Graphene oxide as a catalyst for ring opening reactions in amine crosslinking of epoxy resins

Abstract

The influence of different graphite-based nanofillers on epoxide ring opening reactions, as induced by amines for diglycidyl ether of bisphenol A (DGEBA), is studied. Direct kinetic studies, with full chemical characterization and quantitative evaluation of the low molecular mass products, for reactions of DGEBA with primary and secondary monoamines as well with alcohols, are conducted. Moreover, the kinetic behavior of a commercial epoxy resin based on DGEBA and a diamine, leading to crosslinked insoluble networks, is studied by indirect methods, such as differential scanning calorimetry (DSC) and rheometry. The reported results show a relevant catalytic activity of graphene oxide on epoxy resin crosslinking by amines. For instance, for a graphene oxide content of 3 wt%, the exothermic crosslinking DSC peak is shifted (upon heating at 10 °C min⁻¹) from 113 °C down to 96 °C, while the gel time at 50 °C is reduced by a factor of 2.5. This behavior is due to the ability of graphene oxide to catalyze primary amine–epoxy, secondary amine–epoxy and mainly hydroxyl–epoxy additions.

---

1. Introduction

Polymers reinforced with inorganic fillers of dimensions in the nanometer range, known as nanocomposites, have attracted significant interest from researchers, due to the significant enhancement of properties obtained with a very low amount of nano-dispersed filler. Most studied fillers for polymer nanocomposites are clays,^1–5 silica,^6,7 carbon nanotubes, nanographite and graphene nanoplatelets.^8–23

Graphene is generally dispersed in various polymer matrices as reduced graphene oxide. Polymer matrices reinforced with graphene and nano-graphite platelets present higher electrical and thermal conductivity sharing with those reinforced with other layered nanofillers improved strength, modulus, heat distortion temperature and barrier properties.^23–26

In previous works from some of us^25,27,28 a significant effort was directed to preparation of epoxy nanocomposites reinforced with nanographite and to the analysis of their structural, rheological, mechanical, and thermal behaviour. In particular, in ref. 28 comparative thermal analysis of epoxy resins filled with different carbon nanofillers, for different thermal histories, were conducted. The higher T_g values observed at low curing temperatures, for epoxy resins with graphite-based nanofillers, were rationalized by suggesting a catalytic activity of graphene oxide layers on the reaction between the epoxy and amine groups of the resin, leading to higher crosslinking density in milder conditions. This hypothesis, formulated on the basis of the well established catalytic behavior of graphene and graphene oxide for many organic reactions,^29–34 has been supported by experiments on the epoxide ring opening reaction, for few monofunctional epoxide and amine reactants.²⁸ However, analogous data have been rationalized by different hypothesis: restricted molecular mobility of the epoxy matrix by the nanofiller³⁵ or possible crosslinking of graphene oxide layers via epoxy chains.^36,37

Moreover, although additional studies have shown some catalytic activity of graphene oxide layers on curing reactions of different epoxy resins,^38–40 other studies have reported the opposite effect, i.e. the occurrence of retardation effects.^41–43

In this paper, the possible catalytic activity of graphite-based nanofillers on epoxide ring opening reactions, for the bi-functional basic component of most epoxy resins (diglycidyl ether of bisphenol A, DGEBA, Fig. 1), is explored. This study mainly reports kinetics of epoxide ring opening reactions with monofunctional amines, which produce easy to characterize low-molecular-mass products. The possible catalytic behavior is also explored by studying kinetics of crosslinking (cure) of a commercially available DGEBA-based resin, by indirect phenomenological approches,⁴⁴ based on differential scanning calorimetry (DSC) and on rheological evaluation of viscosity increases.

Fig. 1 Chemical formula of diglycidyl ether of bisphenol A (DGEBA), of its oligomers, of isophorondiamine (IPDA), and of the used primary mono-amines (benzylamine and cyclohexylamine) that exhibit amine groups in chemical environments similar to IPDA. Epikote 828 is a mixture of neat DGEBA mainly with its oligomer with n = 1.

2. Experimental

2.1 Materials

Diglycidyl ether of bisphenol A (DGEBA) was supplied by Resolution Performance Products as Epikote 828 and benzylamine, cyclohexylamine, dibenzylamine and benzyl alcohol were supplied by Sigma Aldrich and used without any further purification.

The epoxy resin is characterized by an epoxy equivalent weight (EEW) of 190 g eq.⁻¹ corresponding to an average molecular weight (M_{w Epikote 828}) of 380 g mol⁻¹. Chemical formula of DGEBA and of corresponding oligomers are reported, in the left side of Fig. 1. Since neat DGEBA is characterized by an EEW = 170 g eq.⁻¹, that corresponds to an average molecular weight (M_{w DGEBA}) of 340 g mol⁻¹, Epikote 828 is a mixture of neat DGEBA and of its oligomer with n = 1.

Epikote 828 crosslinking was obtained with isophorondiamine (IPDA) supplied by Aldrich, which is characterized by an Hydrogen Equivalent Weight, HEW = 43. It was selected since it is a typical hardener used for the fabrication of epoxy based composites. According to the stoichiometric ratio, 22 parts of hardener were added to hundred parts of DGEBA (22 phr).

The used primary monoamines (benzylamine and cyclohexylamine) were selected because they exhibit amine groups in chemical environments similar to IPDA. The di-amine IPDA and used primary mono-amines are shown in Fig. 1.

Synthetic Graphite 8427® as trademark with a high surface area (HSAG, of about 308 m² g⁻¹) and a high shape anisotropy of the crystallites,⁴⁵ was purchased from Asbury Graphite Mills Inc. All other standard reagents were bought from Aldrich.

2.2 Synthetic procedures

2.2.1 Preparation of graphite oxide samples. Graphite oxide (GO) samples were prepared by Hummers' oxidation method,⁴⁶ starting from HSAG. 120 mL of sulfuric acid and 2.5 g of sodium nitrate were introduced into a 2000 mL three-neck round bottomed flask immersed into an ice bath and 5 g of graphite were added, under nitrogen, with a magnetic stirring. After obtaining a uniform dispersion of graphite powders, 15 g of potassium permanganate were added very slowly to minimize the risk of explosion. The reaction mixture was thus heated to 35 °C and stirred for 24 h. The resulting dark green slurry was first poured into a copious amount of deionized water, and then centrifuged at 10 000 rpm for 15 min with a Hermle Z 323 K centrifuge. The isolated GO powder was first washed twice with 100 mL of a 5 wt% HCl aqueous solution and subsequently extensively washed with deionized water. Finally, it was dried at 60 °C for 24 h. Afterwards, 3 g of GO powders were first stirred with 250 mL of acetone for 1 h and then sonicated for 30 minutes in an ultrasonic bath at a temperature of 20 °C and power of 80 W.

Exfoliated graphite oxide (eGO) samples were prepared by GO powders introduced in 125 mL ceramic jars (inner diameter of 75 mm) together with stainless steel balls (10 mm in diameter) and dry-milled in a planetary ball mill (Retsch GmbH 5657 Haan) for 2 h with a milling speed of 500 rpm and a ball-to-powder mass ratio of 10 to 1.

2.2.2 Epoxide ring opening reaction with monofunctional amines. The reaction was carried out in a vial. DGEBA (100 mg) and benzylamine (2.4 eq., 82.6 mg, 77 μl) were added to the eGO (3 wt% respect to the reaction mixture, 5.5 mg) and stirred at room temperature. The crude was purified by column chromatography on silica gel in gradient elution with petroleum ether–ethyl acetate to obtain the pure product, which was characterized by ¹H-NMR and ¹³C-NMR, mass spectroscopy and elemental analysis.

All yields reported refer to the isolated products obtained by column chromatography on silica gel in gradient elution with petroleum ether–ethyl acetate. All ¹H NMR and ¹³C NMR spectra were recorded with a DRX 400 MHz Bruker instrument, by using CDCl₃ (d = 7.26 ppm in ¹H NMR spectra and d = 77.0 ppm in ¹³C NMR spectra) as solvent (400.135 MHz for ¹H and 100.03 MHz for ¹³C NMR). Chemical shifts are reported in ppm, multiplicities are indicated by s (singlet), dd (double doublet), bs (broad singlet). Mass spectrometry analysis was carried out using an electrospray spectrometer Waters 4 micro quadrupole. The elemental analyses were performed with FLASH EA 1112 Thermo equipment.

3,3′-(3,3′-(Propane-2,2-diyl)bis(3,1-phenylene))bis(oxy)bis(1-(benzylamino)propan-2-ol) 3. MS: m/z = 555.2 [M + H⁺], 561.9 [M + Li⁺]; ¹H NMR (400 MHz, CDCl₃):δ = 7.32–6.75 (m, 18H), 4.09 (bs, 2H), 3.94–3.77 (m, 8H), 2.96 (–OH, bs), 2.86 (2H, dd), 2.80–2.75 (2H, m), 1.65 (s, 6H) ppm; ¹³C NMR (100 MHz, CDCl₃): δ = 156.9, 143.9, 140.1, 129.6, 128.9, 128.6, 128.2, 127.6, 127.4, 114.3, 70.9, 68.8, 54.2, 51.8, 31.5 ppm. Anal. calcd for C₃₅H₄₂N₂O₄: C: 75.8; H: 7.6; N: 5.1; O: 11.5; found C: 76.5; H: 7.2; N: 5.2; O: 11.1.

2.2.3 Dispersion of graphite-based nanofillers and curing agent in the epoxy resin. The dispersion of the graphite-based filler in the epoxy matrix was obtained by the solvent swelling method, previously developed.^24,27,28 Each filler was dispersed in the polymer, according to the following steps: after the dispersion of each filler in acetone, the filler-solvent mixture was added to the epoxy resin and stirred for 5 h at 80 °C and 400 rpm, until solvent evaporation. The filler-epoxy mixture was subsequently degassed under vacuum and at T = 30 °C. The curing agent (22 phr) was finally added to the mixture filler/resin. In all cases a filler amount equal to 3 wt% was added to the epoxy matrix.

2.3 Characterization techniques

2.3.1 Wide-angle X-ray diffraction. Wide-angle X-ray diffraction (WAXD) patterns were obtained by an automatic Bruker D8 Advance diffractometer, in reflection, at 35 kV and 40 mA, using nickel filtered Cu-Kα radiation (1.5418 Å). The d-spacings were calculated using Bragg's law and the observed integral breadths (β_obs) were determined by a fit with a Lorentzian function of the intensity corrected diffraction patterns. The instrumental broadening (β_inst) was also determined by fitting of Lorentzian function to line profiles of a standard silicon powder 325 mesh (99%). For each observed reflection, the corrected integral breadths were determined by subtracting the instrumental broadening of the closest silicon reflection from the observed integral breadths, β = β_obs − β_inst. The correlation lengths (D) were determined using Scherrer's equation.where λ is the wavelength of the incident X-rays and θ the diffraction angle, assuming the Scherrer constant K = 1.

2.3.2 Differential scanning calorimetry and rheometry. The reactivity of filled and unfilled epoxy resin was measured using a differential scanning calorimeter (DSC Mettler Toledo 622).

Rheological properties of filled and unfilled mixtures were evaluated by monitoring the evolution of the storage modulus G′ and the loss modulus G′′ as function of time. Measurements were made using a strain-controlled rheometer (Ares, Rheometric Scientific) with a parallel plate geometry (12.5 mm radius) at a temperature of 50 °C, with a frequency of 1 Hz and a deformation of 10%. Three runs were made for each sample. The gel time is observed for constant value of the degree of reaction, according to Flory⁴⁷ and hence it represents a fast way to compare the rate of polymerization reactions of thermosetting resins with the same reagents and under different conditions, such as temperature or catalysis.

3. Results and discussion

3.1 X-ray diffraction characterization of the considered graphite-based nanofillers

X-ray-diffraction patterns of the graphite-based fillers used in the present study are reported in Fig. 2. The used high surface area graphite (HSAG with a negligible oxygen content, Fig. 2A) shows an interlayer distance of 0.339 nm, a high shape anisotropy (D_‖/D_⊥ = 3.1).⁴⁵ The X-ray diffraction pattern of the derived GO (with an oxygen content of 32 wt%, excluding water) shows an increase of the interlayer distance from 0.339 nm to 0.84 nm (Fig. 2B). The correlation length perpendicular to the layers (as evaluated from the first 00l reflection) decreases from 9.8 nm to 4.2 nm, while the in-plane correlation length (as evaluated from the 100 reflection) remains almost unchanged (D_‖ ≈ 30 nm), thus leading to an increased shape anisotropy up to D_‖/D_⊥ = 7. The X-ray diffraction pattern of the eGO sample (Fig. 2C), as derived by ball-milling of GO (with an analogous oxygen content: 31.4 wt%, excluding water) shows the complete disappearance of the 001 reflection and maintenance of the 100 reflection, confirming the maintenance of in-plane crystalline order and complete loss of crystalline order perpendicular to the graphitic planes, i.e. a complete GO exfoliation.

Fig. 2 X-ray diffraction patterns (Cu Kα) of the high-surface-area graphite (HSAG, A), of the derived GO (B) and eGO (C).

Additional characterizations of the used nanofillers (elemental analysis, BET surface areas and FTIR spectra) are added as ESI.†

3.2 Nanofillers and DGEBA reactions with primary mono-amines

In a recent paper,²⁸ the catalytic activity of the above described carbon nanofillers toward epoxide ring opening reactions for a monofunctional low-molecular-mass epoxide with different amines has been studied. At room temperature and in the presence of 3 wt% of these nanofillers, substantial yields of the addition products between styrene oxide and benzylamine were achieved while the uncatalyzed reaction does not proceed at all.²⁸

Analogous studies are presently extended to a much more reactive di-epoxide, i.e. DGEBA that is the most common epoxide used in commercial epoxy resins. These studies are conducted in the presence of two different mono-amines (benzylamine and cyclohexylamine), which exhibit amine groups in chemical environments similar to those experienced by the two amine groups of the crosslinking agent used in Epikote 828 (isophorondiamine, IPDA, see Fig. 1). The use of mono-amines rather than di-amines avoids cross-linking, thus making easy characterization of the obtained products. The epoxy ring opening reactions were conducted at room temperature in presence of 3 wt% of carbon nanofiller, in solvent free condition and by using different epoxide/amine ratios. The study of reactions at room temperature rather than in the temperature range 50–180 °C (usual for epoxy resins), is useful to put in evidence the catalytic activity of the used nanofillers.

The considered reactions, both in the absence and presence of nanofillers, are highly regioselective (Fig. 3), due to a SN₂ mechanism with the nucleophile attack to the less substituted end. For molar ratios between epoxide and amine groups close to one, both reactions lead to a single product (3 and 5 in Fig. 3) and the conversion of the di-epoxide is complete already at room temperature, if sufficiently long reaction times are allowed.

Fig. 3 Ring opening reactions of DGEBA and two different monoamines.

Reaction rate is higher with benzylamine and for both amines it is faster in the presence of eGO. In particular, after 7 hours at room temperature, from benzylamine 2, product 3 was obtained with yields of 70% and 30%, for eGO catalyzed and uncatalyzed reactions, respectively. Moreover, after 24 hours at room temperature, from cyclohexylamine 4, product 5 was obtained with yields of 60% and 15%, for eGO catalyzed and uncatalyzed reactions, respectively.

The influence of all the considered graphitic nanomaterials (HSAG, GO and eGO) on reaction rate from 2 to 3, for epoxide/amine molar ratio of 1.0/1.2 and at room temperature, is shown in Fig. 4, as yield of product 3, versus reaction time. All nanofillers increase reaction rates, with maximum increases for graphene oxide (eGO) due to the further activation exerted by its acidic functionalities. In fact, in presence of HSAG the epoxide activation derives only from π-stacking interactions between DGEBA and graphitic surface, as already suggested, also on the basis of DFT calculations, for other reactions.^33,34 In the presence of GO, instead, due to its OH and COOH groups, both mechanism, π-stacking activation and acidic activation could be involved, increasing the rate of the reaction. Moreover, an induction time longer than 1 hour occurs only in the absence of nanofillers. In fact, the typical autocatalytic behavior of epoxy–amine reactions, characterized by a maximum yield rate at time > 0,⁴⁸ is clearly observed only with neat resin. This indicates a catalytic activity of all nanofillers on the epoxide ring opening reaction, which eliminates the induction time and makes irrelevant the formation of OH groups for the reaction beginning.

Fig. 4 Yield for the product 3 of the ring opening reactions of DGEBA with benzylamine, for reactions conducted at room temperature in solvent free conditions, by using epoxide/amine ratio 1/1.2, in the presence of 3 wt% of different carbon nanofillers: none (black), HSAG (magenta), GO (blue) and eGO (green).

To establish if the used nanocarbons can be considered real catalysts, the reusability of both eGO and GO was investigated by repeated tests for the reaction leading to product 3. The nanofillers, recovered after extraction from the aqueous solution and dried at 60 °C overnight, were used without any further treatment. The reaction conditions, room temperature and 3 h, were kept constant for all cycles. Yields, for four consecutive reaction cycles, in the presence of GO and eGO are shown in Fig. 5.

Fig. 5 Yields of consecutive cycles of reaction from 1 and 2 to 3, in the presence of 3 wt% of eGO (red) and GO (blue). The yield in the same conditions, in the absence of nanofiller is close to 10%. Error bars are added for runs 1.

In the presence of eGO, the yield is maximum for the first run and markedly decreases to a nearly constant value, starting from the second run. In the presence of GO, on the contrary, the yield is maximum for the second run and starting from the 4^th run reaches values close to those observed for eGO.

These data can be easily rationalized on the basis of the structural reorganization of the GO and eGO, after the first reaction cycle and removal from the reaction system. In fact, both for the ordered graphite oxide (Fig. 6A) and the disordered graphene oxide (Fig. 6B), relevant changes of the WAXD patterns are observed, which lead to rather similar patterns for both nanofillers (Fig. 6A′ and B′).

Fig. 6 X-ray diffraction patterns (Cu Kα) of GO (A and A′) and eGO (B and B′), before (A and B) and after (A′ and B′) the first reaction cycle and removal from the reaction system.

In particular, the pattern of Fig. 6A′ shows that the 001 graphite oxide peak at d = 0.835 nm is completely lost and replaced by narrower peaks at d = 3.4 nm and d = 1.7 nm. These peaks can be rationalized as 001 and 002 reflections of an intercalate compound of graphite oxide,^49–52 with the used amine (benzylamine). The pattern of Fig. 6B′ shows that the eGO sample, although initially fully exfoliated, after removal from the reaction system partially reorganizes leading to a similar intercalate compound with slightly increased interlayer spacing d = 4.2 nm (only the d₀₀₂ = 2.1 nm is clearly apparent).

The yield data for consecutive reaction runs of Fig. 5 clearly confirm that the influence of carbon nanofillers on epoxide ring opening reaction corresponds to a real catalytic action. However, the yield reduction (after the first run for eGO and after the second run for GO), can be attributed to a partial loss of functional groups, mainly due to amine reaction with epoxy groups on the GO layers.⁵³ This conclusion is mainly supported by the formation of GO/benzylamine intercalate structures, as shown by the WAXD patterns of Fig. 6A′ and B′.

3.3 Epoxy resin crosslinking

A classical method to evaluate the catalytic activity of nanofillers in crosslinking of thermosetting materials is based on DSC heating scans. In particular, DSC scans at heating rate of 10 °C min⁻¹, performed on the neat epoxy resin and on resins filled with 3 wt% of the three considered carbon nanofillers are compared in Fig. 7. The exothermic peaks present similar shapes and areas (ΔH in the range 310–380 J g⁻¹) but their maxima are shifted when nanofillers are added, i.e. the peak temperature is lowered of 1 °C, 9 °C and 17 °C in presence of HSAG, GO and eGO, respectively (see Table 1). The observed shifts of the exothermic peaks indicate a relevant catalytic activity of graphite oxide (mainly of eGO) and a nearly negligible catalytic activity of HSAG.

Fig. 7 DSC scans (heating rate = 10 °C min⁻¹, exotherms downward) for the pure epoxy resin (black) and for composite resins wit 3 wt% of: HSAG (purple); GO (blue) and eGO (green).

Table 1 Delta H values and peak temperature (T_p) of DSC scans (heating rate = 10 °C min⁻¹) for the pure epoxy resin and for composite resins with 3 wt% of: HSAG; GO and eGO

Sample	Delta H [J g^−1]	T Peak [°C]
Epoxy	313 ± 35	113
Epoxy + HSAG	339 ± 35	112
Epoxy + GO	373 ± 35	104
Epoxy + eGO	375 ± 35	96

---

The cure reaction was also studied by isothermal rheological measurements, at 50 °C. The ratio between the modulus of complex viscosity |η*| and the complex viscosity measured at t = 0 of the neat resin and of the resin filled with 3 wt% with the three considered nanofillers, are reported in Fig. 8 as a function of time.

Fig. 8 Modulus of the complex viscosity |η*|/|η^*₀| (determined by forced harmonic oscillation measurements) versus time for the DGEBA–IPDA epoxy resin, pure (black) or with 3 wt% of different carbon nanofillers: HGSA (red); GO (blue); eGO (green).

The dynamic-mechanical properties of a curing system depend on the degree of reaction, dramatically changing when the gel point is approached. The time to gelation or gel time (t_g) can be determined according to one of the different criteria, as reviewed by Laza et al.⁵⁴ In this paper we have chosen as gel times those corresponding to |η*|/|η^*₀| = 1000. The measured values of gel time and the corresponding complex viscosity measured at t = 0 (|η^*₀|), for the neat resin and the resin filled with 3 wt% with the three considered nanofillers, have been collected in Table 2. The gel time of the neat resin (t_gel = 2776 s) is only slightly reduced in the presence of HSAG (t_gel = 2560 s) while it is strongly reduced in the presence of graphite oxide (t_g = 1650 s) and graphene oxide (t_gel = 1550 s). A similar trend is obtained adopting any of the others criteria described in ref. 55. This result can be again rationalized by a relevant catalytic activity of graphite oxide (mainly of eGO) and a nearly negligible catalytic activity of HSAG toward the whole process of epoxy resin crosslinking.

Table 2 Gel time (t_gel) at 50 °C for the neat resin and the resin filled with 3 wt% of the three considered nanofillers. The last column reports the corresponding complex viscosity, as measured at t = 0 (|η^*₀|)

Sample	tgel (|η*|/|η^*0| = 1000) [s]	|η^*0| [Pas]
Epoxy	2776	0.136
Epoxy_HGSA	2560	0.239
Epoxy_GO	1650	0.596
Epoxy_eGO	1550	0.564

---

3.4 Nanofillers and DGEBA reactions with secondary mono-amines and alcohols

Our data relative to reactions of DGEBA with monoamines (Fig. 3–5) indicate that epoxy ring opening reactions by primary amines are catalyzed both by high-surface-area graphite and graphene oxide. Our data on crosslinking of DGEBA with a diamine, based on temperature of exothermic peaks in DSC heating scans (Fig. 7) as well as on gelation times (Fig. 8), confirm a relevant catalytic activity of nanofillers based on graphene oxide but show a nearly negligible catalytic activity of graphite based nanofillers.

These apparently conflicting results can be rationalized by considering beside primary amine–epoxy additions also the other two major reactions involved in amine curing of epoxy resins: secondary amine–epoxy and hydroxyl–epoxy additions (as schematically shown in Fig. 9).^43,55 The latter reaction, of course, becomes significant only after an adequate increase of OH group concentration, as a result the first two reactions.

Fig. 9 Three major reactions during amine curing of epoxy resins.

The model reactions of DGEBA with monoamines (Fig. 3–5) have demonstrated catalytic activity of all the considered nanofillers on reaction 1 of Fig. 9. Additional model reactions are conducted to evaluate the possible influence of the considered graphitic nanofillers on secondary amine–epoxy addition (reaction 2 in Fig. 9) and on hydroxyl–epoxy additions (reaction 3 in Fig. 9). In particular, we have compared reactions of DGEBA with a secondary monoamine (dibenzylamine, Fig. 10) and with an alcohol (benzyl alcohol, Fig. 11), in the presence of 3 wt% of carbon nanofillers in solvent free condition and by using molar ratios between reacting groups close to one.

Fig. 10 Ring opening reactions of DGEBA with secondary monoamine.

Fig. 11 Ring opening reactions of DGEBA with benzyl alcohol.

The obtained results for the secondary amine–epoxy addition of Fig. 10, are rather similar to those obtained for primary amine–epoxy addition of Fig. 3, showing a similar catalytic activity of graphene oxide and graphite on the ring opening reaction. For instance, after 19 h, in the presence of eGO as well as of HSAG the product 7 is obtained in a 43% yield while in absence of nanofillers the reaction only proceeds up to 20% yield.

Completely different are the results obtained for the model reaction for the hydroxyl–epoxy addition (Fig. 11), showing a relevant influence of graphene oxide but a negligible influence of graphite on the ring opening reaction. In particular, reactions conducted in the presence of graphene oxide after 24 h lead to 20% and 50% of overall yield of 9 and 10, when conducted at 90 °C and 110 °C, respectively. For the same reactions, if conducted without nanofillers or in the presence of HSAG, no traces of products are detected. This result can be explained based on the reduced nucleophilicity of alcohol with respect to the amine group, that needs a more activated electrophile to provide the desired product. In fact, only in the presence of acidic activation the etherification reaction is catalyzed.

In summary, according to the results of ring-opening model reactions of present section, graphene oxide is able to catalyze all three major reactions that are involved in amine curing of epoxy resins (Fig. 9), while graphite-based nanofillers are only able to catalyze amine–epoxy additions (reactions 1 and 2, in Fig. 9). Because only nanofillers based on graphene oxide (but not those based on graphite) are able to catalyze crosslinking of epoxy resins (Section 3.3), the present results indicate, in agreement with a literature suggestion,⁵⁵ that the hydroxyl–epoxy addition (etherification reaction) is the rate-determining step of epoxy resin crosslinking.

4. Conclusions

The influence of different graphite-based nanofillers (high-surface-area graphite, graphite oxide and graphene oxide) on epoxide ring opening reactions as induced by different amines on diglycidyl ether of bisphenol A (DGEBA, the most used component of commercial epoxy resins), has been studied.

Direct kinetic studies of epoxide ring opening reaction for DGEBA in the presence of different mono-amines (benzylamine, cyclohexylamine and dibenzylamine), with full chemical characterization and quantitative evaluation of low molecular mass products, have been conducted. The reported results indicate a strong influence of carbon nanofillers, mainly of graphene oxide but also of graphite oxide and high-surface-area graphite, on epoxide ring opening reactions by amines. For instance, for the reaction with benzylamine for the epoxy/amine group molar ratio of 1.0/1.2, at room temperature, the addition of 3 wt% of nanofillers leads to elimination of the reaction induction time, which is instead longer than 1 hour in the absence of the nanofillers.

The occurrence of a real catalytic activity of these carbon nanofillers on amine–epoxy additions has been clearly proved by repeated tests for the reaction between the di-epoxide and a mono-amine (benzylamine), as conducted in the presence of recovered nanofillers. After few runs the yields become similar for graphite oxide and graphene oxide. This phenomenon is mainly due to the exfoliation of graphite oxide in the reaction conditions, already occurring during the first reaction run.

Kinetics of crosslinking of a commercial epoxide resin with a diamine have confirmed a relevant catalytic activity of graphite oxide (mainly of graphene oxide) while have shown a negligible catalytic activity of graphite. In particular, for a nanofillers content of 3 wt%, the exothermic DSC peak corresponding to the crosslinking reaction (at a heating rate of 10 °C min⁻¹) is shifted of 17 °C and 1 °C for eGO and HSAG, respectively. Analogously, gel time, as evaluated by rheological data at 50 °C in the presence of 3 wt% of nanofillers, is reduced of nearly 45% in the presence of eGO while it is reduced of less than 10% for high-surface-area graphite.

These results have been rationalized by showing, for model reactions of DGEBA with an alcohol, that only graphene oxide (but not graphite) is able to catalyze hydroxyl–epoxy additions (etherification reactions). The latter results also suggest that etherification reaction is the rate-determining step in amine curing of epoxy resins.

In summary, nanofillers based on graphene oxide reduce temperature and/or time of crosslinking of epoxy resins by amines, by catalyzing primary amine–epoxy, secondary amine–epoxy and mainly hydroxyl–epoxy additions.

Acknowledgements

We thank Dr Luca Giannini and Dr Marco Mauro for useful discussions. Financial support by Pirelli Tyre Research Center and “Ministero dell' Istruzione, dell' Università e della Ricerca” is gratefully acknowledged.

References

1. J. Park and S. C. Jana, Polymer, 2004, 45, 7673.
2. E. M. Arndt, M. D. Gawryla and D. A. Schiraldi, J. Mater. Chem., 2007, 17, 3525.
3. A. Greco, C. Esposito Corcione, A. Strafella and A. Maffezzoli, J. Appl. Polym. Sci., 2010, 118, 3666.
4. C. Esposito Corcione, A. Cavallo, E. Pesce, A. Greco and A. Maffezzoli, Polym. Eng. Sci., 2011, 51, 1280.
5. S. S. Rahatekar, M. Zammarano, S. Matko, K. K. Koziol, A. H. Windle, M. Nyden, T. Kashiwagia and J. W. Gilman, Polym. Degrad. Stab., 2010, 95, 870.
6. F. Lionetto, E. Calò, F. Di Benedetto, D. Pisignano and A. Maffezzoli, Compos. Sci. Technol., 2014, 96, 47.
7. A. Greco, F. Lionetto and A. Maffezzoli, Polym. Compos., 2015, 36, 1096.
8. A. Terenzi, C. Vedova, G. Lelli, J. Mijovic, L. Torre, L. Valentini and J. M. Kenny, Compos. Sci. Technol., 2008, 68, 1862–1868.
9. F. Pan, F. Peng and Z. Jiang, Chem. Eng. Sci., 2007, 62, 703.
10. A. Yu, P. Ramesh, M. E. Itkis, E. Bekyarova and R. C. Haddon, J. Phys. Chem. C, 2007, 111, 7565.
11. J. Li, L. Vaisman, G. Marom and J. K. Kim, Carbon, 2007, 45, 744–750.
12. S. Steiner and P. Ermanni, Carbon, 2012, 50, 1713.
13. S. Gangulia, A. K. Roya and D. P. Anderson, Carbon, 2008, 46, 806.
14. A. Yu, P. Ramesh, X. Sun, E. Bekyarova, M. E. Itkis and R. C. Haddon, Adv. Mater., 2008, 20, 4740.
15. N. Jovic, D. Dudic, A. Montone, M. Vittori Antisari, M. Mitric and V. Djoković, Scr. Mater., 2008, 58, 846.
16. L. Monica Veca, M. J. Meziani, W. Wang, X. Wang, F. Lu, P. Zhang, Y. Lin, F. Robert, J. W. Connell and Y.-P. Sun, Adv. Mater., 2009, 21, 2088.
17. J. Liang, Y. Wang, Y. Huang, Y. Ma, Z. Liu, J. Cai, C. Zhang, H. Gao and Y. Chen, Carbon, 2009, 47, 922.
18. I. Zaman, T. T. Phan, H.-C. Kuan, Q. Meng, L. T. B. La, L. Luong, O. Youssf and J. Ma, Polymer, 2011, 52, 1603.
19. M. Zhang, L. Yang, S. Zhiqiang and W. Gang, Polym. Chem., 2015, 6, 6107.
20. K. M. F. Shahil and A. A. Balandin, Nano Lett., 2012, 12, 861.
21. M. Abdalla, D. Dean, P. Robinson and E. Nyairo, Polymer, 2008, 49, 3310.
22. S. Longo, M. Mauro, C. Daniel, P. Musto and G. Guerra, Carbon, 2014, 77, 896.
23. L. Guadagno, M. Raimondo, L. Vertuccio, M. Mauro, G. Guerra, K. Lafdi, B. De Vivo, P. Lamberti, G. Spinelli and V. Tucci, RSC Adv., 2015, 5, 36969.
24. C. Esposito Corcione and A. Maffezzoli, Thermochim. Acta, 2009, 485, 43.
25. C. Esposito Corcione, F. Freuli and A. Maffezzoli, Polym. Eng. Sci., 2013, 53, 531.
26. A. Yasmin and I. M. Daniel, Polymer, 2004, 45, 8211.
27. C. Esposito Corcione and A. Maffezzoli, Polym. Test., 2013, 32, 880.
28. M. Mauro, M. R. Acocella, C. Esposito Corcione, A. Maffezzoli and G. Guerra, Polymer, 2014, 55, 5612.
29. M. H. Sarvari and H. Slarghi, Helv. Chim. Acta, 2005, 88, 2282.
30. D. R. Dreyer, K. A. Jarvis, P. J. Ferreira and C. W. Bielawski, Macromolecules, 2011, 44, 7659.
31. S. Verma, H. P. Mungse, N. Kumar, S. Choudhary, S. L. Jain, B. Sain and O. P. Khatri, Chem. Commun., 2011, 47, 12673.
32. D. R. Dreyer and C. W. Bielawski, Adv. Funct. Mater., 2012, 12, 247.
33. M. R. Acocella, M. Mauro, L. Falivene, L. Cavallo and G. Guerra, ACS Catal., 2014, 4, 92–96.
34. M. R. Acocella, M. Mauro and G. Guerra, ChemSusChem, 2014, 7, 3279.
35. X. J. Shen, X. Q. Pei, S. Y. Fu and K. Friedrich, Polymer, 2013, 54, 1234.
36. B. K. Ahn, J. Sung, Y. Li, N. Kim, M. Ikenberry, K. Hohn, N. Mohanty, P. Nguyen, T. S. Sreeprasad, S. Kraft, V. Berry and X. S. Sun, Adv. Mater., 2012, 24, 2123.
37. B. Shen, W. Zhai, M. Tao, D. Lu and W. Zheng, Compos. Sci. Technol., 2013, 77, 87.
38. D. R. Bortz, E. G. Heras and I. Martin-Gullon, Macromolecules, 2011, 45, 238.
39. J. K. Park and D. S. Kim, Polym. Eng. Sci., 2014, 54, 969.
40. S. L. Qiu, C. S. Wang, Y. T. Wang, C. G. Liu, X. Y. Chen, H. F. Xie, Y. A. Huang and R. S. Cheng, eXPRESS Polym. Lett., 2011, 5, 809.
41. S. H. Ryu, J. H. Sin and A. M. Shanmugharaj, Eur. Polym. J., 2014, 52, 88.
42. C.-C. Teng, C.-C. M. Ma, C.-H. Lu, S.-Y. Yang, S.-H. Lee, M.-C. Hsiao, M.-Y. Yen, K. C. Chiou and T.-M. Lee, Carbon, 2011, 49, 5107.
43. G. D. Dilini, J. Galpaya, F. S. Fernando, L. Rintoul, N. Motta and E. R. Waclawik, Polymer, 2015, 71, 122.
44. M. Opalicki, J. M. Kenny and L. Nicolais, J. Appl. Polym. Sci., 1996, 61, 1025.
45. M. Mauro, V. Cipolletti, M. Galimberti, P. Longo and G. Guerra, J. Phys. Chem. C, 2012, 116, 24809.
46. S. W. R. Hummers and E. Offeman, J. Am. Chem. Soc., 1958, 80, 1339.
47. P. J. Flory, in Principles of polymer chemistry, Cornell University Press, Ithaca, NY, USA, 1953.
48. J. M. Kenny, J. Appl. Polym. Sci., 1994, 51, 761.
49. Y. Matsuo, T. Niwa and Y. Sugie, Carbon, 1999, 37, 897.
50. Y. Matsuo, T. Miyabe, T. Fukutsuka and Y. Sugie, Carbon, 2007, 45, 1005.
51. M. Mauro, M. Maggio, V. Cipolletti, M. Galimberti, P. Longo and G. Guerra, Carbon, 2013, 61, 395.
52. M. Mauro, M. Maggio, A. Antonelli, M. R. Acocella and G. Guerra, Chem. Mater., 2015, 27, 1590.
53. B. Xue, J. Zhu, N. Liu and Y. Li, Catal. Commun., 2015, 64, 105.
54. J. M. Laza, C. A. Julian, E. Larrauri, M. Rodriguez and L. M. Leon, Polymer, 1998, 40, 35–45.
55. N. A. St John and G. A. George, Polymer, 1992, 33, 2679.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: 10.1039/c6ra00485g

---