Introduction
Among the renewable resources, cardanol extracted from cashew nut shell liquid (CNSL) is regarded as a very useful raw material because of its availability, its low cost, and its structure, which comprises both phenol and long-chain hydrocarbon moieties [ 1, 2]. CNSL is derived from the cashew nut, which is produced in Brazil, India, Vietnam, West Africa, and other regions. Cardanol has been used to make numerous specialty organics and polymers. Much of the past work has been summarized in recent review articles [ 2‒ 4].
One of the popular modifications of cardanol is epoxidization, and there are at least two ways to achieve it. One way is to functionalize the cardanol phenol moiety. Several papers have reported the reaction of phenol with epichlorohydrin to form the glycidyl derivative [ 5‒ 9]. Another way is to generate the epoxide from the olefins on cardanol long-chain alkyl [ 10‒ 12]. This has been accomplished through enzyme catalysis [ 10] with meta-chloroperoxybenzoic acid [ 11], or with peracetic acid with the help of a cationic resin [ 12]. A recent paper conducted epoxidation reactions on both phenolic and hydrocarbon olefin moieties [ 11]. In most cases, the epoxide formed can be polymerized or reacted with appropriate reagents to form additional cardanol derivatives.
Ionic liquids (ILs) are salts frequently based on alkyl-substituted imidazolium and pyridinium cations with appropriate anions. In recent years, these materials have attracted a lot of attention, and many publications covering a wide range of applications have appeared [ 13‒ 15]. Some of the advantages of ILs include 1) good solvency for many organic, inorganic, and polymeric materials; 2) non-volatility and resistance towards air oxidation; 3) Large working temperature range (‒40 to 200 °C) for different ILs; 4) flexibility in end-use because different ILs exhibit different properties, e.g., acidity/basicity, water/organic solubility, and melting points. Many processes have been developed to use ILs as solvents and catalysts [ 16‒ 19]. In this work we have endeavored to use an IL as a catalyst. Because of our prior experience [ 19], 1-methylimidazolium tetrafluoroborate was chosen, although related ILs can likely be used as well.
The purpose of this work was to carry out the epoxidation of the olefins on cardanol long-chain hydrocarbon and the derivatization of the resulting epoxide with an aromatic amine. Interestingly, there was no reaction when cardanol epoxide and an aromatic amine were heated together, but the reaction proceeded readily in the presence of an IL to form an aminohydrin moiety. As far as we know, this is the first report of the synthesis of an aminohydrin adduct attached to the long-chain hydrocarbon moiety of cardanol.
Experimental
Materials
Cardanol (stabilized, 88.5%) was purchased from Shanghai Meidong Biological Material Co., Ltd. (Shanghai, China) and used after distillation at 240‒250 °C under 3‒5 Torr. Amberlite® IR120 cation exchange resin, glacial acetic acid, 30% hydrogen peroxide solution, aniline, toluene, ethyl acetate, formaldehyde, and deuterochloroform all came from Sigma Aldrich (Milwaukee, Wisconsin) and were used without further purification.
Synthesis of cardanol epoxide (CEO)
To a 40-dram glass vial with stir bar, the following compounds were added in turn: cardanol 1.0 g, 3.26 mmol, glacial acetic acid (1 mL, 17.5 mmol), Amberlite® IR120 cation exchange resin (1.15 g), and toluene (2 g). The vial, with a stir bar and an adaptor for reflux condenser, was placed in a Reacti-thermTM set at 55 °C. When the temperature reaches 55 °C, 30% hydrogen peroxide (0.8 mL, 7.8 mmol for one sample) was added drop-wise by needle and syringe just above the solution. The reaction mixture was stirred and allowed to heat at 55 °C for 7 h. The heat was removed, and the solution filtered with a Buchner funnel and house vacuum to remove the Amberlite. The sample was transferred with toluene to a separatory funnel, and washed with water and a saturated solution of sodium carbonate. The aqueous solution was discarded, and the toluene solution was transferred to a beaker, dried with anhydrous magnesium sulfate, and and filtered into a pre-weighted vial. It was evaporated to dryness using a rotary evaporator operating at 80 °C and under vacuum. Yield varied from 50%‒80%.
Synthesis of cardanol aminohydrin
Aniline (0.245 g, 2.6 mmol) was added to an 8-dram glass vial containing epoxidized cardanol (0.821 g, 2.6 mmol) and 0.0125 g of the ionic liquid, 1-methylimidazolium tetrafluoroborate. The mixture was stirred at 105 °C for 2.5 h, transferred to a 125 mL separatory funnel with 30 mL of ethyl acetate, and washed three times with 50 mL of water. The water layer was discarded; the ethyl acetate layer was transferred to a 100-mL beaker, where magnesium sulfate was added, and the mixture filtered. The filtrate was collected in a round-bottom flask, from which ethyl acetate was evaporated under vacuum using a rotary evaporator. The product was called sample A1. The various water washes were collected and combined in a round-bottom flask. This water was evaporated under vacuum using a rotary evaporator, and the ionic liquid was recovered.
Reaction of cardanol aminohydrin with formaldehyde
Cardanol aminohydrin (0.823 g) and concentrated HCl (0.25 g) were mixed together in a water bath kept strictly at room temperature. About 0.6 g formaldehyde solution was added dropwise over 30 minutes. The product became very thick but was still fully soluble in chloroform; it was called sample F1. A portion of the sample was then heated to 65 °C for 30 min, whereupon an insoluble material was found.
Nuclear magnetic resonance (NMR) spectroscopy
NMR data were acquired on a Bruker ARX-500 spectrometer (Bruker, Rheinstetten, Germany) at a frequency of 125 MHz, a 5-mm dual probe, and standard operating procedures. The sample solutions were prepared in deuterochloroform (99.8% deuterated, Cambridge Isotope Laboratories, Inc., Andover, MA, USA).
Thermal analysis
Pressurized differential scanning calorimetry (PDSC) experiments were carried out using a DSC 2910 thermal analyzer from TA Instruments (New Castle, DE, USA). Typically, a 2 µL samples was placed in an aluminum pan hermetically sealed with a pinhole lid and oxidized in the presence of reactant gas (dry air). Dry air (Gateway Airgas, St. Louis, MO, USA) was used to pressurize the module at a constant pressure of 1379 kPa (200 psi). A heating rate of 10 °C/min from room temperature to 300 °C was used during the experiments. The oxidation onset temperature was calculated from the exotherm in each case.
Results and discussion
Epoxidation of cardanol
The 1H NMR spectrum of cardanol is given in Fig. 1(a). The assignments of the 1H spectrum were previously reported [ 11, 20]; the peak assignments for the triene structure are noted in Fig. 1(a). The NMR peaks at 5.1 and 5.8 ppm are due to the triene terminal olefin, and the cluster of peaks at 5.4‒5.6 ppm due to the other olefins.
Epoxidation of the olefins was carried out (Reaction 1 in Scheme 1) according to the procedure reported by Rao and Palanisamy [ 12], involving acetic acid and hydrogen peroxide in the presence of a cation exchange resin (Amberlite® IR120). The 1H NMR spectra of three samples (E1, E2, E3) obtained at different degrees of epoxidation are shown in Figs 1(b‒d), and quantitative results in Table 1.
Tab.1 Examples of cardanol epoxidation under different experimental conditionsa) |
| Sample | Volume of 30% H2O2 /mL | Reaction time at 55 °C /h | Epoxidation /% (from NMR) |
|---|---|---|---|
| E1 | 0.38 | 7 | 30 |
| E2 | 0.53 | 5 | 55 |
| E3 | 0.80 | 5 | 80 |
a)In each case 1 g cardanol was used, together with toluene, acetic acid and cation exchange resin. |
With the addition of 0.38 mL of 30% H2O2 (sample E1), 30% of the olefins are epoxidized, and the epoxy peaks at 3.0‒3.2 ppm appear on the 1H NMR spectrum reaction. At the same time, the peak at 5.4 ppm decreases, and new peaks show up at 5.2 and 5.9 ppm (due to terminal olefin adjacent to epoxide) and 5.6 ppm (internal olefin adjacent to epoxide) (Fig. 1(b)). With 0.53 mL H2O2 added to cardanol (sample E2), more epoxidation is observed, the peak at 5.4 decreases further in intensity (Fig. 1(c)). With 0.8 mL H2O2 added to cardanol (sample E3), the peaks at 5.4‒5.6 ppm mostly disappear and only the terminal olefins at 5.2 and 5.9 ppm are left (Fig. 1(d)). (We also attempted the Prilezhaev reaction to produce the epoxide with meta-chloroperoxybenzoic acid on cardanol, and the 1H NMR spectrum-not shown is similar to that of Fig. 1(d).)
The 13C NMR spectra for cardanol and several of its derivatives are shown in Fig. 2. For convenience, we concentrate on the 50‒160 ppm region in the discussion below. The peaks for cardanol (Fig. 2(a)) can be assigned as follows: phenolic carbons, P1 155.3 ppm, P3 145.0 ppm, P5 129.5 ppm, P4 121.1 ppm, P6 115.5 ppm, and P2 112.6 ppm; mono-ene, M14 129.9 ppm, M15 130.0 ppm; diene, D14 130.0 ppm, D15 128.2 ppm, D17 128.1 ppm, D18 130.2 ppm; triene, T20 136.9 ppm, T14 130.5 ppm, T15 129.4 ppm, T17 127.7 ppm, T18 126.9 ppm, T21 114.8 ppm.
Fig.3 13C NMR spectra of cardanol and its derivatives: (a) cardanol; (b) cardanol epoxide sample E3, where two large olefinic peaks (*) are found at 133.1 and 117.5 ppm; (c) cardanol aminohydrin sample A1, where the aniline peaks (N) are assigned; (d) reaction product of cardanol aminohydrin and formaldehyde sample F1. Note that E= residual ethyl acetate; S= solvent, CDCl3 |
The 13C NMR spectrum for cardanol epoxide (sample E3) is given in Fig. 2(b). The epoxy peaks can be clearly seen at 54‒58 ppm. In the 110‒160 ppm region the peaks for the carbons on the phenol rings (P1‒P6) are unchanged in their chemical shifts or intensities; however, there is a notable decrease in the intensities of most of the olefinic carbons. There are a number of small olefin peaks, but two major olefinic peaks can be observed at 133.1 and 117.5 ppm. These two peaks can be assigned to the terminal olefins adjacent to an epoxide:
Thus, the 13C NMR data confirm the 1H NMR results that terminal olefins are present. It is of interest that relative to the other olefins, the terminal olefins with adjacent epoxide appears to have the least reactivity towards epoxidation under the reaction conditions employed.
Formation of cardanol aminohydrin
The next task was to react cardanol epoxide with aniline (Reaction 2 in Scheme 1). Note that by themselves these two compounds did not react even with sustained heating. However, it was discovered that the use of a small amount of an ionic liquid, 1-methylimidazolium tetrafluoroborate, catalyzed the reaction. The same ionic liquid was shown earlier to catalyze the amine-epoxide reaction in epoxidized methyl oleate even at 0.5% by weight [ 19]. In this work, about 1.0%‒1.5% by weight of this ionic liquid was used. After 2.5 h of reaction time at 105 °C, virtually all the epoxide was converted to the aminohydrin (sample A1).
The 13C NMR spectrum of the cardanol aminohydrin (sample A1, Fig. 2(c)) shows clearly the aminohydrin carbons at 68‒75 ppm (carbon attached to OH) and 56‒61 ppm (carbon attached to nitrogen). The epoxide peaks at 54‒58 ppm have completed disappeared. For this sample, aniline was in excess and only about a half of the aniline added was needed to react with epoxide. The C1 (ipso) carbon of aniline can be found at 146.2 ppm (unreacted) and 148.5 ppm (reacted). The C2, C3, and C4 aniline peaks are found at 115.4, 129.4 and 118.8 ppm, respectively.
Possible applications
We have explored a possible application of the cardanol aminohydrin as an antioxidant or heat stabilizer. An appropriate test is to add different levels of the potential antioxidant to a poly(α-olefin) fluid (PAO 8) and then run the pressurized differential scanning calorimeter (PDSC) to observe the onset temperatures of oxidation changes. Figure 3 shows the oxidation onset temperature for three concentrations of aniline-modified cardanol epoxide. It is encouraging that the cardanol aminohydrin increased the onset oxidation temperature of PAO 8 by roughly 14 degrees.
Furthermore, the new cardanol aminohydrin derivative contains the highly reactive aromatic amine functionality, and it can be a useful synthon for further product development. As a test of principle, we have chosen to react the cardanol aminohydrin with formaldehyde (Scheme 2). In the literature aniline-formaldehyde resins are known, and their synthetic procedures well documented [ 21‒ 23]. However, cardanol also contains the phenolic functionality, which can undergo phenol-formaldehyde resin formation [ 24]. Thus, we need to be careful to distinguish the reactions of formaldehyde with aniline versus phenol in the formaldehyde reaction.
As it turned out, the reaction temperature served to distinguish the formaldehyde reactions between aniline and phenol. Through appropriate cooling, we kept the reaction between cardanol aminohydrin and formaldehyde at room temperature. The resulting reaction mixture visibly increased in viscosity indicating the presence of resin formation. The 13C NMR spectrum corresponding to the room temperature reaction of formaldehyde with cardanol-epoxy-aniline adduct (sample F1) is shown in Fig. 2(d). The 13C shifts of the phenol carbons stay the same (labelled as P1–P6), but the peaks for aniline carbons have all moved because many different structures have been produced. Thus, the phenol moiety stays unreacted, but the aniline moiety has reacted with formaldehyde at room temperature (A confirmation experiment was also done where formaldehyde was added to cardanol at room temperature, and no change in 13C NMR spectrum was observed). Because of the excess formaldehyde used and the low temperature, there are several peaks in Fig. 2(d) corresponding to the formaldehyde oligomers at 87‒94 ppm and related methoxy peaks at around 55 ppm. The methoxy peaks are derived from methanol, present in formaldehyde as a stabilizer. These peaks from formaldehyde self-reactions have been previously reported [ 25, 26]. When the same cardanol aminohydrin/formaldehyde reaction mixture was heated to 65 °C, the phenol started to react and an insoluble, crosslinked material was obtained. However, it seems that if we generate the cardanol aminohydrin/formaldehyde adduct at room temperature, it can serve as a prepolymer, which can then be subjected to further polymerization at a higher temperature with the phenol on cardanol (self-condensation) or with a second component to produce a copolymer.
Further applications of the cardanol-aminohydrin-formaldehyde resin may be gleaned from the literature on aniline-formaldehyde resins. Thus, this material may perhaps be used as an adhesive [ 27], curing agent for epoxy resins [ 21, 22], fuel cell material [ 23], or chelating resins [ 28‒ 30], e.g., for removal of metal ions from water [ 29, 30].
Conclusions
In view of the availability and low cost of cardanol, it is useful to develop new reaction pathways. It was shown in this work that an aromatic amine can be attached to the hydrocarbon part of cardanol through a two-step reaction involving epoxidation and ring-opening reaction with the amine. The resulting aminohydrin has antioxidant properties and can be used to make further cardanol-derived products. As proof of principle, the aromatic amine has been reacted with formaldehyde to form a prepolymer resin.