A rapid ultrasound-promoted Horner–Wadsworth–Emmons reaction for the preparation of ferrocene derivatives. Application to ferrocene-modified ITO electrodes

2.1 Measurements

The chemicals were purchased from sigma–Aldrich and Acros and used as received. All the solvents used are of analytical grade from LabScan. Thin layer chromatography (TLC) was performed on Fluka aluminum foils pre-coated with F₂₅₄ silica gel. ¹H, ¹³C and ³¹P NMR spectra were recorded on a Bruker Advance 300 MHz apparatus in deuterated solvents as indicated below. Chemical shifts for ¹H were reported using TMS signals as internal references. The reaction progress was monitored by TLC or ¹H NMR spectroscopy. The ultrasonic irradiation was performed by using a standard commercial cleaning bath operating at a frequency of 35 kHz. Fourier-transform infrared spectra were recorded in ATR mode using a Nicolet 8700 apparatus. The electrochemical experiments were conducted in phosphate saline buffer (PBS, pH = 7.4) solution in a standard three-electrode glass cell using a Metrohm Autolab PGStat M101 electrochemical workstation controlled by Nova 1.8 software. A platinum wire was used as the counter electrode and Ag/AgCl electrode was used as reference electrode. A 2 × 3 cm ITO-coated glass plate was used as the working electrode. GCMS were recorded on a Perkin–Elmer Autosystem XL GC apparatus (Waltham, MA, USA) equipped with a quadrupole Turbo mass detector. The ionization energy was set à 70 eV.

2.2 ITO electrode modification

ITO-coated glass purchased from SOLEMS (Palaiseau, France) cut into (2 × 3 cm) rectangular slides was used as a working electrode. The electrode was cleaned by successive ultrasonic treatment in solutions of NaOH (0.1 M), ultrapure water, and ethanol for 20 min. Then, it was rinsed with ultrapure water and sonicated for further 10 min. To modify the ITO electrode with (E)-2-ferrocenylvinylphosphonic acid 2a, the cleaned ITO electrode was immersed in a 10 mM ferrocenyl vinylphosphonic acid solution for 3 h and sonicated in the same solution for 15 min. The modified ITO electrode was rinsed finally with ultrapure water, dried with N₂ and then used for voltammetric studies.

2.3 General procedure for the ultrasound-assisted reaction

A mixture of ferrocenecarboxaldehyde (54 mg, 0.25 mmol), pastel of sodium hydroxide (70–110 mg, 1.75–4.50 mmol), solvent (1.0 mL), H₂O (0.2 mL, 5 drops) and the desired phosphonate was introduced in a glass vial (volume = 5 mL) and closed with a septum. The sample was irradiated by ultrasound using a Bandelin Sonorex RK25H apparatus operating at 35 kHz (60 W). The bath temperature was maintained at 25–30 °C if not indicated otherwise. The mixture was poured into water and extracted twice with 10 mL of chloroform. The relative yields were determined from the ¹H NMR signal integration. Synthesized products were purified by silica chromatography eluted by (30:70) ethyl acetate/cyclohexane. Compound 2a was prepared according to the literature from the hydrolysis of diethylphosphonate ester 2 using trimethylchlorosilane followed by treatment in hydrochloric acid (1.0 M) (Plazuk et al., 2004).

3.1 Base and solvent effects

Preliminary investigations showed that sonication of a mixture of ferrocenecarboxaldehyde 1 and methylene bis(diethyl-phosphonate) in a given solvent afforded diethyl 2-ferrocenyl-α,β-unsaturated phosphonate 2 in few minutes (5–15 min). However, the same mixture of reagents gives the correspondent product with 30% of yield using heating mode and under stirring for 24 h. The unreacted aldehyde 1 can be easily recovered using separation by silica column chromatography eluted by ethyl acetate/cyclohexane (30:70). Moreover, the reaction progress can be easily monitored by TLC or ¹H NMR spectroscopy. Aldehyde 1 was reacted with methylene bis(diethyl-phosphonate) in the presence of a base affording compound 2 according to Scheme 1. A small amount of water was systematically added to dissolve the mineral bases. The bath temperature was kept at ambient temperature by addition of cold water if it rises above 30 °C. The influence of solvent and base was systematically studied and the results are reported in Table 1. The relative yields of product 2 were determined from the ratio of the integrated ¹H NMR signal of the remaining carbonyl proton of the aldehyde 1 and the α,β vinylic protons of the product 2. All the products were obtained exclusively as (E)-isomers which has been confirmed by comparing ¹H NMR spectra with the literature data (Davis et al., 1998; Plazuk et al., 2004).

In THF and in absence of a base, no conversion of compound 1 was observed even after 180 min of ultrasonic irradiation (Table 1, entry 1). Weak bases such as sodium bicarbonate or sodium carbonate or organic bases such as triethylamine did not yield any HWE products. The starting aldehyde 1 was entirely recovered (entry 2–3, 6). When potassium bicarbonate is used, the yields are still low and do not exceed 15% after 90 min (entry 4–5). For stronger bases such as alkali hydroxides, the yields of 2 were significantly improved (entry 7–11). For instance, a total consumption of the aldehyde and formation of 2 in yields >99% were obtained after 15 min of sonication using NaOH or LiOH and the aldehyde proton signal was no longer detectable in the ¹H NMR spectra. However, when KOH was used, the transformation was still unfinished after 1 h of irradiation suggesting an effect of the metallic cation in the stabilization of the anionic intermediate (Thompson and Heathcook, 1990). After 30 min of irradiation with stronger bases such as tBuONa or tBuOK, the reaction was not complete (entry 12–13). In the water-containing reaction mixture, the tert-butyl alcoholates are converted into the corresponding alcohols rendering the solution highly viscous what is probably the reason for the lower yields of 2.

In acetonitrile (ACN), dimethylformamide (DMF), dimethylsulfoxide (DMSO), tert-butanol (tBuOH) or H₂O as solvents, sodium hydroxide gives always better results than potassium hydroxide and the yields in these solvents are lower than those obtained in THF or in water (entry 15–28). A combination of THF/H₂O–NaOH or H₂O–NaOH seems to be the best choice for this reaction (entry 7 and 27–28).

3.2 Effects of ratio of reagents and sonication time

The reaction rate also depends on the ratio of aldehyde and phosphonate, and the sonication time (Table 2). In fact, for 1.2 equiv. of bisphosphonate (EWG = (EtO)₂P(O)) dissolved in THF, irradiation for 30 min gave product 2 in yields of 90% (entry 1–2). In the presence of 2.0 equiv. of bisphosphonate, conversion of 1 was faster and product 2 was obtained in 81% after 15 min of irradiation, and the reaction was almost complete after 30 min (entry 3–4). Using a threefold to sevenfold excess of bisphosphonate, the yields increased to 83% and more than 99% after 15 min of sonication, respectively (entry 5–7). For the same duration in water, only 2 equiv. of bisphosphonate were needed for a complete conversion (entry 8–11) denoting strong effect of the solvent polarity on the reaction rate (Thompson and Heathcook, 1990).

3.3 Preparation of 2-substituted vinylferrocenes

The optimal conditions (THF/H₂O—NaOH) have been successfully applied to the preparation of a series of 2-substituted vinylferrocenes 2–5 where EWG = P(O)(OEt)₂, CO₂H, CO₂Et, CO₂Me, COCH₃ and CN. Starting from the appropriate dialkyl phosphonate, ultrasonic irradiation in presence of sodium hydroxide in THF or in H₂O afforded the desired products (Table 3). In case of products 3b/c (EWG = CO₂R), when a large excess of sodium hydroxide in water was used (ca. 11 equiv.), the formed products 3b and 3c were not isolated but they were hydrolyzed in situ to their corresponding acid 3a (entry 4, 5). A 4–8 fold excess gave the vinylferrocenes 2–5 in excellent yields. In the case of product 3c using THF/H₂O as solvent and after 15 min of sonication, the yield was only of 72% with the recovery of ∼30% aldehyde. Using only THF did not improve the yields (entry 8–9). Rathke et al. reported that the addition of a chelating cation such Li⁺ or Mg²⁺ ameliorated the reaction yields by stabilizing the carbanions (Rathke and Nowak, 1985). Addition as LiClO₄ did not improve much the yields but when anhydrous magnesium sulfate was added, the reaction was complete in less than 5 min (entry 8, 10–11). A general procedure was established and the compounds 2–5 were characterized by ¹H NMR spectroscopy and GC–MS confirming already published results.

3.4 Modification of ITO electrode

Recently, it was demonstrated that the modification of ITO substrates by self-assembled monolayers (SAMs) of electroactive molecules or oligomers, bearing a terminal phosphonic group (R-P(O)(OH)₂), can be used in optical devices such as polymer light-emitting diodes (PLEDs). The SAMs showed great promises in maximizing the injection of holes into conducting photoactive oligomers and polymers (Hanson et al., 2005; Bardecker et al., 2008). The (E)-2-ferrocenylvinylphosphonic acid 2a was chosen thanks to the good affinity of the phosphonic acid group to ITO substrates (Chockalingam et al., 2011). Before immobilizing the ferrocenylphosphonic acid 3a onto ITO electrode surfaces, the bare electrode was pretreated by a sequential cleaning with NaOH, EtOH and ultrapure water. The electrode was then immersed for 3 h and sonicated for 15 min in a diluted ferrocenic solution and thoroughly washed with ethanol then water to remove nonbonded molecules (Fig. 1).

Close

Figure 1 Schematic representation of the adsorption process of the ferrocenylvinylphosphonic acid and the oxidation of ferrocene-modified SAM on the ITO electrode.

Export to PPT

Cyclic voltammetry (CV), performed in a PBS solution, displayed a quasi-reversible signal (i_pc-to-i_pa ratio close to unity and peak-to-peak separation is close to 60 mV) related to a one electron transfer from the ferrocene moiety which is oxidized to a ferrocenium cation (Fig. 2a). The DPV curve of the ferrocene-modified ITO electrode as a working electrode marks the presence of an oxidation signal localized at 0.16 V which is related to the presence of ferrocenic group confined on ITO electrode surface (Fig. 2b). To assess the presence of the ferrocene binding the ITO, several CVs at various scan rates from 100 to 500 mV s⁻¹ were performed and are displayed in Fig. 2c. The Randles–Sevick plot representing the oxidation current peak versus the scan rate showed a linear relationship suggesting an adsorption-controlled electrode process according to plot (Fig. 2d) (Li et al., 2008; Sahli et al., 2013) and thus proving that the electroactive molecules are confined to the electrode surface. Moreover, these redox wave shapes did not change during continuous potential cycling over several hundred of cycles at room temperature confirming the stability of SAMs.

Close

Figure 2 (a) Cyclic voltammetry recorded at 100 mV/s in a PBS solution of the ferrocene-modified ITO electrode (working electrode), Ag/AgCl (KCl 3 M) as reference electrode and a platinum wire as counter electrode, (b) DPV performed in PBS using the previous three electrodes as electrochemical device, (c) cyclic voltammetry at various scan rates from 100 to 500 mV s⁻¹, (d) Randles–Sevick plot of oxidation peak current versus the scan rates.

Export to PPT