Graphene-based polymer nanocomposites attracted significant scientific and commercial interest in recent years [1
]. The main reason for this interest lies in the outstanding performance that high aspect ratio fillers can induce into a polymer matrix [7
]. This performance increase in comparison with traditional filled systems (e.g., carbon black) allows for a wide range of potential industrial applications [11
], like gas-barriers [13
], supercapacitors [14
], or solar cells [15
The introduction of a filler in a polymer matrix can lead to an increase of mechanical properties (e.g., Young’s modulus) and in the case of carbonaceous fillers an increase of electrical conductivity. Both these properties are tightly linked to the type of filler and polymer combination. Depending on the targeted application the filler amount might be significant (e.g., for application that require high electrical conductivity the amounts of carbon black used can be in the range of 15–40 wt.%), something that has been seen to lead to an increase of the overall viscosity with potential processing and performance difficulties [16
]. The way the filler is incorporated into the polymer matrix is of paramount importance as it affects the overall dispersion and the subsequent properties of interest. Dispersion is further inhibited by the tendency of the filler particles to cluster in larger agglomerates, something that is mainly driven by the existence of attractive interactions between the filler particles [17
]. These effects are augmented in the case of high aspect ratio fillers like graphene by the extensive size of the primary particle. The agglomeration that takes place during mixing can be understood via filler–filler, filler–polymer, and polymer–polymer interactions, where the polymer mobility plays an important role on the facilitation of local and long range arrangements of the filler particles (of all sizes) in the polymer matrix [18
The most economically advantageous method of dispersing a filler in a matrix, especially in a large scale industrial environment, is by melt mixing. Such methods have been successfully used for traditional carbonaceous fillers like carbon black [19
], but due to the inherent agglomeration and extensive clustering of the primary particles the resulting systems tend to exhibit a high electrical percolation threshold [20
] and increased brittleness [23
]. Similarly, for high aspect ratio fillers like graphene the attractive interactions between the filler particles and the resulting level of clustering and agglomeration render melt mixing a real challenge [3
]. A number of attempts of melt mixing of thermally reduced graphene in glassy [1
] thermoplastic [27
] and elastomeric [29
] polymers has been reported, but the majority of the resulting compositions yield systems of significantly worse graphene dispersion in comparison with solvent-based mixing methods [2
Here we must note that although graphene is commonly seen as a pristine monolayer of hexagonal carbon atoms, in reality, there is a whole class of materials that can be considered graphene-like. The initial distinction comes from the different preparation methods that can be classified as bottom-up and top-down. Bottom-up graphene can be produced via a variety of methods that include chemical vapor deposition [30
], epitaxial growth [31
], chemical conversion [32
], and reduction of CO [33
]. Top-down graphene summarizes the different processes that originate from separation and subsequent exfoliation of graphite and its derivatives. These methods offer significant financial advantages for large scale applications since graphite is a material of relative abundance and reasonably low price [2
]. The most promising graphene derivatives originating from this type of production are expanded graphite (EG) and graphene oxide (GO). Thin forms of EG (~10 nm) known as graphene nanoplatelets (GNP) are already commercial and can be produced by different types of graphite intercalation [27
]. In these systems the flexural rigidity and dimensions of the platelets is retained during preparation, thus allowing some element of high aspect ratio properties to be present even without complete exfoliation [27
]. Probably the most promising of alternatives is based on exfoliation and reduction of GO, a process that can be performed using a number of alternative variations of the Staudenmaier [36
] or Hummers [37
] methods, where graphite is oxidized in the presence of strong oxidants followed by a sequence of reduction steps, mainly to restore its electrical conductivity.
Polypropylene (PP) is one of the most commonly-used polyolefins with a myriad of application uses ranging from everyday items like trays and bottles to high performance compounds suitable for medical, automotive, energy transport and storage, and pipe products [38
]. Due to its commercial importance a number of studies have been undertaken on PP/graphene nanocomposites focusing mainly on electrical [27
], thermal [42
], and mechanical [43
] performance. Electrical conductivity and especially the electrical percolation threshold is directly linked to the level of dispersion of the filler in the matrix [46
]. As discussed previously, melt mixing can lead to significant agglomeration and inefficient dispersion. That can be seen by relatively high values of the electrical percolation threshold3
. In previous studies on PP the electrical percolation threshold for melt mixed systems using a variety of mixing protocols and pre/post mixing procedures has been reported in the range 6–15 wt.% [27
]. These values are significantly higher than theoretical predictions [47
] and values obtained for near perfect dispersion of graphene in a variety of matrices using solvents and sonication protocols [2
]. This indicates that the nominal aspect ratio of the filler is being reduced by the mixing procedures leading to significant agglomeration that increases the electrical percolation threshold. For most fillers in their bulk form including different variants of graphene, a significant level of clustering and agglomeration exists, and this has the potential to be still present after mixing. For any such system to be commercially viable, the initial agglomeration level needs to be significantly reduced, thus allowing for a reasonably good dispersion using melt mixing protocols. The use of simple melt mixing protocols is of paramount importance if one considers large-scale commercial applications. This is due to the need for efficient and relatively streamlined procedures in an environment where safety, throughput and material property consistency needs to be evaluated and kept as constant as possible at all times, without forgetting that the overall volumes associated with the process can range from a few kilograms to a few hundred tones.
In this work we demonstrate for the first time an industrially relevant system of polypropylene with a novel hierarchical graphene nanostructure, which is capable to achieve superior levels of dispersion as seen through the observed electrical percolation threshold. Furthermore, the mixing protocol used was kept intentionally simple as to explore the potential industrial scalability of the process. By utilizing these simple mixing protocols, we have managed to achieve dispersion levels that are on par with solution techniques, something that becomes obvious from the location of the electrical percolation threshold at the region of ~1 wt.%. The combination of superior dispersion and the design of the filler (highly de-agglomerated) allows us for the first time to get partial access to the primary graphene particles in a melt mixed system and study the effect it has on its electrical, thermal and mechanical properties. A comprehensive overview of the effect of the filler on the overall morphology, thermal, rheological, and mechanical properties is also presented and discussed.
2. Materials and Methods
De-agglomerated hierarchical thermally reduced graphene oxide was supplied by Cabot Corporation. An overview of the available preparation methods for the different types of graphene can be found elsewhere [37
]. In Figure 1
SEM pictures of the material in bulk form can be seen. From these pictures, we can see that despite the various production steps the material still contains a significant level of de-agglomeration.
A highly isotactic (>90%) polypropylene (Mw = 300 kg/mol, Mw/Mn = 8) provided by Borealis AG was used.
2.1. Sample Preparation
All compositions were prepared using a Brabender mixer Type W50 driven by a Brabender Plasticorder. Initially the polypropylene was added and allowed to melt at 210 °C at 20 rpm for 15 min, followed by addition of graphene. The combined composite was then mixed at 210 °C at 50 rpm for another 15 min. Following compounding the composites were cut into small pellet-like pieces for easier use.
2.2. Electrical Measurements
Electrical conductivity measurements ware performed using a Novocontrol Alpha spectrometer in a frequency range of 10−2 to 107 Hz, at atmospheric pressure and under nitrogen atmosphere. The sample cell consisted of two silver-coated electrodes 40 mm in diameter and the sample with a thickness of 0.1 mm. Each measurement was carried out six times, and average values were recorded. The complex conductivity σ* = σ′ + iσ″, the real part of which is used for the analysis herein, can be deducted from the complex dielectric permittivity ε* as σ* = iωε0ε*, where ε0 is the permittivity of free space. For compositions above the electrical percolation threshold, the conductivity value recorded was the one obtained from the frequency independent part of the curve, whereas for compositions below the electrical percolation the value recorded was the lowest conductivity value we were able to obtain as no conductivity plateau was reached within the experimentally available frequency window.
2.3. Thermal Analysis
DSC measurements were carried out under nitrogen at a rate of 10 °C/min in a temperature range of −50 and 210 °C, using a Mettler Toledo DSC2 equipped with a HSS7 sensor and a TC-125MT intercooler. The average weight of the samples was 3–4 mg.
Thermal Gravimetric Analysis, TGA, measurements were carried out using a Mettler Toledo TGA/DSC 3+ under nitrogen atmosphere with a heating rate 20 °C/min in a temperature range 30 °C −900 °C and kept at 900 °C for 10 min in air atmosphere.
2.4. Thermal Conductivity
Thermal conductivity was investigated using a Hot Disk Thermal Constants Analyser 2500 S following the ISO Standard 22007-2. All the measurements were carried out at room temperature and repeated 5 times for each sample and an average value was taken.
Linear and nonlinear shear rheological tests were performed using an Anton Paar MCR702 TwinDrive (Graz, Austria) rotational rheometer in twin drive mode (separate motor-transducer) equipped with a convection oven (CTD450TD).
2.6. Melt Rheology
A 25 mm parallel plate geometry was used, with the gap set at 1 mm. The tests were performed at a measuring temperature of 200 °C. The nonlinear rheological analysis was performed using Fourier transform (FT) rheology, as a more sensitive method to detect rheological percolation [48
]. Given a sinusoidal strain input, in contrast to linear viscoelastic oscillatory measurements, the shear stress response is non-sinusoidal and therefore higher harmonics are recording in the corresponding Fourier spectrum, Figure 2
a. These higher harmonics give access to material nonlinear parameters, mainly through the use of the third relative higher harmonic, I3/1
, as it contains the dominant nonlinear contribution to the shear stress signal [48
]. The variation of I3/1
during a strain sweep test on the iPP matrix investigated is presented in Figure 2
b. At small strain amplitudes, the measured signal corresponds to instrumentation noise and indicates the sensitivity limits of the torque sensor. This region is called small amplitude oscillatory shear (SAOS). At a critical shear strain amplitude, the nonlinearities become detectable with I3/1 ∝ γ2
, region called medium amplitude oscillatory shear (MAOS) or intrinsic nonlinearity. The large amplitude oscillatory shear regime (LAOS) is reached when the quadratic scaling with the strain amplitude is lost.
2.7. Dynamic Mechanical Analysis (DMTA)
A rectangular fixture (SRF12) was used with pressed samples of 50 mm in length, 10 mm in width and thickness varying between 0.8–1.5 mm. A heating rate of 2 °C/min was applied between −40 °C to 200 °C, the measurements were performed at 1 Hz.
Scanning Electron Microscopy, SEM, was performed with a FEI Quanta 200. Fresh surfaces were prepared with a RMC ultracryomicrotome, and they were subsequently etched for one hour using a solution of 1 wt.% potassium permanganate in 86% ortho-phosphoric acid. The process was terminated by rinsing the samples with deionized water, followed by hydrogen peroxide and finally isopropanol. The etching has been performed in order to show clearly the iPP’s microstructure as well as the filler dispersion. Afterwards, approximately 5 nm thick Pd-Au layer was deposited onto the observed surfaces.
Electric Force Microscopy, EFM, was performed with an Asylum Research MFP-3D Atomic Force Microscope. Conductive tips, ASYELEC-01, were used. The samples were first scanned in AFM mode with free oscillation amplitude of ca. 1V and set point ratio of 0.7. The EFM scans were performed in nap mode. The tip tracked the sample surface, and was afterwards scanned again following the same profile, yet raised of 50 nm. In this second pass a voltage of 3 V was applied to the tip, and the tip scanned with very low amplitude. Electrical interaction between the sample and the DC voltage applied to the tip causes a force gradient which perturbs the oscillation frequency and as a result the phase shift.