Indoles, oxindoles, isatins, and their derivatives are relevant nitrogen-containing heterocycles widely distributed in natural products and pharmaceutical compounds [1]. Oxindole is used as a neurodepressant tryptophan metabolite and physiologically present in mammalian brain/blood, and can influence brain functions [2]. 3-Alkyledene-oxindole derivatives are medicinally important and exist in anticancer kinase inhibitors and anti-inflammatory drugs including Sunitinid and Tedinap, respectively [3-5]. They are also well-known synthetic intermediates of natural products named TMC-95, Gelsemine and biologically appealing spirooxindoles [6-13]. Conventionally, secondary amine (piperidine/pyrrolidine)-assisted Knoevenagel condensation reaction is applied to synthesize 3-alkyl/arylidene-oxindoles from oxindole and carbonyl compounds [14-17]. Lee et al. [18] reported (Z)-selective synthesis of 3-arylidene-oxindoles employing stoichiometric amounts of Ti(OⁱPr)₄/pyridine with unsymmetrical ketones. Villemin et al. [19] reported 3-alkenylation of oxindole with aromatic aldehydes promoted by an excess amount (10 equiv.) of solid base (KF/Al₂O₃) under microwave radiation. Br nsted acidic ionic liquids [20] were reported to promote 3-alkenylation reactions of oxindole with aromatic aldehydes, but this method used an excess amount (> 6.6 equiv.) of the Br nsted acids with respect to the substrates. Apart from these non-catalytic methods, Vankelecom and co-workers [21] showed the catalytic synthesis of 3-benzylidene-oxindole by the reaction of oxindole with benzaldehyde using a Zr-incorporated UiO-66 MOF catalyst. Recently, Zhu and co-workers [22] reported that a homogeneous Re-complex catalyzed dehydrogenative 3-alkenylation of N-substituted oxindoles with benzylic alcohols. However, these two catalytic methods are applicable to activated (aromatic and allylic) aldehydes and suffer from the difficulties in catalyst reuse. Gholamzadeha et al. [23] demonstrated a reusable catalytic method using SO₃H-loaded silica catalyst, but the method is applicable only for activated aromatic aldehydes. Hence, it is desirable to develop a reusable general heterogeneous catalyst for the 3-alkenylation of oxindoles with aldehydes including aliphatic (non-aromatic) aldehydes.

Organic synthesis employing heterogeneous CeO₂ catalysis fascinates much attention in recent years [24-26]. Our group has studied a series of catalytic transformation of carboxylic acid derivatives using heterogeneous CeO₂ catalysts [27-36]. This access has advantages combined with the facility of catalyst preparation and exploitation of a recyclable heterogeneous catalyst. It is anticipated that this catalytic system would find applications to other challenging reactions that produce value-added chemicals in one-pot manner. Herein, we report a simple heterogenous catalytic method using CeO₂ for the C3-alkenylation of oxindole with aldehydes including unactivated aliphatic aldehydes into the corresponding 3-alkenyl-oxindoles.

Commercial compounds (Tokyo Chemical Industry, Sigma-Aldrich or Wako Pure Chemical Industries) were used without further purification. Substrates and products were analyzed by GC (Shimadzu GC-2014) and GCMS (Shimadzu GCMS-QP2010) with an Ultra ALLOY⁺-1 capillary column (Frontier Laboratories Ltd.) using N₂ and He as the carrier. Column chromatography was performed with silica gel 60 (spherical, 40-100 μm, Kanto Chemical Co. Ltd.). ¹H and ¹³C NMR spectra were recorded at ambient temperature on a JEOL-ECX 600 (¹H: 600.17 MHz; ¹³C: 150.92 MHz) spectrometer with tetramethylsilane as an internal standard.

CeO₂ (JRC-CEO-1, 185.3 m² g^–1, supplied by Santoku Co.), MgO (JRC-MGO-3, 19 m² g^–1), and TiO₂ (JRC-TIO-4, 47 m² g^–1) were obtained from Catalysis Society of Japan. SiO₂ (Q-10, 300 m² g^–1) was supplied by Fuji Silysia Chemical Ltd. Different CeO₂ catalysts were prepared with CeO₂ (JRC-CEO-1) by calcining at four different temperatures (500, 600, 800, and 1000 ℃) for 3 h. γ-Al₂O₃ (124 m² g^–1) was prepared by calcination of γ-AlOOH (Catapal B Alumina purchased from Sasol) at 900 ℃ for 3 h. Niobic acid was supplied by CBMM, and Nb₂O₅ (54 m² g^–1) was prepared by calcination of the niobic acid at 500 ℃ for 3 h. ZrO₂ (73 m² g^–1) and SnO₂ (25 m² g^–1) were prepared by calcination (500 ℃, 3 h) of ZrO₂∙nH₂O and H₂SnO₃ (Kojundo Chemical Laboratory Co., Ltd.). Hydroxides of Ca, Zn, and Mg were commercially available. Hydroxides of La were prepared by hydrolysis of La(NO₃)₃·6H₂O with aqueous NH₄OH solution, followed by filtration, washing with distilled water, drying at 100 ℃ for 12 h, and calcination at 500 ℃ for 3 h. Oxides of Ca (22 m² g^–1), Zn (12 m² g^–1) and La were prepared by calcination of these hydroxides at 500 ℃ for 3 h. Hβ-20 (HSZ-940HOA, 530 m² g^–1) was purchased from Tosoh Co. A sulfonic resin (Amberlyst-15, 45 m² g^–1) and montmorillonite K10 clay (mont. K10, 220 m² g^–1) were purchased from Sigma-Aldrich. Scandium(III) trifluoromethanesulfonate Sc(OTf)₃, Ce(NO₃)₄, and Ce₃(PO₄)₄ were purchased from Tokyo Chemical Industry.

The solid catalysts, stored under ambient conditions, were used without any pretreatment. In a typical procedure, oxindole (1 mmol) and each of the aldehydes (1.25 mmol) were subjected to condensation reaction in the presence of 20 mg of catalysts. The reaction mixture was added to a reaction tube (Pyrex pressure tube, 13 mL), with a magnetic stirrer bar and placed in a heated reactor at 100 ℃ under N₂ with stirring at 400 rpm. After completion of the reaction, the reaction mixture was diluted with 2-propanol (3 mL), and n-dodecane (0.2 mmol) was added as an internal standard and the mixture was analyzed by GC-FID to determine the conversion and yield of the products. The GC-FID sensitivity of the product was determined using the isolated 3-alkeny-oxindole. After the reaction, the catalyst was separated through filtration, and the solvent was evaporated from the reaction mixture. The E/Z ratio of the isomers and total yield were determined by ¹H NMR analyses of crude reaction mixture using mesitylene as an internal. The E and Z isomers were assigned and confirmed in comparison with their reported ¹H NMR data [18]. The major product was isolated by column chromatography using silica gel 60 (spherical, 40-100 μm, Kanto Chemical Co. Ltd.) with hexane/ethyl acetate (95/5 to 80/20) as the eluting solvent. The isolated product was analyzed by ¹H NMR, ¹³C NMR, and GC-MS equipped with the same column as GC-FID.

In situ IR spectra were recorded using a JASCO FT/IR-4200 spectrometer equipped with a mercury cadmium telluride detector. For the benzaldehyde adsorption IR study, a closed IR cell surrounded by a Dewar vessel was connected to an evacuation system. During the IR measurement, the IR cell was cooled with a freezing mixture of dry-ice/acetone in the Dewar vessel, and the thermocouple near the sample showed T = (–50 ± 5) ℃. The sample was pressed into a self-supporting wafer (40 mg) and mounted into the IR cell with CaF₂ windows. Spectra were measured by accumulating 15 scans at a resolution of 4 cm^–1. After in situ pre-evacuation of the sample at 500 ℃ for 0.5 h, a reference spectrum of the sample disc was measured at T = (–50 ± 5) ℃. The sample was then exposed to benzaldehyde (1 μL) under 650 Pa at T = –50 ℃ for 600 s, followed by evacuation for 600 s. A differential IR spectrum, with respect to the reference spectrum, was then recorded at T = (–50 ± 5) ℃. XRD measurements were conducted using a Rigaku Miniflex with a Cu-K_α radiation source. N₂ adsorption measurements were carried out by using AUTOSORB 6AG (Yuasa Ionics Co.). X-ray photoelectron spectroscopy (XPS) measurements were carried out using a JEOL JPS-9010MC with an Mg K_α anode operated at 10 mA and 10 kV. Binding energies were calibrated with respect to C 1s at 284.2 eV.

To find the optimal catalyst, we employed a series of potential catalysts for the 3-alkenylation of oxindole (1a) with benzaldehyde (2a) into the corresponding 3-benzylidene-oxindole product (3a). A series of acidic or basic heterogeneous and homogeneous catalysts (20 mg) were tested for the 3-alkenylation of 1a (1 mmol) and 2a (1.2 mmol) under N₂ at 100 ℃ for 10 h. The conversions of 1a and the yield of 3a based on 1a using different catalysts are summarized in Table 1. Without catalyst (entry 1), only 4% yield of 3a was observed. Among the metal oxides (entries 2–11), CeO₂ (entry 2) showed the highest yield of 3a (97%). Lewis acidic oxides including Nb₂O₅, TiO₂, and ZrO₂ gave 3a in good yields (54%–73%), and SiO₂, SnO₂, and amphoteric oxides Al₂O₃ and ZnO showed low yields of 3a (19%–32%). Basic oxides such as CaO, MgO, and La₂O₃ (entries 10–12) provided 3a in low yields (16%–27%). Commercially available solid acids, including proton-exchanged zeolite Hβ-20 (entry 13), montmorillonite K10 clay (entry 14), and a sulfonic resin Amberlyst-15 (entry 15), gave 3a in 15%–40% yields. Homogeneous Lewis acid Sc(OTf)₃ (entry 16) afforded 3a in 52% yield. We also screened Ce salts (entries 17 and 18), Ce(NO₃)₄ and Ce₃(PO₄)₄, which gave 3a in 57% and 49% yield, respectively. Screening test in Table 1 exhibits that CeO₂ is the most effective catalyst for the 3-alkenylation of 1a with 2a.

Admitting CeO₂ as the best catalyst, we optimized the reaction conditions for the model 3-arenylation reaction as shown in Table 2. For the reaction at 80 ℃, the yield of 3a depends on the molar amount of 2a; the reaction with 1.25 equiv. of 2a gave higher yield (78%) than that with 1 equiv. of 2a (68%). The reactions of 1a with 1.25 equiv. of 2a at different temperatures (80–100 ℃) under solvent-free conditions were tested (entries 2–4). The reaction at 100 ℃ gave the highest yield of 3a (99%). The reactions of 1a with 1.25 equiv. of 2a at 100 ℃ in different solvents (entries 5–7) gave slightly lower yields (93%–96%) than the solvent-free conditions. Under the optimized conditions (100 ℃, 20 mg CeO₂, 1 mmol 1a, and 1.25 mmol 2a), the time course of the reaction (Fig. 1) shows that a reaction time of 10 h was sufficient to obtain the maximum yield of 3a (99%).

We studied the reusability of CeO₂ for the alkenylation of 1a with 2a under the optimized conditions (Fig. 2). After completion of the reaction of each cycle, 2-propanol (3 mL) was added to the reaction mixture and CeO₂ was separated by centrifugation, followed by consecutive washing with 2-propanol (3 mL) and acetone (3 mL), and drying at 110 ℃ for 6 h. The recovered catalyst was then used for the next reaction cycle. The result shows that the catalyst was reusable for five cycles with slight gradual decrease in the 3a yield (99%–92%). Initial product formation rates were examined (at conversion less than 30%) for each cycle in addition to the final yields (Fig. 2). A significant loss of the initial rate was not observed, but a slight decrease in the yield was observed. The yield was recovered after calcination of the CeO₂ catalyst at 400 ℃ for 3 h. This indicates that the decrease in the final yield is not due to changes in the catalyst structure but due to carbonaceous compounds on the surface. The result in Fig. 2 shows that CeO₂ is a robust heterogeneous catalyst. XRD pattern of the catalyst after the fifth cycle is nearly consistent with that of fresh catalyst (Fig. S1), which suggests no change in the catalyst structure after the reaction.

Under the optimized reaction conditions, we investigated the substrate scope for the C3-alkenylation of oxindole with various aromatic and aliphatic aldehydes by CeO₂. The yields of alkylidene-oxindoles are shown in Scheme 1. Note that the yields of the major products (E isomers) were estimated from the weight of the isolated E isomers, while the total yields of E and Z isomers and the E/Z ratios in parenthesis were determined by ¹H NMR. The benzaldehydes, including electron-donating (methyl, t-butyl, N, N-dimethyl, and methoxy at the para positions, 2b–2e; hydroxy at ortho position, 2f) and electron-withdrawing groups (chloride and nitro at the para positions, 2g and 2h) were successfully transformed into the corresponding 3-benzylidene-oxindoles (3a–3h) in high total yield (88%–99%). Regardless of the electronic nature, substrates with substituents at the para position selectively afford E isomers. Sterically congested substrate 2f with hydroxy group at the ortho position exceptionally formed Z isomer (84%, E/Z = 8:92, 3f) as a major product. Benzaldehyde homologue, i.e., 1-naphthaldehyde, is also well tolerated in this E-selective 3-arenylation reaction and affords high yield of 3-nepthalen-1-ylmethylene-oxindole (87%, E/Z = 93:7, 3i). Heteroaromatic aldehydes (e.g., isonicotinyl and furanyl groups) were successfully transformed into the corresponding 3-arylidene-oxindoles in high yield (90%, E/Z = 84:16, 3j; and 93%, E/Z = 85:15, 3k). An allylic aldehyde also underwent alkenylation to give 3-phenylallylidene-oxindole 3l in high yield and high E selectivity. Aliphatic aldehydes including linear and branched substituents at α-carbon atoms were successfully converted to the corresponding 3-alkelidene-oxindoles (3m–3p) in high yields and high E selectivity. To our knowledge, this is the first example of the catalytic C3-alkenylation of oxindole with aliphatic aldehydes.

We also demonstrated the gram scale reaction for this CeO₂-promoted 3-alkenyl-oxindole synthesis using different aldehydes (12.5 mmol) with 10 mmol of 1a and 20 mg of CeO₂ catalysts. The yields of the corresponding 3-alkenyl-oxindole products are shown in Scheme 2. The reaction of 1a with benzaldehyde 2a, p-methyl-benzaldehyde 2b, p-chloro-benzal- dehyde 2g, isonicotinaldehyde 2j, and 1-octanal 2o gave the corresponding 3-alkenyl-oxindoles in 82%–93% yield.

We prepared CeO₂ catalysts calcined at four different temperatures (500, 600, 800, and 1000 ℃), and the structure and catalytic activity of these comparative catalysts were compared with the standard CeO₂ catalyst (as received JRC-CEO-1), which was originally calcined at 300 ℃. X-ray diffraction patterns of the CeO₂ catalysts (Fig. 3) show that the samples show the same crystal phase (fluorite structure). Average size of CeO₂ crystals, calculated from the half-width of the peak at 28.5° using the Scherrer equation, was plotted versus calcination temperature in Fig. 5(B). The crystal size increases with the calcination temperature. The BET surface area (Table S1) estimated using the N₂ adsorption isotherms (Fig. 4) is plotted in Fig. 5(B) as a function of calcination temperature. The surface area of the CeO₂ catalysts decreased with the calcination temperature. We carried out catalytic tests for the standard reaction of 1a and 2a using these CeO₂ catalysts to obtain the rates of 3a formation per catalyst weight (Table S1). As shown in Fig. 5(A), the rate per catalyst weight decreases with increase in the calcination temperature. The rates per catalyst weight were divided by the BET surface areas to give the rate per surface area. As shown in Fig. 5(A), the rate per surface area of the CeO₂ catalysts increases with the calcination temperature. Clearly, the intrinsic catalytic activity of CeO₂ increases with decrease in surface area of CeO₂. This suggests that defect-free CeO₂ surface, possibly the most stable (111) surface, is the active site for this alkenylation reaction.

To check the presence of surface Ce³⁺, we studied X-ray photoelectron spectroscopy (XPS) analysis for CeO₂ catalysts (Fig. 6). A strong peak due to Ce⁴⁺ species (918 eV) [37, 38] is observed in the Ce 3d XPS spectra for all the catalysts. A weak shoulder peak at 884 eV due to Ce³⁺ species [37, 38] is observed for the catalyst calcined at 300–800 ℃, but another characteristic peak for Ce³⁺ species at 904 eV is not clearly observed for these catalysts. These results suggest that Ce species in all the catalysts are basically in Ce⁴⁺ state and Ce³⁺ species are present at the surface as a minor Ce species.

The high activity of CeO₂ in the present reaction may be due to acid-base bifunctional properties [27, 35, 36] of CeO₂ as shown in Scheme 3. First, we studied Lewis acid-base interaction between surface Ce cation and carbonyl oxygen of benzaldehyde using in situ IR experiment of benzaldehyde adsorbed on CeO₂ (calcined at 300, 600, and 800 ℃) at –50 ℃ (Fig. 7). The benzaldehyde adsorbed on CeO₂ catalysts gave the C=O stretching band at lower wavenumber (1682–1685 cm^–1) than that on non-Lewis acidic oxide, SiO₂ (1690 cm^–1). The IR band intensity of the C=O stretching band slightly increases with calcination temperature, which is nearly the same for these three CeO₂ catalysts. Considering that the surface area of CeO₂ decreases with calcination temperature, the IR result suggests that the density of Lewis acid sites increases with calcination temperature. Hence, the steep increase in the catalytic activity from 600 to 800 ℃ in Fig. 5(A) may be due to the higher density of the Lewis acid sites. This indicates that Lewis acid-base interaction between the carbonyl group (Lewis base) and the Ce⁴⁺ Lewis acid site increases nucleophilicity of the adsorbed aldehyde on CeO₂. We have previously shown that the basic sites of CeO₂ are effective for aldol condensation of ketones to form α, β-unsaturated ketone [39]. In the present condensation reaction, the basic sites of CeO₂ (i.e., surface oxygen atom) should abstract the proton of α-carbon to the carbonyl group in oxindole to give an enolate intermediate. Finally, nucleophilic attack of the enolate to the aldehyde activated over the Ce⁴⁺ site and subsequent dehydration results in the formation of the final product as illustrated in Scheme 3.

CeO₂ was found to be an efficient catalyst for synthesis of E selective 3-alkenyl-oxindoles from oxindole and aldehydes. The present method is the first general heterogeneous catalytic C3-selective alkenylation reaction of oxindole using both aromatic/aliphatic (unactivated) aldehydes. The study on the structure-activity relationship suggests that defect-free CeO₂ surface, possibly the most stable (111) surface, is the active site for this alkenylation reaction. The specific catalysis of CeO₂ in the present reaction can be due to its Lewis acid-base bifunctional property and where defect-free CeO₂ surface is active site for this reaction.

The authors are indebted to the technical division of the Institute for Catalysis (Hokkaido University) for manufacturing experimental equipment.