The 3M Scotch-WeldTM epoxy adhesive 2216 B/A, used in this study, is a flexible, two-part, room temperature curing epoxy with high peel and shear strength supplied by 3M Company (Saint-Paul, Minnesota, USA). The graphene nanoplatelets (GNPs) were provided by XG Sciences. Inc (Lansing, Michigan, USA). GNPs consist of short stacks of graphene sheets having a platelet shape with an average thickness of approximately 15 nanometers and a typical diameter of 5 microns, while their surface area was about 50 to 80 m²/g. Multi-walled carbon nanotubes (MWCNTs) are tubular structures that produced by catalyzed chemical vapor deposition (CVD) and were supplied by ARKEMA (Colombe, France). Their diameters were 10 - 15 nm, their typical length was more than 500 nm long and their surface area was approximately 230 m²/g. Finally, a unidirectional carbon prepreg (SIGRAPREG CU150-0/NF-E340/38%) supplied by SGL CARBON (Wiesbaden, Germany), was used for manufacturing single lap joints for shear testing.

The dispersion of nanoparticles in the adhesive material was carried out using three-roll milling technique, which is described analytically by Kostagiannakopoulou, C. et al. [32]. The preparation of nano-reinforced blends took place in a glove box for safety reasons, where the nanoparticles and Part A of the adhesive were mixed by hand stirring for 2 min. Then, the nanomodified mixtures were fed in the calandar machine, as represented in Figure 1, following specific protocol regarding the milling process. In particular, three cycles of milling were repeated for each gap setting during the processing of nano-blends. Finally, the speed of the apron roll was maintained at 270 rpm throughout the duration of milling procedure. After the milling process, part B of the adhesive was added to the produced blend and the final mixture was poured in silicon rubber molds and cured in an autoclave oven, according to manufacturer recommended curing cycle, at 93˚C for 1 h and 6 bars pressure. These bulk materials developed for the measurement of thermal and electrical conductivity. Following this process three different nanocomposites resulted: 1) with 8%wt. GNPs, 2) with 2%wt. MWCNTs and 3) with the combination of 1%wt. MWCNTs and 8%wt. GNPs. Finally, neat epoxy adhesive samples were also produced following the above process to have the reference values in order to investigate the influence of nanoparticles on the material properties.

Moreover, for the single lap shear tests, coupons were manufactured according to ASTM D5868-01. Both the adherends were taken from UD ([0]₁₆) carbon fiber composite plates, which were cured according to the manufacturer’s instructions in an autoclave oven (6 bar pressure, 2 hours at 120˚C). Coupons with dimensions of 100 mm length by 25 mm width were cut from the CFRP plates using a water-cooled diamond coated disc cutter. After cutting, the samples were properly dried. The areas, where the adhesive was to be applied, were prepared by wet abrasion (SiC paper, P600) and cleaned thoroughly using acetone. Immediately after preparation, adhesive was applied onto the one of the adherends, in such a way that a 25 mm by 25 mm joint to be achieved. Using appropriate shims, the second adherend was placed on top of the first, and alignment was ensured. The system was then placed in a conventional oven for curing as per the adhesive manufacturer’s instructions (93˚C for 1h). Reference specimens were produced this way, using as-received adhesive, as well as the nano-modified, using the developed formulas containing GNPs, MWCNTs and the hybrid system.

Firstly, for the electrical characterization of the developed materials DC electrical conductivity measurements were performed. A digital multimeter (Keithley 2002) was used for the through thickness measurements of specific electrical resistance of samples (Figure 2(a)). Four specimens were measured for each material type. The dimensions of the measured samples were (25 mm × 25 mm × 5 mm).

Then, measurements of thermal conductivity of the developed materials were per-formed by using a thermal conductivity analyzer (TCi Mathis), as depicted in Figure 2(b). TCi Mathis Analyzer measures directly and rapidly the through thickness thermal conductivity of sample at room temperature, providing a detailed overview of its thermal characteristics. Five specimens were measured for each tested material type. The dimensions of the measured samples were (25 mm × 25 mm × 5 mm).

For the single lap shear tests, a hydraulic universal testing machine (Instron 8802) was used (Figure 2(c)). Single lap shear tests were carried out according to ASTM D 5868-01 standard. The crosshead speed was set at 13 mm/min, while five specimens were tested for each produced material. Finally, the fracture surfaces of the tested coupons were carefully examined under a LEO SUPRA 35VP Scanning Electron Microscopy (SEM).

The measured electrical conductivity of the reference material and the developed formulations are shown in Figure 3. An electrical conductivity of about 10⁻¹² S/m was measured in case of unmodified epoxy adhesive, which is expected due to its typical insulating behavior. The electrical conductivity of the epoxy adhesive is increased by several orders of magnitude, when the conductive nano-fillers are introduced into the reference material. More specifically, the last column of Table 1 reports the increases, in order of magnitude, caused in electrical conductivity of

the adhesive by the addition of carbon nanoparticles. As it can be seen, the highest increase, 9 orders of magnitude, was observed in the case of nano-modified adhesive doped with 2%wt. MWCNTs. The addition of 8%wt. GNPs into the epoxy adhesive improved its electrical conductivity by 6 orders of magnitude, while the simultaneous integration of MWCNTs and GNPs did not lead to further improvement of the property.

Considering the results, MWCNTs proved more efficient nano-fillers for the enhancement of the electrical conductivity of adhesives in comparison with GNPs. According to literature, MWCNTs cause significant improvements in electrical conductivity of epoxy resins (polymer and adhesives) converting them from insulators to conductors or semi-contactors by achieving low percolation thresholds [20] [21] [22] [32] [33]. It is well known that one of the main factors that considerably affects the electrical properties of the nano-reinforced materials is the filler geometry, [33] - [39]. The superiority of MWCNTs relative to graphene nano-filler for the improvement of electrical conductivity is due to the 2D nature of GNPs that may cause greater scattering to the electron transfer compared to the 1D-geometry MWCNTs. The 2D geometry leads to lesser mean free path of electrons into the material [39], and thus to the inability of forming the appropriate conductive channels. So, enhancement of electrical conductivity is limited in the case of the graphene-based materials in relation to MWCNTs.

At this point it is worth mentioning that the hybrid material containing both MWCNTs and GNPs did not cause further improvement in electrical conductivity than the corresponding materials doped with the individual nanoparticles. It is obvious, that the improvement that was achieved in hybrid system was mainly due to the presence of multi-walled carbon nanotubes into the epoxy adhesive. This fact proved that no synergetic effect between the nano-fillers for further increase of electrical properties of the adhesive was observed.

Figure 4 depicts thermal conductivity results of base and nano-reinforced adhesives for each filler type. Table 2 also summarizes the resulted average values of thermal conductivity measurements and their standard deviation. Furthermore, Table 2 presents the percentage increase that was achieved by adding carbon nanofillers to the adhesive relative to the base material. It is obvious that the thermal conductivity of adhesive increased with the incorporation of carbon nanofillers. By adding a reinforcement content of 8%wt. GNPs, the thermal conductivity enhanced by 207%. The hybrid material resulted to an almost similar enhancement and increased the thermal conductivity by 203%, while the addition of MWCNTs caused the lowest improvement of 17%.

As already mentioned, the major enhancement of the thermal conductivity was succeeded by GNPs. According to literature [14] - [22], multilayer graphene is a very promising nanomaterial for the development of thermally conductive polymers for two reasons. Firstly, the 2D geometry of graphene-based materials

provides a 2D path, which leads to a more efficient phonon transmission between adjacent nanoparticles compared to the corresponding 1D path formed by carbon nanotubes [40]. Then, the geometry of GNPs in contrast to MWCNTs allows the development of high content nano-modified mixtures without introducing limitations during the manufacturing process due to the increased viscosity. That is a very important fact, considering that fillers contents higher than 5%wt. are needed to successfully increase the property of thermal conductivity. At lower concentrations the distance between fillers increases and leads to an increase in the thickness layer of the insulating polymer, forming between them preventing the direct transmission of phonons.

Finally, the hybrid material did not show any further increase in thermal conductivity, which indicates that there was no cooperation between MWCNTs and GNPs to facilitate the phonon transmission inside the material. Obviously, the reported enhancement in the case of hybrid system is only due to the presence of GNPs into the epoxy adhesive.

The single lap shear strength of pure and nano-reinforced epoxy adhesives is presented in Figure 5. In this figure and in the Table 3, it is observed that the incorporation of nanofiller into the epoxy adhesive decreased the single lap shear strength of the reference material. The most significant reduction is observed with the addition of 2%wt. MWCNTs and this was at the level of 74%. The incorporation of GNPs caused a decrease of 17%, while the hybrid system presented the lowest decrease (10%). The quite high decrease of the single lap shear strength that occurred in nano-reinforced adhesives with MWCNTs resulted due to the high viscosity of the prepared adhesive that introduced difficulties in handling and applying the adhesive material. Obviously, the reduction of 74% proves that the addition of high contents of MWCNTs into the adhesives carries the risk of losing the adhesive’s operational properties.

In the following figures, SEM micrographs of the fracture surfaces of nano-reinforced polymers are presented. More specifically, representative SEM micrographs of nano-doped materials filled with 2%wt. MWCNTs, 8%wt. GNPs and the combination of 1%wt. MWCNTs & 8%wt. GNPs are shown in Figures 6-8, respectively. In Figure 6(a), the dispersion of multi-walled carbon nanotubes

into the epoxy adhesive is shown. It is obvious that there is insufficient homogeneity in the dispersion of MWCNTs into the epoxy adhesive. The presence of agglomerates is evident in Figure 6(b) at higher magnification. Measurements of the size of representative agglomerates are presented in Figure 6(c), and the average size of the agglomerates is 2.065 μm. This intense agglomeration into the epoxy adhesive in case of MWCNTs possibly led to high viscosity, which in turns caused difficulties in handling of the material. Also, the presence of agglomeration is responsible for the strong decrease (74%) of the single lap shear strength after the integration of MWCNTs into the epoxy adhesive.

Next, it can be seen in Figure 7(a) that the introduction of GNPs into the epoxy adhesive showed greater homogeneity in comparison with that of MWCNTs. More specifically, in Figure 7(b) the high content of GNPs into the epoxy adhesive, in combination with their 2D geometry, led to the successful formation of appropriate conductive paths (green arrows) for the phonon transmission, which is responsible for the increase of the thermal conductivity of the reference material.

In this study, the simultaneous introduction of GNPs and MWCNTs into the epoxy adhesive did not report much different values in thermal and electrical conductivity than the corresponding values resulted from the addition of the individual nanoparticles. The combined use of MWCNTs and GNPs increases slightly (1 order of magnitude) the electrical conductivity compared to the pure GNPs, while reduces marginally the thermal conductivity compared again to the pure GNPs. However, the combined use of the specific carbon based nanofillers keeps very high values of single lap shear strength values. This fact proved that although there is no synergy between MWCNTs and GNPs on the thermal and electrical conductivity, the addition of both carbon nanoparticles presents a combined effect. Based on these data, the hybrid GNPs/MWCNTs material presented a better combination of properties than the neat composites, leading to the successful development of multifunctional adhesives by utilizing filler’s advantages for the improvement of each property.

Until now, the relatively published results have been controversial. Li, Y. et al.

[41] published a remarkable synergistic effect between CNTs and GO sheets, which improved the mechanical properties of the epoxy. Yue, L. et al. [42] reported that a hybrid nanofiller of CNTs and GNPs at a mass ratio of 8:2 synergistically contributed to increased flexural properties and a reduction of the electrical percolation threshold in epoxy composites. It was reported that the hybrid nanofiller demonstrated a reduced percolation threshold (pc) of 0.62%wt. in contrast to the CNT and GNP single-filled composites, which exhibited pc values of 0.84%wt. and 0.88%wt., respectively. This was attributed to the more efficient formation of conductive pathways between the carbon nanoparticles. In contrast, Im, H. et al. [43] did not find a synergistic effect on the storage modulus of epoxies reinforced with GO nanoplatelets and MWCNTs. However, the same author showed a synergistic effect for the thermal conductivity of the studied composites. Han, S. and coworkers et al. [23] reported a synergistic effect of graphene and carbon nanotubes on lap shear strength and electrical conductivity of epoxy adhesives. Kopsidas, S. et al. [25] measured a synergetic increase in both electrical conductivity and fracture toughness of epoxy polymers doped simultaneously with GNPs and CNTs at a mass ratio of 9:1. However, no significant differences were observed in the tensile properties and the glass transition temperature for the carbon-modified epoxies.

In the following SEM micrographs (Figures 8(a)-(d)), it can be seen the simultaneous dispersion of two fillers into the epoxy adhesive. In Figure 8(b), the squares indicate the representative GNPs, while the corresponding MWCNTs were circled. Figures 8(b)-(d) present the simultaneous integration of fillers into the epoxy adhesive at different magnifications. There is no clear cooperation between them for the creation of conductive channels and paths, which are needed for the improvement of electrical and thermal conductivity of the adhesives, respectively. This is probably attributed to the weak interfacial adhesion between the matrix and the carbon nanofillers. The presence of strong interfacial adhesion is necessary for effective reinforcement as it allows stress transfer from the high modulus filler to the low modulus polymer matrix. However, this can only be achieved when the two parts are chemically compatible. For this reason, efforts have been focused on functionalizing the carbon nanoparticles [28], without implying that there will be the desired results [27].