Nanoheterogeneous catalytic hydrogenation of N-, O

Inorganica Chimica Acta 357 (2004) 3099–3103 www.elsevier.com/locate/ica

Note

Nanoheterogeneous catalytic hydrogenation of N-, Oor S-heteroaromatic compounds by re-usable aqueous colloidal suspensions of rhodium(0) Vincent Mevellec, Alain Roucoux

*

UMR CNRS 6052 ‘‘Syntheses et Activations de Biomolecules’’ Ecole Nationale Superieure de Chimie de Rennes, Institut de Chimie de Rennes Avenue du General Leclerc, 35700 Rennes, France Received 31 December 2003; accepted 17 April 2004 Available online 19 May 2004

Abstract The hydrogenation of various nitrogen-, oxygen- or sulfur-heterocyclic aromatic compounds by various surfactant-stabilized aqueous rhodium(0) colloidal suspensions was investigated. The nanocatalysts in the size range of 2.1–2.4 nm have been synthesized by reducing RhCl3 Æ 3H2 O with sodium borohydride and were stabilized by highly water soluble N,N-dimethyl-N-cetyl-N-(2-hydroxyethyl)ammonium bromide or chloride salts. The catalytic reactions were performed under mild reaction conditions, namely room temperature and under atmospheric hydrogen pressure. The influence of the bromide or chloride nature of the surfactant counter-ion on the recycling of the aqueous phase containing the Rh(0) particles was studied. 2004 Elsevier B.V. All rights reserved. Keywords: Nanoparticles; Rhodium; Hydrogenation; Heteroaromatic compounds

1. Introduction Today, catalysis is one of the most essential research area of metal nanoparticles. Most of those nanoheterogeneous catalysts are largely used in alkene or arene derivatives hydrogenation reactions [1–4]. Nevertheless, only few reports show the catalytic activities of metal nanoparticles in heteroaromatic compounds hydrogenations [5–8]. Generally, the catalytic transformations of mono- or poly-heteroaromatic compound were performed with rhodium, iridium or ruthenium homogeneous or heterogeneous catalysts [8–13]. The total or partial hydrogenation of heteroaromatic derivatives is largely described in the literature as industrial catalytic challenge, in particular for catalytic hydrodenitrogenation (HDN) and hydrodesulfurization (HDS) of petroleum [14–16]. Actually, processes were performed over most of the group VI–VIII metals deposited on inorganic supports. These systems require in many if not *

Corresponding author. Tel.: +2-23-23-80-37; fax: +2-23-23-81-99. E-mail address: [email protected] (A. Roucoux).

0020-1693/$ - see front matter 2004 Elsevier B.V. All rights reserved. doi:10.1016/j.ica.2004.04.015

most cases drastic reaction conditions (300–450 C, 35– 170 bar) [17–23]. Several homogeneous rhodium or ruthenium catalysts coordinated by water-soluble phosphines such as TPPMS-Na or SULPHOS-Na have also been developed for selective hydrogenation of aromatic compounds containing sulfur and nitrogen in biphasic conditions [8,12,14,24,25]. Moreover, homogeneous systems have been found for the enantioselective hydrogenation of heteroaromatic compounds. Excellent results have been obtained in terms of enantiomeric excesses (up to 90%) [10–14] or diastereoselective excesses (up to 95%) when the heterocyclic was coupled to chiral auxiliaries [26–29]. In the past year, we have described the efficient activity for the catalytic hydrogenation of various aromatic derivatives by aqueous suspensions of metal nanoparticles in biphasic media and under mild conditions [30–32]. In this system, the catalytically active suspension is made of metal rhodium(0) particles prepared by reducing rhodium trichloride with sodium borohydride in dilute aqueous solutions of hydroxyethylammonium salts. Based on our benzene derivatives

3100

V. Mevellec, A. Roucoux / Inorganica Chimica Acta 357 (2004) 3099–3103

studies, here, we describe the original hydrogenation of various nitrogen-heterocyclic aromatic compounds such as pyridine derivatives, oxygen-heterocyclic aromatic compounds such as furan derivatives and sulfur-heterocyclic aromatic compounds such as thiophene in biphasic liquid–liquid systems by surfactant-protected rhodium(0) nanoparticles. We investigate the effect of counter-ion (chloride or bromide) of the ammonium salt on the nanoparticle size distributions and its influence on the catalytic hydrogenation. Finally, we demonstrate that the protected colloid suspensions can be re-used with a satisfactory recycling process.

2. Experimentals 2.1. Materials Rhodium chloride hydrate was obtained from Strem chemicals. Sodium borohydride, N,N-dimethylethanolamine, chloro or bromohexadecane and all heteroaromatic substrates were purchased from Aldrich or Acros and were used without further purification. Water was deionized before use by conventional method. The cationic surfactants (bromide and chloride salts) were prepared and fully characterized as previously reported [30,31]. 2.2. Analytical procedures The transmission electronic cryo-microscopic studies were conducted using a PHILIPS CM 12 transmission electron microscope at 100 keV. Samples were prepared by a dropwise addition of the stabilized colloid in water onto a copper sample mesh covered with carbon. The colloidal dispersion was removed after 1 min using cellulose, then the samples were quickly frozen in liquid ethane before transfer to the microscope. The size distribution of the particles was determinated by the measurement of about 300 particles. Gas chromatography was performed on a Carlo Erba GC 6000 with FID detector equipped with an Altech AT1 column (30 m long, 0.25 nm inner diameter). 2.3. Synthesis of the aqueous Rh0 colloidal suspensions The suspensions were prepared at 20 C in the air. Sodium borohydride (36 mg, 9.5 · 104 mol) was added to an aqueous solution of the appropriate surfactant (95 ml, 7.6 · 103 mol l1 ). This solution was quickly added under vigorous magnetic stirring to an aqueous solution (5 ml) of the precursor RhCl3 Æ 3H2 O (100 mg, 3.8 · 104 mol) providing an aqueous Rh0 colloidal suspension (100 ml). The initial red solution darkened immediately and after one night, the suspensions obtained remain stable under stirring for a long time.

2.4. General hydrogenation procedure All hydrogenations with Rh0 nanoparticles were carried out under standard conditions (20 C, 1 bar H2 ). 2.4.1. Hydrogenation with Rh0 nanoparticles A round-bottom flask (25 ml), charged with the chosen aqueous suspension of Rh0 (10 ml) and a magnetic stirrer, was connected to a gas burette (500 ml) with a flask to balance the pressure. The flask was closed with a septum and the system was filled with hydrogen. The appropriate heteroaromatic substrate (3.8 · 103 mole for a ratio ¼ 100) was injected through the septum and the mixture was stirred (1500 min1 ). The reaction was monitored by volume of gas consumed and by gas chromatography. At the end of the reaction, the two phases were separated by extraction and the aqueous phase was re-used in a second run. The turnover frequencies (TOFs) were determined for 100% conversion.

3. Results and discussion The aqueous suspensions are made of metallic rhodium(0) colloids prepared at room temperature and under vigorous stirring from rhodium trichloride chemical reduction with sodium borohydride in dilute aqueous solutions of N,N-dimethyl-N-cetyl-N-(2-HydroxyEthyl)Ammonium salts HEA16X (X ¼ Cl or Br). The synthesis of bromide and chloride surfactants has fully been characterized as previously reported [30,31]. The particle size of the Rh–HEA16X systems has been determined by transmission electron cryomicroscopic studies. The mean advantage of this observation technique was the absence of induced disturbances on the sample because it was quickly frozen in liquid ethane before transfer to the microscope, consequently, the distribution of the dispersed particles is not modified by specimen preparation. The histograms of the size distribution (Fig. 1) were estimated once the original negative had been digitalized and expanded for more accurate resolution. The average particles size was 2.1 and 2.4 nm for Rh(0)–HEA16Br and Rh(0)–HEA16Cl, respectively. The mean diameter of the Rh(0) nanoparticles was calculated from ensembles of about 300 particles found in an arbitrary chosen area of the enlarged micrographs using an automatic counting objects program based on shape recognition [33]. These comparative TEM studies show that the counter-ion of the surfactant does not seem to present a major influence on the average particle size of colloidal suspensions. The particles are homogeneous in size: 54% of the particles have a mean diameter between 1.4 and 2.8 nm. In preliminary studies, the aqueous suspensions Rh– HEA16X (X ¼ Br, Cl) have been optimized [30,31].


3101

Fig. 1. Cryo-TEM and size distribution histogram of rhodium(0) suspensions stabilized with (a) HEA16Br and (b) HEA16Cl.

Consequently, the aqueous suspensions have directly been used for reduction of heterocyclic compounds at atmospheric hydrogen pressure and 20 C. We have investigated the hydrogenation of various nitrogen-heterocyclic aromatic compounds such as pyridine derivatives, oxygen-heterocyclic aromatic compounds such as furan or benzofuran and sulfurheterocyclic aromatic compounds such as thiophene (Table 1). In all cases, the nitrogen- and oxygen-heterocyclic rings were reduced as usually described in the literature [12,15,18–25]. In the particular cases of bicyclic substrates such as benzofuran, quinoline and Nmethylindole, we have observed that the heterocyclic aromatic ring containing N- or O-atoms was exclusively reduced in heterocycle-hydrogenated products. The reaction was monitored by the volume of hydrogen consumed and the TOF defined as number of mole of consumed H2 per mole of rhodium per hour was determinated for the first run and successive runs after

separation of the aqueous phase. We have observed that aromatic substrates containing sulfur atom such as thiophene or benzothiophene were not reduced (Entries 16 and 17). The absence of conversion may be attributed to the affinity of sulfur atoms for its classical g1 -S adsorption onto the surface of transition metal(0) catalysts. Moreover, under the experimental conditions investigated in this work, benzothiophene is selectively adsorbed in the g1 -S mode but it is neither hydrogenated due to its unfavorable orientation, nor C–S inserted. The same result has been described by Bianchini et al. [8] with the Ru(0)/SiO2 heterogeneous catalyst. The catalytic suspension of rhodium(0) seems to be very active for hydrogenation of nitrogen substrates under mild reaction conditions. A total hydrogenation of the heterocyclic ring is observed for pyridine, picoline derivatives, quinoline and N-methylindole. Complete regioselectivity was obtained with bicyclic substrates containing heterocyclic and carbocyclic rings. These

3102


Table 1 Hydrogenation of heteroaromatic compoundsa Entry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18

Substrate

Pyridine Pyridine Pyridine 2-Picoline 2-Picoline 3-Picoline Quinoline Quinoline N-Methylindole Furan Furan Furane Furan Furan Benzofuranf Thiophene Benzothiophene 1,3,5-Triazine

Surfactant

HEA-16-Br HEA-16-Cl HEA-16-Cl HEA-16-Br HEA-16-Cl HEA-16-Cl HEA-16-Br HEA-16-Cl HEA-16-Cl HEA-16-Br HEA-16-Cl HEA-16-Cl HEA-16-Cl HEA-16-Cl HEA-16-Cl HEA-16-Cl HEA-16-Cl HEA-16-Cl

Substrate/Rh0

100 100 200 100 100 100 100 100 100 100 100 100 500 1000 100 100 100 100

Product (%)b

First run

Piperidine (100) Piperidine (100) Piperidine (100) 2-Methylpiperidine (100) 2-Methylpiperidine (100) 2-Methylpiperidine (100) 1,2,3,4-Tetrahydroquinoline (100) 1,2,3,4-Tetrahydroquinoline (100) 1,2-N-Methyldihydroindole (100) 1,2,3,4-Tetrahydrofuran (100) 1,2,3,4-Tetrahydrofuran (100) 1,2,3,4-Tetrahydrofuran (100) 1,2,3,4-Tetrahydrofuran (100) 1,2,3,4-Tetrahydrofuran (100) 1,2-Dihydrobenzofuran (100) 1,2,3,4-Tetrahydrothiophene 1,2-Dihydrobenzothiophene 1,3,5-Hexahydrotriazine (100)

Second run

t (h)

TOFc (h1 )

t (h)

TOFc (h1 )

4.3 3.9 7 9.8 9.1 30 10.4 9.1 5.9 2.4 2.1 1 10.3 22 2.5 – – 1.7

70 77 85 31 33 10 19 22 17 83 95 200 97 91 40 – – 176

5.3d 5.1d 8d 11d 10.3d 50d 11.2 10.0 6.7d 2.9d 2.4d 1.3 ndg 23.2 2.6 – – nrh

57 59 75 27 29 6 17 20 15 69 83 153 nd 86 38 – – nr

a

Conditions are catalyst: 3.8 · 105 mole; surfactant: 7.6 · 105 mole; water 10 ml; pressure of H2 : 1 bar; temperature: 20 C: stirred: 1500 min1 . Determined by GC analysis. c Turnover frequency defined as mole of H2 per mole of rhodium per hour. d After heptane extraction. e T ¼ 50 C. f Under 40 bar of H2 . g Not determined. h Not recycled. b

results can be explained by the favorable g1 -nitrogen adsorption mode onto the surface of nanoparticles [8,34]. Under our reaction conditions (1 bar H2 , 20 C), 5,6,7,8-tetrahydroquinoline produced via the g6 -C6 adsorption of quinoline onto the nanoparticles and decahydroquinoline were not obtained. The 1,2,3,4-tetrahydroquinoline may selectively be produced via direct 1,4-hydrogen transfer. The colloidal suspensions stabilized by bromide or chloride ammonium salts give analogous activities during the first run. The catalytic activities can be correlated with the particle sizes because the various histograms of the size distribution show a similar average. In fact, we have observed the similar TOF values such as 70 and 77 h1 for the pyridine hydrogenation (Entries 1 and 2) or 31 and 33 h1 for 2-picoline hydrogenation (Entries 4 and 5). The results show that the counter-ion (X stabilizer) poorly influences the catalytic activities. In most cases, the nitrogen reaction products such as piperidine or methylpiperidine were soluble in aqueous phase justifying the extraction and decantation with an organic solvent such as heptane before the recycling of the aqueous catalytic phase for a second run. Unfortunately, the slightly formation of aggregates and their gravity sedimentation in the flask were observed (Entries 1–6) and can be explained by the decrease of the surface charge due to the displacement of adsorbed chloride or bromide anions by

a more strongly binding neutral adsorbate such as pyridine and its derivatives. This phenomenon was previously demonstrated by the addition of pyridine to a colloidal solution of gold particles [35,36]. The interactions between water soluble product formed during the hydrogenation and the surfactant can also be responsible of the destabilization of the catalytic suspensions. Accordingly, the catalytic activities during the second run dramatically decrease and justify the recycling test stop. Finally, we have observed the rapid formation of 1,3,5-hexahydrotriazine by reduction of the 1,3,5-triazine (TOF ¼ 176 h1 ); here the catalyst cannot be recycled, since the colloidal suspension was insufficiently stable and generally aggregated (Entry 18) . The agglomeration was not observed with furan derivatives and the aqueous suspensions were still active after a second run after extraction of the reaction product. No problem of catalyst deactivation is detected as usually observed with palladium for example [37]. The Rh–HEA16Cl gave good results and the conversion was complete after 2 h for a ratio S/Rh ¼ 100 (TOF ¼ 95 h1 ; Entry11). The highest TOF value (97 h1 ) is observed for a ratio furan/Rh ¼ 500. The hydrogenation was performed at 50 C to observe the thermal stability of the catalytic system Rh–HEA16Cl (Entry 12). The 1,2,3,4-tetrahydrofuran was obtained after 1 h giving a TOF ¼ 200 h1 . The sedimentation of particles was not


visually detected, consequently, the catalytic suspension can be used for the hydrogenation of O-heterocyclic substrates up to 50 C.

4. Conclusion The results presented herein demonstrate that the Rh(0) nanoparticles protected by N,N-dimethyl-N-cetylN-(2-hydroxyethyl)ammonium salts can be used for the reduction of oxygen-heterocyclic aromatic rings. These experiments show that the stabilized nanoheterogeneous colloidal catalysts can be a real alternative to homogeneous or supported catalytic systems. The best activities are obtained with nanoparticles stabilized by the highly water soluble ammonium chloride salt. We have also observed good transformation in nitrogen heterocyclic compounds, but in few cases the aqueous suspension is insufficiently stable and aggregate. For these substrates, it will be needed to deposit nanoparticles on support. This research area is an alternative way to reduce this structural element which is present in some micropollutants such as atrazine. The denaturation of this molecule is still a challenge in the effluent water treatment. Finally, no catalytic activities were detected with sulfur compounds such as thiophene.

Acknowledgements We thank the Region Bretagne for financial support. We also thank Dr. Rolland (University of Rennes 1) for TEM experiments. References [1] [2] [3] [4] [5]

L.N. Lewis, Chem. Rev. 93 (1993) 2693. A. Roucoux, J. Schulz, H. Patin, Chem. Rev. 102 (2002) 3757. J.A. Widegren, R.G. Finke, J. Mol. Catal. A 191 (2003) 187. J.A. Widegren, R.G. Finke, Inorg. Chem. 41 (2002) 1558. J.-L. Pellegatta, C. Blandy, V. Colliere, R. Choukroun, B. Chaudret, P. Cheng, K. Philippot, J. Mol. Catal. A 178 (2002) 55. [6] K. Pelzer, O. Vidoni, K. Philippot, B. Chaudret, V. Colliere, Adv. Funct. Mater. 213 (2003) 118. [7] D. Tabuani, O. Monticelli, A. Chincarini, C. Bianchini, F. Vizza, S. Moneti, S. Russo, Macromolecules 36 (2003) 4294. [8] C. Bianchini, V. Dal Santo, A. Meli, S. Moneti, M. Moreno, W. Oberhauser, R. Psaro, L. Sordelli, F. Vizza, J. Catal. 213 (2003) 47.

3103

[9] A.F. Borowski, S. Sabo-Etienne, B. Donnadieu, B. Chaudret, Organometallics 22 (2003) 1630. [10] C. Bianchini, P. Barbaro, G. Scapacci, E. Farnetti, M. Graziani, Organometallics 17 (1998) 3308. [11] S.A. Rynor, J.M. Thomas, R. Raja, B.F.G. Johnson, R.G. Bell, M.D. Mantle, Chem. Commun. (2000) 1925. [12] I. Rojas, F.L. Linares, N. Valencia, C. Bianchini, J. Mol. Catal. A 144 (1999) 1. [13] R. Kuwano, K. Sato, Y. Ito, Chem. Lett. (2000) 428. [14] M. Studer, C. Wedemeyer-Exl, F. Sindler, H.U. Blaser, Monatsh. Chem. 131 (2000) 1335. [15] C. Bianchini, P. Frediani, V. Herrera, M.V. Jimenez, A. Meli, L. Rincon, R. Sanchez-Delgado, F. Vizza, J. Am. Chem. Soc. 117 (1995) 4333. [16] J. Cinibulk, Z. Vit, Appl. Catal. A 180 (1999) 15. [17] K. Miga, K. Stanczyk, C. Sayag, D. Brodzki, G. Djega-Mariadassou, J. Catal. 183 (1999) 63. [18] H. Topsoe, B.S. Clausen, F.E. Massoth, Hydrotreating Catalysis, Springler, Berlin-Heidelberg, 1996. [19] J. Scherzer, A.J. Gruia, Hydrocracking Science and Technology, Marcel Dekker, New York, 1996. [20] M.L. Occelli, R. Chianelli, Hydrotreating Technology for Pollution Control, Dekker, New York, 1996. [21] M. Egorova, R. Prins, J. Catal. 221 (2004) 11, and references therein. [22] J.E. Shaw, Phillips Petroleum Company, US Patent No. 4739063, 1988. [23] Y. Yoshimura, H. Yasuda, T. Sato, N. Kijima, US Patent No. 6498279B1, 2002. [24] J.E. Shaw, Phillips Petroleum Company, US Patent No. 4739064, 1988. [25] C. Bianchini, A. Meli, Aqueous-Phase Organomet. Cat. Ed. B (1998) 477. [26] D.E. Paez, A. Andriollo, R.A. Sanchez.Delgado, N.D.V. Valencia, R.E. Galiasso, F.A. Lopez, INTEVEP S.A., US Patent No. 5753584, 1998. [27] V. Hada, A. Tungler, L. Szepesy, Appl. Catal. A 210 (2001) 165. [28] L. Hegedus, V. Hada, A. Tungler, T. Mathe, L. Szepesy, Appl. Catal. A 201 (2000) 107. [29] M. Besson, F. Delbecq, P. Gallezot, S. Neto, C. Pinel, Chem. Eur. J. 6 (2000) 949. [30] A. Roucoux, J. Schulz, H. Patin, Adv. Synth. Catal. 345 (2002) 222. [31] J. Schulz, A. Roucoux, H. Patin, Chem. Eur. J. 6 (2000) 618. [32] V. Mevellec, A. Roucoux, E. Ramirez, K. Philippot, B. Chaudret, Adv. Synth. Catal. 346 (2004) 72. [33] J.-P. Rolland, P. Bon, D. Thomas, Comput. Appl. Biosci. 13 (1998) 563. [34] G. Cerichelli, L. Luchetti, G. Mancini, G. Savelli, Tetrahedron 51 (1995) 10281. [35] G. Schmid, Cluster and Colloids: From Therory to Applications, VCH, New York, 1994, p. 465. [36] C.G. Blatchford, J.R. Campbell, J.A. Creighton, Surf. Sci. 120 (1982) 435. [37] S.D. Jackson, A.S. Canning, E.M. Vass, S.R. Watson, Ind. Eng. Chem. Res. 42 (2003) 5489.