Dispersability of multiwalled carbon nanotubes in polycarbonate-chloroform solutions

Polymer xxx (2014) 1e10

Contents lists available at ScienceDirect

Polymer journal homepage: www.elsevier.com/locate/polymer

Dispersability of multiwalled carbon nanotubes in polycarbonatechloroform solutions € tschke, Brigitte Voit Ulrike Staudinger*, Beate Krause, Christine Steinbach, Petra Po Leibniz-Institut für Polymerforschung Dresden e.V., Hohe Straße 6, 01069 Dresden, Germany

a r t i c l e i n f o

a b s t r a c t

Article history: Received 12 May 2014 Received in revised form 30 September 2014 Accepted 5 October 2014 Available online xxx

The dispersion of commercial multiwalled carbon nanotubes (MWCNTs, Nanocyl™ NC7000) in chloroform and in polycarbonate (PC)-chloroform solutions was investigated by variation of the polymer concentration, MWCNT amount and sonication time and compared with PC/MWCNT composites, which were processed by melt mixing, subsequently dissolved in chloroform and dispersed via sonication under the same conditions. The sedimentation behaviour was characterised under centrifugal forces using a LUMiSizer® separation analyser. The space and time resolved extinction profiles as a measure of the stability of the dispersion and the particle size distribution were evaluated. Sonication up to 5 min gradually increases the amount of dispersed particles in the solutions. A significant improvement of the MWCNT dispersion in chloroform was achieved by the addition of PC indicating the mechanism of polymer chain wrapping around the MWCNTs. In dispersions of melt mixed PC/MWCNT composites the dispersion of MWCNTs is significantly enhanced already at a low sonication time of only 0.5 min due to very efficient polymer wrapping during the melt mixing process. However, the best dispersion quality does not lead to the highest electrical conductivity of thin composite films made of these PC/MWCNT dispersions. © 2014 Elsevier Ltd. All rights reserved.

Keywords: Carbon nanotubes Dispersion Polymer solutions

1. Introduction For the development of electrically conductive carbon nanotube (CNT) filled polymer nanocomposites a sufficient dispersion of the CNTs in the polymer matrix is essential. A good quality of dispersion strongly depends on various parameters like the type of CNT and its geometrical dimensions, the surface properties of CNTs and polymer, the matrix viscosity and the dispersion method, to name but a few. The dispersion of CNTs in thermoplastic polymers via melt mixing and the significant impact of melt mixing parameters like temperature, screw speed or mixing time on the dispersion of CNT and therefore on the electrical properties of the composites have been intensively studied [1e7]. Further the aspects of polymer viscosity [8,9] and choice of CNT type were discussed by several authors [8,10e13]. To develop thin films of electrically conducting polymer/CNT nanocomposites processing from solution like solution casting, dip coating or spin coating is required. The dispersion of CNTs in organic solvents is supported by sonication, a process which generates cavitation to break the CNT agglomerates, instead of shear forces generated by the mixing elements during melt

* Corresponding author. Tel.: þ49 351 4658 646; fax: þ49 351 4658 565. E-mail address: [email protected] (U. Staudinger).

mixing. Parameters like sonication time, frequency and amplitude significantly influence the degree of CNT dispersion. For many applications surface modification of CNTs is an essential tool to effectively disperse CNTs in solvents and polymers [14]. A number of studies have been published within the last 10 years, characterising the dispersability of CNTs in organic solvents in correlation with their surface properties, considering the solubility parameters of nanotubes and solvents. Up to now, these efforts have mainly concentrated on dispersions with singlewalled carbon nanotubes (SWCNTs) [15e18]. The dispersion of multiwalled CNTs (MWCNTs) in various organic solvents was discussed by Detriche et al. [19,20]. Few studies focus on the dispersion of CNTs in polymer solutions [21,22]. Some authors discussed polymer wrapping at CNTs, which was specifically used to improve the dispersion of CNTs in water [23] or organic solvents [24]. Theoretical studies reported the effective wrapping of polymers containing aromatic phenyl rings due to a high binding energy of the polymer resulting in high interfacial adhesion with SWCNTs [25] and postulated a dependency of the interaction strength of the CNT with the polymer on the number of repeating units in the polymer [26]. The group of Szleifer and Yerushalmi-Rozen suggested the decoration of SWCNTs with end-tethered polymers as the entropic repulsion among the tethered chains generates a free energy barrier that prevents the SWCNTs from approaching [27,28].

http://dx.doi.org/10.1016/j.polymer.2014.10.012 0032-3861/© 2014 Elsevier Ltd. All rights reserved.

Please cite this article in press as: Staudinger U, et al., Dispersability of multiwalled carbon nanotubes in polycarbonate-chloroform solutions, Polymer (2014), http://dx.doi.org/10.1016/j.polymer.2014.10.012

2

U. Staudinger et al. / Polymer xxx (2014) 1e10

The quality of the CNT dispersion in a polymer solution strongly depends on the type of polymer and solvent which are used, the polymer concentration and CNT content in the solvent, and the ultrasonic treatment of the dispersion. In the present work the dispersion of non-functionalized MWCNTs in chloroform and polycarbonate(PC)-chloroform solutions was investigated and compared with PC/MWCNT composites, which were pre-treated via melt mixing and subsequently dissolved in chloroform. This pre-treatment was performed to further improve the MWCNT dispersion and consequently enhance the electrical conductivity of the prepared films. To qualify the dispersion stability and the particle size distribution of the MWCNTs in the solutions a centrifugal separation analysis was performed. This well-established method [29,30] was already applied by Krause et al. [31] to investigate the dispersability of different types of CNTs in aqueous surfactant solutions using a LUMiSizer® centrifugal separation analyser. Thereby a correlation between the stability of the aqueous MWCNT dispersions and the dispersability of the nanotubes in polymer melts of polyamide 6.6 was found. A conformance of the sedimentation behaviour of aqueous MWCNT dispersions with the MWCNT dispersion in polycarbonate melts was also shown by Pegel et al. [10]. A further study focussed on the dispersability of four different commercially available MWCNT materials in aqueous surfactant solutions as a function of ultrasonic treatment time [32]. In that work best dispersabilities were found for MWCNTs of type Nanocyl™ NC7000 and Future Carbon CNT-MW. During the last years centrifugal analysis has become a favoured tool to study various types of CNT dispersions [33e35]. 2. Experimental An overview about the experimental study is given in Fig. 1. The experimental details are described in the following sections. 2.1. Materials Commercially available multiwalled carbon nanotubes NC7000 from Nanocyl™ (Sambreville, Belgium) were used for this study.

They were produced in an industrial large-scale catalytic vapour deposition process and have an average diameter of 10 nm, an average length of 1.3 mm [12], a carbon purity of 90% and a surface area of 250e300 m2/g [36]. As found in former studies NC7000 are characterised by a good dispersability in various thermoplastic polymer melts [8,10] and in aqueous dispersions [31,32] compared to other types of MWCNTs. The polymer used in this study was a medium viscosity grade PC Makrolon® 2600 from Bayer MaterialScience AG, Germany. The MWCNTs and PC were dried at 120  C for 4 h in a vacuum oven before processing.

2.2. Preparation of MWCNT-polymer-solvent-dispersions Three different types of dispersion were prepared (see Table 1). For the first one, MWCNTs NC7000 were added to polycarbonatechloroform solutions with the polymer concentration in the chloroform varying between 0.25, 0.5 and 1 wt%. The MWCNT content was fixed to 1 wt% referred to the added polymer, i.e. the MWCNT amount in the solution was increasing from 0.037 g/l to 0.74 g/l to 0.148 g/l with increasing polymer concentration. The dispersions were named as PCCL dispersions. The attached number refers to the polymer concentration, e.g. PCCL025 contains 0.25 wt% of PC in chloroform, PCCL050 contains 0.5 wt% and PCCL100 contains 1 wt% of PC in the solvent. As reference, equal MWCNT amounts were added to pure chloroform under the same conditions. These dispersions are named as CL dispersions. To remain consistent in the notation of the samples the numbers 025, 050 and 100 are also used for the CL dispersions but due to absence of polymer they refer to the MWCNT content which is equal to the corresponding PCCL dispersions. For the third type of dispersion, melt mixed PC/MWCNT composites with 1 wt% of MWCNT were solved in chloroform in an analogous manner. The compounding process is described in Section 2.3. The dispersions made from melt mixed PC/MWCNT composites are called as PCcomp dispersions. The amount of solvent used for the dispersion of the MWCNTs was fixed to 10 ml. To disperse the MWCNTs in the solvents and in the polymer solutions sonication was applied at room temperature for 0.5, 1, 2, 5 and 10 min using an UP200S processor (Hielscher Ultrasonics GmbH, Teltow, Germany) with a maximum frequency

Fig. 1. Scheme of the preparation and analysis of the studied MWCNT dispersions.

Please cite this article in press as: Staudinger U, et al., Dispersability of multiwalled carbon nanotubes in polycarbonate-chloroform solutions, Polymer (2014), http://dx.doi.org/10.1016/j.polymer.2014.10.012

U. Staudinger et al. / Polymer xxx (2014) 1e10 Table 1 Characteristics of the investigated CL, PCCL and PCcomp dispersions. Dispersion

Polymer concentration [wt%]

Polymer weight [mg]a

MWCNT weight [mg]b

CL025 CL050 CL100 PCCL025 PCCL050 PCCL100

e e e 0.25 0.5 1

e e e 37 74 148

0.37 0.74 1.48 0.37 0.74 1.48

Composite concentration [wt%]

Composite weight [mg]a/b

0.25 0.5 1

37 74 148

PCcomp025 PCcomp050 PCcomp100 a b

Based on 10 ml CHCl3. 1 wt% MWCNTs related to the added polymer.

of 24 kHz and a power of 200 W, adjusting an amplitude of 20%. In the figures of the present paper the abbreviation US is used for ultrasonication. 2.3. Melt mixing of PC/MWCNT composites Melt mixing of the PC/MWCNT composites was performed applying a masterbatch dilution process using a co-rotating ZE25 twin screw extruder (Berstorff, Germany) having a screw diameter of 25 mm and a barrel length of 900 mm (L/D ¼ 36). First, a masterbatch with 7.5 wt% MWCNT was produced by simultaneously feeding the PC granulate and MWCNT powder into the hopper by the use of gravimetric dosing. Second, pellets of the masterbatch were diluted with PC granulates into varying compositions containing 0.125 to 5 wt% MWCNTs. The extrusion conditions were adjusted according prior related studies [2,3,37,38]. For this study a PC/MWCNT composite with 1 wt% MWCNT is used, named as PCcomp. This composite was found to be electrically conductive [12,38], as the percolation threshold occured at a MWCNT content lower than 0.5 wt%. 2.4. Sedimentation analysis The sedimentation behaviour of the MWCNT dispersions was studied under centrifugal forces using a centrifugal separation analyser (CSA) LUMiSizer 6100-29 (L.U.M. GmbH Berlin). This microprocessor controlled analytical centrifuge simultaneously detects the intensity of the transmitted light of a NIR light source as function of time and position over the entire sample cell length. At the same time up to 12 samples can be analysed. The space and time resolved extinction profiles are evaluated between the bottom and the fluid level of the dispersion to characterise the stability of the dispersion and the particle size distribution. The type of sedimentation can be evaluated from the shape of the transmission profiles. Horizontal profiles refer to a polydisperse dispersion with very low particle concentration, where the sedimentation occurs unhindered and the particles sediment because of their size. At higher particle concentration agglomerates and networks will be formed, which are constraining each other and sediment faster than the primary particles. In this case the transmission profiles consist of parallel vertical curves. Vertical curves are also obtained in the case of monomodal dispersions if rigid spherical particles of same size and density migrate in the centrifugal field [39]. Within this study the MWCNT dispersions were put into the measurement cells immediately after the ultrasonic treatment. The centrifugation was applied for 45 min at room temperature at a centrifugal speed of 1000 rpm. To evaluate the dispersion stability the integration of

3

the transmission profiles was performed in the middle region of the cell between the positions 111 and 124 mm. The initial slope of the integral transmission correspond to the sedimentation rate of the particles. In the case of good particle dispersion the stability of the dispersion is high, resulting in a low sedimentation rate. If the particle dispersion is not sufficient, the remaining large agglomerates settle fast which corresponds to a high sedimentation rate. 2.5. Characterisation of particle size distribution The particle size distribution was calculated from the transmission profiles obtained from the centrifugal analysis. Therefore the transmission values have to be transformed into extinction values by dividing the measured transmission T by the transmission of the sample cell filled with the dispersion media only. The particle size for spherical particles can be calculated for unhindered sedimentation in a laminar regime using Stokes law, considering the fluid density, the dynamic viscosity, starting and measurement position of the particle and the measurement time. By means of the software SEPView the velocity distribution was calculated and the particle size distribution at constant position was evaluated according to ISO 13318. A second method to determine the particle size distribution was to use dynamic light scattering (DLS, Zetasizer Nano SZ, Malvern Instr., UK) according to ISO 22412. This method is used to analyse particles with a size between 0.6 nm and 6 mm [40]. The particles were illuminated with a laser and the scattered light intensity fluctuates depending on the size of the particles. Smaller particles move more rapidly than larger ones, caused by Brownian motion of the particles. The velocity of the Brownian motion is used to calculate the particle size using the StokeseEinstein relationship. 2.6. Morphological characterisation For transmission electron microscopy (TEM) studies thin films of selected PC/MWCNT composites were prepared by spin coating of the dispersion at 5000 rpm on silicon substrates for 30 s directly after sonication. A part of the film was transferred to a TEM grid by the use of a flotation technique. Small squares of 2 mm  2 mm were scribed on the polymer film and removed from the silicon wafer by etching the native oxide layer of the substrate with a 1 M NaOH solution (4 g NaOH in 100 ml H2O). The floated film was picked up by a perfect loop and placed onto a TEM grid. To remove the NaOH completely, the film on the grid was cleaned with Millipore water for three times. To investigate the MWCNT dispersion in the composite film a LIBRA® 120 microscope, Carl Zeiss AG, Germany with an acceleration voltage of 120 kV was used. 2.7. Electrical characterisation The electrical characterisation was performed on films made from reference solutions of the PCCL and PCcomp dispersions. For each polymer concentration three identical solutions were prepared to study the reproducibility of the dispersion quality and therefore of the electrical conductivity of the composite films. After ultrasonic treatment of 0.5, 1, 2, 5 and 10 min of each dispersion thin films of ca. 5e10 mm were prepared by drop casting of 100 ml of the dispersion on a glass substrate. To measure the surface resistivity of the films the Loresta electrometer from Mitsubishi with a measuring range between 103 and 107 U was used for samples with resistances < 107 U. The Hiresta electrometer from Mitsubishi with a measuring range between 104 and 1013 U was used for samples with resistances > 107 U. The Hiresta was connected to a ring electrode (URS) with the inner electrode having a diameter of 5.9 mm and the exterior electrode having an inner and outer

Please cite this article in press as: Staudinger U, et al., Dispersability of multiwalled carbon nanotubes in polycarbonate-chloroform solutions, Polymer (2014), http://dx.doi.org/10.1016/j.polymer.2014.10.012

4

U. Staudinger et al. / Polymer xxx (2014) 1e10

MWCNT dispersion in the polymer solution. For the PCCL025 dispersion a marginal sedimentation and the formation of a sediment were determined. For the PCcomp025 dispersion no significant changes in the transmission profiles with centrifugation time were detectable. The transmission profiles obtained were horizontal caused by very broad particle size distributions. This dispersion contains particles which settle separately dependent on their size with their own sedimentation velocity. In Fig. 3 the integral transmission of the investigated dispersions, which is the integration of the transmission profiles in the middle position between 111 and 124 mm, is plotted as a function of centrifugation time. On the left the CL dispersions are shown. In the middle the corresponding PCCL dispersions are presented. On the right the PCcomp dispersions are illustrated. From top to bottom diagram the MWCNT concentration varies between 0.37, 0.74 and 1.48 mg/10 ml. Each diagram includes the integral transmission curves of five dispersions varying in their sonication times between 0.5 and 10 min. For PCcomp dispersions only four curves for sonication times between 0.5 and 5 min are presented. It was abdicated to apply sonication of 10 min to pre-treated dispersions as no further improvement of the dispersion quality was expected. Four parameters, which are influencing the integral transmission and therefore the sedimentation behaviour of the dispersions can be specified from these results: (a) the MWCNT concentration, (b) the sonication time, (c) the addition of polymer and (d) the pre-dispersion of MWCNTs in melt.

diameter of 11 mm and 18 mm respectively, whereas the Loresta was used with a 4-point electrode (ESP) with an electrode distance of 5 mm and an electrode diameter of 2 mm. On each film three measurements were conducted and the arithmetic average and the standard deviation were calculated. 3. Results 3.1. Sedimentation behaviour of MWCNTs in chloroform and polymer solutions The sedimentation behaviour of MWCNTs in chloroform and in polymer solutions was studied using the CSA. During the CSA measurements the transmission profiles were recorded every 10 s as a function of position. In Fig. 2 the transmission profiles of a CL025 dispersion after 1 min of sonication are plotted and compared with the transmission profiles of a PCCL025 dispersion and a PCcomp025 dispersion. For the CL025 dispersion a combination of horizontal and vertical profiles was observed (Fig. 2a). The successive increase of transmission during centrifugation describes a polydisperse dispersion with a gradual sedimentation of the particles depending on their size. This process is superposed by the presence of bigger CNT agglomerates and networks which are constraining each other and sediment faster than the primary particles causing vertical curves in the transmission profiles. Additionally a sediment was formed which was compressed with the centrifugation time. In contrast, the levels of the transmission profiles of the PCCL025 and PCcomp025 dispersions were significantly lower (Fig. 2b, c), which can be attributed to a very good MWCNT dispersion with marginal sedimentation of the particles concluding a high stability of the

90 80 70 60

last profile

50 40 30 20 10 0

100

CL025 - 1 min US

a

normalised transmission (%)

normalised transmission (%)

100

(a) The MWCNT concentration distinctly affects the transmission level of the dispersion as it can be observed for all three types of investigated dispersions. The less the MWCNT concentration in the solvent the larger the transmission of

PCCL025 - 1 min US

b

80 70 60 50 40 30

first profile

last profile

20 10

first profile 112

90

114

116

118

120

122

0

124

112

114

position (mm) 100

normalised transmission (%)

90

116

118

120

122

124

position (mm)

c

PCcomp025 - 1 min US

80 70 60 50 40 30 20 10 0 112

114

116

118

120

122

124

position (mm) Fig. 2. Normalised transmission profiles of MWCNT-chloroform and PC/MWCNT-chloroform dispersions with a MWCNT concentration of 0.37 mg/10 ml or 1 wt% referred to the polymer respectively after sonication of 1 min applying 45 min of centrifugation at 1000 rpm. a) CL dispersion b) PCCL dispersion, c) PCcomp dispersion.

Please cite this article in press as: Staudinger U, et al., Dispersability of multiwalled carbon nanotubes in polycarbonate-chloroform solutions, Polymer (2014), http://dx.doi.org/10.1016/j.polymer.2014.10.012

U. Staudinger et al. / Polymer xxx (2014) 1e10

0.5 min US 1 min US

2 min US 10 min US 5 min US

2500

3000

1 min US 2 min US 10 min US 5 min US

2500

3000

1000

1500

2000

2500

centrifugation time (sec)

60 55 50 45 40 35 30 25 20 15 10 5 0

60

0.5 wt% PC in 10 ml CHCl3 + 0.74 mg MWCNT

55

PCCL050

0.5 min US

1 min US 2 min US 5-10 min US

0

500

1000

1500

2000

3000

2500

1500

2000

2500

3000

0.5 wt% of PC/1wt% MWCNT composite in 10ml CHCl3

50 PCcomp050 45 40 35 30 25 20 15 10

0.5 - 5 min US

5 0

3000

60 55 1 wt% PC in 10 ml CHCl3 + 1.48 mg MWCNT 50 PCCL100 45 40 35 30 25 20 0.5 min US 15 10 1-10 min US 5 0 0 500 1000 1500 2000 2500 3000

centrifugation time (sec)

1000

centrifugation time (sec)

0

500

1000

1500

2000

2500

3000

centrifugation time (sec) 60 55

integral transmission (%)

2 min US 1 min US 0.5 min US 5 min US 10 min US

integral transmission (%)

integral transmission (%)

CL100

500

500

centrifugation time (sec)

1.48 mg MWCNT in 10 ml CHCl3

0

0

integral transmission (%)

0.5 min US

centrifugation time (sec) 60 55 50 45 40 35 30 25 20 15 10 5 0

60 0.25 wt% of PC/1wt% MWCNT composite in 10ml CHCl3 55 50 PCcomp025 45 40 35 30 25 20 15 10 0.5 - 5 min US 5 0

centrifugation time (sec)

integral transmission (%)

integral transmission (%)

centrifugation time (sec) 60 55 0.74 mg MWCNT in 10 ml CHCl3 50 CL050 45 40 35 30 25 20 15 10 5 0 0 500 1000 1500 2000

PCcomp dispersions

60 55 0.25 wt% of PC in 10 ml CHCl3+ 0.37 mg MWCNT 50 PCCL025 45 40 35 0.5 min US 30 25 20 1 min US 15 2 min US 10 5 min US 5 10 min US 0 0 500 1000 1500 2000 2500 3000

integral transmission (%)

PCCL dispersions

60 55 0.37 mg MWCNT in 10 ml CHCl3 50 CL025 45 40 35 30 25 20 15 10 5 0 0 500 1000 1500 2000

integral transmission (%)

integral transmission (%)

CL dispersions

5

50

1 wt% of PC/1wt% MWCNT composite in 10ml CHCl3

PCcomp100

45 40 35 30 25 20 15 10 0.5 - 5 min US

5 0

0

500

1000

1500

2000

2500

3000

centrifugation time (sec)

Fig. 3. Integral transmission versus centrifugation time of the investigated CL dispersions (left), PCCL dispersions (middle) and PCcomp dispersions (right), with increasing MWCNT or/ and PC concentration from top to bottom.

the dispersion due to a more efficient dispersion of the MWCNTs. As discussed in literature for SWCNTs [15,41] a limit of dispersion can be calculated as a measure of the dispersability of CNTs in a solvent. Above a critical CNT concentration the agglomeration of CNTs dominates the dispersion, which is considered to be the dispersion limit of the CNTs. Furthermore it has to be considered, that the energy input per particle caused by sonication is influenced by the particle concentration in the solvent. At fixed sonication conditions a better dispersion will be expected in dispersions with lower particle concentration. (b) The sonication time strongly influences the MWCNT dispersion. As it significantly occurs in CL and PCCL dispersions, the increase in sonication time causes a distinct decrease in the integral transmission, which is attributed to a higher content of dispersed MWCNT in the solution followed by an enhanced stability of the dispersion. A longer sonication treatment leads to a larger amount of broken agglomerates and individualized particles. Sonication up to 5 min gradually increases the amount of dispersed particles in the (polymer) solutions. Additional sonication treatment up to 10 min does not lead to a further decrease of the integral transmission, which indicates a maximum stability

which is reached after 5e10 min of sonication. In contrast, in PCcomp dispersions the dispersion of MWCNTs is significantly enhanced already at a low sonication time of only 0.5 min as it can be observed from the low integral transmission in Fig. 3 (right). It has to be noted, that sonication can also cause shortening of individualized MWCNTs, which will lead to a further improvement of the MWCNT dispersion. (c) The addition of the polymer to the solvent results in a considerable improvement of the MWCNT dispersion already at low sonication times as it is illustrated by the distinct decrease of the integral transmission of PCCL and PCcomp dispersions compared to CL dispersions. It can be assumed that polycarbonate chains wrap around the dispersed MWCNTs which is attributed to the high binding energy of the polymer containing aromatic phenyl rings resulting in a high interfacial adhesion with the MWCNTs as discussed by Linton et al. [25] for polystyrene wrapping at SWCNTs. Covered by a polymer layer, the MWCNTs are not able to rebundle due to an entropic repulsion among the polymer-decorated MWCNTs which exceeds the van der Waals interactions of the nanotubes [27]. (d) The pre-dispersion of MWCNTs in melt results in a distinctly improved dispersion quality of the PCcomp

Please cite this article in press as: Staudinger U, et al., Dispersability of multiwalled carbon nanotubes in polycarbonate-chloroform solutions, Polymer (2014), http://dx.doi.org/10.1016/j.polymer.2014.10.012

U. Staudinger et al. / Polymer xxx (2014) 1e10

2500

sedimentation rate (%/hrs)

a

CL025 CL050 CL100

2000 1500 1000 500

b 300

sedimentation rate (%/hrs)

6

PCCL025 PCCL050 PCCL100 PCcomp025 PCcomp050 PCcomp100

250 200 150 100 50 0

0 0

2

4

6

8

0

10

2

4

6

8

10

sonication time (min)

sonication time (min)

Fig. 4. Sedimentation rate in (a) CL dispersions and (b) PCCL and PCcomp dispersions as a function of sonication time.

In Fig. 4 the sedimentation rate is displayed as a function of sonication time for CL dispersions (Fig. 4a) and PCCL and PCcomp dispersions (Fig. 4b). The sedimentation rate of the MWCNTs can be characterised by determining the initial slope of the integral transmission curves in Fig. 3 and is given in %/hrs. A low initial slope indicates a low sedimentation rate during the measurement corresponding to a high stability of the dispersion [39]. In polymer solutions the sedimentation rate of MWCNTs was found to be

particle diameter D50 (nm) (LUMiSizer)

a

CL025 CL050 CL100

1200 1000 800 600 400 200 0

0

2

4

6

sonication time (min)

8

10

significantly lower than in pure chloroform, indicating a higher stability of the MWNT dispersions caused by the addition of the polymer. As expected from the previous results an increasing sonication time leads to a distinct decrease of the sedimentation rate and therefore to a significant stabilisation of the MWCNT dispersion. The sedimentation rate of the MWCNT particles in all PCcomp dispersions ranges between 0.17 and 0.65 %/hrs, which is very low compared to the sedimentation rates of the PCCL dispersions varying between 3.66 %/hrs for PCCL100 after 5 min of sonication and 241 %/hrs for PCCL025 after 0.5 min of sonication. Due to the very low values no correlation between the sedimentation rate and sonication or even the composite concentration in the solvent could be drawn for the PCcomp dispersions which have a high dispersion stability already after 0.5 min of sonication. 3.2. CNT agglomerate size The average particle diameter (D50) obtained from the intensity weighted size distribution measured by CSA for MWCNT agglomerates in CL, PCCL and PCcomp dispersions is plotted in Fig. 5 as a function of sonication time. The D50 value of the intensity weighted particle size distribution shows a distinct dependence on the sonication time of dispersions in chloroform and in PC solutions as the particle size shows a strong decrease between 0.5 and 2 min of sonication. The effect of this particle size reduction is more pronounced in CL dispersions (Fig. 5a) than in PCCL dispersions (Fig. 5b) and is larger at lower MWCNT concentration. However, the level of particle size after 2 min of sonication is about 1.5 times higher in chloroform than in PC solutions. Sonication up to 5e10 min does not lead to a further distinct reduction of D50.

b particle diameter D50 (nm) (LUMiSizer)

dispersions compared to PCCL dispersions. PCcomp dispersions show very low and stable transmission during centrifugation already after 0.5 min of sonication. These results indicate a very effective MWCNT dispersion and polymer wrapping mechanism around the MWCNTs, taking place during the melt mixing process as already proposed by Hunley et al. [42] for MWCNTs in polyurethane melts. Furthermore, it has to be considered, that during melt mixing shortening of the MWCNTs takes place as it was quantified by Krause et al. [43] who measured the nanotube length of NC7000 and Baytubes® C150HP before and after melt processing. A significant nanotube shortening after melt processing with PC up to 30% of the initial values was found. Socher et al. [9] investigated the influence of matrix viscosity on MWCNT dispersion and electrical properties in different thermoplastic nanocomposites, varying the type of polymer and its viscosity. In polycarbonate they found a MWCNT (Baytubes® C150P) shortening to 45e54 % of the initial length, depending on the viscosity of the matrix. Both, pre-dispersion of the MWCNTs in the polycarbonate melt and the additional shortening of the nanotubes during melt processing, form the basis of the high dispersion quality in the solvent.

PCCL025 PCCL050 PCCL100 PCcomp025 PCcomp050 PCcomp100

1200 1000 800 600 400 200 0

0

2

4

6

8

10

sonication time (min)

Fig. 5. Particle diameter D50 as a function of sonication time evaluated from CSA measurements of (a) CL dispersions and (b) PCCL and PCcomp dispersions.

Please cite this article in press as: Staudinger U, et al., Dispersability of multiwalled carbon nanotubes in polycarbonate-chloroform solutions, Polymer (2014), http://dx.doi.org/10.1016/j.polymer.2014.10.012

U. Staudinger et al. / Polymer xxx (2014) 1e10

For PCcomp dispersions D50 values could not be evaluated for all samples. Due to the high stability of these dispersions only the sedimentation of bigger particles was detected as reflected by the particle size values shown in Fig. 5b. The measured particle size is partially quite large and could be attributed to few big remaining MWCNT agglomerates, or even to air bubbles or dust particles present in the dispersions. As the particle size measurement is conducted at four constant positions of the sample cell, particles which remain very stable during centrifugation will not be detected. Consequently, the D50 values evaluated by CSA represent a mean MWCNT agglomerate size disregarding very small particles which are stabilized in the dispersion. Additionally, it has to be considered that the detection of the particle velocity could be constrained by the very low transmission level of these dispersions as they appear as “black solutions”. To obtain a more precise estimation about the MWCNT agglomerate size in the dispersions the particle size distribution of selected PCCL and PCcomp dispersions were additionally measured by DLS. For PCCL dispersions the polydispersity index (PDI) and the intensity weighted size distribution for samples with 5 and 10 min of ultrasonic treatment, which show the largest dispersion stability from the centrifugal analysis, are summarized and compared with the D50 value evaluated from CSA as presented in Table 2. The PDI describes the size distribution of the colloidal particles in a system. At ideal monodisperse size distributions the PDI is zero. Because of the high polydispersity of the MWCNT dispersions it is reasonable to discuss the size distribution weighted by intensity, rather than the z-average particle diameter, which is calculated by the cumulant method assuming a monomodal distribution of the particles, which is not the case in this study. The mean particle size dave,I, weighted by intensity, increases with increasing MWCNT concentration in the polymer solution. A second and even third peak could be measured as the MWCNT concentration rises up to 1.48 mg/10 ml, considering also very large MWCNT agglomerates of a mean particle size up to 5 mm (however, which can be partly attributed to individual dust particles or air bubbles) and a small fraction with a size of ca. 76 nm in the case of the 10 min sonicated PCCL100. The large PDI values in the PCCL dispersions varying between 0.29 and 0.51 indicate the size heterogeneity of the MWCNT agglomerates present in the dispersions. The mean particles size values measured by DLS were distinctly higher than the D50 values measured by CSA. As already discussed by Krause et al. [32] in CSA the particles will be oriented in the centrifugal field and the obtained sedimentation velocity is related to the lowest diameter of the non-spherical dispersed MWCNT particles. In contrast, in DLS larger particles and aggregates are overestimated and it has to be considered that this may also apply

Table 2 Particle size distribution and polydispersity index (PDI) in PCCL dispersions obtained from DLS; D50 value obtained from CSA for comparison (see also Fig. 5b). Dispersion

Sonication time (min)

dave,I (nm)

% by intensity

PDI

D50 (nm)

PCCL025

5 10

267.7 283.0 4224 323.0 4631 342.0 4957 468.6 4576 75.9 444.9 5162

100 90.6 9.4 88.7 11.3 95.9 4.1 97.0 3.0 2.2 92.5 5.3

0.35 0.38

161 196

0.51

234

PCCL050

5 10

PCCL100

5 10

0.38

221

0.29

208

0.36

318

7

to the non-spherical MWCNTs and MWCNT agglomerates studied in this case. It was found that an increase of the sonication time from 5 to 10 min has no significant influence on the particle size. This finding correlates with the values of the sedimentation rate (see Fig. 4b) and the integral transmission (see Fig. 3 middle). Thereby a plateau was reached for the MWCNT dispersion quality and stability which could not be further improved by additional energy input by sonication treatment. This effect was also described by Krause et al. [32] for aqueous surfactant solutions containing MWCNTs. The mean particle size measured by DLS and the PDI of PCcomp025 dispersions for different sonication times is shown in Table 3. For each sonication time two samples were measured for repeat determination. The good agreement of the measured values shows the good repeatability of the DLS measurement. Compared to PCCL025 dispersions the PDI of PCcomp025 dispersions was found to be significantly reduced to values of 0.17e0.20, which is attributed to the better dispersability of the MWCNTs in the PCcomp dispersions due to the previous melt mixing process. The particle size was found to be independent on the applied sonication time of 0.5e5 min. This correlates with the values of the sedimentation rate (see Fig. 4b) and the integral transmission (see Fig. 3 right) which remain on a very low level already after 0.5 min of sonication and do not change significantly with further sonication treatment. In comparison to the PCCL dispersions, which contain the same MWCNT and PC content, the MWCNT dispersion stability in the PCcomp dispersions is distinctly enhanced due to the pre-dispersion of the MWCNTs in PC via melt mixing.

3.3. Electrical surface resistivity of PC/MWCNT composite films The results of the electrical surface resistivity measurements of thin films made from the PCCL and PCcomp dispersions are shown in Fig. 6. Assuming a sufficient dispersion of the MWCNTs in the polymer solution the prepared composite films should have a MWCNT content of ca. 1 wt%, which can be calculated from the polymer and CNT concentration in the solvent. It is expected that the composite films are electrically conductive, because the percolation threshold of comparable melt mixed PC/MWCNT composites had been found at a MWCNT content of 0.5 wt% [12,38]. To achieve high electrical conductivity in the composites the dispersion of the primary nanotube agglomerates on one hand and the formation of a conductive network by so-called secondary agglomeration on the other hand is required as discussed by Alig et al. [44]. Whereas in Ref. [44] the secondary agglomeration is described as a process occurring in the quiescent melt or under shear deformation, in this case it occurs in the process of solution casting during the evaporation of the solvent.

Table 3 Particle size distribution and PDI in PCcomp025 dispersions for different sonication times, obtained from DLS; measurement of two samples for each sonication time (repeat determination). PCcomp025 Sonication (min)

Sample no.

dave,I (nm)

PDI

0.5

1 2 1 2 1 2 1 2

237.5 228.7 207.4 214.6 221.6 211.2 266.8 264.4

0.19 0.17 0.19 0.19 0.20 0.19 0.20 0.20

1 2 5

Please cite this article in press as: Staudinger U, et al., Dispersability of multiwalled carbon nanotubes in polycarbonate-chloroform solutions, Polymer (2014), http://dx.doi.org/10.1016/j.polymer.2014.10.012

8

U. Staudinger et al. / Polymer xxx (2014) 1e10 14

surface resistivity (Ω/sq.)

10 13 10 10

PCCL025 PCCL050 PCCL100 PCcomp100

10

10

9

10

8

10

7

10

6

10

5

10

4

0

2

4

6

8

10

sonication time (min) Fig. 6. Surface resistivity of solution cast films of PCCL and PCcomp dispersions.

In Fig. 6 the surface resistivity of the PC/MWCNT films is plotted versus sonication time comparing the PCCL dispersions at the three different polymer concentrations. For each type of composite an increasing sonication time tends to result in a decrease of the surface resistivity due to the better distribution and dispersion of the MWCNTs in the polymer solution. In PCCL025 the high dispersion quality dominates the medium level of the electrical resistivity of the composite film. However, significant fluctuations of the resistivity can be observed in films of PCCL050 dispersions. Larger MWCNT concentration in PCCL050 dispersions reduces the energy input per particle during sonication and therefore the dispersability. This results in the reduction of the dispersion and in partial hindrance of the MWCNT network formation during film preparation. Thus, the composite films have a quite inhomogeneous distribution of the MWCNTs, causing the measurement of partly high resistivity values combined with a large standard deviation. With further increase of the MWCNT content in the solution (PCCL100) the larger volume of CNT agglomerates suffices to form a conductive MWCNT network despite declined dispersion quality. Moreover it has to be considered that the film thickness slightly increases with increasing polymer concentration in the solvent

which can also influence the resistivity values. Lowest resistivity values of about 5*104 U/sq could be obtained for composite films of PCCL100 dispersions after 2 min of sonication. In films of PCcomp dispersions the generation of a conducting MWCNT network is constrained due to their very good dispersion, including polymer wrapping around the MWCNTs, which separates the MWCNTs from each other, and the shortening of the MWCNTs during melt mixing. All investigated melt mixed composites exhibit high surface resistivity values of >1013 U/sq. Films of PCcomp025 and PCcomp050 dispersions exhibit even higher resistivity values which are outside the measuring range of the Hiresta electrometer. The excellent MWCNT dispersion results in the formation of homogeneous films with good transparency. TEM images of spincoated PC/MWCNT films are illustrated in Fig. 7. On the left a thin film made of a PCCL100 dispersion after 5 min of sonication is shown. The image exemplifies the composite structure, which contains not only regions with well dispersed MWCNTs but also areas of agglomerated MWCNTs. On the right the well dispersed MWCNTs in a composite film of a PCcomp100 dispersion after 5 min of sonication are presented. Extended experimental work on the development of conductive films revealed that an increase of the MWCNT content to 2 wt% in the composite is sufficient to reduce the surface resistivity to about 105 U/sq, while maintaining the well dispersed MWCNT network and thus the transparency of the composite film. Further details on that will be the issue of a future publication.

4. Conclusion In the present study the dispersion of commercial MWCNTs NC7000 in pure chloroform and in PC-chloroform solutions was investigated by variation of the polymer concentration, MWCNT concentration and sonication time and compared with PC/ MWCNT composites, which were processed by melt mixing, subsequently dissolved in chloroform and dispersed via sonication under the same conditions. The sedimentation behaviour of the dispersions was studied using a LUMiSizer® separation analyser. The integral transmission which is directly attributed to the sedimentation behaviour of the dispersions was found to be significantly influenced by four parameters: the MWCNT concentration, the sonication time, the addition of polymer and the pre-dispersion of MWCNTs in melt. A significant improvement of

Fig. 7. TEM of spin coated films of a PCCL100 dispersion (left) and a PCcomp100 dispersion (right); 5 min sonication.

Please cite this article in press as: Staudinger U, et al., Dispersability of multiwalled carbon nanotubes in polycarbonate-chloroform solutions, Polymer (2014), http://dx.doi.org/10.1016/j.polymer.2014.10.012

U. Staudinger et al. / Polymer xxx (2014) 1e10

the MWCNT dispersion in chloroform was achieved by the addition of PC, indicating the mechanism of polymer wrapping at the MWCNTs. Sonication up to 5 min gradually increased the amount of dispersed particles in the polymer solution and hence the stability of the dispersion. However, in PCcomp dispersions the dispersability of MWCNTs is significantly enhanced already at a low sonication time of only 0.5 min due to the pre-dispersion in melt. This implies that the polymer wrapping of PC chains around the CNTs is promoted by the melt mixing which, together with the process of MWCNT shortening during melt mixing, lead to better dispersability of the MWCNTs in solution. The dispersion quality of the MWCNTs in solution directly affects the electrical properties of thin solution casted composite films made of these MWCNT/polymer dispersions. Lowest surface resistivity values of about 5*104 U/sq were obtained for composite films of PCCL100 dispersions after 2 min of sonication. Despite the very good MWCNT dispersion in PCcomp dispersions, which results in the formation of homogeneous films with good transparency, all investigated melt mixed composite films exhibit high surface resistivity values of >1012 U/sq. The very well-dispersed and polymer wrapped MWCNTs are separated from each other and not able to form a percolating network. Thus, for the development of conducting composites a compromise between retaining high nanotube length, good dispersion, and a certain amount of secondary agglomeration seems to be favourable. Acknowledgement Special thanks go to Manuela Heber for her help in TEM investigations and to the colleagues from the institute's technical centre for melt compounding of the PC/MWCNT composites. The authors thank the German Federal Ministry of Education and Research (BMBF) for financial support within the project of Nano System Integration Network of Excellence (nanett) “Application of nano technologies for energy efficient sensor systems” with the support code 03IS2011P. References € tschke P, Halley P, Murphy M, Martin D, Bell SEJ, et al. Poly[1] McNally T, Po ethylene multiwalled carbon nanotube composites. Polymer 2005;46(19): 8222e32. €tschke P, Pegel S, Ha €ussler L, Kretzschmar B. Influence of twin[2] Villmow T, Po screw extrusion conditions on the dispersion of multi-walled carbon nanotubes in a poly(lactic acid) matrix. Polymer 2008;49(16):3500e9. € ldel A, Po € tschke P. Influence of processing conditions in small[3] Kasaliwal G, Go scale melt mixing and compression molding on the resistivity and morphology of polycarbonate-MWNT composites. J Appl Polym Sci 2009;112(6):3494e509. €tschke P, Ha €ußler L. Influence of small scale melt mixing condi[4] Krause B, Po tions on electrical resistivity of carbon nanotube-polyamide composites. Compos Sci Technol 2009;69(10):1505e15. €tschke P. Influence of feeding condi[5] Müller MT, Krause B, Kretzschmar B, Po tions in twin-screw extrusion of PP/MWCNT composites on electrical and mechanical properties. Compos Sci Technol 2011;71:1535e42. € tschke P, Villmow T, Krause B. Melt mixed PCL/MWCNT composites pre[6] Po pared at different rotation speeds: characterization of rheological, thermal, and electrical properties, molecular weight, MWCNT macrodispersion, and MWCNT length distribution. Polymer 2013;54:3071e8. €tschke P. Influence of material and pro[7] Kasaliwal G, Villmow T, Pegel S, Po cessing parameters on carbon nanotube dispersion in polymer melts. In: €tschke P, editors. Polymer-carbon nanotube composites. 1 ed. McNally T, Po Cambridge: Woodhead Publishing Limited; 2011. p. 92e132. € tschke P. Melt mixed [8] Socher R, Krause B, Boldt R, Hermasch S, Wursche R, Po nano composites of PA12 with MWNTs: influence of MWNT and matrix properties on macrodispersion and electrical properties. Compos Sci Technol 2011;71:306e14. €tschke P. The influence of matrix [9] Socher R, Krause B, Müller MT, Boldt R, Po viscosity on MWCNT dispersion and electrical properties in different thermoplastic nanocomposites. Polymer 2012;53:495e504. €tschke P, Petzold G, Alig I, Dudkin SM, Lellinger D. Dispersion, [10] Pegel S, Po agglomeration, and network formation of multiwalled carbon nanotubes in polycarbonate melts. Polymer 2008;49(4):974e84.

9

€schner C, Oswald S, Gruner W, Leonhardt A, et al. [11] Krause B, Ritschel M, Ta Comparison of nanotubes produced by fixed bed and aerosol-CVD methods and their electrical percolation behaviour in melt mixed polyamide 6.6 composites. Compos Sci Technol 2010;70(1):151e60. [12] Castillo FY, Socher R, Krause B, Headrick R, Grady BP, Prada-Silvy R, et al. Electrical, mechanical, and glass transition behavior of polycarbonate-based nanocomposites with different multi-walled carbon nanotubes. Polymer 2011;52:3835e45. [13] Guo J, Liu Y, Prada-Silvy R, Tan Y, Azad S, Krause B, et al. Aspect ratio effects of multi-walled carbon nanotubes on electrical, mechanical, and thermal properties of polycarbonate/MWCNT composites. J Polym Sci Part B Polym Phys 2013;52:73e83. [14] Kim SW, Kim T, Kim YS, Choi HS, Lim HJ, Yang SJ, et al. Surface modifications for the effective dispersion of carbon nanotubes in solvents and polymers. Carbon 2012;50(1):3e33. [15] Cheng QH, Debnath S, O'Neill L, Hedderman TG, Gregan E, Byrne HJ. Systematic study of the dispersion of SWNTs in organic solvents. J Phys Chem C 2010;114(11):4857e63. [16] Bergin SD, Nicolosi V, Cathcart H, Lotya M, Rickard D, Sun ZY, et al. Large populations of individual nanotubes in surfactant-based dispersions without the need for ultracentrifugation. J Phys Chem C 2008;112(4):972e7. [17] Bergin SD, Sun ZY, Rickard D, Streich PV, Hamilton JP, Coleman JN. Multicomponent solubility parameters for single-walled carbon nanotube-solvent mixtures. ACS Nano 2009;3(8):2340e50. [18] Lee K, Lim HJ, Yang SJ, Kim YS, Park CR. Determination of solubility parameters of single-walled and double-walled carbon nanotubes using a finite-length model. RSC Adv 2013;3(14):4814e20. [19] Detriche S, Nagy JB, Mekhalif Z, Delhalle J. Surface state of carbon nanotubes and Hansen solubility parameters. J Nanosci Nanotechnol 2009;9(10): 6015e25. [20] Detriche S, Zorzini G, Colomer JF, Fonseca A, Nagy JB. Application of the Hansen solubility parameters theory to carbon nanotubes. J Nanosci Nanotechnol 2008;8(11):6082e92. [21] Liu C-X, Choi J-W. Improved dispersion of carbon nanotubes in polymers at high concentrations. Nanomaterials 2012;2(4):329e47. [22] Grady BP. Recent developments concerning the dispersion of carbon nanotubes in polymers. Macromol Rapid Commun 2010;31(3):247e57. [23] O'Connell MJ, Boul P, Ericson LM, Huffman C, Wang YH, Haroz E, et al. Reversible water-solubilization of single-walled carbon nanotubes by polymer wrapping. Chem Phys Lett 2001;342(3e4):265e71. [24] Zhao W, Liu YT, Feng QP, Xie XM, Wang XH, Ye XY. Dispersion and noncovalent modification of multiwalled carbon nanotubes by various polystyrene-based polymers. J Appl Polym Sci 2008;109(6):3525e32. [25] Linton D, Driva P, Sumpter B, Ivanov I, Geohegan D, Feigerle C, et al. The importance of chain connectivity in the formation of non-covalent interactions between polymers and single-walled carbon nanotubes and its impact on dispersion. Soft Matter 2010;6(12):2801e14. [26] Saha LC, Mian SA, Jang J. Molecular dynamics simulation study on the carbon nanotube interacting with a polymer. Bull Korean Chem Soc 2012;33(3): 893e6. [27] Shvartzman-Cohen R, Nativ-Roth E, Baskaran E, Levi-Kalisman Y, Szleifer I, Yerushalmi-Rozen R. Selective dispersion of single-walled carbon nanotubes in the presence of polymers: the role of molecular and colloidal length scales. J Am Chem Soc 2004;126(45):14850e7. [28] Szleifer I, Yerushalmi-Rozen R. Polymers and carbon nanotubes - dimensionality, interactions and nanotechnology. Polymer 2005;46(19):7803e18. [29] Detloff T, Sobisch T, Lerche D. Particle size distribution by space or time dependent extinction profiles obtained by analytical centrifugation. Part Part Syst Charact 2006;23(2):184e7. [30] Detloff T, Sobisch T, Lerche D. Particle size distribution by space or time dependent extinction profiles obtained by analytical centrifugation (concentrated systems). Powder Technol 2007;174(1e2):50e5. € tschke P. Correlation of carbon nanotube [31] Krause B, Petzold G, Pegel S, Po dispersability in aqueous surfactant solutions and polymers. Carbon 2009;47(3):602e12. € tschke P, Petzold G. Dispersability and particle size [32] Krause B, Mende M, Po distribution of CNTs in an aqueous surfactant dispersion as a function of ultrasonic treatment time. Carbon 2010;48:2746e54. [33] Nadler M, Mahrholz T, Riedel U, Schilde C, Kwade A. Preparation of colloidal carbon nanotube dispersions and their characterisation using a disc centrifuge. Carbon 2008;46(11):1384e92. [34] Arnold MS, Suntivich J, Stupp SI, Hersam MC. Hydrodynamic characterization of surfactant encapsulated carbon nanotubes using an analytical ultracentrifuge. ACS Nano 2008;2(11):2291e300. [35] Azoubel S, Magdassi S. The formation of carbon nanotube dispersions by high pressure homogenization and their rapid characterization by analytical centrifuge. Carbon 2010;48(12):3346e52. [36] NANOCYLTM. NC7000 series - Product Datasheet. Edition V05. 10 March 2009. www.nanocyl.com. € tschke P. Influence of screw configuration, [37] Villmow T, Kretzschmar B, Po residence time, and specific mechanical energy in twin-screw extrusion of polycaprolactone/multi-walled carbon nanotube composites. Compos Sci Technol 2010;70:2045e55. € tschke P. Influence of dry [38] Krause B, Villmow T, Boldt R, Mende M, Petzold P, Po grinding in a ball mill on the length of multiwalled carbon nanotubes and

Please cite this article in press as: Staudinger U, et al., Dispersability of multiwalled carbon nanotubes in polycarbonate-chloroform solutions, Polymer (2014), http://dx.doi.org/10.1016/j.polymer.2014.10.012

10

U. Staudinger et al. / Polymer xxx (2014) 1e10

their dispersion and percolation behaviour in melt mixed polycarbonate composites. Compos Sci Technol 2011;71:1145e53. [39] Lerche D. Dispersion stability and particle characterization by sedimentation kinetics in a centrifugal field. J Disper Sci Technol 2002;23(5):699e709. [40] Zetasizer nano series user manual, MANO0317, issue 2.1. England: Malvern Instruments Ltd.; July 2004. [41] Giordani S, Bergin SD, Nicolosi V, Lebedkin S, Kappes MM, Blau WJ, et al. Debundling of single-walled nanotubes by dilution: observation of large populations of individual nanotubes in amide solvent dispersions. J Phys Chem B 2006;110(32):15708e18.

€tschke P, Long TE. Melt dispersion and electrospinning of non[42] Hunley MT, Po functionalized multiwalled carbon nanotubes in thermoplastic polyurethane. Macromol Rapid Commun 2009;30:2102e6. €tschke P. A method for determination of length distri[43] Krause B, Boldt R, Po butions of multiwalled carbon nanotubes before and after melt processing. Carbon 2011;49(4):1243e7. €tschke P, Lellinger D, Skipa T, Pegel S, Kasaliwal GR, et al. Estab[44] Alig I, Po lishment, morphology and properties of carbon nanotube networks in polymer melts. Polymer 2012;53(1):4e28.

Please cite this article in press as: Staudinger U, et al., Dispersability of multiwalled carbon nanotubes in polycarbonate-chloroform solutions, Polymer (2014), http://dx.doi.org/10.1016/j.polymer.2014.10.012