In spite of extensive studies conducted on carbon nanotubes and silicate layers for their polymer-based nanocomposites, the rise of graphene now provides a more promising candidate due to its exceptionally high mechanical performance and electrical and thermal conductivities. The present study developed a facile approach to fabricate epoxy–graphene nanocomposites by thermally expanding a commercial product followed by ultrasonication and solution-compounding with epoxy, and investigated their morphologies, mechanical properties, electrical conductivity and thermal mechanical behaviour. Graphene platelets (GnPs) of 3.57 ± 0.50 nm in thickness were created after the expanded product was dispersed in tetrahydrofuran using 60 min ultrasonication. Since epoxy resins cured by various hardeners are widely used in industries, we chose two common hardeners: polyoxypropylene (J230) and 4,4′-diaminodiphenylsulfone (DDS). DDS-cured nanocomposites showed a better dispersion and exfoliation of GnPs, a higher improvement (573%) in fracture energy release rate and a lower percolation threshold (0.612 vol%) for electrical conductivity, because DDS contains benzene groups which create π–π interactions with GnPs promoting a higher degree of dispersion and exfoliation of GnPs during curing. This research pointed out a potential trend where GnPs would replace carbon nanotubes and silicate layers for many applications of polymer nanocomposites.

The interface is well known for taking a critical role in the determination of the functional and mechanical properties of polymer composites. Previous interface research has focused on utilising reduced graphene oxide that is limited by a low structural integrity, which means a high fraction is needed to produce electrically conductive composites. By using 4,4′-diaminophenylsulfone, we in this study chemically modified high-structural integrity graphene platelets (GnPs) of 2–4 nm in thickness, covalently bonded GnPs with an epoxy matrix, and investigated the morphology and functional and mechanical performance of these composites. This covalently bonded interface prevented GnPs stacking in the matrix. In comparison with unmodified composites showing no reduction in electrical volume resistivity, the interface-modified composite at 0.489 vol% GnPs demonstrates an eight-order reduction in the resistivity, a 47.7% further improvement in modulus and 84.6% in fracture energy release rate. Comparison of GnPs with clay and multi-walled carbon nanotubes shows that our GnPs are more advantageous in terms of performance and cost. This study provides a novel method for developing interface-tuned polymer/graphene composites.