Destruction and formation of a carbon nanotube network in polymer melts: Rheology and conductivity spectroscopy
Graphical abstract
Introduction
It is known over decades that filling of electrically insulating polymers even with small amounts (only few percent) of conducting particles results in an increase of the electrical conductivity of composites by orders of magnitude and can lead to a mechanical reinforcement [1], [2], [3]. Carbon nanotubes (CNTs) as fillers were found to improve electrical and mechanical properties of polymer matrices similar to carbon black (CB) particles with the advantage that for building up the conductive percolation network much lower weight content of CNT is needed [4], [5], [6], [7], [8], [9], [10], [11]. This is related to the high aspect ratio (ratio between length and diameter) of CNT (about 100–1000) compared to more spherical CB particles. This geometrical advantage as well as the huge nanotube stiffness and their high thermal and electrical conductivities makes CNT-based materials attractive for new applications. Therefore, carbon nanotube–polymer composites belong to a fast-developing field of material science, which is in close contact with the industrial needs.
The enhancements of thermal, electrical and mechanical characteristics of nanotube–polymer composites are attributed to the formation of a network of interconnected filler particles which can either conduct heat and electrical current or relax mechanical stress without a large matrix deformation [1], [2], [3]. Studies on thermoplastic polymers, melt compounded with both singlewalled (SWNT) and multiwalled carbon nanotubes (MWNTs), show electrical percolation at concentrations ranging from 0.05 vol% towards 5 wt% [4], [5], [6], [7], [8], [9], [10], [11], [12], [13], [14], [15]. The percolation concentrations were found to depend strongly on the nanotube length, diameter, degree of purification, bundling, type of matrix polymer and on the processing of the composite (e.g. temperature and mixing conditions, for example, see Refs. [12], [14], [16]). All these factors result in a wide variation in electrical conductivity and other material properties of the finished plastic products and appear to be one of the major restraints for a broad market acceptance of this new class of polymer nanocomposites.
Some years ago, our group reported on the influence of the extrusion conditions on the electrical conductivity of polycarbonate–MWNT mixtures using conductivity spectroscopy [11]. The conductivity measurements have shown the influence of screw speed and mixing time on the dispersion of the nanotubes. We assumed that even small geometrical changes in the local contact regions between the nanotubes – which are usually separated by polymer chains – can lead to considerable changes in the contact resistance and contact capacity. More recently it was shown by time-resolved conductivity measurements during isothermal annealing of pressed plates (polypropylene containing 2 wt% MWNT) that the thermal treatment above the melting temperature leads to an increase of the conductivity by about 10 orders of magnitude in 10 h [16]. This is an indication of the formation of a conductive network in the melt during annealing. A similar conductivity recovery was observed in a slit die for polypropylene containing 2 wt% MWNT which was flanged to the outlet of an extruder, after stopping extrusion for some time [17]. For a similar in-line setup, we found such conductivity recovery as well for polycarbonate and polyamide at different melt temperatures [18]. Zhang et al. [19] found a similar conductivity recovery during melt annealing of polyethylene/poly(methyl)methacrylate blends containing carbon black and carbon fibres in one of the phases. Using a combined rheological-dielectric setup the conductivity recovery was detected after a short shear deformation (shear rate dγ/dt = 1 s−1 for 10 s) for MWNT in polycarbonate in a well-defined laboratory experiment [20]. The decrease of the electrical conductivity with increasing shear rate was first reported by Kharchenko et al. [21] for polypropylene containing MWNT. Obrzut et al. also observed a shear-induced conductor–insulator transition of CNT in PP melts during shear [22]. These experiments have been performed under steady shear conditions. They also reported on a conductivity recovery after steady state shearing (dγ/dt = 6.3 s−1) was stopped. Using polarized light-scattering experiments on a weakly elastic melt, Hobbie et al. [23] showed that the tubes orient along the direction of flow already at low shear stresses, with a transition to vortices' alignment above a critical shear stress. More recently, Hobbie and Fry [24] measured the rheological properties of carbon nanotubes suspended in low-molecular-mass polyisobutylene using a polysuccinimide dispersant over a range of nanotube volume fractions. Using controlled strain rate and controlled stress measurements of yielding in shear flow, they proposed a universal scaling of both the linear viscoelastic and steady-shear viscometric responses.
The aim of this paper is to study the kinetics of destruction and reformation of a conductive CNT network in a polymer melt by simultaneous time resolved measurements of electrical conductivity and dynamic shear modulus during thermal annealing and after well-defined (short) shear deformations. For a quantitative description of the conductivity recovery after shear deformation a combined model of cluster aggregation and electrical percolation was developed [20]. For the agglomeration of CNT a second order kinetics was used, which was proposed for agglomeration of filler particles in a polymer matrix by Heinrich et al. [25]. The idea of conductive filler agglomeration was reported before by Schüler and co-workers [26], [27] for reactive epoxy mixtures containing carbon black.
Here we performed studies on polycarbonate filled with 0.6 vol% of MWNT, which is close to the concentration of electrical percolation. We expect that besides its practical interest for polymer–CNT composites, this study may contribute to the understanding of agglomeration of fillers in polymer melts in general.
Section snippets
Sample characterization and preparation prehistory
Polycarbonate composites filled with 0.6 vol% (corresponding to 0.875 wt%) of multiwalled carbon nanotubes (MWNTs) were prepared by melt dilution of a masterbatch containing 15 wt% MWNT (Hyperion Catalysis International, Cambridge, MA) using a DACA Microcompounder (DACA Instruments, Goleta, USA) as described in Ref. [13]. The extruded strands were compression molded into sheets with a thickness of about 600 μm at 265 °C for about 1 min at 50 kN [20]. The nanotubes were originally produced by chemical
Electrical conductivity
Fig. 3a and b presents the time evolution of the DC conductivities of two identical PC–MWNT samples containing 0.6 vol% MWNT measured for the “single” and “double shear” experiments, respectively. The time region t < 7200 s represents isothermal annealing of the fresh samples at 230 °C for 2 h. After annealing the DC conductivities for both samples are similar and reach values of about 10−4 S/cm.
At t ≈ 7200 s the “single shear pulse” or the “double shear pulse” was applied. During this time, in both
Summary and conclusions
In this paper a combined investigation of the electrical and rheological properties of PC–MWNT melts was performed using simultaneous dielectric and mechanical measurements during annealing of the as-received samples and in the rest time after different shear deformations. The shear deformation applied to the annealed composites containing 0.6 vol% MWNT leads to a tremendous decrease in the DC conductivity by 6 orders of magnitude as well as modulus decrease by a factor of 20 (i.e. down to the
Acknowledgments
This project was funded by the Bundesministerium für Wirtschaft und Arbeit via the Arbeitsgemeinschaft industrieller Forschungsgesellschaften (AiF Projects Nos. 122 ZBG and 144454N). We would like to thank Sven Pegel (IPF Dresden) for sample preparation and Wolfgang Böhm (DKI) for performing the combined DRS-DMA experiments. Furthermore, we thank Martin Engel (DKI and TU Darmstadt) for pre-experiments and for his contribution to the model. IA would like to thank Gert Heinrich for his hints and
References (55)
- et al.
Compos Sci Technol
(2005) - et al.
Polymer
(2003) - et al.
Polymer
(2005) - et al.
Polymer
(2003) - et al.
Polymer
(2004) - et al.
Polymer
(2008) - et al.
Polymer
(2006) - et al.
Phys Status Solidi B
(2007) Philos Mag B
(1987)- et al.
Polymer
(2004)
Macromolecules
Phys Rev B
Rubber Chem Technol
Adv Polym Sci
Rubber Chem Technol
Polym Compos
Adv Mater
Macromol Mater Eng
Adv Funct Mater
AIP Conf Proc
AIP Conf Proc
J Appl Polym Sci
Gummi Fasern Kunstst
Rubber Fibres Plast Int
Polymer
Nat Mater
Phys Rev B
Phys Fluids
Cited by (237)
Carbon nanotube network formation and configuration/morphology on reinforcing and conductive performance of polymer-based nanocomposites
2023, Composites Science and TechnologyPolymer nanocomposites for sensing applications
2023, Fundamentals of Sensor Technology: Principles and Novel DesignsSelective localization of carbon nanotubes and its effect on the structure and properties of polymer blends
2021, Progress in Polymer Science