Synthesis, curing kinetics, mechanical and thermal properties of novel cardanol-based curing agents with thiourea

Abstract

Two new flexible cardanol-based epoxy curing agents with cross-linkable thiourea groups ([2-(2-amino-ethylamino)-ethyl]-thiourea of modified cardanol and [2-(2-amino-ethylamino)-ethyl]-thiourea of modified cardanol 2,3-dihydroxy propyl ether, ETC and TCP for short, respectively) were synthesized using a Mannich reaction. The structure of the prepared curing agents was analyzed by a nuclear magnetic resonance (NMR) and Fourier transform infrared spectroscopy (FT-IR). The curing kinetics of ETC (or TCP) cured diglycidyl ether of biphenyl A (DGEBA) epoxy resins was investigated using the Kissinger, Flynn–Wall–Ozawa and Crane methods by non-isothermal differential scanning calorimetry (DSC). Thermal properties of the cured epoxy resins were evaluated with thermogravimetric analysis (TGA) and DSC. The morphology and mechanical properties of the cured epoxy resins were also studied. The results showed that the epoxy curing agents had a high-reactivity and a low-viscosity, especially at low temperatures. The activation energy of ETC/DGEBA and TCP/DGEBA calculated by the Kissinger method were 53.6 and 60.5 kJ mol⁻¹, respectively. The toughness of the cured epoxy resin was noticeably improved by the introduction of two new flexible curing agents. The glass transition temperature (T_g) decreased with the introduction of cardanol-based curing agents in an epoxy resin system. In addition, the cured epoxy materials exhibited good mechanical properties.

---

1. Introduction

Biomass, which is considered economical and environmentally friendly, has attracted interest from both academic and industrial researchers to reduce the usage of petroleum-based chemicals for the production of commercial products.^1–4 Among these, cashew nut shell liquid (CNSL) is a kind of sustainable, low-cost and largely available natural resource,⁵ which is obtained as a byproduct of the cashew processing industry.⁶ Cardanol is readily separated from the other components of CNSL by the vacuum distillation method. It is a phenol derivative, which has a meta-substituent of a C₁₅ unsaturated hydrocarbon chain with one to three double bonds.^7–9 This unique structure exhibits promising properties for making tailor-made polymers of high value.¹⁰ Cardanol has been used in phenolic resins,^11–13 benzoxazine resins,¹⁴ epoxy resins,^15,16 vinyl ester polymers,¹⁷ and soft nanomaterials.¹⁸

Although epoxy resins are widely used in industrial applications, one of main drawbacks of epoxy resins is their brittleness.¹⁹ There are many methods described in the literatures to improve the low impact strength and high-brittleness of epoxy resins.^20–24 One of most common approaches is the introduction of toughening curing agents into the epoxy network.²⁵ Another disadvantage of epoxy resins is their high viscosity, especially under low temperatures, which is unfavorable to the processing. Besides structural modifications for epoxy resins, the introduction of a curing agent with a low viscosity is an alternate approach to decrease the viscosity of the curing systems.

Phenalkamines are widely used as epoxy curing agents. These phenalkamines are Mannich-based and generally prepared by phenolic compounds, polyamines and aldehydes. Phenalkamines overcome the shortcomings of aliphatic amines as curing agents for epoxies. Those epoxies are irritating, toxic and easily absorb moisture in air.²⁶ From a structural point of view, cardanol is used as phenolic compounds in the Mannich reaction. The side C₁₅ aliphatic molecular chain of the product is responsible for low viscosity, high toughness and water resistance. Moreover, the phenolic hydroxyl group makes a curing agent highly active even at low temperatures.²⁷ Therefore, phenalkamines are generally used at room or low temperatures for wet surfaces under a high humidity condition. However, phenolic hydroxyl of phenalkamines leads to a poor chemical resistance. The cured epoxy resins are dark brown in color,⁵ thus the application is limited.

In this study, two different cardanol-based curing agents with high reactivity were prepared. The curing agents were synthesized by a Mannich reaction from cardanol (or cardanol 2,3-dihydroxy-propyl ethers), paraformaldehyde and [2-(2-amino-ethylamino)-ethyl]-thiourea. It was noted that using cardanol 2,3-dihydroxy-propyl ethers as synthesis materials could greatly improve the brittleness of epoxy resins. The cured epikote resin appeared a light yellow in color. This novel cardanol-based curing agent was applied as top coating materials. The curing agent possessing the propylene glycol ether groups had a higher glass transition temperature (T_g). The introduction of thiourea group in the structure of phenalkamines accelerated the curing reaction at room temperature (20 °C) or a low temperature (0 °C). In addition, the cured epoxy materials exhibited high fracture toughness and good impact resistance.

2. Experimental

2.1. Materials

Cardanol was purchased from the Shanghai Dura Chemical Corporation Ltd and used after vacuum distillation (at 75 °C and 5 kPa in 3 h). Diethylenetriamine (DETA, ≥98%) and thiourea were obtained from the Sinopharm Chemical Reagent Corporation Ltd. 3-Chloro-1,2-propanediol (≥99%), ethanol (99.7%), methanol (99.5%), dichloromethane (99.5%), sodium hydroxide solution (98%) and sodium sulfate anhydrous (99%) were purchased from the Aladdin Industrial Corporation. Paraformaldehyde (≥95%) was acquired from Guangzhou Import and Export Chemicals Company Ltd. Diglycidyl ether of biphenyl A DGEBA (commercial name: E51) was supplied by the China Petrochemical Corporation.

2.2. Instrumentation

Infrared spectra were recorded by a Tensor 27 Bruker FT-IR spectrometer (KBr disk). ¹H NMR and ¹³C NMR spectra were measured on a DRX-400 spectrometer using tetramethylsilane (TMS) as an internal reference. The compounds were ionized in the electrospray ionization (ESI) and determined by a Bruker maXis impact high-resolution mass spectrometer. Viscosity measurements were carried out using a NDJ-1 Rotational Viscometer. Curing kinetics of the ETC/DGEBA and TCP/DGEBA systems (5.0–8.0 mg) were conducted by the DSC method (Perkin Elmer diamond DSC analyzer) at different heating rates of 5, 10, 15, 20 °C min⁻¹, under N₂ atmosphere at a constant flow of 50 mL min⁻¹. The glass transition temperature (T_g) was determined by the same DSC thermal analyzer. Thermogravimetric analyzer (TGA) thermograms were obtained with a TG 209 F3 analyzer under N₂ atmosphere at a heating rate of 10 °C min⁻¹ from 50 to 800 °C. Tensile strength, lap shear strength and compressive strength were measured on a WDW-3020 microcomputer control electronic universal testing machine with a crosshead speed of 5 mm min⁻¹. The charpy impact strength of the samples was examined with a XJU Izod Impact Tester. Five samples of each compound were measured, and the average values were recorded. The fracture surfaces of the impact samples were tested by an S-4300 15 kV Hitachi scanning electron microscopy (SEM).

2.3. Synthesis of curing agent

2.3.1 Synthesis of [2-(2-amino-ethylamino)-ethyl]-thiourea (DET). 61.8 g (0.60 mol) of diethylenetriamine and 39.5 g (0.52 mol) of thiourea were added into a flask equipped with a mechanical agitator, a thermometer and an ammonia absorption device. The contents were heated to 55–60 °C with mechanical stirring. After the complete dissolution of thiourea, the reactant mixture was incubated at 140 °C for 4 h. Then the temperature was lowered to 50–55 °C for 1 h. A clear yellow liquid product was obtained at room temperature.

IR (KBr, cm⁻¹): 3206, 3315 (N–H), 1225 (C–N), 1037 (C=S). ¹H NMR (400 MHz, DMSO-d₆, ppm): 2.47–3.68 (–H₂C–N), 1.93 (–NH–). ¹³C NMR (400 MHz, DMSO-d₆, ppm): 182.7, 183.8 (C=S), 41.0–51.7 (N–C). Mass (m/z): calculated C₅H₁₄N₄S⁺, C₁₀H₂₅N₇S₂Na⁺, C₁₅H₃₆N₁₀S₃⁺: 163.1012, 330.1505, 453.2359; found 163.1012, 330.1502, 453.2350.

2.3.2 Synthesis of cardanol 2,3-dihydroxy-propyl ether (CPE). A flask equipped with a thermometer, a magnetic stirrer, a reflux condenser and a constant pressure funnel, was charged with 22.50 g of cardanol (0.075 mol), 9.62 g of 3-chloro-1,2-propanediol (0.087 mol) and 50 mL ethanol. The mixture was heated to 45 °C for 30 min. Then an aqueous solution of sodium hydroxide (20% wt, 17.0 g) was added using a constant pressure funnel. The reaction was continued at 75 °C for 3 h. The salt was filtered and the product was washed by distilled water and dichloromethane until neutralized. The oil layer was purified using rotary evaporators under a vacuum to remove unreacted 3-chloro-1,2-propanediol and the solvent. Then a brown liquid was obtained.

IR (KBr, cm⁻¹): 3364 (O–H), 1625 (C=C), 1187 (C–O). ¹H NMR (400 MHz, DMSO-d₆, ppm): 6.71–7.17 (aromatic protons), 4.83–5.81 (–HC=CH–), 4.68 (O–H), 3.45–3.95 (–CH₂–O–), 0.83–2.85 (aliphatic side chain of cardanol). ¹³C NMR (400 MHz, CDCl₃, ppm): 114.7–158.4 (aromatic), 111.4–130.3 (C=C), 77.0, 70.4, 63.7 (C–O), 13.7–35.9 (aliphatic side chain of cardanol).

2.3.3 Synthesis of [2-(2-amino-ethylamino)-ethyl]-thiourea of modified cardanol (ETC). The compound ETC was synthesized by a Mannich reaction, as illustrated in the synthesis sequence shown in Fig. 1. The experimental procedure is as follows: (1) charge 39.2 g (0.13 mol) of cardanol and 30.3 g (0.17 mol) of DET into a flask equipped with a magnetic stirrer, a thermometer and a Dean–Stark water trap. (2) After the reagents were homogeneously mixed at 55 °C under a N₂ atmosphere, 4.3 g (0.14 mol) of paraformaldehyde was added in three intervals, over a time period of 1.5 h. (3) The temperature of mixture was raised to 95 °C to continue the reaction in 3 h. (4) The product was washed by distilled water, and heated by rotary evaporators under a vacuum to remove the solvent. (5) The product was purified by the silica gel column chromatography (the eluent mixture of dichloromethane and methanol (30 : 70)) to give a reddish-brown liquid. The resulting product had an amine value (the number of milligrams of KOH equivalent to 1 g of curing agent) of 280 mg KOH per g and a viscosity at 25 °C of 1050 mPa s.

Fig. 1 Synthesis sequence adopted for the preparation of DET, CPE, ETC and TCP.

IR (KBr, cm⁻¹): 3386 (O–H), 1627 (C=C), 1256 (C–N), 1014 (–C=S). ¹H NMR (400 MHz, DMSO-d₆, ppm): 7.95 (Ar-OH), 6.71–7.17 (aromatic protons), 4.95–5.82 (–HC=CH–), 3.74–3.93 (N–CH₂-Ar), 3.38–3.50 (N–CH₂–), 2.49–2.89 (N–CH₂–), 2.09 (–NH–), 0.85–1.99 (aliphatic side chain of cardanol). ¹³C NMR (400 MHz, DMSO-d₆, ppm): 182.7, 183.8 (C=S), 114.8–157.4 (aromatic), 112.5–129.9 (C=C), 40.1–51.7 (N–C), 13.9–35.2 (aliphatic side chain of cardanol).

2.3.4 Synthesis of [2-(2-amino-ethylamino)-ethyl]-thiourea of modified cardanol 2,3-dihydroxy propyl ether (TCP). The compound TCP, according to the aforementioned procedure, was treated with 19.8 g (0.053 mol) of CPE, 12.2 g (0.069 mol) of DET and 1.8 g (0.061 mol) of paraformaldehyde at 95 °C for 3 h. The product was washed in distilled water and vacuum distilled to remove the solvent. Purification was obtained using the eluent mixture of dichloromethane and methanol (50 : 50) by the silica gel column chromatography. The product obtained was a bright yellow liquid, having an amine value of 245 mg KOH per g and a viscosity at 25 °C of 890 mPa s.

IR (KBr, cm⁻¹): 3328 (O–H), 1663 (C=C), 1186 (C–N), 1040 (C=S). ¹H NMR (400 MHz, DMSO-d₆, ppm): 6.73–7.17 (aromatic protons), 4.96–5.82 (–HC=CH–), 3.81–4.03 (–CH₂–O–), 3.76–3.91 (N–CH₂-Ar), 3.38–3.44 (N–CH₂–), 2.76 (C=C–CH₂–), 2.49 (N–CH₂–), 0.83–1.97 (aliphatic side chain of cardanol). ¹³C NMR (400 MHz, DMSO-d₆, ppm): 182.7, 183.8 (C=S), 114.4–161.3 (aromatic), 111.5–129.9 (C=C), 61.0–70.5 (C–O), 40.9–52.2 (N–C), 12.8–35.2 (aliphatic side chain of cardanol).

2.4. Preparation of samples

(1) Mixture: for the curing system, to obtain the optimal performance of the composite, the number of epoxy groups should equal the number of amine hydrogens at the stoichiometric point.²⁶ ETC/DGEBA, TCP/DGEBA and DETA/DGEBA were mixed at the ratio of 48.5/100, 59.4/100 and 12.5/100 by weight, respectively, and degassed with a vacuum pump to eliminate air bubbles. (2) Pre-curing: the samples were cured at 30 °C for 8 h for pre-curing, and the prepolymers were used to study the kinetic of epoxy resin curing processes. (3) Post-curing: then the mixtures were heated at 80 °C for another 4 h for post-curing. The cured samples were analyzed for their thermal and mechanical property. The tensile and impact specimens were prepared in accordance with ASTM D638-03 and ASTM D256, respectively. The samples for testing lap shear strength and compressive strength were prepared according to GB/T 2567 and GB/T 7124. The impact fractured surfaces of the samples were cleaned (with a soft banister brush) and coated with thin layer of gold for scanning electron microscopy examinations.

3. Results and discussion

3.1. Synthesis and characterization

The synthesis sequence of DET, CPE, ETC and TCP are shown in Fig. 1. DET, a thiourea-modified polyamine, was prepared from diethylenetriamine and thiourea by a condensation reaction with release of NH₃, in accordance with the method described by Fernando.²⁸ CPE, which is ether of cardanol with a diol, was synthesized by Suresh and Kishanprasad²⁹ from epichlorohydrin in two steps. In our experiment, 3-chloro-1,2-propanediol was used as raw materials in the presence of 20% sodium hydroxide aqueous solution in ethanol to obtain the product in one step. In addition, the post-treatment was less complicated.

A novel phenalkamine containing thiourea groups in the structure, ETC, has high-activated hydrogens in the molecules. It was synthesized by a Mannich reaction of cardanol, paraformaldehyde and thiourea-modified diethylenetriamine. The primary amines of thiourea groups in the structure can rapidly cure epoxy resin at an ambient temperature. Thiourea group is the N, S nucleophile and sulfur is more nucleophilic than nitrogen,³⁰ so the primary amines of thiourea groups have the greater nucleophilicity and the higher activity relative to the aliphatic amines.³¹ Thiourea group can react very quickly and exothermically with epoxide groups in the curing process even at a low temperature (0 °C). The epoxy resin cured by ETC also had excellent mechanical properties.

TCP, a bright yellow liquid, is a unique Mannich base modified from ETC by the replacement of the phenol's hydroxyl group with the propylene glycol ether group. The instability of the phenolic hydroxyl group in the air can be improved. The cured epoxy resin with TCP is more transparent and lustrous as well. The viscosity of TCP is lower than that of ECP. This can be attributed to a steric hindrance effect and an intermolecular hydrogen bonding. The propylene glycol ether group is much larger and longer than the phenolic hydroxyl group. FT-IR, ¹H NMR and ¹³C NMR were employed to characterize the structure of DET, CPE, ETC and TCP.

The structures of the products were supported by FT-IR spectrum shown in Fig. 2. The intense broad peak at 3315 cm⁻¹ in the FT-IR spectrum of DET is the indicative of stretching vibration of N–H group. Signs of stretching vibration of C–N group and stretching vibration of C=S group are observed at 1225 cm⁻¹ and 1037 cm⁻¹, respectively. The C–O–C stretching vibration is characterized by a peak at 1187 cm⁻¹ in the FT-IR spectrum of CPE. The peaks at 1589 cm⁻¹ and 1456 cm⁻¹ are attributed to aromatic stretching bends. In the FT-IR spectrum of ETC peaks are observed at 3386 cm⁻¹ (stretching vibration of O–H group), at 3045 cm⁻¹ (C–H stretching vibration of the unsaturated group in cardanol), at 1256 cm⁻¹ (stretching vibration of C–N group), at 1014 cm⁻¹ (stretching vibration of C=S group). In the FT-IR spectrum of TCP peaks are observed at 3328 cm⁻¹ (stretching vibration of O–H group), at 3052 cm⁻¹ (C–H stretching vibration of the unsaturated group in cardanol), 1663 cm⁻¹ (stretching vibrations of C=C group), at 1186 cm⁻¹ (stretching vibration of C–N group), at 1040 cm⁻¹ (stretching vibration of C=S group).

Fig. 2 FT-IR spectrum of DET, CPE, ETC and TCP.

The structures of the products were further confirmed by the ¹H NMR and ¹³C NMR spectra (Fig. 3 and 4). In the ¹H NMR spectra of DET, the peaks from 2.42 to 2.71 ppm, at 3.42 ppm, and at 3.68 ppm are attributed to the –CH₂–N structure, while the peak observed at 1.93 ppm is assigned to the N–H structure. The peaks from 3.79 to 4.03 ppm (for CPE and TCP) are attributed to the –O–CH₂– structure.³² In the ¹H NMR spectra of CPE, ETC and TCP, the peaks from 5.04 to 5.82 ppm are assigned to the –CH=CH– structure.³³ The peaks observed from 3.74 to 3.93 ppm (for ETC and TCP) are the characteristic chemical shift for the Ar-CH₂–N structure of Mannich bases.³⁴ The peaks at 3.44 ppm (for ETC and TCP), at 2.73 and 2.89 ppm (for ETC), and at 2.49 ppm (for ETC and TCP) were assigned to the N–CH₂– structure. In the ¹H NMR spectra of ETC, the peaks at 7.95 and 2.09 ppm are attributed to the Ar-OH and the N–H structures, respectively. In the ¹³C NMR spectra, the peaks at 182.7 and 183.8 ppm (for DET, ETC and TCP) are ascribed to the C=S structure. The peaks from 40.1 to 52.2 ppm are attributed to the C–N structure of the modified polyamines. The C–O structure is characterized with the absorption peaks from 61.0 to 70.5 ppm (for CPE and TCP). The signals of aromatic rings are observed from 114.4 to 161.3 ppm. The peaks from 111.4 to 136.7, and from 12.8 to 35.9 ppm are attributed to the unsaturated alkyl side chain of the cardanol.²⁹

Fig. 3 ¹H NMR spectrum of DET, CPE, ETC and TCP.

Fig. 4 ¹³C NMR spectrum of DET, CPE, ETC and TCP.

3.2. Curing kinetics of ETC/DGEBA and TCP/DGEBA systems

The curing kinetics of ETC/DGEBA and TCP/DGEBA systems was studied by DSC analysis at different heating rates of 5, 10, 15, 20 °C min⁻¹. Fig. 5 and 6 show the DSC thermograms of the curing agent/DGEBA systems at different heating rates. There is a single exothermic peak on the DSC curve at each of four different heating rates. These shifted to a higher temperature with the increase of heating rate due to thermal hysteresis.²⁶ The peak of temperature (T_p) can be regarded as the maximum conversion rate in the DSC curves and the suitable curing temperature for cuing the epoxy resin.³⁵ The T_ps of curing exotherms at four different heating rates are summarized in Table 1. In the DSC curves for TCP/DGEBA system, the T_p is higher than that of ETC/DGEBA system at every heating rates. This can provide an explanation that ETC is more reactive than TCP.

Fig. 5 DSC thermograms for the ETC/DGEBA system at heating rates of (a) 5 °C min⁻¹; (b) 10 °C min⁻¹; (c) 15 °C min⁻¹; (d) 20 °C min⁻¹.

Fig. 6 DSC thermograms for the TCP/DGEBA system at heating rates of (a) 5 °C min⁻¹; (b) 10 °C min⁻¹; (c) 15 °C min⁻¹; (d) 20 °C min⁻¹.

Table 1 Peak temperature (T_p, °C) of curing exotherms for samples at different heating rates (β, °C min⁻¹)

Curing agent/epoxy resin system	β (°C min^−1)
	5	10	15	20
ETC/DGEBA	105.4	119.2	127.4	135.4
TCP/DGEBA	116.3	128.8	138.1	144.3

---

To qualitatively evaluate the curing kinetics of ETC/DGEBA and TCP/DGEBA systems, the Kissinger,³⁶ Flynn–Wall–Ozawa^37,38 and Crane³⁹ methods were considered as the simple ways to deal with the complex thermoset curing process.^26,40,41 The activation energy (E_a) and curing reaction order (n) can be calculated using the following methods.

Kissinger's equation can be expressed as follows

where T_p is the temperature of exothermic peak in the DSC curve, E_a is the activation energy, β, R and A are heating rate, ideal gas constant and Arrhenius constant, respectively. The activation energy (E_a) and Arrhenius constant (A) can be obtained from the slope of ln(β/T_p) vs. 1/T_p plot and the intercept of the corresponding straight line (Fig. 7).

Fig. 7 Plots for determination of the E_a and A by the Kissinger method for ETC/DGEBA and TCP/DGEBA systems.

Flynn–Wall–Ozawa's equation is

in eqn (2), E_a is the activation energy, T_p is the temperature of the exothermic peak in the DSC curve, where β is heating rate, R is ideal gas constant, A is Arrhenius constant, and α is the conversion of the curing reaction. On the basis of Flynn–Wall–Ozawa's equation, E_a can be obtained from the slope of ln β vs. 1/T_p plot (Fig. 8).

Fig. 8 Plots for determination of the E_a by the Flynn–Wall–Ozawa method for ETC/DGEBA and TCP/DGEBA systems.

Furthermore, the order of curing reaction (n) can be obtained by applying the Crane method, using the following equation:

where, E_a is the activation energy calculated by the Kissinger method, n can be derived from the slope of ln β vs. 1/T_p plot when E_a is much higher than 2T_p.

According to the experiment, results from the dynamic DSC study, parameters E_a, A and n obtained by the Kissinger, Flynn–Wall–Ozawa and Crane methods are summarized in Table 2. The activation energy (E_a) values were calculated using the Kissinger method (53.6 and 60.5 kJ mol⁻¹) and the Flynn–Wall–Ozawa method (57.4 and 63.1 kJ mol⁻¹). These are slightly higher than the result of other epoxy resins and phenalkamines systems (around 51–55 kJ mol⁻¹).²⁶ This may be attributed to the steric hindrance based on the long aliphatic side chain in structure of the curing agent, which may reduce the molecular mobility and activity during the curing process.⁴¹ As seen in Table 2, the E_a values obtained from the Kissinger method are lower than that calculated values obtained from the Flynn–Wall–Ozawa method. The different results may be caused by the different assumptions of equation.⁴² In addition, there are a few differences in E_a between ETC/DGEBA and TCP/DGEBA systems. This implies that the curing agents have a different reactivity factor, and reconfirms that ECT shows higher activity toward the epoxy resin. The result may be due to the phenolic hydroxyl groups in the structure of the phenalkamine with a high catalytic efficiency during the curing reaction,⁴³ and the aliphatic hydroxyl group exhibits catalytic effects only at high temperatures. The ETC/DGEBA system has a higher cross-linking density at an earlier curing stage. This would cause the cross-linking density of TCP/DGEBA system to grow faster in the later curing process with aliphatic hydroxyl groups taking part in the reaction. Therefore, hydroxyl groups and amino groups cooperate to affect the curing reaction of epoxy resin.

Table 2 Curing kinetic parameters of curing agent/epoxy resin systems determined by the Kissinger, Flynn–Wall–Ozawa and Crane methods

Curing agent/epoxy resin systems	Kissinger method		Flynn–Wall–Ozawa method	Crane method
	Ea (kJ mol^−1)	A (s^−1)	Ea (kJ mol^−1)	n
ETC/DGEBA	53.6	5.7 × 10^3	57.4	0.887
TCP/DGEBA	60.5	2.7 × 10^4	63.1	0.911

---

3.3. Thermal properties of the cured epoxy resins by ETC and TCP

3.3.1 Determination of glass-transition temperature (T_g). The samples were first heated from 0 °C to 150 °C to eliminate thermal history. The glass-transition temperatures were measured from the second heating curves with DSC (Fig. 9). The cured polymer of ETC/DGEBA shows lower glass-transition temperature (46.5 °C) values than those of the cured polymer of TCP/DGEBA (80.1 °C). The obvious difference of T_gs between ETC/DGEBA and TCP/DGEBA may be attributed to the structure of the curing agents and the cross-linking density of the cured epoxy resins.⁴⁴ Both ETC and TCP having plenty of high-activated hydrogens can react quickly and exothermally with epoxide groups at the early stage of curing process. A large number of the secondary hydroxyl groups are generated in the ring-opening reaction. Then other activated hydrogens of ECT or TCP take place the reaction with them to form a more complicated network. In contrast with ETC, TCP having two aliphatic hydroxyl groups can increase the cross-linking density of the cured epoxy at the late stage. Although phenolic hydroxyl group promotes the active amino reacting with epoxide group at early stage, the cross-linking density of the epoxy resin cured by ECT is lower than that of the polymer cured by TCP. Therefore, the number of aliphatic and phenolic hydroxyl groups in the curing agents can greatly affect the glass-transition temperature of the cured materials. The particular molecular structure of a curing agent has been studied with increases of epoxy functionality due to adjusting the cross-linking density.

Fig. 9 DSC thermograms of the cured epoxy resins by ETC and TCP under N₂ at a heating rate of 10 °C min⁻¹.

3.3.2 Thermal stabilities of the cured polymer. Fig. 10 shows the TGA thermograms of the ETC/DGEBA and TCP/DGEBA systems. The thermal parameters are summarized in Table 3. T_{d, 5%}, and T_max, denote the temperature corresponding to 5% weight loss and the maximum decomposition temperature, respectively. Y_c (wt%) stands for the char yield at 800 °C. Both ETC/DGEBA and TCP/DGEBA systems show the poor thermal stability. It may be attributed to the long aliphatic side chain in structure of the curing agents. This leads to the low cross-linking density of the cured epoxy resins, so the thermodynamics performance is greatly affected by it. Although the materials have the poor thermal stability, they can be applied in the chemical grouting and coating fields due to the high reactivity and the low viscosity. The maximum decomposition temperature of the ETC/DGEBA system (394.7 °C) is close to the temperature of the TCP/DGEBA system (393.6 °C). Compared with the mass loss rate at maximum decomposition temperature of the ETC/DGEBA system, the TCP/DGEBA system has a higher mass loss rate. The degradation of early period may be attributed to the cleavage of the long aliphatic side chain (C₁₅) and carbon–nitrogen bond in the structure of the curing agents. Then the later-stage of degradation was phenalkamine cured epoxy polymer degradation.³⁴ The TG curve of ETC/DGEBA system is very close to that of TCP/DGEBA system, which may attributed to the similar structure of the two cured epoxy resins.⁵

Fig. 10 TGA thermograms of the cured polymers in N₂, 10 °C min⁻¹.

Table 3 Thermal properties parameters for the cured epoxy resins

The cured epoxy resin	Tg (°C)	Td, 5% (°C)	Tmax (°C)	Yc% (800 °C)	Mass loss rate at Tmax (% min^−1)
ETC/DGEBA	46.5	238.5	394.7	6.3	−8.8
TCP/DGEBA	80.1	255.1	393.6	7.0	−10.1

---

3.4. Mechanical properties

The mechanical properties of the cured epoxy resin are of critical importance to industrial applications. The comparison of mechanical properties among the ETC/DGEBA system, the TCP/DGEBA system and the DETA/DGEBA system is summarized in Table 4. As evidenced in Table 4, the toughness of the epoxy resins is improved by the synthesized curing agents. The toughness of cured epoxy resins can be increased with the introduction of a soft segment alkyl side chain to reduce the internal stress of the material.²⁶ Although the ETC/DGEBA system (or the TCP/DGEBA system) has a lower tensile strength and compressive strength, mixing with aliphatic amines in the system is a good approach to obtain materials with excellent comprehensive performance. The ETC/DGEBA system shows a slightly rigid and better compressive property (59.4 MPa) than the TCP/DGEBA system (43.7 MPa). It is interesting to note that both of two cured epoxy polymer samples resumed their original shapes a few seconds after a compressive strength test. Moreover, the samples having been compressed once were used to do a second compressive test. The data (56.8 MPa and 42.2 MPa) collected in the second time were close to the first data (59.4 MPa and 43.7 MPa) gathered. Therefore, the materials have strong elastic under the heavy weight, and can be applicable to repair expansion joint in the fields of bridge and road. The TCP/DGEBA system shows a better impact strength (39.8 J m⁻¹), which is nearly 28% higher than that of the ETC/DGEBA system (31.1 J m⁻¹). This can be attributed to the reduction of internal stress due to the stress relaxation by flexible molecular chains of a curing agent.²⁵ This data suggests that the mechanical properties of the cured epoxy resin can be correlated with the structure of a curing agent.

Table 4 Mechanical properties of the cured epoxy resins by DETA, ETC and TCP

The cured epoxy resin	Tensile strength (MPa)	Tensile strain (%)	Impact strength (J m^−1)	Lap shear strength (MPa)	Compressive strength (MPa)
ETC/DGEBA	42.6 ± 0.8	14.5 ± 0.7	31.1 ± 1.1	6.8 ± 0.7	59.4 ± 1.7
TCP/DGEBA	34.7 ± 1.0	26.8 ± 0.4	39.8 ± 1.5	8.9 ± 0.5	43.7 ± 0.9
DETA/DGEBA	73.2 ± 1.4	10.7 ± 0.3	23.5 ± 0.6	6.2 ± 0.7	83.1 ± 2.3

---

3.5. Fractography of the cured polymers

The corresponding microfractographs are shown in Fig. 11–13. From the photographs it is evident that the durability of a cured epoxy resin is improved when two flexible curing agents are introduced. The SEM photographs of the cured DETA/DGEBA system show a smooth glassy surface with a regular texture revealing typical characteristics of brittle fracture (Fig. 11). They possess rigid and brittle features and are homogeneous with no expectations of a phase separation. The fracture surfaces of ETC and TCP modified epoxy resins are relatively rough (Fig. 12 and 13). The bulge and dent marks of the epoxy matrix and tortuous cracks on the surfaces indicate that energy is absorbed with the impact force during plastic deformation.⁴⁵ The crack propagation path of the rough fracture surface is deficient, which means that more energy would be required to produce more cracks on the surface.

Fig. 11 SEM photographs for fracture surfaces of the DETA/DGEBA system under the magnification of 500× and 1000×.

Fig. 12 SEM photographs for fracture surfaces of the ETC/DGEBA system under the magnification of 500× and 1000×.

Fig. 13 SEM photographs for fracture surfaces of the CTP/DGEBA system under the magnification of 500× and 1000×.

Compared with the ETC/DGEBA system, it can be seen that the fracture surfaces of TCP cured epoxy resins are relatively rough. This may due to the poor solubility of TCP in the epoxy resin. Hence some small white lumps or spots are formed and inlayed into an epoxy matrix. The inlays may be the unreacted TCP (or ETC) and the incompletely cured TCP/DGEBA (or ETC/DGEBA) system, which can absorb more energy due to the less crack propagation path around them. Although both the ETC/DGEBA system and the TCP/DGEBA system have a slightly incompatibility, the cured epoxy resins possess the high impact strength, the high broken elongation and the strong elasticity under the heavy weight.

4. Conclusions

In this paper, two new cardanol-based epoxy curing agents (ETC and TCP) containing thiourea groups were successfully synthesized with a Mannich reaction. New epoxy polymers were prepared by polymerization of ETC (or TCP) with DGEBA. The SEM results demonstrated that the toughness of cured epoxy resin was improved greatly by two newly introduced flexible curing agents compared to DETA. The T_g of the TCP/DGEBA system is higher due to a 2,3-dihydroxy propyl ether substitution of phenol hydroxyl. Both ETC and TCP have a high reactivity in the epoxy resin especially at low temperatures. The cured epoxy resins exhibit greater mechanical properties. We expect that our results will provide insight into the utilization of a new cardanol-based epoxy curing agent containing thiourea groups for architecture and coating applications.

Acknowledgements

The research was funded by Science and Technology Plan Projects of Guangdong Province, China (Grant No. 2015A010105001) and Science and Technology Plan Projects of Guangzhou, China (Grant No. 201510010184).

References

1. O. Kreye, T. Tóth and M. A. R. Meier, J. Am. Chem. Soc., 2011, 133, 1790–1792.
2. A. Van, K. Chiou and H. Ishida, Polymer, 2014, 55, 1443–1451.
3. A. A. Koutinas, A. Chatzifragkou, N. Kopsahelis, S. Papanikolaou and I. K. Kookos, Fuel, 2014, 116, 566–577.
4. C. R. Arza, P. Froimowicz and H. Ishida, RSC Adv., 2015, 5, 97855–97861.
5. K. Huang, Y. Zhang, M. Li, J. Lian, X. Yang and J. Xia, Prog. Org. Coat., 2012, 74, 240–247.
6. K. I. Suresh and V. S. Kishanprasad, Ind. Eng. Chem. Res., 2005, 44, 4504–4512.
7. V. S. Balachandran, S. R. Jadhav, P. K. Vemula and G. John, Chem. Soc. Rev., 2013, 42, 427–438.
8. D.-G. Kim, H. Kang, Y.-S. Choi, S. Han and J.-C. Lee, Polym. Chem., 2013, 4, 5065–5073.
9. M. Sultania, J. S. P. Rai and D. Srivastava, Eur. Polym. J., 2010, 46, 2019–2032.
10. C. Voirin, S. Caillol, N. V. Sadavarte, B. V. Tawade, B. Boutevin and P. P. Wadgaonkar, Polym. Chem., 2014, 5, 3142–3162.
11. E. Papadopoulou and K. Chrissafis, Thermochim. Acta, 2011, 512, 105–109.
12. A. Devi and D. Srivastava, Mater. Sci. Eng., A, 2007, 458, 336–347.
13. R. Yadav and D. Srivastava, Mater. Chem. Phys., 2007, 106, 74–81.
14. E. Calo, A. Maffezzoli, G. Mele, F. Martina, S. E. Mazzetto, A. Tarzia and C. Stifani, Green Chem., 2007, 9, 754–759.
15. L. K. Aggarwal, P. C. Thapliyal and S. R. Karade, Prog. Org. Coat., 2007, 59, 76–80.
16. R. P. Kumar, J. Coat. Technol., 2011, 8, 563–575.
17. M. Sultania, J. S. P. Rai and D. Srivastava, Mater. Chem. Phys., 2012, 132, 180–186.
18. V. S. Balachandran, S. R. Jadhav, P. K. Vemula and G. Joh, Chem. Soc. Rev., 2013, 42, 427–438.
19. G. M. Roudsari, A. K. Mohanty and M. Misra, ACS Sustainable Chem. Eng., 2014, 2, 2111–2116.
20. L. Pan, S. Lu, G. X. Xiao, Z. He, C. Zeng, J. Gao and J. Yu, RSC Adv., 2015, 5, 3177–3186.
21. X. Cheng, Y. X. Chen, Z. L. Du, P. X. Zhu and D. C. Wu, J. Appl. Polym. Sci., 2011, 119, 3504–3510.
22. D. M. Dhevi, S. N. Jaisankar and M. Pathak, Eur. Polym. J., 2013, 49, 3561–3572.
23. J. P. Yang, Z. K. Chen, Q. P. Feng, Y. H. Deng, Y. Liu, Q. Q. Ni and S. Y. Fu, Composites, Part B, 2012, 43, 22–26.
24. X. Yang, W. Huang and Y. Yu, J. Appl. Polym. Sci., 2012, 123, 1913–1921.
25. G. Yang, S. Y. Fu and J. P. Yang, Polymer, 2007, 48, 302–310.
26. Y. Liu, J. Wang and S. Xu, J. Polym. Sci., Part A: Polym. Chem., 2014, 52, 472–480.
27. S. K. Pathak and B. S. Rao, J. Appl. Polym. Sci., 2006, 102, 4741–4748.
28. F. J. V. E. Oliveira, E. C. da Silva Filho, M. A. Melo Jr and C. Airoldi, Surf. Sci., 2009, 603, 2200–2206.
29. K. I. Suresh and V. S. Kishanprasad, Ind. Eng. Chem. Res., 2005, 44, 4504–4512.
30. D. S. Baranov, B. Gold, S. F. Vasilevsky and I. V. Alabugin, J. Org. Chem., 2013, 78, 2074–2082.
31. P. Ramesh, L. Ravikumar and A. Burkanudeen, J. Therm. Anal. Calorim., 2014, 115, 713–722.
32. N. Rekha and S. K. Asha, J. Appl. Polym. Sci., 2008, 109, 2781–2790.
33. B. S. Rao and A. Palanisamy, Eur. Polym. J., 2013, 49, 2365–2376.
34. S. K. Pathak and B. S. Rao, J. Appl. Polym. Sci., 2006, 102, 4741–4748.
35. X. H. Zhang, S. Chen, Y. Q. Min and G. R. Qi, Polymer, 2006, 47, 1785–1795.
36. H. E. Kissinger, Anal. Chem., 1957, 29, 1702–1706.
37. T. Ozawa, Bull. Chem. Soc. Jpn., 1965, 38, 1881–1886.
38. J. H. Flynn and L. A. Wall, J. Polym. Sci., Part C: Polym. Lett., 1966, 4, 323–328.
39. L. W. Crane, P. J. Dynes and D. H. Kaelble, J. Polym. Sci., Polym. Lett. Ed., 1973, 11, 533–540.
40. R. Ambrožič, U. Šebenik and M. Krajnc, Polymer, 2015, 76, 203–212.
41. W. B. Liu, Q. H. Qiu, J. Wang, Z. C. Huo and H. Sun, Polymer, 2008, 49, 4399–4405.
42. Y. Lu, J. Chen, Y. Lu, P. Gai and H. Zhong, J. Appl. Polym. Sci., 2013, 127, 282–288.
43. R. Thomas, S. Durix, C. Sinturel, T. Omonov, S. Goossens, G. Groeninckx, P. Moldenaers and S. Thomas, Polymer, 2007, 48, 1695–1710.
44. Q. Ma, X. Liu, R. Zhang, J. Zhu and Y. Jiang, Green Chem., 2013, 15, 1300–1310.
45. P. Wang, X. Zhang, G. Lim, H. Neo, A. A. Malcolm, Y. Xiang, G. Lu and J. Yang, J. Mater. Sci., 2015, 50, 5978–5992.

---