Effect of SDS Decoration of Graphene on the Rheological and Electrical Properties of Graphene-filled Epoxy/Ag Composites

Transcription

1 Effect of SDS Decoration of Graphene on the Rheological and Electrical Properties of Graphene-filled Epoxy/Ag Composites Josh Trinidad a,b,c, Behnam Meschi Amoli a,b,c, Wei Zhang a,b, Rajinder Pal a and Boxin Zhao a,b,c * adepartment of Chemical Engineering, b Waterloo Institute of Nanotechnology, c Institute for Polymer Research, University of Waterloo, 200 University Avenue West, Waterloo ON, N2L 3G1, Canada Corresponding Zhaob@uwaterloo.ca Abstract The rheological behavior and viscosities of epoxy composites containing graphene decorated with SDS were investigated. Graphene filled epoxy composites were mixed and dispersed without solvent using a planetary shear mixer (PSM). It was found that SDS decorated graphene (Gr(s)) exhibits lower viscosities when compared to non-decorated graphene. Further findings from the viscosity results show that the addition of Gr(s) to an intrinsically shear thickening (dilatant) epoxy resin not only suppressed this rheological behavior, but also began to exhibit shear-thinning behavior past a weight concentration of 0.5%. The morphology and electrical conductivity of the ECAs synthesized from this method have also been investigated and show bulk resistivity values in the range of Ω cccc. This work demonstrates the possibility of using Gr(s) in ECA applications, as well as for any composites that desire the suppression of shear thickening behavior for better workability and fluid flow. Keywords: Surfactant, Viscosity, Shear-thinning, Graphene, conductive adhesives 1

2 1. Introduction Graphene is an atomically thin sheet of carbon bonded in bonds arranged as a hexagonal (honeycomb-like) lattice and can primarily be seen as either a monolayer of graphite or an unwinded single walled carbon nanotube [1 4]. As a two-dimensional nanomaterial, graphene holds great potential for a variety of applications owing to its excellent electrical (primarily stemming from its sp 2 -bonds [1]), mechanical and thermal conductivity, high aspect ratio and high specific area [2,3,5,6] as well as high carrier mobility (even higher than semiconductors at room temperature), high mechanical stiffness and optical transparency etc [1]. Despite possessing these wonderful properties, graphene comes with limitations and challenges such as obtaining homogeneous dispersion and the preservation of its exfoliated 2-D single-layer structure when introduced into a polymer matrix [2 4,7]. There are two common approaches that are used to disperse graphene: a covalent and noncovalent approach [3,5,8]. The covalent approach involves decorating graphene with nanoparticles (for example, silver) to prevent aggregation, and at the same time keep the graphene exfoliated [8]. However, this approach uses harsh chemicals, complicated chemistry, and has an effect that disturbs the ππ-electrons in graphene resulting in the hindrance of graphene s excellent electrical properties [5]. The non-covalent approach uses an anionic surfactant such as SDS as the primary agent that allows for the graphene to be dispersed in ethanol solvent. This method was successful at keeping it exfoliated even during its integration into the composite, and is shown to exhibit remarkable electrical conductivity [5]. Although the use of a solvent as a means to disperse the carbon filler is effective, it is also known to negatively impact the thermal, electrical and mechanical properties of the nanocomposites [9,10]. To add to this, both trace amounts of solvent and 2

3 bubbles formed during the curing are unavoidable even with long desiccation times [9,10]. Furthermore, the yield that is expected from such a process is expected to be very low because for a small amount of material, a large amount of solvent is needed [5]. M. L. Gupta et al addressed a similar problem when it came to making composites, however, instead of using graphene, they used carbon nanotubes (CNTs) as the nanofiller. They successfully determined how to disperse and mix the nanofiller without solvent when it came to forming composites through the use of a planetary shear mixer (PSM), as well as other high shear mixing machines [9]. Investigating the rheological properties of the material is important. Previous works such by T. Filleter et al and J. Lin et al suggest that single layer graphene possesses properties that allow it to behave as a solid lubricant [11,12]. This concept is useful as J. Coo et al and R. Bowman et al explain that frictional force (that lubricants decrease) is proportional to viscosity [13,14]. These rheological studies for graphene and epoxy composites are in specific useful for electrically conductive adhesives (ECAs); a material intended for replacing traditional alloy-based solders [15]. The viscosity and rheological properties of an ECA is key to determining its performance and feasibility as it is applied to stencil and screen-printing techniques [15]. In one of our recent works, a non-covalent approach using sodium dodecyl sulfate (SDS) surfactant was employed to exfoliate graphene and overcome some of the concerns mentioned above [5]. However, this solvent assisted method could only produce small quantities, which is an obstacle for large-scale production. In this work, we report the use of a PSM to carry out a solvent-free in situ polymerization method in order to disperse nanofillers and synthesize hybrid ECAs as well as investigate nanocomposites that were 3

4 not loaded with silver flakes. This work aims to systematically investigate how adding graphene and SDS decorated graphene will affect the viscosity and the electrical performance of the composite while using a solvent-free mixing approach to disperse the nanofillers. The way this mixing method affects the ECA properties will be briefly discussed by examining the morphology of the ECA using SEM. We will also determine the electrical conductivity of the material as it is suspected that the absence of the solvent will directly affect the dispersion of graphene and so, also affect its ability to decrease bulk resistivity. 2. Experiments and Methods 2.1. Functionalization and stabilization of graphene nanosheets with Sodium Dodecyl Sulfate (SDS) Graphene nanosheets (Gr) were purchased from ACS Materials (USA) and used as received. The size of the graphene nanosheets according to the supplier is 0.5 to 5 µm and is confirmed using TEM by Amoli et al to be around 1 µm [5]. SDS was purchased and used as received from Sigma Aldrich. Although the procedure before the nanocomposites preparation requires the use of solvents (ultra-pure water and HPLC Ethanol) in order to properly functionalize and stabilize the Gr with surfactant, the mixing procedure for preparing the nanocomposites is solvent-free. The procedure for functionalizing graphene using a non-covalent method via surfactant was inspired by our previous work [5] and is illustrated below in Fig. 1, where a liquid solution is prepared by mixing SDS powder with ultra-pure water at a concentration of 60 mmol L -1, which is 20 mmol L -1 lower than the critical micelle concentration (CMC) of 4

5 SDS in water [16,17]: the point where the excess surfactant forms micelles. The aim of the SDS decoration is not to make micelles, but rather, to decorate and self-assemble onto the graphene and non-covalently exfoliate the graphene sheets. This solution was then placed in low power sonic bath (Branson 2510R-MT) to undergo ultrasonication for 30 minutes, and allows for the hydrophobic backbone of the SDS to adsorb onto graphene s surface. The solution is centrifuged at 8500 RPM for 10 minutes and the supernatant is removed. The remaining precipitate is washed four times using HPLC ethanol as its dispersant which is then sonicated for 5 minutes. This process is repeated three times in order to completely remove all un-bonded SDS from the precipitate. Finally, after completing the repetitions, the precipitate is then placed in a vacuum oven at 75 C overnight to allow any remaining solvent to evaporate wherein the leftover precipitate of SDS stabilized graphene (Gr(s)) is harvested Preparation of nanocomposites Liquid epoxy resin (D.E.R. 331) Diglycidyl Ether Bisphenol-A (DGEBA) and curing agent (D.E.H. 24) triethylenetetraamine (TETA) were purchased and used as received from DOW Chemical Company (USA). Silver flakes (~10 μm) were purchased from Sigma Aldrich and used as received. The following combinations denoted in table 1 were combined in one container where they were mixed at 2000 RPM for 4 minutes and defoamed at a speed of 2200 RPM afterwards for 1 minute using a Thinky Mixer (ARE-310). Table 1: List of the combinations of compositions used for nanocomposites Sample Composition Ag flakes [wt %] Gr/Gr-s [wt %] 1 Epoxy Epoxy + Silver flakes Epoxy + Graphene , 0.5, 1, 1.5 and 2 5

6 4 Epoxy + Graphene-SDS , 0.5, 1, 1.5 and 2 5 Epoxy + Silver flakes + Graphene Epoxy + Silver flakes + Graphene-SDS Viscosity measurements A cone & plate viscometer (Brookfield CAP2000H+, USA) was used to determine the effects of filler concentration on the viscosity of the system at room temperature. Cone & plate viscometers operate under the principle of measuring a torque felt by a cone shaft undergoing a constant shear rate and known spindle dimensions. As seen in Fig. 2, this viscometer is comprised of a flat plane where the sample is located and a rotating shaft (housing a cone shaped spindle at the end). The cone itself has an obtuse angle propagating away from the apex of the shaft; the gap between the cone and plate once the apex touches the plane is filled with the sample where the linear velocity is proportionate to the radial distance resulting in a constant shear rate as the spindle rotates [18]. Shear stress ττ is denoted below as a ratio between torque G felt by the spring and radius r of the cone [18]: ττ = 3GG (1) 2ππrr 3 Moreover, the shear rate γγ is determined through the ratio between the velocity of the cone Ωrr and cone gap C: γγ = Ωrr CC = Ω ψψ (2) The viscosity ηη obtained using a cone & plate viscometer is defined as the ratio between shear stress and shear rate: ηη = ττ γγ = 3GG 2ππrr 3 Ω (3) ψψ 6

7 As can be seen above, the equation for determining the viscosity of a material under a cone & plate viscometer is heavily dependent on the both the torque felt by the spring as well as the rotation speed of the cone. In this viscometer, a cone CAP-05 was used, having a radius rr = cccc, a cone angle ψψ = , and finally, a cone gap of CC = cccc Morphological Characterization of Nanocomposites using SEM A Field Emission Scanning Electron Microscope (FE-SEM, LEO-Ultra, Gemini, Germany) was used in order to carry out a qualitative study on the morphology and degree of dispersion of graphene nanosheets within the nanocomposites produced. The samples were placed on a 90-degree stub to view the cross-section of the ECA Measurements for Electrical Conductivity The resulting composites that contained silver flakes were casted into a glass slide that had a square mold (7mm x 7mm x 0.5 mm) made of adhesive tape. A clean, flat copper sheet was placed on top of the paste in the mold to ensure a flat surface and consistent thickness. The sample was then inserted into the oven and pre-cured at 60 C for 30 minutes, and then cured at 150 C for two hours. Once the curing is complete, the copper sheet and adhesive tape are removed and the in order to determine the bulk resistivity of the sample using a sheet-resistance four-point probe measurement setup. The configuration consists of the probe unit (Cascade Microtech Inc.) and the digital multimeter & function generator unit (Keithley A SourceMeter, Keithley Instruments Inc.). The sheet resistance readings acquired from this configuration can then be converted into bulk resistance readings using the following equation [8,19]: 7

8 ρρ = FFFF ππ llll 2 VV (4) II where t is the thickness of the sample, I is the applied current, and V is the voltage drop measured by the SourceMeter. The F in equation (4) is a correction factor for samples that have a finite thickness that is defined by the ratio between the sample thickness t and probe spacing s (1 mm). The correction factor F can be approximated to 1 if 0.4 < tt < 1. ss The mold is 0.5 mm while the probe spacing is 1 mm, which leads to a ratio of 0.5: a value that allows F to be approximated to Results and Discussion 3.1. Viscosity and Flow Behavior of Uncured Nanocomposites The viscosities of six nanocomposites samples, all containing varying compositions were measured at room temperature at 10 RPM or shear rate of 33.3s -1 (listed in the Table 2) in order to do quantified studies on the behavior and characteristics of the uncured nanocomposites. As expected and observed, the addition of any form of fillers results in an increase in viscosity. Compared to the 60 wt% silver, the added 1% graphene has a very pronounced effect on the viscosity of the composite, reflecting the fluffy nature of graphene nanosheets. As such, the fourth sample containing Gr(s) exhibits lower viscosities as opposed to the third containing Gr, indicating that SDS plays an important role in the viscosity behavior of the composite. The main function of SDS decoration is to improve the dispersion of graphene nanosheets with in the epoxy resin [5]. However, some studies have shown that the presence of an anionic surfactant molecule such as SDS can disrupt the inter-chain contact or entanglement of the hydrophobic polymer chains (in our case, 8

9 Bisphenol A diglycidyl ether) causing a decrease in viscosity [20]. This decrease in viscosity can also be explained by considering that the surfactant (that self-assembles on the polymer and graphene) participates in an untying process that severs the connections between different polymer chains [21]. This means that the presence of SDS in our system may have not only prevented the aggregation of the graphene but also the entanglement of the epoxy polymer, thus leading to lower viscosities. Table 2: Viscosities for the different compositions of the nanocomposites at 10 RPM Sample Composition Viscosity [cp] 1 Epoxy Epoxy + Silver flakes (60 wt%) Epoxy + Graphene (1 wt%) Epoxy + Graphene-SDS (1 wt%) Epoxy + Silver flakes + Graphene Above Epoxy + Silver flakes (60 wt%)+ Graphene-SDS (1 wt%) The effect of Gr/Gr(s) on the viscosities of the composites were systematically investigated by varying the amount of added Gr/Gr(s) and the shear rate, with a particular goal of observing the role SDS decoration. Fig. 3 uses bar graphs to show the viscosities of epoxy with Gr (Fig. 3a) and epoxy with Gr(s) (Fig. 3b) nanocomposites as a function of increasing weight concentration at a constant 20 RPM (or a shear rate of 66.7 s -1 ). It is seen that the Gr(s) graphs show much lower values when compared to the Gr graphs. The maximum value in Gr(s) graphs is close to 60,000 cp whereas the maximum value of the Gr graphs is doubled, almost exceeding 120,000 cp. The dependence of viscosity on the weight concentration of pristine Gr is an exponential increase; this same trend was not seen for Gr(s). 9

10 Fig. 4 and Fig. 5 show plots of viscosity as a function of shear rate in log base 2, providing a different perspective on what happens to the viscosity of the system as the rotational speed of the cone & plate is increased, where Fig. 4 is Gr and Fig. 5 is Gr(s). It can be observed that the way shear rate affects the viscosity depends on the amount graphene added to epoxy. For the samples of lower wt% (< 0.5 wt%), the viscosity increased with the shear rate, following a known behavior where the non-aggregated solid particles show high viscosities at low shear rates, but then begin showing lower viscosities upon higher shear rates [22]. In contrast, for the samples of higher wt% (> 0.5 wt% Gr or > 1 wt % Gr(s)), the viscosity first decreased with the shear rate but exhibited spikes (i.e. sharp increase) that thus far appears after going past what appears to be a critical shear rate of 167 s -1. The dotted lines indicate spikes in the viscosity that can no longer be measured by the cone and plate viscometer as the viscosity rapidly increased at those high shear rates. This type of spike was observed in Fig. 4 and Fig. 5, occurring at higher shear rates for sample loadings of 1-2 wt%. The effects of increasing shear rate on the viscosity of the nanocomposites system are further interpreted in terms of the concepts of shear thickening and shear-thinning. Shear thickening is a behavior seen in fluids that experience a sharp increase in viscosity above a critical shear rate, where the fluid often transitions from a liquid-like dispersion into a solid-like material as a result of the particles forming clusters [23]. Conversely, fluids that behave in the opposite manner are known as shear-thinning fluids; for example, colloidal dispersions are typically shear-thinning fluids [24]. When this type of fluid is subjected to flow, the colloidal particles begin to form layers resulting in lower flow resistance as well 10

11 as lower viscosity [23]. As shown in Fig. 4 and Fig. 5, the viscosity of the pure epoxy (D.E.R. 331 DGEBA) increases with the shear rate, behaving as a shear thickening liquid. Note that DGEBA can render either shear thickening or shear-thinning behavior depending on the supplier [25]. The viscosity of the epoxy nanocomposites with 0.25 increased only slightly with higher shear rate; and the viscosity of the composite with 0.5 wt % Gr and Gr(s) are almost constant with higher shear rate. An explanation for this shear-thickening behavior of the nanocomposites containing low loading of graphene (< 0.5 wt%) is that the composite is in a regime where the system is dominated by the molecular strands of the epoxy resin, which is similar to the one-dimensional particle arrangement discussed by D. Smith et al. [26]. As a result of the viscometer applying force, the polymer strands become entangled and take a three-dimensional arrangement, resulting in higher viscosities [22] within the higher shear rate region. The behavior of the epoxy composite with higher loadings (1-2 wt%) of Gr and Gr(s) starts to show signs of shear-thinning, hinting the possibility that the addition of graphene nanosheets suppressed the shear thickening properties of the epoxy. It appears to have a transition from shear-thickening to shearthinning around wt % Gr or Gr(s) loadings. It is known in literature that the presence of a layered structure (in our case, graphene) can suppress the onset of shear thickening as it reduces the efficacy of the shear flow to create the hydrodynamic clusters [24]. Similar results indicating that shear-thinning of polymer-filled graphene at higher weight concentrations is found in D. Wu et al s work where they reported that graphene at 1 wt% showed stronger shear-thinning behavior, and further reported that graphene nanosheets were responsible for the suppression of the shear flow of their polylactide polymer chains [27]. 11

12 Fig. 5 revealed a stronger shear-thinning behavior in Gr(s) when compared to Fig. 4, which happens at higher loadings of the Gr: in particular, at the 1 wt %, suggesting the some influence resulting from the SDS molecules decorated on the graphene surface. The SDS molecules may have formed additional intermolecular bonds between the graphene and epoxy [28], which resulted in the increase of viscosity at low shear rates, however, applied force at high shear rates break the bonds thus resulting in the shear-thinning behavior. Other than the 1 wt% Gr(s) curve, all other curves at higher loadings above 0.5 wt % in Figure 4 and 5 show an initial shear-thinning behavior, followed by a sharp increase in viscosity at higher shear rates. This phenomenon is similar to the typical Order-Disorder flow curve of stable suspensions of Brownian hard particles [23,29]. The Order-Disorder theory describes the sliding of particles over one another, forming layers during shearing; after exceeding a critical shear rate, the hydrodynamic forces would pull out the particles from their ordered layers to enter a different regime where the particles agglomerate with one another to form a disordered state resulting in a sharp increase in viscosity [23]. Overall, Figure 4 and 5 reveal that the addition of SDS-decorated graphene Gr(s) led to lower viscosities for the epoxy nanocomposites than the pristine graphene Gr. This finding has profound implications for the future applicability on these conductive composites as some printing techniques that already focus on dispensing conductive ink are concentrating on ways to improve their printing process (which involves the use of shear-thinning conductive inks) [30]. Furthermore, conductive adhesives used for surface mount assemblies are recommended to exhibit strong shear-thinning, highly stable yet low 12

13 viscosities at high shear rates, and good recovery of the pastes after the process is complete [31]. This suggests that SDS decorated graphene may be useful as a dilatant-suppressor when formulating ECAs that are dispensed through printing techniques. On another note, it has been reported that negative effects such as excess deposition or nonsmooth flow across the stencil are associated with conductive adhesives that do not fall within the desired viscosity ranges (recommending a range of 100,000 cp, although the conductive adhesive used in that work was 70,000 cp) [32]. It can be seen that sample 6 is reaching the range of the recommended viscosity and that the use of exfoliated graphene as a lubricant is indeed successful. Further work for optimization can result in reducing the viscosity to even lower, more acceptable levels, which is especially relevant for other applications such as dispensing through printing technology [15] Morphology and electrical conductivity of the cured nanocomposites We characterized the morphology and conductivity of the cured nanocomposites for its potential applications in electrically conductive adhesives. The SEM images in Fig. 6 show the morphologies of the epoxy and Gr/Gr(s) nanocomposites at different magnifications. Fig. 6a and 6b compare the difference between low weight percent and high weight percent for Gr as seen in Fig. 6c and 6d. It is seen that the increase of filler content definitely changes the morphology from a smooth, flat surface to a flakier and mountainous appearance. As such, the higher weight percent Gr contains a morphology that is overall sharper, and higher in surface area even when compared to Gr(s) at the same weight percent. Fig. 6e is the SEM image of the nanocomposite with 2 wt% Gr(s), showing a similar morphologies to that of 2 wt % Gr, containing moderately hilly or mountainous 13

14 terrain. The transparent-looking sheets at the edge of the flakes found in the high magnification zoom in Fig. 6f indicate that the Gr(s) are better exfoliated. The electrical conductivities for the synthesized hybrid ECAs were investigated using a four-point probe method. Fig. 7 plots the electrical bulk resistivity of the ECAs as a function of the weight loading of two nanofillers Gr and Gr(s) added into the composites without the use of solvent (identified as solvent-free method and denoted by solid lines) in this work and with the assistance of ethanol solvent used in our previous work (identified as solvent-assisted and denoted by dashed lines) [5]. The electrical resistivity values of the ECAs filled with Gr(s) are in the same order of magnitude as those of the ECAs filled with Gr; in general, the resistivity decreased with the addition of graphene. This observation confirms the positive effect of adding graphene on improving the electrical conductivity when compared to conventional ECAs and suggests that the SDS decoration of graphene has a negligible effect on the conductivity improvement. However, a slight increase in the resistivity between 1-1.5% wt Gr was noticed, perhaps because of the uneven dispersion of the Gr at higher concentrations. The bulk resistivity values for the solvent-free method are in the range of Ω cccc, which is one magnitude higher compared to the solvent assisted method. This is reasonable considering the role of using solvent was to improve the filler dispersion and deliver better conductivity. It is important to note that we see this solvent effect even for convention ECAs that do not contain graphene nanofillers. Even though the solvent-free method may hold the advantage in the simplicity of the manufacturing process and the workability of the paste itself, the graphene-filled ECAs prepared in this work may be suitable for the applications where only moderate 14

15 conductivity is needed. Further work is necessary to improve the electrical conductivity and as a result, broaden its potential applications. 4. Conclusion The rheological and electrical properties of graphene/silver-filled composites were systematically investigated, with a particular focus on elucidating the effect of SDS decoration of graphene. It was found that Gr(s) exhibits lower viscosities when compared to Gr (with or without the presence of silver), revealing a positive effect of SDS on reducing the viscosity. Our research findings also show that weight concentrations from both Gr and Gr(s) exhibit shear thickening from 0 to 0.5 wt% owing to the potential transition of the resin going from 2 dimensional to 3 dimensional conformation as well as the intrinsic properties of the epoxy chosen. However, past 0.5 wt%, it is suggested that due to the presence of a layered structure suppressing the shear thickening behavior of the system, the Gr and Gr(s) viscosities began exhibiting shear-thinning behavior. Both systems follow the order-to-disorder theory and curves at higher weight concentrations. Gr(s) has been shown to follow the typical curve for stable suspension of Brownian hard particles after a high enough concentration is present within the epoxy matrix. As a result, the presence of graphene in general has been shown to play a key role in the shear-thinning behavior of these composites. Furthermore, this work revealed new rheological behavior upon the addition of surfactant SDS on graphene. The decoration of graphene is usually exploited to improve dispersion; however, in our case it has an added unique effect. It is clear that the multiple effects of surfactant should be considered in the development of ECAs: not simply for the improvement of filler dispersion, but also the suppression of dilatant 15

16 behavior. Future work can incorporate this finding into the optimization of ECA fabrication processes, as well as any other endeavors that require a shift in rheological behavior. 5. Acknowledgements The authors want to thank the Refined Manufacturing Acceleration Process Network (ReMAP), Natural Sciences and Engineering Research Council of Canada (NSERC) for their financial support on this project. 6. References 1. F. Akbar, M. Kolahdouz, S. Larimian, B. Radfar, and H. H. Radamson, J. Mater. Sci. Mater. Electron. 26, 4347 (2015). 2. J. Du and H.-M. Cheng, Macromol. Chem. Phys. 213, 1060 (2012). 3. V. Mittal, S. Roy, and S. K. Srivastava, Macromol. Mater. Eng. 300, 346 (2015). 4. V. K. Rana, M. Choi, J. Kong, G. Y. Kim, M. J. Kim, S. Kim, S. Mishra, R. P. Singh, and C. Ha, Macromol. Mater. Eng. 296, 131 (2011). 5. B. Meschi Amoli, J. Trinidad, G. Rivers, S. Sy, P. Russo, A. Yu, N. Y. Zhou, and B. Zhao, Carbon N. Y. 91, 188 (2015). 6. K. Liu, S. Chen, Y. Luo, D. Jia, H. Gao, G. Hu, and L. Liu, Compos. Sci. Technol. 88, 84 (2013). 7. N. W. Pu, Y. Y. Peng, P. C. Wang, C. Y. Chen, J. N. Shi, Y. M. Liu, M. Der Ger, and C. L. Chang, Carbon N. Y. 67, 449 (2013). 8. B. Meschi Amoli, J. Trinidad, A. Hu, Y. N. Zhou, and B. Zhao, J. Mater. Sci. Mater. Electron. 26, 590 (2014). 9. M. L. Gupta, S. a. Sydlik, J. M. Schnorr, D. J. Woo, S. Osswald, T. M. Swager, and D. Raghavan, J. Polym. Sci. Part B Polym. Phys. 51, 410 (2013). 10. K. Lau and M. Lu, Compos. Sci. Technol. 65, 719 (2005). 11. T. Filleter, J. L. McChesney, A. Bostwick, E. Rotenberg, K. V. Emtsev, T. Seyller, K. Horn, and R. Bennewitz, Phys. Rev. Lett. 102, 1 (2009). 12. J. Lin, L. Wang, and G. Chen, Tribol. Lett. 41, 209 (2011). 13. J. Koo and C. Kleinstreuer, Int. J. Heat Mass Transf. 47, 3159 (2004). 14. R. M. Bowman and K. B. Eisenthal, Chem. Phys. Lett. 155, 99 (1989). 15. M. Pudas, N. Halonen, P. Granat, and J. Vähäkangas, Prog. Org. Coatings 54, 310 (2005). 16. F. H. Quina, P. M. Nassar, J. B. S. Bonilha, and B. L. Bales, J. Phys. Chem. 99, (1995). 17. in IUPAC Compend. Chem. Terminol. (IUPAC, Research Triagle Park, NC, 2014), p R. McKennell, Anal. Chem. 28, 1710 (1956). 19. W. Zhang, Y. Zhou, K. Feng, J. Trinidad, A. Yu, and B. Zhao, Adv. Electron. Mater. 1 (2015). 20. S. Biggs, J. Selb, and F. Candau, 838 (1992). 16

17 21. S. Nilsson, K. Thuresson, and P. Hansson, 5647, 7099 (1998). 22. A. Zupani-it, R. Lapasinb, and M. Turnera, 30, (1997). 23. a. Srivastava, A. Majumdar, and B. S. Butola, Crit. Rev. Solid State Mater. Sci. 37, 115 (2012). 24. J. Bender, J. Rheol. (N. Y. N. Y). 40, 899 (1996). 25. M. Wang, X. Fan, W. Thitsartarn, and C. He, 58, 43 (2015). 26. D. E. Smith, Science (80-. ). 283, 1724 (1999). 27. D. Wu, Y. Cheng, S. Feng, Z. Yao, and M. Zhang, Ind. Eng. Chem. Res. 52, 6731 (2013). 28. B. Briscoe, P. Luckham, and S. Zhu, Polymer (Guildf). 41, 3851 (2000). 29. W. J. Frith, P. Haene, R. Buscall, and J. Mewis, J. Rheol. (N. Y. N. Y). 40, 531 (1996). 30. J. Noh, D. Yeom, C. Lim, H. Cha, J. Han, J. Kim, Y. Park, V. Subramanian, and G. Cho, IEEE Trans. Electron. Packag. Manuf. 33, 275 (2010). 31. H.-V. Nguyen, E. Andreassen, H. Kristiansen, and K. E. Aasmundtveit, IEEE Trans. Components, Packag. Manuf. Technol. 3, 1084 (2013). 32. H. J. Lewis and a Ryan, J. Adhes. Sci. Technol. 22, 893 (2008). 17

18 Fig. 1 Schematic illustration of the decoration or functionalization of the graphene nanosheets with surfactant SDS 18

19 Fig. 2 a Optical image of the cone & plate viscometer; b zoom-in of cone & plate loaded with epoxy resin; c cross-sectional diagram of the cone & plate with the terms used in the 19

20 viscosity equation Fig. 3 a Viscosity as a function of weight loading of the pristine graphene (Gr) in epoxy resin at 20 RPM; b viscosity as a function of weight loading of SDS-decorated graphene Gr(s) in epoxy resin at 20 RPM. 20

21 Fig. 4 Viscosity in log scale as a function of Shear Rate of Gr and Epoxy where dotted lines indicate viscosity readings too high for the viscometer to display; Note: All values in graphs that contain error bars represent standard error. 21

22 Fig. 5 Viscosity in log scale as a function of Shear Rate of Gr(s) and Epoxy where dotted lines indicate viscosity readings too high for the viscometer to display; Note: All values in graphs that contain error bars represent standard error. 22

23 Fig. 6 SEM images of the epoxy composites with 60% silver and Gr or Gr(s); a Low zoom view of Graphene and epoxy [Gr] at 0.25 wt%; b Highest zoom view of Gr 0.25 wt% on a large flake; c Low zoom view of Gr 2 wt% showing a mountainous morphology; d Highest zoom view of Gr 2 wt% taken of an area from 6c; e Low zoom view of Gr(s) 2 wt% showing a similar morphology to 6c; f Highest zoom view of Gr(s) 2 wt% showing smoother morphology likely as a result of the SDS decoration 23

24 Fig. 7 Bulk resistivity as a function of weight percent comparison graph between Gr and Gr(s). The solid lines denote (solvent-free) results from our work. The dotted lines show results from previous work [5] that used ethanol to assist in the dispersion of filler content. 24