decreased with formation of a new peak at 642 nm which eventually disappears after 30 minutes.

Fig. 15 Visible spectral change in 10⁻⁴ mol L⁻¹1–2-propanol solution as a function of time at 82 °C without (A) and with NaOH (B).

In view of the Mie theory⁴⁷ that allows the reckoning of the spectrum of a ruthenium colloid, we may safely claim that a Ru(0) colloid has been quantitatively formed in the above solution.

When the same test was repeated with addition of a solution of 0.2 M of NaOH, identical transformation was realized at a much higher rate. This is shown in Fig. 15B, the original peak at 664 nm disappeared within minutes.

The change in the absorption of the above solutions could be monitored even with the naked eye. In Fig. 16A one can see the change of color of the 1-IPA solution from initial green to brown yellow after 15 minutes corresponding to the formation of nanoparticles. In the presence of NaOH (Fig. 16B) the conversion is faster. After 2 minutes the green color changed to brown yellow and after 10 minutes to dark brown. The latter color is attributed to assemblies of nanoparticles.

Fig. 16 Color changes of 10⁻³ mol L⁻¹ RuCl₂(PPh₃)₃–IPA solution as a function of time at 85 °C with (A) and without (B) NaOH.

When the same experiment was run at 25 °C we could detect in the visible spectrum the presence of an intermediate peak at λ_max = 642 nm that after merely 5 seconds replaced the original peak at λ_max = 664 nm and then disappeared after 20 minutes. This is shown in Fig. 17. We attribute the intermediate peak to a ruthenium hydride complex which is a plausible intermediate in the reduction process of 1 to Ru(0).

Fig. 17 Visible spectral change of 10⁻⁴ mol L⁻¹ solution of 1 in 2-propanol with NaOH as a function of time at 25 °C.

UV absorption spectroscopy of 1 in 2-propanol was carried out similarly at lower concentration (10⁻⁴ M). Results are shown in Fig. 18A. The original peak at 254 nm disappears completely after 5 minutes at 82 °C with formation of a new peak at 265 nm which reached its maximum intensity after 15 minutes. In the presence of NaOH the original peak at 254 nm vanished instantly and a new absorption maximum at 265 nm was formed (Fig. 18B). Since Ru(0) colloids are not expected to have a UV absorption⁴⁷ we verified that the peak at 265 nm actually fits into free uncomplexed triphenylphosphine. The spectrum of the latter is shown in Fig. 18B for comparison.

Fig. 18 UV absorption spectra of 10⁻⁴ M RuCl₂(PPh₃)₃–IPA solution as a function of time at 85 °C with (A) and without (B) NaOH.

We can thus safely conclude that complex 1 is not stable in boiling 2-propanol and is swiftly reduced to Ru(0) which forms nanoparticles. This transformation is even faster in the presence of a base. In the course of the reduction process the triphenylphosphine ligand is evidently released to solution while part of it is adsorbing to the formed metallic surface.

Based on the above observations we infer that in all the previous studies where RuCl₂(PPh₃)₃ has been used as a catalyst in hydrogen transfer reactions with alcohols as donors, with or without an added base, the true catalyst is nanoparticles of Ru(0). The latter are particularly catalytically active when the particle size is 0.5–1.0 nm. The catalytic activity is gradually lost with agglomeration of the small particles to larger cluster assemblies. The agglomeration process is particularly swift in the presence of NaOH.

Poisoning experiments

Upon addition of mercury to the reaction mixture in Scheme 2, at the start of the pretreatment, we have observed a sharp decrease in the reaction rate with TOF falling from 16.6 to 5 min⁻¹. It is reasonable to assume that soluble nanoclusters are not affected by the presence of elementary mercury and it is only the non-soluble aggregates which are affected.

Hence, the suppression or diminishing of catalysis inflicted by Hg(0) is an additional evidence for the role of a heterogeneous catalyst, as the true catalyst in CTH according to Scheme 2.⁴⁸

Formation of ruthenium nanoparticles

The reduction process of the initial (inactive) precatalyst, with and without a base, is a critical step in the RuCl₂(PPh₃)₃ driven CTH process (Scheme 2). We have identified several key experimental variables which are crucial to the effectual transformation of the homogeneous precatalysts into active nanoparticles. These are listed below:

Inert atmosphere. The performance of the catalytic system (Scheme 2) was very different when the pretreatment of the catalyst was carried out under air instead of under the standard inert nitrogen conditions. Thus when the pretreatment was carried out under nitrogen and the reaction itself under air, the system still maintained 80% of its activity. However, with pretreatment under air, followed by reaction under nitrogen, the maximum rate observed was only 25% of that under the standard conditions. Indeed the sensitivity of the formed nanoparticle to oxidation was also verified in the XPS experiment described above (Fig. 14).

The role of boiling conditions. We perceived that the process of boiling is crucial in the generation of nanoparticles via reduction of the original complex 1. Thus at 82 °C, where reflux is taking place, the reaction (Scheme 2) is rapid as described above (see Fig. 19). However at 70 °C, without reflux, and surprisingly also in a closed vessel at 82 °C (also without reflux) the reaction is sluggish and hardly proceeds. We confirmed that boiling conditions are vital only to the first step where nanoparticles are evidently formed. Thus when the pretreatment is done at 82 °C and the following CTH reaction at 70 °C, the process ensues. In the opposite scheme (where the pretreatment is carried out at 70 °C and upon addition of the base and substrate the system is heated to 82 °C) the reaction is slow moving. This is shown in Fig. 19.

Fig. 19 Influence of temperature at different stages of the process in Scheme 2.

The critical role of boiling situation was also established with the application of another secondary alcohol, 2-butanol, as hydrogen donor in Scheme 2. From the chemical point of view one should not expect any major difference in the activity of 2-butanol in comparison to 2-propanol. Nonetheless, running the pretreatment and the reaction with 2-butanol at 82 °C resulted in a very slow reaction where the maximum measured TOF was only 3.0 min⁻¹ (see Fig. 20).

Fig. 20 TOF of different hydrogen donors as a function of temperature.

This rate of reaction with 2-butanol almost did not change at 90 °C and at 95 °C attaining TOF of 3.3 and 3.5 min⁻¹ respectively but upon boiling at 100 °C the maximum observed TOF spiked to 16 min⁻¹ (identical to the rate of 2-propanol at 82 °C). Although aliphatic primary alcohols are not active as hydrogen donors in Scheme 2, benzyl alcohol is fairly effectual. Again the reaction rate is significant only at the boiling point with clear reflux conditions (180 °C), where a maximum TOF of 35 min⁻¹ was measured. At 175 °C, however, the maximum rate measured was TOF = 6 min⁻¹ and at 120 °C the reaction did not proceed at all. These results are summarized in Fig. 20. Note the sharp increase in TOF observed upon boiling in the three alcohols tested. For comparison the maximum observed TOF without pretreatment at the boiling point is also displayed. To further corroborate the critical role of boiling, Scheme 2 was tested with a 1 : 1 mixture of THF/2-propanol as a solvent which boils at 70 °C. Indeed, at this temperature the measured TOF attained a significant number of 12.5 min⁻¹. At the same temperature, reaction in pure 2-propanol did not proceed at all (see Fig. 19).

The exact role of reflux circumstances in the formation process of the catalytically active nanoparticles is not clear at this point. We hypothesize that the phenomenon is related to microscale mechanical forces inflicted on the initial nanocrystal seeds by the bubble formation in the boiling process. Our attempts to achieve similar results by applying ultrasound irradiation on the reaction mixture, at different temperatures below the boiling point, failed.

Stabilization of catalytically active NP

The catalytic activity of the ruthenium nanoparticles formed in situ in the above CTH reaction is evidently very high upon their formation. The activity is however rapidly lost with the agglomeration and formation of non-soluble and far less effective metallic clusters which is particularly noticeable in the presence of a base. This undesired phenomenon can potentially be avoided or at least slowed down, by applying means to stabilize the nanoparticles in their active state. For that aim we have examined several known methods for stabilization of nanoparticles. These include stabilization by surface active agents, by supporting the nanoparticles on a high surface area solid carrier or by microencapsulation. We have thus examined the activity of 1 in the presence of Aliquat 336, supported on active carbon and encapsulated in a polyurea matrix.⁴⁹ Results are summarized in Table 1 where the above formulations were tested, each in three consecutive runs and compared with the performance of the original free catalyst.

Table 1 Comparison of TOF for RuCl₂(PPh₃)₃ stabilized by different methods

Method of stabilization	TOF batch 1/min^−1	TOF batch 2/min^−1	TOF batch 3/min^−1
a With 2.86% of the Ru content in PU.
Free 1	16.6	2.5	0.7
With addition of Aliquat 336	12.1	9.9	3.3
Absorbed on active carbon	16.6	12	9
Encapsulated in polyurea matrix^a	0.19	0.19	0.19

---

It is clearly observed that when the unmodified RuCl₂(PPh₃)₃ precatalyst is used, the catalytic activity is decreasing by 90% at the second batch and the catalyst is almost dead at the third batch. With an Aliquat 336 stabilizer, the rate at the first batch is lower by 25% but the activity is decreasing slower (22% at the second batch and 66% at the third batch). Active carbon is a superior stabilizing agent which does not alter the original rate at the first batch while losing only 25% of the activity in each of the second and third runs. Interestingly, with the Ru nanoparticles encapsulated in a polyurea matrix the rate is decreasing by two orders of magnitude relative to the free catalyst at the first round, but this low activity is fully preserved in the second and the third batches. We may conclude that Aliquat 336 and active carbon slow down the deactivation of the nanoparticles by agglomeration while the polyurea encapsulation totally prohibits the agglomeration although at a cost of much inferior activity.

Conclusions

We have unequivocally confirmed that the RuCl₂(PPh₃)₃ is not catalytically active in transfer hydrogenation reactions between alcohols and ketones. The role of this, allegedly homogeneous, catalyst is evidently to provide a source for the formation of catalytically active ruthenium metal nanoparticles. Several methods were used to authenticate the presence and to determine the role of nanoparticles in the catalysis shown in Scheme 2.

Experimental section

General considerations

All experiments were carried out under an inert nitrogen atmosphere using standard Schlenk techniques. Solvents were distilled from the appropriate drying agents and degassed before use.

RuCl₂(PPh₃)₃ was prepared according to the method described in the literature.⁵⁰

Catalytic activity measurements

Samples from kinetic runs were analyzed by a Thermo Focus gas chromatograph equipped with a 95%-dimethylpolysiloxane-5%-diphenyl packed column (30 m length and 0.25 mm) and an FID detector. Helium was used as a carrier gas at a pressure of 50 kPa.

Catalyst characterization

Several experimental techniques were used to characterize the various samples tested.

Transmission Electron Microscopy (TEM) was performed by the high resolution TEM Tecnai F20 G². The microscope combines ultra-high resolution performance (point resolution 0.24 nm, line resolution 0.1 nm, limit of information 0.15 nm, HR STEM resolution <0.2 nm) with extended analytical abilities and equipped with an energy dispersive X-ray spectroscopy EDS detector.

UV-vis spectra were measured by a UV-vis (Varian EL-03097225) spectrophotometer using isopropanol as the reference.

H-NMR spectra were recorded on a 500 MHz Burker Avance II instrument. Diffusion rates are measured by COSY-NMR.

Typical procedure for catalytic transfer hydrogenation of cyclohexanone

RuCl₂(PPh₃)₃ (9.5 mg, 0.01 mmol) was dissolved in degassed propan-2-ol (10 ml) and the mixture was heated at 82 °C for a defined pretreatment time under nitrogen. Cyclohexanone (0.5 ml, 5.0 mmol) was added to the refluxing mixture and after additional 10 minutes a solution of NaOH (9.5 mg, 0.237 mmol) in propan-2-ol (2 ml) was added dropwise. The mixture was stirred at 82 °C for 1 h. The conversions were determined by GC and the products could be isolated by silica gel column chromatography.

Procedure for catalytic transfer hydrogenation of cyclohexanone in the presence of Aliquat 336

To solid RuCl₂(PPh₃)₃ (9.5 mg, 0.01 mmol), after evacuation and purging with nitrogen (×3), was added degassed propan-2-ol (5 ml) containing Aliquat 336 (0.5 g, 1 mmol) and the mixture was stirred at 82 °C for 30 min under N₂. Cyclohexanone (0.98 g, 10.0 mmol) dissolved in degassed propan-2-ol (3 ml) was added to the refluxing mixture. Then, a solution of NaOH (9.5 mg, 0.237 mmol) in propan-2-ol (2 ml) was added dropwise. The reaction was continued for 1 hour.

Synthesis of Ru(0) nanoparticles and their application as catalysts in Scheme 2

Aliquat 336 (0.1 g, 0.25 mmol) and 5 g of toluene were mixed in a 50 mL flask. To this solution 0.01 mmol (0.05 g) of RuCl₃ was added and the mixture was stirred at 50 °C for 10 min. Then, this mixture was added to 10 ml of water containing 0.01 g (0.25 mmol) of sodium borohydride. Instant and rapid reduction was observed and stirring was continued for 30 min. This solution was used directly as catalyst in CTH reaction by addition of 15 ml of 2-propanol and cyclohexanone (0.98 g, 10.0 mmol). Then, a solution of NaOH (9.5 mg, 0.237 mmol) in propan-2-ol (2 ml) was added dropwise.

Synthesis of nanoparticles adsorbed on active carbon and their application in Scheme 2

To solid RuCl₂(PPh₃)₃ (9.5 mg, 0.01 mmol), after evacuation and purging with nitrogen (×3), degassed propan-2-ol (5 ml) containing 100 mg of active carbon was added and the mixture was heated at 82 °C for 30 min under N₂ and at room temperature for additional 24 h with stirring. The product is separated by filtration and is directly applied as catalyst in Scheme 2.

Synthesis of Ru(0) nanoparticles encapsulated in polyurea

RuCl₂(PPh₃)₃ (95 mg, 0.1 mmol) was stirred at 82 °C for 20 min in 100 ml 2-propanol and the mixture was evaporated under vacuum to obtain 1 ml of solution. The latter was added to a solution of Aliquat 336 (0.5 g, 1 mmol) in 5 g of toluene which was stirred at 50 °C for 10 min. Then, 1.2 g (5 mmol) of PMDI was added to the above solution and the mixture was stirred for 5 min. This obtained solution was slowly added to 100 ml of water containing 0.1 g (1 mmol) of diethylenetriamine. Instant and rapid polymerization was observed and stirring was continued for 1 h using an overhead mechanical stirrer at 200 rpm at room temperature. The resulting microcapsules were filtered through a polyethylene frit (20 micron porosity), washed by de-ionized water and by dichloromethane and dried at 100 °C for 4 h to obtain the polyurea encapsulated catalyst, a grey solid powder.

Notes and references

1. M. Gagliardo, P. A. Chase, S. Brouwer, G. P. M. van Klink and G. van Koten, Organometallics, 2007, 26, 9.
2. S. E. Clapham, A. Hadzovic and R. H. Morris, Coord. Chem. Rev., 2004, 248, 2201.
3. J. E. Backvall, J. Organomet. Chem., 2002, 652, 105.
4. R. A. W. Johnstone and A. H. Wilby, Chem. Rev., 1985, 85, 129.
5. G. Brieger and T. J. Nestrick, Chem. Rev., 1974, 74, 567.
6. Y. Sasson and J. Blum, J. Org. Chem., 1975, 49, 1887.
7. A. Aranyos, G. Csjernyik, K. J. Szabo and J. E. Backvall, Chem. Commun., 1999, 351.
8. (a) E. Hodsonand S.J. Simpson, Polyhedron, 2004, 23, 2695; (b) R. V. Wisman, J. G. de Vries, B.-J. Deelman and H. J. Heeres, Org. Process Res. Dev., 2006, 10, 423.
9. K. J. Haack, S. Hashiguchi, A. Fujii, T. Ikariya and R. Noyori, Angew. Chem., Int. Ed., 1997, 36, 285.
10. Z. Strassberger, M. Mooijman, E. Ruijter, A. H. Alberts, C. De Graff, R. V. A. Orru and G. Rothenberg, Appl. Organomet. Chem., 2009, 24, 142.
11. W. Baratta, G. Chelucci, S. Gladiali, K. Siega, M. Toniutti, M. Zanette, E. Zangrando and P. Rigo, Angew. Chem., Int. Ed., 2005, 44, 6214.
12. S. Gladiali and E. Alberico, Chem. Soc. Rev., 2006, 35, 226.
13. A. Bøgevig, I. M. Pastor and H. Adolfsson, Chem.–Eur. J., 2004, 10, 294.
14. (a) W. Ye, M. Zhao, W. Du, Q. Jiang, K. Wu, P. Wu and Z. Yu, Chem.–Eur. J., 2011, 17, 4737; (b) M. Zhao, Z. Yu, S. Yan and Y. Li, Tetrahedron Lett., 2009, 50, 4624.
15. J. S. M. Samec, J. E. Backvall, P. G. Andersson and P. Brandt, Chem. Soc. Rev., 2006, 35, 237.
16. R. L. Chowdhury and J. E. Backvall, J. Chem. Soc., Chem. Commun., 1991, 1063.
17. R. Grigg, T. R. B. Mitchell and S. Sutthivaiyakit, Tetrahedron, 1981, 37, 4313.
18. H. Imai, T. Nishiguchi and K. Fukuzumi, J. Org. Chem., 1976, 41, 2688.
19. (a) A. Camus, G. Mesroni and G. Zassinovich, J. Mol. Catal., 1979, 6, 231; (b) G. Z. Wang and J. E. Backvall, J. Chem. Soc., Chem. Commun., 1992, 337; (c) M. L. S. Almeida, M. Beller, G. Z. Wang and J. E. Backvall, Chem.–Eur. J., 1996, 2, 1533.
20. (a) K. Eceraere, J. L. Scheffler, A. Mortreux and J. F. Carpentier, Tetrahedron Lett., 2001, 42, 1899; (b) R. Castarlenas, M. A. Esteruelas and E. Onate, Organometallics, 2008, 27, 3240; (c) H. Zhang, C. B. Yang, Y. Y. Li, Z. R. Dong, J. X. Gao, H. Nakamura, K. Murata and T. Ikariya, Chem. Commun., 2003, 142.
21. A. B. R. Mayer, J. E. Mark and S. H. Hausner, Angew. Makromol. Chem., 1998, 259, 45.
22. W. Yu, M. Liu, H. Liu and J. Zheng, J. Colloid Interface Sci., 1999, 210, 218.
23. L. D. Pachon and G. Rothenberg, Appl. Organomet. Chem., 2008, 22, 288.
24. A. Roucoux, J. Schulz and H. Patin, Chem. Rev., 2002, 102, 3778.
25. (a) H. Hirai, Y. Nakao and N. Toshima, J. Macromol. Sci., Chem., 1978, A12, 1117; (b) H. Hirai, Y. Nakao and N. Toshima, J. Macromol. Sci., Chem., 1979, A13, 727; (c) H. Hirai, J. Macromol. Sci., Chem., 1979, A13, 633.
26. A. Borsla, A. M. Wilhelm and H. Delmas, Catal. Today, 2001, 66, 389.
27. W. Yu, M. Liu, H. Liu and J. Zheng, J. Colloid Interface Sci., 1999, 210, 218.
28. (a) T. Teranishi and M. Miyake, Chem. Mater., 1998, 10, 594; (b) T. Teranishi, M. Hosoe and M. Miyake, Adv. Mater., 1997, 9, 65; (c) G. W. Busser, J. G. Van Ommen and J. A. Lercher, in Advanced Catalysis and Nanostructured Materials, ed. W. R. Moser, Academic Press, San Diego, 1996, p. 213.
29. Y. Xia, Y. Xiong, B. Lim and S. E. Skrabalak, Angew. Chem., Int. Ed., 2009, 48, 60.
30. (a) H. F. Liu, H. j. Niu and J. G. Gao, Chin. J. Polym. Sci., 2003, 21, 1; (b) Y. Zhang, J. Jiu Yu, H. Niu and H. Liu, J. Colloid Interface Sci., 2007, 313, 503.
31. (a) M. L. Rossi and G. Machado, J. Mol. Catal., 2009, 298, 69; (b) K. Pelzer, O. Vidoni, K. Philippot, B. Chaudret and V. Colliere, Adv. Funct. Mater., 2003, 13, 118.
32. (a) J. A. Widegren and R. G. Finke, J. Mol. Catal., 2003, 198, 317; (b) J. A. Widegren and R. G. Finke, J. Mol. Catal., 2003, 191, 187.
33. L. S. Ott and R. G. Finke, Coord. Chem. Rev., 2007, 205, 1075.
34. J. P. Collman, L. S. Hegedus, J. R. Norton and R. G. Finke, Principles and Applications of Organotransition Metal Chemistry, 1987, vol. 10.
35. R. G. Finke, Met. Nanopart., 2002, 2, 17.
36. (a) M. A. Watzky and R. G. Finke, J. Am. Chem. Soc., 1997, 119, 10382; (b) M. A. Watzky and R. G. Finke, Chem. Mater., 1997, 9, 3083; (c) J. A. Widegren, J. D. Aiken, S. Ozkar and R. G. Finke, Chem. Mater., 2001, 13, 312; (d) E. E. Finney and R. G. Finke, J. Colloid Interface Sci., 2008, 317, 351; (e) C. Besson, E. E. Finney and R. G. Finke, J. Am. Chem. Soc., 2005, 127, 8179; (f) E. E. Finney and R. G. Finke, Chem. Mater., 2009, 21, 4468.
37. (a) H. Junger and M. Beller, Tetrahedron Lett., 2005, 46, 103.
38. From JCPDS Power Diffraction File, the standard database for X-ray Powder Diffraction Data for natural and synthetic materials.
39. (a) S. J. Gibbs and C. S. Johnson, J. Magn. Reson., 1991, 93, 395; (b) K. F. Morris and C. S. Johnson, J. Am. Chem. Soc., 1992, 114, 3139; (c) C. S. Johnson, Prog. NMR Spectrosc., 1999, 34, 203.
40. (a) E. O. Stejskal and J. E. Tanner, J. Chem. Phys., 1965, 42, 288; (b) W. S. Price, Concepts Magn. Reson., 1997, 9, 299.
41. (a) K. Nicolay, K. P. J. Braun, R. A. de Graaf, R. M. Dijkhuizen and M. J. Kruiskamp, NMR Biomed., 2001, 14, 94; (b) Y. Cohen, L. Avram and L. Frish, Angew. Chem., Int. Ed., 2005, 44, 520.
42. (a) I. Favier, E. Teuma and G. Montserrat, C. R. Chim., 2009, 12, 353; (b) G. Uccello-Barretta, et al. , J. Organomet. Chem., 2009, 694, 1813.
43. Y. Cohen, L. Avram and L. Frish, Angew. Chem., Int. Ed., 2005, 44, 52.
44. (a) D. E. Linn and J. Halpern, J. Am. Chem. Soc., 1987, 109, 2969; (b) R. U. Kirss, T. C. Eisenschmid and R. Eisenberg, J. Am. Chem. Soc., 1988, 110, 8564.
45. C. W. Wagner, W. M. Riggs, L. E. Davis, J. F. Moulder and B. E. Muilenberg, Handbook of X-Ray Photoelectron Spectroscopy, Perkin-Elmer, Eden Prairie, 1979, p. 106.
46. (a) A. I. Kirkland, P. P. Edwards, D. A. Jefferson and D. G. Duff, Annu. Rep. Prog. Chem., 1990, 87, 247; (b) J. A. Creighton and D. G. Eadon, J. Chem. Soc., 1991, 99, 3881; (c) N. Toshima, M. Harada, T. Yonezawa, K. Kushihashi and K. Asakura, J. Phys. Chem., 1991, 95, 7448.
47. J. A. Creighton and D. G. Eadon, J. Chem. Soc., Faraday Trans., 1991, 87, 3881.
48. (a) G. M. Whitesides, M. Hackett, R. L. Brainard, J. P. P. M. Lavalleye, A. F. Sowinski, A. N. Izumi, S. S. Moore, D. W. Brown and E. M. Staudt, Organometallics, 1985, 4, 1819; (b) Y. Lin and R. G. Finke, Inorg. Chem., 1994, 33, 4891.
49. J. Toubiana, M. Chidambaram, A. Santo and Y. Sasson, Adv. Synth. Catal., 2008, 350, 1230.
50. T. A. Stephenson and G. Wilkinson, J. Inorg. Nucl. Chem., 1966, 28, 945.

---