The aldol condensation is one of the most widely used reactions employed to form a new carbon-carbon bond. The development of new asymmetric methodologies for this reaction, has contributed to increasing scope and applications even further [1, 2]. In general, these methodologies belong to one of the following categories: (a) use of a chiral inductor (chiral auxiliary) in stoichiometric quantities or in catalytic amounts (as is the case for organocatalysis); (b) use of chiral Lewis acid- or base-catalysts; (c) use of a chiral heterobimetallic catalyst; and (d) use of enzymes of the aldolase type as catalysts [2–4].
Organocatalysis has become a field of central importance for organic synthesis, by employing small chiral organic compounds free of metals as catalysts  for the asymmetric synthesis of desired molecules [6, 7]. New methods and the use of alternative techniques, such as microwave irradiation, sonication and non-conventional reaction media (water, ionic liquids and supercritical carbon dioxide), are increasing the scope and possibilities of this kind of asymmetric catalysis . Chiral amino acids, especially l-proline and its derivatives, have been successfully employed as organocatalysts for the aldol reaction, leading to high yields with high stereoselectivity [3, 4, 8, 9].
There are, however, certain disadvantages in the aldol reaction: low selectivity and the use of stoichiometric amounts of aqueous base that needs to be neutralized and separated from the product after the completion of the reaction, leading to a high E factor (a high waste/product ratio) [10, 11]. Evaluation of new reaction media, which could lead to an increase in selectivity and reduce the waste generated in the reaction, is therefore essential .
Supercritical carbon dioxide (sc-CO2) is an environmentally benign solvent with an accessible critical point (Tc=304.2 K, Pc=7.38 MPa) and unique properties, that make it an interesting reaction media [12–15]. Its association with ionic liquids, forming a polar gas expanded phase, increases even further its possible applications . There are, however, only a few reports in the literature regarding the use of these reaction media for the aldol reaction [1, 11, 17].
In this work, the use of l-proline derivatives as organocatalysts for the aldol condensation reaction between acetone and 4-nitrobenzaldehyde (Scheme 1), in sc-CO2 and in biphasic systems containing sc-CO2 and ionic liquids derived from imidazole (1-allyl-3-alkyl-imidazolium chlorides) was investigated. l-Proline was used as a control, in order to compare the results of this work with those reported previously in the literature .
The influence of pressure, temperature, presence of ionic liquid, and type of catalyst on the yield and enantiomeric composition of the addition product (a β-hydroxy ketone) and on the yield of the elimination product (a α,β-unsaturated ketone) was evaluated.
Two ionic liquids were evaluated for the biphasic systems; 1-allyl-3-methyl-imidazolium chloride [IL1, compound (5)] and 1-allyl-3-hexyl-imidazolium chloride [IL2, compound (6)] (Scheme 2).
2 Experimental section
All chemicals were obtained from commercial sources (Sigma-Aldrich, St. Louis, MO, USA and Merck KGaA, Darmstadt, Germany), except the two ionic liquids which were a gift from Professor Omar A. El Seoud (IQ-USP). Compound purification was carried out by column chromatography over silica gel (230–400 mesh).
Enantiomeric excesses (e.e.) were determined by HPLC on a CBM-10A (Shimadzu Corporation Kyoto, Japan) chromatograph, equipped with a AS-H Chiralpak chiral column. The mobile phase was hexane/isopropanol (9:1), with a 1 h run time at 1 ml/min. Detection was performed by UV-Vis at a wavelength of 254 nm. The peak positions of the R and S enantiomers were determined using an authentic sample of the R-enantiomer. The e.e. refers to the excess of this enantiomer (R).
1H and 13C NMR spectra were measured on a Bruker Avance 3 spectrometer (Bruker BioSpin, Billerica, MA, USA) (200 MHz for proton) with samples dissolved in CDCl3. Chemical shifts are presented in ppm using tetramethylsilane (TMS) as reference.
All of the reactions in sc-CO2 were performed in a 25 ml stainless steel micro reaction apparatus (Parr Instrument Company model 4591, Moline, IL, USA), equipped with a heating jacket, a magnetic drive stirrer and a PID control unit (Parr Instrument Company model 4848). This unit has a type J thermocouple, a tachometer and a pressure transducer, to measure/control temperature (within ±0.3°C) and rotation speed (within ±5 rpm) and to measure pressure (within ±5 psi).
Supercritical CO2 was delivered to the reactor at the desired pressure using a syringe pump (Teledyne ISCO model 260D, Lincoln, NE, USA) and the appropriate valves and connections (HIP).
The IUPAC names were obtained using the ChemBio-Draw Ultra 12.0.1 software package (PerkinElmer, Cambridge, MA, USA).
2.1 General procedure for the aldol condensation reaction in organic solvents
4-Nitrobenzaldehyde (0.075 g, 0.50 mmol) and catalyst (20–30 mol%) were added to a solution of acetone (1.0 ml) in DMSO (4.0 ml) and the resulting mixture was stirred at room temperature for 4–72 h. The reaction mixture was treated with a saturated ammonium chloride aqueous solution and extracted several times with ethyl acetate. The organic extracts were combined, dried with anhydrous MgSO4 and the solvent was removed on a rotary evaporator. The pure aldol products were obtained by flash column chromatography on silica gel, eluted with a mixture of hexane/ethyl acetate 2:1 .
Addition product: 4-hydroxy-4-(4-nitrophenyl)butan-2-one (Scheme 3).
1H NMR: (200 MHz, CDCl3, ppm), δ: 2.23 (s, 3H); 2.86 (d, J=6 Hz, 2H); 3.72 (s, 1H); 5.27 (t, J=6 Hz, 1H); 7.54 (m, 2H); 8.2 (m, 2H).
Elimination product: (E)-4-(4-nitrophenyl)but-3-en-2-one (no Z-isomer was observed) (Scheme 4).
1H RMN: (200 MHz, CDCl3, ppm), δ: 2.43 (s, 3H); 6.82 (d, J=16 Hz, 1H); 7.54 (d, J=16 Hz, 1H); 7.7 (m, 2H); 8.27 (m, 2H).
2.2 General procedure for the aldol condensation reaction in sc-CO2 or in biphasic systems containing ionic liquids
The 25 ml stainless steel Parr reaction apparatus was charged with acetone (1.0 ml), 4-nitrobenzaldehyde (0.075 g, 0.50 mmol) and the appropriate amount of one of the catalysts evaluated (20–30 mol%): l-proline, 0.034 g; l-thioproline, 0.026 g; dimethylphenylsilyloxy-l-proline, 0.053 g; butyldimethylsilyloxy-l-proline, 0.049 g; and peracetylated glucosamine-l-proline, 0.048 g. In some experiments, an ionic liquid (1.0 ml) or a base (TEA, 0.07 ml) were also added. The reactor temperature was set to the desired value at the control unit, while the reactor was gently purged with a very small flow of CO2 for 1 min. It was then charged with CO2 at the desired pressure using the Teledyne syringe pump. The stirrer speed was set to 500 rpm (verified with the tachometer) and the reactor was kept at the set conditions for the desired time. The reactor was then slowly depressurized. All the CO2 exiting the reactor was passed through a glass trap containing 20 ml of ethyl acetate, in order to collect any material that might have been carried out by the gas. The reactor was then thoroughly washed with ethyl acetate twice (each time with approximately 20 ml of solvent), to ensure that no material remained attached to the walls, or other parts of the vessel. All of the ethyl acetate fractions were combined and the solvent was removed under reduced pressure in a rotary evaporator. In the experiments where an ionic liquid was used, the washing of the reactor was performed with acetone in the same way as described before. The acetone was removed under reduced pressure and the resulting ionic liquid was extracted three times with ethyl acetated (around 10 ml each time), the fractions were combined and the solvent removed under reduced pressure in a rotary evaporator.
The organic fraction obtained after the solvent removal, was submitted to flash column chromatography in silica gel, eluted with hexane:ethyl acetate (2:1), in order to separate products from unreacted reagents. The recovered 4-nitrobenzaldehyde was weighed and used to determine the conversion. The isolated product obtained was weighed and used to determine the yield.
2.3 Catalyst preparation
2.3.1 (R)-thiazolidine-4-carboxylic acid (thio-l-proline) (Scheme 5)
The procedure was based on that described by Kaupp et al. . Cysteine (0.48 g, 4 mmol) and p-formaldehyde (0.12 g, 4 mmol) were mixed in a scintillation vial and the mixture was stirred for 4 h at room temperature. The system was then kept under vacuum at 80°C for another 30 min. The product was purified by column chromatography. 1H NMR (200 MHz, CDCl3, ppm), δ: 2.31 (dd, J=6 Hz, 2H); 4.29 (q, J=10 Hz, 2H); 4.61 (dd, J=6 Hz, 1H). 13C NMR (50 MHz, D2O, ppm), δ: 32.0, 48.4, 62.2, 169.7.
2.3.2 (2S,4R)-4-(dimethyl(phenyl)silyl)oxy-pyrrolidine-2-carboxylic acid (dimethylphenylsilyloxy-l-proline) (Scheme 6)
The procedure was based on that described by Opalka et al. . Dimethylphenyl silane (1.86 g) was added to a solution of (E)-4-hydroxy-l-proline (1.39 g) in acetonitrile (6 ml), in a round bottom flask. This mixture was cooled to 0°C, and DBU (1.95 ml) was added. The reaction was allowed to warm to room temperature and stirred for 24 h and then extracted three times with pentane. The pentane extracts were combined and concentrated. Methanol (32 ml), tetrahydrofuran (THF) (16 ml), water (16 ml), and a 2m NaOH aqueous solution (24 ml) were added to the resulting oil. The mixture was stirred at room temperature for 90 min and then the pH was adjusted to 6, using an aqueous 1m HCl solution. The solvents were then removed under reduced pressure, leaving a yellow clear liquid. 1H NMR (200 MHz, DMSO-d6, ppm), δ: 2.37 (s, 3H); 0.29 (s, 6H); 1.15 (t, J=14 Hz, 1H); 1.96 (s, 1H); 4.01 (q, J=22 Hz, 1H); 7.42 (m, 7H). 13C NMR (50 MHz, DMSO-d6, ppm), δ: 0.53, 0.73, 14.03, 20.68, 59.71, 127.73, 128.96, 129.37, 132.69, 132.90, 139.05, 170.24.
2.3.3 (2S,4R)-4-(tert-butyldimethylsilyloxy)pyrrolidine-2-carboxylic acid (t-butyl-dimethylsilyloxy-l-proline) (Scheme 7)
The procedure was based on that described by Opalka et al. . Tert-butyl dimethyl silane (1.86 g) was added to a solution of (E)-4-hydroxy-l-proline (1.39 g) in acetonitrile (6 ml), in a round bottom flask. This mixture was cooled to 0°C, and DBU (1.95 ml) was added. The reaction was allowed to warm to room temperature and stirred for 24 h and then extracted three times with pentane. The pentane extracts were combined and concentrated. Methanol (32 ml), THF (16 ml), water (16 ml), and a 2m NaOH aqueous solution (24 ml) were added to the resulting oil. The mixture was stirred at room temperature for 90 min and then the pH was adjusted to 6 using an aqueous 1m HCl solution. The solvents were then removed under reduced pressure until a white precipitate just began to form. At this point, water was added until all the precipitate went into solution. The solution was then allowed to sit until crystals formed. The crystals were filtered and washed with diethyl ether to afford white crystals . 1H NMR (200 MHz, MeOD, ppm), δ: 0.14, 0.15 (s, 3H); 0.93 (s, 8H); 2.07 (m, 1H); 2.35 (m, 1H); 3.16 (m, 1H); 3.42 (m, 1H); 4.17 (q, J=18 Hz); 4.66 (m, 1H).
2.3.4 (2R,3R,4R,5R,6S)-2-(acetoxymethyl)-6-(pyrrolidine-2-carboxamido)tetrahydro-2H-pyran-3,4,5-triyl triacetate (peracetyl-glucosamine-l-proline) (Scheme 8)
The procedure was based on the method described by Chandrasekhar et al. . N-benzyloxycarbonyl-protected proline (2 g, 7.9 mmol) was dissolved in dry dichloromethane (10 ml). HOBt (1.28 g, 9.48 mmol) was added and the reaction mixture was stirred for 15 min. The solution was cooled to 0°C and 1-ethyl-3-(3-dimethylaminopropyl) carbodiimide (EDCI) (3.02 g, 15.8 mmol) was added, followed by a solution of peracetylated glucosamine (2.55 g, 7.9 mmol) and diisopropylethylamine (DIEPA) (7.9 mmol, 1.10 ml) in dry dichloromethane (10 ml). The mixture was stirred overnight at room temperature. After completion of the reaction (monitored by TLC), water (20 ml) was added, and the two layers were separated. The organic layer was washed with aqueous ammonium chloride (10 ml) and then aqueous sodium bicarbonate (15 ml), dried over sodium sulfate, and the solvent was removed under reduced pressure. The crude product was purified by column chromatography. The deprotection of the compound was carried out by hydrogenation with 20% Pd(OH)2/C (50 mg) . 1H NMR (200 MHz, CDCl3, ppm), δ: 2.03 (m, 17H); 3.53 (m, 1H); 3.84 (m, 1H); 4.18 (m, 1H); 5.20 (m, 2H); 5.95 (d, J=10 Hz, 1H); 7.50 (m, 5H).
3 Results and discussion
The behavior of the l-proline derivatives as organocatalysts for the aldol condensation of acetone with 4-nitrobenzaldhyde was evaluated using several reaction conditions in different reaction media. The results were reported as conversions (percentage of reagent converted to products), yields of the addition product (percentage of the addition product formed) and enantiomeric excesses of the addition product (excess of the R-enantiomer). The most promising catalysts were explored more thoroughly.
3.1 (S)-pyrrolidine-2-carboxylic acid (l-proline)
Initial tests with l-proline were performed to compare the data of this work with existing literature results . The temperature and pressure were chosen based on the literature, however, the pressure had to be set at a lower value (150 bar), due to equipment limitation.
The results obtained agree with those previously reported on the literature (Entry 1 – Table 1), but with a lower enantiomeric excess. The use of the ionic liquid (IL1) in combination with sc-CO2 (Entry 2 – Table 1) resulted in slightly higher enantiomeric excess, but also increased the amount of the elimination product (higher conversion for the same yield of the addition product).
3.2 (2S,4R)-4-(dimethyl(phenyl)silyl)-oxy-pyrrolidine-2-carboxylic acid (dimethylphenylsilyloxy-l-proline)
This catalyst was not as effective as l-proline, as can be seen from the results in Table 2. Using exactly the same reaction conditions resulted in slightly lower yields and enantiomeric excesses with the silyloxy derivative. Reducing the reaction time resulted in lower yields and e.e. (Entry 2 – Table 2). An increase in the reaction time resulted in much lower yields and e.e.s (Entry 3 – Table 2), probably as a result of racemization by a retro-aldol reaction (as evidenced by the reduced e.e.) and catalyst degradation observed for longer reaction times. The use of an ionic liquid (IL1) actually prevented any reaction from proceeding (Entry 4 – Table 2) and thus no further tests with the other ionic (IL2) were performed.
3.3 (R)-thiazolidine-4-carboxylic acid (thio-l-proline)
This catalyst was also not as effective as l-proline, as can be seen by the results in Table 3. Using DMSO as a solvent, the yields were low, even though the e.e.s were higher than the ones obtained with l-proline. In sc-CO2, the results were worse, with very low yields, probably because of the very low solubility of the catalyst in sc-CO2. The presence of a base (0.07 ml of TEA) made the results even worse.
3.4 (2R,3R,4R,5R,6S)-2-(acetoxymethyl)-6-(pyrrolidine-2-carboxamido)tetrahydro-2H-pyran-3,4,5-triyl triacetate (peracetyl-glucosamine-l-proline)
The derivatization of l-proline with a peracetylated sugar (a CO2-philic group) increased the solubility of the proline derivative in sc-CO2. However, this had almost no effect on the catalytic properties of the compound, as can be seenfrom the results in Table 4. The yields and e.e.s were low and the catalyst was not as effective as l-proline.
3.5 (2S,4R)-4-(tert-butyldimethylsilyloxy)-pyrrolidine-2-carboxylic acid (t-butyl-dimethylsilyloxy-l-proline)
This catalyst provided the best results. Using DMSO as a solvent resulted in a very good yield, 88.5%, and a good e.e., 77.9% (Entry 1 – Table 5). The presence of a base (TEA) decreased both the yield and the e.e. (Entry 2 – Table 5). The reaction time can be reduced to 2 h, with a low decrease in yield and no effect in the e.e. (Entry 4 – Table 5). Dilution of the reagents, using the same total volume of the high pressure reactor (25 ml), in order to compare the results with the ones in sc-CO2, resulted in a significant decrease in yield (49.0%), but no effect on the e.e. (Entry 3 – Table 5).
Using sc-CO2 as a solvent resulted in lower yields and a very low e.e. for longer reaction times (Entry 6 – Table 5). Decreasing the reaction time, however, resulted in good yields, around 46.0%, comparable to those observed for the reaction in diluted conditions in DMSO (same volume as the sc-CO2 phase, 25 ml) (Entry 12 – Table 5). The e.e., however, was very low.
The best results for sc-CO2 were obtained in the presence of the ionic liquid (IL1), with a yield of 54.0% and a very good e.e. of 79.0% (Entry 13 – Table 5). This is probably the result of the higher polarity of the ionic liquid phase. Our group is currently investigating this issue in detail in order to understand the effect created by the presence of the ionic liquid.
One of the catalysts tested, the tert-butyldimethylsilyloxy-l-proline, was very effective, allowing lower reaction times, resulting in better yields and e.e. than l-proline itself, both in organic solvent (DMSO) and in sc-CO2 or a mixture of sc-CO2 and ionic liquids. The use of the ionic liquid resulted in better yields and e.e. in sc-CO2.
The other catalysts tested were not as effective, resulting in yields and e.e.s similar or inferior to the ones obtained with l-proline.
FAPESP, CNPq, CAPES and INCT de Estudos do Meio Ambiente for financial support and fellowships. We thank Professor Omar A. El Seoud for the gift of the ILs.
Matsui K, Kawanami H, Ikushima Y, Hayashi H. Chem. Commun. 2003, 2502–2503.Google Scholar
Trost BM, Brindle CS. Chem. Soc. Rev. 2010, 39, 1600–1632.Google Scholar
Sakthivel K, Notz W, Bui T, Barbas CF. J. Am. Chem. Soc. 2001, 123, 5260–5267.Google Scholar
Bisai V, Bisai A, Singh VK. Tetrahedron 2012, 68, 4541–4580.Google Scholar
Wade EO, Walsh KE. J. Chem. Educ. 2011, 88, 1152–1154.Google Scholar
Chandrasekhar S, Johny K, Reddy CR. Tetrahedron: Asymmetry 2009, 20, 1742–1745.Google Scholar
List B, Lerner RA, Barbas CF. J. Am. Chem. Soc. 2000, 122, 2395–2396.Google Scholar
Mestres R. Green Chem. 2004, 6, 583–603.Google Scholar
Stevens JG, Bourne RA, Poliakoff M. Green Chem. 2009, 11, 409–416.Google Scholar
Oakes RS, Clifford AA, Rayner CM. J. Chem. Soc., Perkin Trans. 1 2001, 917–941.Google Scholar
Licence P, Ke J, Sokolova M, Ross SK, Poliakoff M. Green Chem. 2003, 5, 99–104.Google Scholar
Skouta R. Green Chem. Lett. Rev. 2009, 2, 121–156.Google Scholar
Jessop PG, Subramaniam B. Chem. Rev. 2007, 107, 2666–2694.Google Scholar
Liu L, Liu Z-T, Liu Z-W, Xue D. Sci. China Chem. 2010, 53, 1586–1591.Google Scholar
Kaupp G, Schmeyers J, Boy J. Tetrahedron 2000, 56, 6899–6911.Google Scholar
Opalka SM, Steinbacher JL, Lambiris BA, McQuade DT. J. Org. Chem. 2011, 76, 6503–6517.Google Scholar
About the article
Rafael F. Cassaro
Rafael Frascino Cassaro studied chemistry at IBILCE/UNESP in São José do Rio Preto, SP, Brazil, where he obtained a BS in Environmental Chemistry in 2009. In 2010, he joined the Instituto de Química, USP, where he is currently pursuing a PhD in Chemistry, under the supervision of Dr. Reinaldo C. Bazito, working with organocatalysis in supercritical carbon dioxide.
Lelaine C. de Oliveira
Lelaine Conceição de Oliveira studied chemistry at Instituto de Química, USP, in São Paulo, SP, Brazil, where she obtained a BS in Environmental Chemistry, in 2012. In that year, she spent 6months as a “scientific initiation student” under the supervision of Dr. Reinaldo C. Bazito, working with organocatalysis in supercritical carbon dioxide.
Rogerio A. Gariani
Rogério Aparecido Gariani studied chemistry at the Universidade Federal do Paraná (Paraná, Brazil), where he obtained his BSc degree in 2004 and his MSc degree in 2006. In 2006, he joined the Instituto de Química, USP (São Paulo, Brazil) where he received his PhD degree, in 2010. He has been an assistant professor at the Chemistry Department of the Universidade do Estado de Santa Catarina, UDESC (Joinville, Brazil), since 2012. His current research interests comprise chemical synthesis, C-C coupling reactions with the use of nanoparticles as catalysts and supercritical dioxide carbon as an alternative solvent.
Claudio A.O. do Nascimento
Claudio Augusto Oller do Nascimento studied chemical engineering at the Escola Politécnica da USP (São Paulo, SP), where he obtained his BSc, in 1975 and his MSc in Chemical Engineering, in 1979. In 1982, he obtained his PhD in Chemical Engineering at University of Salford (Salford, UK). He is currently a full professor at the Escola Politécnica da USP (São Paulo, SP), where he develops research on chemical process control, modeling and optimization.
Reinaldo C. Bazito
Reinaldo Camino Bazito studied chemistry at the Instituto de Química, USP (São Paulo, SP), where he obtained his BSc in chemistry, in 1992, his MSc in 1997 and his PhD in 2001, both in Organic Chemistry. He is an assistant professor at the Instituto de Química da USP, where he develops research on green chemistry, especially in the applications of supercritical carbon dioxide as an alternative solvent. His current research interests include the study of chemical reactions in supercritical carbon dioxide (aldol reaction, Morita-Baylis-Hillman, dehydrohalogenation, among others) and the development of surfactants and amphiphilic polymers for drug encapsulation in this alternative solvent.
Published Online: 2013-02-02
Published in Print: 2013-02-01