lable at ScienceDirect

Polymer 70 (2015) 88e99

Contents lists avai

Polymer

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

Thermal properties, rheology and foams of polystyrene-block-poly(4-vinylpyridine) diblock copolymers

Maria Schulze a, Ulrich A. Handge a, *, Sofia Rangou a, Jelena Lillep€arg a, Volker Abetz a, b

a Helmholtz-Zentrum Geesthacht, Institute of Polymer Research, Max-Planck-Strasse 1, 21502 Geesthacht, Germanyb University of Hamburg, Institute of Physical Chemistry, Grindelallee 117, 20146 Hamburg, Germany

a r t i c l e i n f o

Article history:Received 4 March 2015Received in revised form5 June 2015Accepted 7 June 2015Available online 10 June 2015

Keywords:Polymer foamsDiblock copolymersRheologyMicrophase separation

* Corresponding author.E-mail address: ulrich.handge@hzg.de (U.A. Handg

http://dx.doi.org/10.1016/j.polymer.2015.06.0050032-3861/© 2015 Elsevier Ltd. All rights reserved.

a b s t r a c t

In this study, the thermal and rheological properties of polystyrene-block-poly(4-vinylpyridine) (PS-b-P4VP) diblock copolymers are investigated in order to get information about the optimum foamingtemperature. Foams of these diblock copolymers were prepared using the technique of batch foamingwith carbon dioxide as environmentally-benign blowing agent. The tailored PS-b-P4VP diblock co-polymers with different molecular weights and a cylindrical morphology were prepared and analysedregarding their thermal stability. High-pressure differential scanning calorimetry exposes the plasticisingeffect of the blowing agent which yields a decrease of the glass transition temperature of the polystyreneand the poly(4-vinylpyridine) blocks. Sorption measurements were performed in order to measure theuptake of carbon dioxide in the diblock copolymer. Additionally, rheological experiments in the oscil-latory mode were conducted which confirmed a microphase-separated structure of the PS-b-P4VPdiblock copolymers by a plateau of the storage and loss moduli in the temperature range of processing. Inshear and melt elongation, the transient shear viscosity and the transient elongational viscosity weremuch smaller than the linear viscoelastic prediction at later times. The analysis of the foam morphologyrevealed that the foam density of the diblock copolymers as measured via Archimedes' principle exhibitsthe lowest foam density at a molecular weight in the order of 160 kg mol�1.

© 2015 Elsevier Ltd. All rights reserved.

1. Introduction

Block copolymers have attracted much interest as promisingmaterials due to their ability of microphase separation. Thesematerials combine the characteristics of thermoplastic polymersand the possibility of creating self-assembled structures on themicroscopic level [1e4]. By employing the technique of livinganionic polymerisation, the design of tailored block copolymerswith a defined molecular weight and a small polydispersity isrealisable [5e7]. Hence, block copolymers can be applied innumerous fields in industry, e.g., as insulation, absorbants ormembranes [8,9]. The covalent bonding of different types ofmonomers in block copolymers allows for combining differentproperties, e.g., for enhancing the toughness of polymers byincorporation of a soft block or for the compatibilisation of poly-mer blends [10e13]. In addition, the phenomenon of microphase

e).

separation enables the design of an isoporous active layer inmembranes [14] and the preparation of structured cell walls inpolymer foams [15]. The effect of block copolymer micelles onnucleation of foam cells in thermoplastic polymers was alsoexplored [16]. A systematic study on the influence of blends of astyrene-co-acrylonitrile copolymer and a poly(2,6-dimethyl-1,4-phenylene ether) on the foaming properties was performed inRef. [17].

New ways for fabrication of high-performance polymer foamshas aroused large interest due to the wide range of applications,e.g., in membrane technology or tissue engineering [18e21]. In thefield of membrane technology, new environmentally-benignmethods, especially foaming of thermoplastics, are of high inter-est [22e24] since the common fabrication of membranes byapplying the phase inversion process for polymer solutions re-quires a high amount of organic solvents [25,26]. Two mainmethods for manufacturing polymer foams have been established:batch foaming and foam extrusion [27e29]. Generally, both pro-cesses allow the use of physical blowing agents in order to generatea cellular structure. Polystyrene has been widely used in research

M. Schulze et al. / Polymer 70 (2015) 88e99 89

studies of the foam extrusion process to understand the complexityof extrusion parameters and the influence of specific polymerproperties [30e33]. To achieve a homogenous and well-definedfoam structure it is important to investigate and optimise theprocessing parameters such as foaming temperature and time [34].A suitable candidate as blowing agent to foam polymers is carbondioxide (CO2) because of its environmental-friendliness and plas-ticisation abilities for thermoplastics combined with a reduction ofviscosity [35e38].

The effect of carbon dioxide on the viscoelastic properties of thepolymer can be determined by rheological measurements in shearand elongation providing information in regard with the foamingbehaviour. The decrease of the steady-state viscosity of polystyrenemelts caused by loading with carbon dioxide was experimentallyinvestigated by several authors [39,40]. Viscosity measurements inelongational flows were performed byWang et al. [41]. In the workof Chaudhary and Jayaraman [42] rheological experiments in meltelongation and foaming of polypropylene clay nanocompositeswere performed. The authors showed that an increasing degree ofstrain-hardening along with slower crystallization of poly-propylene was associated with a smaller cell size and a larger celldensity. In Ref. [43], the reduction of the transient shear viscosityand the plasticisation effect, respectively, in polystyrene meltscaused by carbon dioxide was investigated. In general, homopoly-mers and blends of homopolymers have been studied more indepth compared to block copolymers. By way of example, poly-lactides showed an increase in viscosity and shear sensitivity bymodifying the homopolymer through chain extenders for foamapplications [44,45]. The same effect was observed for differentpolypropylene homopolymers [46] and blends of them with ther-moplastic olefins [47]. In addition, the influence of strain-hardening was examined which resulted in an optimised cellstructure of the foams. This effect was studied in detail by Stangeand Münstedt [48] for various polypropylenes with a differentmolecular structure. Whereas in previous studies the relation be-tween thermal and viscoelastic properties and foams of homopol-ymers has been elucidated, fundamental research on the foamingbehaviour has been executed considerably less for diblock co-polymers. In particular, the influence of molecular parameters (e.g.,molecular weight) on the foam properties needs to be investigatedfurther. So far, the influence of the morphology and molecular ar-chitecture of block copolymers on the viscoelastic properties hasbeen investigated by several groups on systems such as poly-styrene-block-poly(methyl methacrylate), polystyrene-block-poly-isoprene or polystyrene-block-polyisoprene-block-polystyreneblock copolymers [15,49,50].

The present study focuses on the influence of thermal andrheological properties of PS-b-P4VP diblock copolymers on theirfoaming ability. The diblock copolymers were chosen because ofthe high value of FloryeHuggins interaction parameter and theirability to create isoporous structures from precisely ordered andaligned morphologies in order to manufacture membranes[51,52]. Therefore, tailored PS-b-P4VP diblock copolymers withnumber average molecular weights between 50 kg mol�1 and220 kg mol�1 are synthesised using anionic polymerisation. In-formation about the thermal properties of the diblock copolymersuch as thermal stability and the influence of the blowing agentCO2 on the glass transition temperature of the microphases isrevealed. Besides, sorption measurements are performed todetermine the diffusion coefficient and the uptake of carbon di-oxide. These data and rheological properties are related to thefoam density to reveal the influence of the molecular weight onthe foaming ability and the characteristics of the foams. Addi-tionally, the optimum foaming temperature is determined byanalysing the foam density.

2. Experimental section

2.1. Synthesis of PS-b-P4VP diblock copolymers

PS-b-P4VP diblock copolymers were synthesised via sequentialanionic polymerisation of styrene and 4-vinylpyridine [51,53]. Allreactants were purified by the following procedures: Tetrahydro-furan (THF) (Th. Geyer, Renningen, Germany) was successivelydistilled under argon atmosphere. Styrene (Sigma Aldrich, Tauf-kirchen, Germany) was purified with dibutyl magnesium (MgBu2)and lately distilled before use. 4-vinylpyridine (Sigma Aldrich) wasdistilled and kept on calcium hydride (CaH2). After purifying twicewith ethylaluminum dichloride (EtAlCl2), 4-vinylpyridine wasdistilled once again.

The reaction was carried out at �78 �C with THF as solvent andsec-butyl lithium (sec-BuLi) (Sigma Aldrich) as initiator. THF wasput into the reaction vessel with lithium chloride (LiCl) (SigmaAldrich) and stirred overnight. The first block was polymerised byadding styrene in a first and sec-BuLi in a second step to the solu-tion. This was left under stirring for 4 h. After adding 4-vinylpyridine, the mixture was stirred further overnight. Toterminate the reaction, the solution was quenched with purifiedmethanol. Finally, THF was removed under reduced pressure andthe polymer solution was precipitated into water before filteringand drying the obtained polymer powder.

2.2. Molecular and morphological characterisation

In order to determine the composition of the diblock co-polymers, 1H nuclear magnetic resonance spectroscopy (1H NMR)was carried out using a Bruker AV-300 FT-NMR spectrometer(Bruker Biospin, Rheinstetten, Germany) at 500 MHz. The solventwas deuterated chloroform (CDCl3) with the internal standardtetramethylsilane (TMS).

The number and the weight average of the molar mass (Mn andMw) and the polydispersity (PDI) were analysed using size exclusionchromatography (SEC). The measurement was performed at 50 �Cwith dimethylacetamide (DMAc) in the presence of LiCl as solventand a flow rate of 1.0 mL min�1 (VWR Hitachi L2130 pump) using apre-column and two main columns (PSS Gran 1000 Å 10 mm andPSS Gran 3000 Å 10 mm). Polystyrene standards were employed forcalibration and a Shodex RI-101 refractive-index detector foranalysis.

The molecular weight Mn of the diblock copolymer was calcu-lated using the molecular weight of the polystyrene precursor andthe evaluated composition of the diblock copolymers from 1H NMRdata. The polydispersity PDI is given by the ratio of the weightaverage (Mw) and the number average (Mn) of the molecularweight. Furthermore, the degree of polymerisation N was given bythe ratio of the number average (Mn) of the molecular weight andthe molar mass of the monomers (M0) for each block. The sum ofboth is the total degree of polymerisation N.

Information about the morphological structure of the diblockcopolymers was obtained by transmission electron microscopy(TEM) on a FEI Tecnai G2 F20 (FEI, Eindhoven, The Netherlands).Therefore, samples of the diblock copolymer with a cylindricalshape (preparation see Section 2.4) were cut in sections ofapproximately 50 nm thickness at room temperature using a LeicaUltramicrotome EM UCT (Leica Microsystems, Wetzlar, Germany)equipped with a diamond knife. The sections were stained withiodine vapour to selectively contrast the P4VP microdomains. Themeasurements were performed at 120 kV in bright field mode. Thediameter and the distance of the centres of approximately 50 P4VPmicrophases were evaluated based on the micrographs.

M. Schulze et al. / Polymer 70 (2015) 88e9990

2.3. Thermal and sorption characterisation

Information about the thermal stability of the PS-b-P4VPdiblock copolymers was obtained by thermal gravimetric analysis(TGA) using a TG 209 F1 Iris instrument (Netzsch, Selb, Germany).The temperature range was set from 25 �C up to 900 �C with aheating rate of 10 K min�1 in an argon atmosphere.

The glass transition temperature (Tg) of the PS and P4VP blocksof the diblock copolymer was analysed via differential scanningcalorimetry (DSC) using a calorimeter DSC 1 (Mettler Toledo, Gie-ßen, Germany). Approximately 7 mg of polymer were transferredinto a 40 mL aluminium panwith a monoperforated lid. The heatingrate was 10 K min�1 in a temperature range between 30 �C and250 �C in a nitrogen atmosphere due to the expected glass transi-tions of approximately 105 �C for polystyrene and 150 �C for poly(4-vinylpyridine). A heating-cooling-heating cycle was applied, andthe value of Tg was determined using the second heating interval.

Besides conventional DSC, high-pressure differential scanningcalorimetry (HP-DSC) was conducted on a calorimeter HP-DSC 1(Mettler Toledo) to investigate the influence of the blowing agentcarbon dioxide (CO2) on the diblock copolymer. Therefore,approximately 7 mg of polymer were placed in a 40 mL aluminiumpan with a multiple perforated lid. The measurement was carriedout under CO2-atmosphere, whereupon the CO2-pressure wasvaried (1 bar, 10 bar, 20 bar). The sample was heated from 30 �C upto 250 �C at a heating rate of 10 K min�1 and then cooled down to50 �C. CO2-pressure was applied and the temperature was setisothermal at 50 �C for 2 h with a flow rate of 60 mL min�1 beforethe sample was heated up again to 250 �C under CO2-pressurewithout flow. The duration of 2 h was large enough to guaranteecomplete saturation of the diblock copolymer powder at 50 �C.Since a single powder particle has an approximate radius ofR ¼ 200 mm and the time scale of saturation is given by tsat ¼ R2/Dwith the diffusion coefficient D z 10�7 cm2/s, tsat roughly is 1.1 hand thus smaller than 2.0 h, see also Ref. [43].

To evaluate the uptake of CO2 in the PS-b-P4VP diblock co-polymers and to calculate the diffusion coefficient D, sorptionmeasurements of CO2 were performed using a gravimetric sorptionanalyser IsoSORP® Static (Rubotherm, Bochum, Germany) [54]. Theavailable equipment allowed determining the gas phase non-idealites, i.e. the true density, through the weighing in situ of aninert sample. The pressure accuracy was 0.006 bar. Casted and pre-dried films of PS-b-P4VP diblock copolymers with a thickness be-tween approximately 250 mm and 500 mmwere evacuated at 80 �Covernight to remove dissolved fluids. The sorption equilibrium datawere determined at 30 �C and a pressure of 40 bar. The specificuptake of CO2 was analysed considering gas buoyancy [55]. For theevaluation of the diffusion coefficient D, the kinetic sorption curvewas fitted using the theory of Fickian diffusion for values Mt/M∞ < 0.6:

Mt

M∞¼ 4

ffiffiffiffiffiffiffiDtpl2

r(1)

where Mt is the amount of gas adsorbed in the polymer at time t,M∞ is the mass of gas adsorbed at equilibrium (t / ∞) and l is thethickness of the sample [56]. The saturation time was estimated bytsat ¼ 0.202 l2/D with the thickness l of the sample and the exper-imentally obtained value of the diffusion coefficient D [43].

2.4. Rheological measurements

To prepare samples for rheological experiments, cylindricalsamples of the dried diblock copolymer were compression moul-ded at a temperature of 180 �C using a pillar press PW 10 H (W/O/

Weber, Remshalden, Germany) to assure a defined geometry(diameter d ¼ 8 mm, thickness h ¼ 2 mm). In this regard, thediblock copolymer powder was prepressed at room temperatureand afterwards molten at 180 �C under slight pressure for 180 sbefore applying vacuum for 90 s and increasing the pressure up to45 kN for 300 s. By applying vacuum and a temperature of 180 �Cduring compression moulding, the remaining solvent (THF) wastotally removed. The compression-moulded tablets were cooleddown for at least 30 min. All samples were additionally dried in avacuum oven at 50 �C.

The rheological measurements were performed using the rota-tional rheometer MCR 502 (Anton Paar, Graz, Austria) in theoscillatory mode to receive information about the viscoelasticproperties of the PS-b-P4VP diblock copolymers. A plateeplategeometry was employed with a plate diameter of 8 mm. The gapwas set slightly smaller than the thickness of the sample at roomtemperature (approximately 2 mm) in order to guarantee a com-plete contact between the plate and the sample. The experimentswere carried out under nitrogen purge, with a strain amplitude g0of 5% and an angular frequency u ranging between 10�2 and102 rad s�1. The temperature varied from 140 �C to 220 �C in steps of20 �C. Throughout the measurements the temperature was keptconstant. The experimental results were analysed using the soft-ware LSSHIFT developed by Honerkamp and Weese [57] to obtainmaster curves by applying the time-temperature superpositionprinciple. To determine the region of linear viscoelasticity, ampli-tude sweeps were conducted in the range from g0 ¼ 1%e10% at anangular frequency u of 10 rad s�1.

Dynamic-mechanical thermal analysis (DMTA) was performedusing the PS80-b-P4VP20220 diblock copolymer with a molecularweight of 220 kg mol�1. A plateeplate geometry with a platediameter of 8 mmwas applied. Themeasurements were carried outin the temperature range from 100 �C to 300 �C in a nitrogen at-mosphere. The measurement was started at the highest tempera-ture of 300 �C with a cooling rate of 0.5 K min�1. The angularfrequency was u ¼ 0.02 rad s�1 and the shear amplitude wasg0 ¼ 10%.

In order to investigate the behaviour of the diblock copolymerunder steady shear flow, stress-growth experiments were con-ducted additionally using the rotational rheometer MCR 502(Anton Paar). Therefore, the same geometry of samples and toolswere used as well as working under nitrogen purge. The experi-ments were carried out at 220 �C, varying the shear rate from0.1 s�1 to 1 s�1.

The apparent elongational viscosity he;app in elongation underuniaxial loading was determined using the UXF tool (Anton Paar).Compression-moulded samples with a thickness of 0.8mm, awidthof 10.0 mm and a length of 25.0 mm were used for the in-vestigations. The measurements were performed at a temperatureof 220 �C. Three different Hencky strain rates _ε0 were chosen, i.e._ε0 ¼ 0:01 ; 0:10 and 1:00 s�1: The maximum Hencky strain wasεmax ¼ 2.0 and the melting time was 8 min. The measured data inelongation were compared with the results of shear viscositymeasurements (see Fig. 7) assuming a Trouton ratio of 3 (i.e.multiplication of the transient shear viscosity with a factor of 3) aswell as with the linear viscoelastic prediction. The linear visco-elastic prediction was obtained by fitting a relaxation time spec-trum to the data of the dynamic moduli using the software NLREG[58,59].

2.5. Preparation and characterisation of diblock copolymer foams

The technique of batch foaming was applied, where cylindricaltablets, prepared like the samples for rheological measurements,were provided into the high pressure reaction vessel highpreactor

M. Schulze et al. / Polymer 70 (2015) 88e99 91

BHM-500 (Berghof, Eningen, Germany). The sorption of the diblockcopolymer with the blowing agent CO2 (Linde Gas, Pullingen,Germany, purity of 99.995%) was maintained at a gas pressure of40 bar and a temperature of 30 �C for a loading time of 5 days. Toexamine the loading pressure and temperature, a BTC-3000 device(Berghof) was utilised. The foam structure was generated by apressure drop down to ambient pressure and a contemporaneoustemperature raise to the foaming temperature which varied be-tween 90 �C and 120 �C.

After batch foaming, the vacuum dried samples were charac-terised with respect to the density applying Archimedes' principleusing a density determination kit (Mettler Toledo, Greifensee,Switzerland). The auxiliary liquid for the measurement was theperfluoro-compound FC-77 (3M Pharmaceuticals, St. Paul, Minne-sota, USA). The density of the auxiliary liquid was calibrated beforethe measurements of the foams were performed.

Besides the evaluation of the foam density, scanning electronmicroscopy (SEM) investigations were performed using a LeoGemini 1550 VP (Zeiss, Oberkochen, Germany) to obtain informa-tion about the distribution of the cells and their sizes as well as thecharacteristics of the cell wall surface. The foams were brokenunder cryogenic conditions and then fixed with a conductive pasteon a sample holder before they were sputtered with an approxi-mately 2 nm thick platinum coating. The cross sections of the foamswere investigated at an acceleration voltage between 5 kV and20 kV. The cell diameters were determined by converting the mi-crographs into black and white images and afterwards using thesoftware analySIS (Olympus) to evaluate minimum 50 cells persample. Furthermore, the cell density N was calculated usingN¼ (n/A)3/2 [42] where n is the number of cells and A is a definedarea of the SEM micrograph.

3. Results

3.1. Diblock copolymer characterization

The composition, the number average molecular weightMn, thepolydispersity index PDI and the degree of polymerisation N of thediblock copolymers are summarised in Table 1. Two groups ofdiblock copolymers were synthesised: The first group had acomposition of 80 wt% polystyrene (PS) and 20 wt% poly(4-vinylpyridine) (P4VP) with a molecular weight Mn between50 kg mol�1 and 220 kg mol�1, whereas the second group had acomposition of approximately 77 wt% PS and 23 wt% P4VP with amolecular weight Mn between 50 kg mol�1 and 134 kg mol�1.

Expectedly, the polydispersity of the diblock copolymers is closeto unity due to the synthesis mechanism of living anionic poly-merisation which does not allow tuning the degree of polymeri-sation. This is indicated by a narrow distribution of the molecularweight and thus, a similar length of polymer chains for each diblockcopolymer. However, the PDI of the diblock copolymer with the

Table 1Synthesised diblock copolymers and the results of molecular characterisation.

Diblock copolymera PS [wt%] P4VP [wt%] Mn [kg mol�1] PDI N

PS80-b-P4VP2050 80 20 50 1.05 479PS80-b-P4VP20164 80 20 164 1.06 1572PS80-b-P4VP20183 80 20 183 1.07 1754PS80-b-P4VP20220 80 20 220 1.15 2108PS76-b-P4VP2450 76 24 50 1.05 479PS77-b-P4VP23106 77 23 106 1.09 1016PS78-b-P4VP22116 78 22 115 1.08 1111PS77-b-P4VP23134 77 23 134 1.09 1284

a Subscript numbers indicate the weight percentage of each block and superscriptnumbers indicate the total number averaged molecular weight in kg mol�1.

highest molecular weight of 220 kg mol�1 is slightly higher thanthe PDI of the other diblock copolymers.

The morphological structure of four PS-b-P4VP diblock co-polymers with a molecular weight between 106 and 220 kg mol�1

was investigated by transmission electron microscopy (TEM)(Fig. 1). It was aimed to achieve diblock copolymers with a cy-lindrical morphology. In Fig. 1, the bright part is the PS matrix andthe dark parts are the P4VP domains. The micrographs of the threediblock copolymers with molecular weights of 106, 134 and164 kg mol�1 confirm a cylindrical structure without long-rangeorder in agreement with the phase diagram of Matsen and Bates[60], whereas the diblock copolymer with the highest molecularweight created a structure on the threshold between cylindricaland spherical morphology with a stronger tendency to assemblyin a cylindrical structure. The slight difference in morphology ofthe 220 kg mol�1 diblock copolymer is most probably caused bythe long relaxation time. Hence the diblock copolymer had notenough time to relax completely in an equilibrated cylindricalmorphology.

Furthermore, the analysis of the TEM micrographs proves thatan increase of the molecular weight leads to larger P4VP micro-phases in the PS matrix. Fig. 2(a) presents the average diameter ofthe P4VPmicrophases and Fig. 2(b) the distance of the centre of theP4VP microphases which were determined using the two-dimensional sections. The examined results for the diameters are30, 42, 47 and 59 nm for the PS-b-P4VP diblock copolymers with amolecular weight of 106, 134, 164 and 220 kg mol�1, respectively.Noticeable is the fact that the diblock copolymer with a molecularweight of 220 kg mol�1 has the highest deviation by the reasonmentioned earlier. A similar trend is observable for the distance ofthe centre of the P4VP microphases. This distance (long-period)increases with the molecular weight from approximately 60 nm(Mn ¼ 106 kg mol�1) to 126 nm (Mn ¼ 220 kg mol�1).

By fitting the evaluated distances with a power-law relationshipyields a power-law exponent of 0.87 ± 0.11 (Fig. 2(a)). Formerstudies on the relation between the diameter of the domains andthe molecular weight revealed power-law exponents of 0.643 forhigh molecular weight diblock copolymers with a similar compo-sition of both blocks [61] and 0.58 for cylindrical and sphericalstructures [62]. Additionally, the power-law relationship differswith the conformation of the polymer chains. It was found thathighly entangled polymer chains have a power-law exponent of 1/3whereas elongated polymer chains have a power-law exponent of 1[62].

3.2. Thermal behaviour and sorption properties

The synthesised diblock copolymers were analysed using ther-mal gravimetric analysis (TGA) in order to get information aboutthe stability and the suitable processing parameters for foaming. Ingeneral, the PS-b-P4VP diblock copolymers are stable up to 300 �C(Fig. 3). At higher temperatures, decomposition of the P4VP blocksstarted which was observable by a change of mass which approx-imately corresponds to the calculations based on the 1H NMR data.At around 400 �C, the PS microphase decomposed until the samplewas completely converted into the gaseous state at 450 �C. Allsamples showed a change of mass of less than 3% at approximately130 �C which may be a result of remaining water in the P4VPmicrophase because of its hydrophilic properties even if the sam-ples were vacuum dried at 40 �C for at least one week before themeasurement.

The influence of the physical blowing agent CO2 on the glasstransition temperature Tg was investigated by applying conven-tional DSC and high-pressure DSC (HP-DSC) (Fig. 4). Three PS-b-P4VP diblock copolymers with molecular weights of 50, 134 and

Fig. 1. Transmission electron micrographs of PS-b-P4VP diblock copolymers with a number average molecular weight of a) 106, b) 134, c) 164 and d) 220 kg mol�1. P4VPmicrophases appear dark, whereas the PS microphase appears bright.

M. Schulze et al. / Polymer 70 (2015) 88e9992

220 kg mol�1 were analysed. For comparison, the data for a pristinePS homopolymer (Polystyrene PS 158K, BASF SE, Ludwigshafen,Germany, Mn ¼ 120 kg mol�1) and a pristine P4VP homopolymer(Sigma Aldrich,Mw ¼ 60 kg mol�1) is shown as well. The evaluationof the Tg for the diblock copolymers was only feasible for the PSmicrophase up to CO2-pressures of 20 bar due to weak signals athigher CO2-pressures. The glass transition temperature Tg for theP4VP microphase in the diblock copolymers was not analysablebecause of the lower P4VP fraction compared to the PS fraction inthe diblock copolymer.

The value of Tg of the PS microphase decreased when increasingthe pressure of the CO2-atmosphere from ambient pressure to

Fig. 2. a) Diameter of the P4VP microphases in the PS matrix and b) distance of the centredetermined by analysis of the two-dimensional ultrathin sections.

20 bar. Considering the results, CO2 acts as a plasticiser by causingsoftening of the diblock copolymers. Especially pristine PS showsthis effect at which the Tg value dropped down from 105.5 �C atambient air to 90 �C at 20 bar whereas the Tg drop for the diblockcopolymers was from approximately the same Tg as the pristine PSdown to around 93.5 �C for the two diblock copolymers with 20 wt% P4VP and 95 �C for the diblock copolymer with 23 wt% P4VP. At20 bar CO2 pressure, the value of Tg increased with the molecularweight of the diblock copolymer. Consequently, with increasingmolecular weight of the diblock copolymer the plasticising effect ofCO2 is less visible [63]. The pristine P4VP showed a reduction of Tgfrom approximately 150 �C down to 134 �C. This information is

of P4VP microphases as a function of the molecular weight of the diblock copolymer

Fig. 3. TGA thermogram under inert atmosphere of a PS-b-P4VP diblock copolymerwith a molecular weight of 164 kg mol�1 and its decomposition of the P4VP micro-phase (first step) and the PS microphase (second step). The heating rate was10 K min�1.

M. Schulze et al. / Polymer 70 (2015) 88e99 93

important for the adjustment of the loading temperature in batchfoaming and the optimum foaming temperature.

The measurements with the sorption balance (see Table 2)exhibited an average specific CO2-uptake of 6.4 wt% CO2 in eachdiblock copolymer with a standard deviation of 0.15 wt% whichmeans that the amount of dissolved CO2 in the diblock copolymerwas constant in the range of the investigated molecular weightwithin experimental scatter. Compared with the pure homopoly-mers with specific CO2-uptakes of 5.9 wt% for PS and 7.2 wt% forP4VP, respectively, the amount of adsorbed gas results from a ruleof mixture of the values of the homopolymers dependent on thecomposition of the diblock copolymer.

A similar trend was observable for the diffusion coefficient ofthe diblock copolymers since the values were all in the samemagnitude between 0.45� 10�7 cm2 s�1 (Mn ¼ 220 kg mol�1) and2.00 � 10�7 cm2 s�1 (Mn ¼ 50 kg mol�1). With an increase in mo-lecular weight of the diblock copolymer, the density of the diblockcopolymer decreases due to the increase of the free volume.However, it seems that a lower molecular weight of the diblockcopolymer leads to a higher diffusion coefficient. A reason might bethe faster swelling, a separation of the polymer chains initiated bythe penetration of CO2-molecules, of a low molecular weightdiblock copolymer compared to high molecular weight diblockcopolymer caused by the plasticization effect of CO2 [64]. The

Fig. 4. Glass transition temperature Tg as a function of pressure p in a CO2-atmosphere of aweights of 50, 134 and 220 kg mol�1 and pristine PS and b) pristine poly(4-vinylpyridine)

diffusion coefficients of the pure homopolymers are in the samerange of magnitude as well, 0.72� 10�7 cm2 s�1 for PS and0.43 � 10�7 cm2 s�1 for P4VP. These values correlate with theliterature where the diffusion coefficients are 0.58 � 10�7 cm2 s�1

for PS (at 25 �C) [56] and 0.13 � 10�7 cm2 s�1 for P4VP (35 �C and8.25 bar) [65]. The diffusion coefficient of the diblock copolymers islarger than the values for the pristine homopolymers which doesnot indicate the validity of a simple mixing rule for the diffusioncoefficient.

The saturation time employing an average diffusion coefficientof 1.07� 10�7 cm2 s�1 is 21.1 h, i.e. less than one day. With regard tothe evaluated saturation time, the loading time of five days forbatch foaming guarantees a complete saturation of the samples.

3.3. Rheological properties of the diblock copolymers

The results of the DMTA measurements are presented in Fig. 5.The measurements were performed in the range from 300 �C (starttemperature) to 100 �C (end temperature). Because of the platee-plate geometry, measurements at a lower temperature could not beperformed due to insufficient adhesion between plates and sample.At the temperature of approximately 100 �C, the pronounceddecrease of storage modulus G0 and the maximum of the lossmodulus G00 can be anticipated which correspond to the glasstransition of the polystyrene block. At higher temperatures, thedynamic moduli decrease and the rubbery plateau of the poly-styrene matrix appears. In the temperature range from approxi-mately 110 �Ce140 �C, the polystyrene microphase reacts rubberyand the poly(4-vinylpyridine) domains are in the glassy state. Themaximum of G00 at a temperature of 144 �C indicates the glasstransition of the poly(4-vinylpyridine) block. At higher tempera-tures, G0 and G00 do not change with temperature. At these hightemperatures, we do not observe the terminal flow regime forhomopolymers (Maxwell fluid behaviour). On the contrary, thedata of Fig. 5 reveal that the diblock copolymers of this work are inthe microphase-separated state up to a temperature of 300 �C. Thisresult agrees with the determination of the order-disorder transi-tion temperature in previous studies [66].

Small amplitude shear oscillations were performed at severaltest temperatures in order to probe the frequency-dependentresponse of the PS-b-P4VP diblock copolymers of this study.Diblock copolymers consist of two different types of monomerswhich are associated with different time-temperature shift factors.Therefore, the time-temperature superposition principle is gener-ally not valid for this class of polymers. However, if the shift factor

) the polystyrene microphase for three PS-b-P4VP diblock copolymers with molecularusing high-pressure DSC measurement.

Table 2Specific uptake of carbon dioxide in the diblock copolymers and diffusion coefficientof the diblock copolymers at a pressure of 40 bar and a temperature of 30 �C. Forcomparison, measurements of pure PS and P4VP are shown.

Diblockcopolymer

Densitya

[g cm�3]SpecificCO2-uptake[wt%]

Diffusioncoefficient[10�7 cm2 s�1]

PS76-b-P4VP2450 1.0967 6.21 2.00PS77-b-P4VP23106 1.0950 6.72 1.03PS77-b-P4VP23134 1.0917 6.33 1.32PS80-b-P4VP20164 1.0897 6.46 0.90PS80-b-P4VP20183 1.0883 6.45 0.45PS80-b-P4VP20220 1.0840 6.48 0.68PS100120 1.0708 5.89 0.72P4VP10060 1.1884 7.18 0.43

a Determined via Archimedes' principle.

M. Schulze et al. / Polymer 70 (2015) 88e9994

for the homopolymers of these two different monomer species areapproximately equal (e.g., if the glass transition temperatures of thetwo homopolymers are similar), then “imperfect” master curvescan be created. The difference of the Tg values of both blocks isapproximately 40 �C. Furthermore, since the major componentpolystyrene is associated with a weight fraction in the order of80 wt% and consequently strongly influences the temperature-dependent response, the existence of such “imperfect” mastercurves (which mask the mismatch of the moderately different Tgvalues) is not surprising. Fig. 6 presents the results of the applica-tion of the time-temperature superposition of two PS-b-P4VPdiblock copolymers with different molecular weights(Mn ¼ 50 kg mol�1 and 183 kg mol�1). The reference temperaturewas 160 �C. The contribution of the rubbery plateau of the poly-styrene matrix can be clearly seen. The molecular weight of thediblock copolymer strongly influences the width of this entangle-ment regime. At lower frequencies, a transition regime appears. Atthe lowest experimentally accessible frequencies the plateaucaused by the phenomenon ofmicrophase separation can be clearlyseen, see also the work of Gil Haenelt et al. [49]. A close inspectionof the master curve indicates that we only deal with an “imperfect”master curve such that the time-temperature superposition prin-ciple is not valid in a strict sense for this class of materials. Analysisof the data at the lowest accessible frequency indicates that a lowermolecular weight of the diblock copolymers leads to a higher valueof the storage modulus.

Fig. 5. Results of dynamic-mechanical thermal analysis: Storage modulus G0 and lossmodulus G00 as a function of temperature T of a PS-b-P4VP diblock copolymer with amolecular weight of 220 kg mol�1. Measurement was performed at a frequency u of0.02 rad s�1.

The frequency sweeps were performed in the linear viscoelasticregime. In order to probe the flow properties in the nonlinearrange, stress-growth tests with a constant shear rate were per-formed. The applied shear rate varied between 0.1 s�1 and 1.0 s�1.The data for one selected PS-b-P4VP diblock copolymer with amolecular weight of 134 kg mol�1 are shown in Fig. 7. Fig. 7(a)shows the time-dependent viscosity for different shear rates. Atlow times, the viscosity increases with time because of the visco-elastic response of the diblock copolymers. At larger times, asteady-state value of viscosity is achieved. Fig. 7(a) clearly showsthat the value of the steady-state viscosity decreases with shearrate.

The influence of molecular weight on the transient viscosity isshown in Fig. 7(b). The data demonstrate that the steady-stateviscosity increases with increasing molecular weight for allapplied shear rates. The pronounced shear thinning at such a lowshear rate of 0.1 s�1 (indicating a still microphase-separatedmorphology) suggests that the shell of polystyrene chains whichare covalently bonded to the cylindrical poly(4-vinylpyridine)blocks strongly influences the value of the steady-state viscosity.

The steady-state value of the viscosity as a function of shear rateis shown in Fig. 8. Two power-law functionswith an exponent of�1and �0.5 are also shown. A clear inspection of the data reveals thetrend that the power-law behaviour of the stationary viscosityvaries with molecular weight. At low molecular weights, thepower-law exponent�1 describes the trend of the data, whereas atthe largest molecular weight the power-law exponent of �0.5corresponds to a better fit.

The rheological results in oscillatory and in steady shear canbe explained in the following way, see Fig. 9. At low deformationand deformation rates, the elastic interactions caused by themicrophase-separated morphology dominate. At higher stresses,the shells of polystyrene chains which are covalently bonded tothe poly(4-vinylpyridine) blocks slide on each other. The size ofthe shell increases with molecular weight, and the interactionscaused by the polystyrene chains dominate the viscousresponse.

During expansion of foam cells, a biaxial elongational flow iscreated in the polymer shell which surrounds an expanding foambubble. Consequently, the rheological properties in melt elonga-tion are highly relevant for polymer foaming. Fig. 10 presents thetime-dependent apparent elongational viscosity he;app of twodifferent diblock copolymers at a temperature of 220 �C. The re-sults of the shear viscosity measurements (cf. Fig. 7) which aremultiplied by the Trouton ratio of 3 are also presented. Further-more, the linear viscoelastic prediction of the elongational vis-cosity is shown.

The transient elongational viscosity depicts a similar trend forthe diblock copolymers of this study. Because of the plateau ofthe storage modulus G

0at low frequencies (see Fig. 6), the linear

viscoelastic prediction does not lead to a steady-state elonga-tional viscosity, but continuously increases with time in theinvestigated time interval. The measured shear viscosity (multi-plied by a factor of 3) nearly agrees at low times with the linearviscoelastic prediction, but then achieves a constant value atlarger times. The experimental data of the extensional viscositynearly agree with the shear viscosity data up to the point ofrupture in elongation. The agreement of the measured exten-sional and transient shear viscosity data indicate that similarmorphological rearrangements take place in shear and elonga-tion. At larger times, the transient shear viscosity and the elon-gational viscosity attain lower values than the linear viscoelasticprediction. In particular, the extensional viscosity does not depicta pronounced strain hardening as observed for longchainbranched polyethylenes.

Fig. 6. Master curves of G0 and G00 as a function of angular frequency u of two different PS-b-P4VP diblock copolymers with molecular weights of a) 50 kg mol�1 and b) 183 kg mol�1.The reference temperature was 160 �C and the shift factor is denoted by aT. The entanglement plateau is mainly caused by the contribution of the PS microphase.

Fig. 7. Results of stress-growth tests with a constant shear rate: Transient viscosity h as a function of time t for a) a selected diblock copolymer with a molecular weight of134 kg mol�1 and b) as a function of the molecular weight for two different shear rates.

M. Schulze et al. / Polymer 70 (2015) 88e99 95

3.4. Batch foaming

The principle of the batch foaming process is illustrated inFig. 11. Generally, batch foaming is a two-step process where thefirst step is the saturation of the polymer with a blowing agent

Fig. 8. Steady-state value of the shear viscosity as a function of shear rate _g0. The testtemperature was 220 �C.

(here CO2) and the second step is the foaming of the polymer bynucleation of foam cells and expansion of the dissolved gas throughpressure drop and/or increase in temperature. Therefore, the pro-cessing parameters were chosen to guarantee a saturation of thesample with CO2 at a pressure of 40 bar provided by the gas source.

The achieved density of the PS-b-P4VP diblock copolymer foamsstrongly depends on the foaming temperature and the molecularweight of the foamed diblock copolymers (Fig. 12). Low densities inthe order of 0.15 g cm�3 are obtained for diblock copolymers with ahighmolecular weight at a foaming temperature of 110 �C. Foaming

Fig. 9. Scheme of the deformation of the morphology during shear flow. The cylin-drical domains of the P4VP blocks are surrounded by a shell of PS blocks. At high shearrates, the cylinders with the shell slide on each other. This deformation mechanismdominates the viscous response.

Fig. 10. Apparent elongational viscosity he;app as a function of time for three different Hencky strain rates _ε0 and two diblock copolymers of this study ((a) PS77-b-P4VP23106 and(b) PS80-b-P4VP20220 diblock copolymers). The measurement temperature was 220 �C. The linear viscoelastic prediction of the elongational viscosity (LVE) and the results of thetransient shear viscosity measurements (multiplied by the Trouton ratio) are also shown. The solid symbols are the data in elongation and the open symbol correspond to the sheardata.

M. Schulze et al. / Polymer 70 (2015) 88e9996

temperatures of 100 �C and 105 �C lead to relatively low densities of0.4 g cm�3 down to 0.3 g cm�3 as well whereas diblock copolymerswhich were foamed below 100 �C and above 110 �C show higherdensities of more than 0.4 g cm�3. Especially diblock copolymerswith a low molecular weight such as 50 kg mol�1 exhibit highdensities of more than 0.7 g cm�3 at foaming temperatures be-tween 90 �C and 120 �C. Consequently, at the chosen foamingtemperatures a larger molecular weight yields a lower foam den-sity. This effect can be explained as follows: A larger molecularweight leads to a larger domain spacing. At a larger molecularweight, foam cells which are nucleated in the PS microphase cangrow to a larger extent without being hindered by the P4VP do-mains. The expanding foam cells deform the microphase-separatedstructure. In the case of a lowmolecular weight this deformation ofthe microstructure requires more energy.

The cell density of the foams in Fig. 12 was analysed using thescanning electron micrographs. For our experimental parameters,the cell density of the PS77-b-P4VP23 diblock copolymer with amolecular weight of 106 kg/mol was approximately28.7 � 106 cells/cm3. An increase of the molecular weight to 164and 220 kg/mol led to lower cell densities of 0.5 � 106 and2.0 � 106 cells/cm3 for the two PS80-b-P4VP20 diblock copolymers,

Fig. 11. Scheme of a two-step batch foaming process with the processing parametersT0 as loading temperature, p0 as saturation pressure, Dt as loading time, Tfoam asfoaming temperature, Dp as pressure drop down to ambient pressure and tfoam asfoaming time.

respectively. This effect is associated with the larger averagediameter of the foams with a molecular weight of 164 and 220 kg/mol.

Furthermore, investigations of the foam structure and the sur-face of the cell walls using scanning electronmicroscopy (SEM) giveinformation about the quality of the foamed diblock copolymers.SEM micrographs of the foam structure (Fig. 13(b)) reveal that afoaming temperature of 110 �C led to the most regular distributionof the cells with a rather smooth surface. In contrast, the foamsprepared at 100 �C and 120 �C (Fig. 13(a) and (c)) exhibit a moreirregular distribution of the cells with a slightly rougher surface at100 �C and a very rough surface at 120 �C. The high standard de-viation for the samples foamed at 100 �C resulted from the com-bination of diffusion of CO2 off thematrix and the high viscosity (i.e.long relaxation times of the diblock copolymer) leading to a lesscontrollable process. The time-dependent expansion of foam cellsis strongly affected by the relatively slow time-dependent increaseof viscosity at low times (cf. Fig. 7) and the fast diffusion of CO2 andconsequently is strongly sensitive to fluctuations at 100 �C.

The effect of the foaming temperature on the cell structure isconfirmed by evaluation of the cell size distribution of the foameddiblock copolymers. Fig. 14(a)e(c) shows the cell size distributionof the PS-b-P4VP diblock copolymer with a molecular weight of220 kgmol�1 with regard to the foaming temperature. The polymerfoamed at 120 �C developed small cells (mean cell diameterdc ¼ 42.1 mm) with a narrow distribution whereas the polymersfoamed at 100 �C and 110 �C display a broader distribution withlarger mean cell diameters of 73.3 mm and 83.9 mm, respectively.The cell size distribution of the sample foamed at 100 �C is broaderthan the sample foamed at 110 �C. In summary, the foaming tem-perature influences the viscosity of the melt during expansion offoam cells.

The molecular weight of the foamed diblock copolymers led todifferences in the foam structure. As seen in the SEM micrographs,the diblock copolymer with a molecular weight of 164 kg mol�1

(Fig. 13(b)) generated on the average larger cells in contrast to thediblock copolymer with a molecular weight of 220 kg mol�1

(Fig. 13(e)). In contrast, the block copolymer with a molecularweight of 106 kg mol�1 (Fig. 13(d)) exhibited small cells. Thecomparison of Fig. 14(b) and (d) reveals the influence of the mo-lecular weight at a foaming temperature of 110 �C. It is clearly seen(Fig. 14(b)) that a higher molecular weight (here 220 kg mol�1) ledto a more narrow cell size distribution in comparison to a broadercell size distribution of the diblock copolymer with a molecularweight of 164 kg mol�1 (Fig. 14(d)). Remarkably, both investigated

Fig. 12. Foam density of the PS-b-P4VP diblock copolymers as a function of the foaming temperature for the block copolymers with (a) a PS fraction of 77 wt% and (b) a PS fraction of80 wt%. The number average Mn of the molecular weight is indicated.

Fig. 13. Scanning electron micrographs displaying i) the influence of the foaming temperature on a 164 kg mol�1 PS-b-P4VP diblock copolymer foamwith temperatures of a) 100 �C,b) 110 �C, c) 120 �C and ii) the influence of the molecular weight of PS-b-P4VP foams with d) 106 kg mol�1, b) 164 kg mol�1, e) 220 kg mol�1 at a foaming temperature of 110 �C.

M. Schulze et al. / Polymer 70 (2015) 88e99 97

diblock copolymers had a similar foam density at a foaming tem-perature of 110 �C (see Fig. 12).

Whereas the effect of foaming temperature is clearly seen by itsinfluence onmelt viscosity and foam density, the explanation of theeffect of molecular weight is less obvious. Our rheological experi-ments in shear and elongation suggest that it is not themacroscopicviscosity which influences the final foam density, but the micro-rheological deformation of the cylindrical structure which appearswhen a gas bubble expands in the microphase separated structure.Such a microrheological analysis is beyond the scope of this study.

4. Conclusions

The present study elucidates the influence of thermal andrheological properties on the foaming behaviour of PS-b-P4VPdiblock copolymers using CO2 as blowing agent. The synthesis vialiving anionic polymerization led to diblock copolymers withdefined molecular weight and composition and allows the inves-tigation of the impact of morphology on the foaming process.DMTA experiments and transmission electron microscopy

confirmed the microphase separated structure of the diblockcopolymers.

The plasticising effect of CO2 on the glass transition temperatureTg was revealed by high-pressure DSC measurements. As a result,the viscosity of the diblock copolymer is reduced which plays animportant role to adjust the processing parameters of the appliedfoaming methods. Furthermore, sorption experiments showed anaverage CO2-uptake of 6.4 wt% CO2 at 30 �C and a pressure of 40 barin each diblock copolymer independent of the investigated range ofthe molecular weight. The influence of the molecular weight on thefoam structure of the diblock copolymers was observable by anal-ysis of the foam density. The foam density decreased with highermolecular weight due to larger cell sizes determined by scanningelectron microscopy. A larger molecular weight is associated withlarger domain spacing and thus foam cells which are nucleated inthe PS microphase can expand more easily. Besides the molecularweight, the foaming temperature played an important role in orderto generate large cell structures. The optimum foaming tempera-ture for PS-b-P4VP diblock copolymers of this study is 110 �C withthe lowest foam density and a regular cell size distribution.

Fig. 14. Cell size distribution of PS-b-P4VP diblock copolymer foams with molecular weights of (a)e(c) 220 kg mol�1 and (d) 164 kg mol�1 processed at foaming temperaturesbetween 100 �C and 120 �C indicated in the graphs. The mean cell diameter dc and its deviation are denoted.

M. Schulze et al. / Polymer 70 (2015) 88e9998

In conclusion, in spite of the microphase-separated structurediblock copolymers with a cylindrical morphology can be utilisedfor foam applications with similar low densities as homopolymers.The advantage of using diblock copolymers is the complexity of thesystem resulting in a variety of tailored properties which can beemployed for the fabrication of high-performance polymer foams,e.g., insulations with water vapour permeability or isoporousmembranes fabricated without the use of solvents.

Acknowledgement

The authors thank Dr. Volkan Filiz and Brigitte Lademann for thesupport in the synthesis of the diblock copolymers. They also thankClarissa Abetz and Sofia Dami (scanning electron microscopy), Sil-vio Neumann and Thomas Emmler (nuclear magnetic spectros-copy), Maren Brinkmann (size exclusion chromatography) andIvonne Ternes (thermal analysis and rheological measurements) fortheir experimental support.

References

[1] F.S. Bates, G.H. Fredrickson, Annu Rev. Phys. Chem. 41 (1) (1990) 525e557.[2] F.S. Bates, G.H. Fredrickson, Phys. Today 52 (2) (1999) 32.[3] V. Abetz, P.F.W. Simon, Phase behaviour and morphologies of block co-

polymers, Adv. Polym. Sci. 189 (2005) 125e212.[4] V. Abetz, A. Boschetti-de-Fierro, 7.02 e Block copolymers in the condensed

state, in: K. Matyjaszewski, M. M€oller (Eds.), Polymer Science: a Compre-hensive Reference, Elsevier, Amsterdam, 2012, pp. 3e44.

[5] N. Hadjichristidis, H. Iatrou, S. Pispas, M. Pitsikalis, J. Polym. Sci. Part A Polym.Chem. 38 (18) (2000) 3211e3234.

[6] D. Uhrig, J.W. Mays, J. Polym. Sci. Part A: Polym. Chem. 43 (24) (2005)6179e6222.

[7] S. Ndoni, C.M. Papadakis, F.S. Bates, K. Almdal, Rev. Sci. Instrum. 66 (2) (1995)1090e1095.

[8] D. Sophiea, D. Klempner, V. Sendijarevic, B. Suthar, K.C. Frisch, Inter-penetrating polymer networks as energy-absorbing materials, in:D. Klempner, L.H. Sperling, L.A. Utracki (Eds.), Interpenetrating Polymer Net-works, Vol. 239, 1994, pp. 39e75.

[9] M. Stumpf, A. Sp€orrer, H.-W. Schmidt, V. Altst€adt, J. Cell. Plastics 47 (6) (2011)519e534.

[10] Z. Oommen, S.R. Zachariah, S. Thomas, G. Groeninckx, P. Moldenaers, J. Mewis,J. Appl. Polym. Sci. 92 (1) (2004) 252e264.

[11] R. Adhikari, S.K. Khatri, S. Adhikari, G.H. Michler, F.J.B. Calleja, Macromol.Symp. 290 (1) (2010) 166e174.

[12] C.W. Macosko, H.K. Jeon, T.R. Hoye, Prog. Polym. Sci. 30 (8e9) (2005)939e947.

[13] H. Ruckd€aschel, V. Altst€adt, A.H.E. Müller, Cell. Polym. 26 (6) (2007) 367e380.[14] M. Radjabian, J. Koll, K. Buhr, U.A. Handge, V. Abetz, Polymer 54 (7) (2013)

1803e1812.[15] G.L. Chakkalakal, C. Abetz, U. Vainio, U.A. Handge, V. Abetz, Polymer 54 (15)

(2013) 3860e3873.[16] P. Spitael, C.W. Macosko, R.B. McClurg, Macromolecules 37 (18) (2004)

6874e6882.[17] H. Ruckd€aschel, P. Gutmann, V. Altst€adt, H. Schmalz, A.H.E. Müller, Foaming of

microstructured and nanostructured polymer blends, in: A.H.E. Müller,H.W. Schmidt (Eds.), Complex Macromolecular Systems I, Vol. 227, SpringerBerlin, 2010, pp. 199e252.

[18] M. Gargiulo, L. Sorrentino, S. Iannace, High performance polymeric foams, in:D. Acierno, A. D'Amore, L. Grassia (Eds.), IVth International Conference onTimes of Polymers and Composites, AIP Conference Proceedings, vol. 1042,2008, pp. 109e111.

[19] B. Krause, K. Diekmann, N.F.A. van der Vegt, M. Wessling, Macromolecules 35(5) (2002) 1738e1745.

[20] X. Wang, W. Li, V. Kumar, Biomaterials 27 (9) (2006) 1924e1929.[21] X. Liao, H. Zhang, T. He, J. Nanomater. 2012 (2012) 12.[22] Q. Huang, D. Paul, B. Seibig, Desalination 144 (1e3) (2002) 1e3.[23] B. Krause, N.F.A. van der Vegt, M. Wessling, Desalination 144 (1e3) (2002)

5e7.[24] V. Abetz, Macromol. Rapid Commun. 36 (1) (2015) 10e22.[25] K.-V. Peinemann, V. Abetz, P.F.W. Simon, Nat. Mater. 6 (12) (2007) 992e996.

M. Schulze et al. / Polymer 70 (2015) 88e99 99

[26] D.R. Lloyd, S.S. Kim, K.E. Kinzer, J. Membr. Sci. 64 (1e2) (1991) 1e11.[27] M. Sauceau, J. Fages, A. Common, C. Nikitine, E. Rodier, Prog. Polym. Sci. 36 (6)

(2011) 749e766.[28] R. Kl€otzer, D. Paul, B. Seibig, Extrusion of microcellular foams and application,

in: Antec'97-Plastics Saving Planet Earth, Conference Proceedings, Vols. 1e3,1997, pp. 2042e2045.

[29] L.J.M. Jacobs, M.F. Kemmere, J.T.F. Keurentjes, Green. Chem. 10 (7) (2008)731e738.

[30] X. Han, K.W. Koelling, D.L. Tomasko, L.J. Lee, Polym. Eng. Sci. 42 (11) (2002)2094e2106.

[31] C. Zhang, B. Zhu, L.J. Lee, Polymer 52 (8) (2011) 1847e1855.[32] K.-M. Lee, E.K. Lee, S.G. Kim, C.B. Park, H.E. Naguib, J. Cell. Plastics 45 (6) (2009)

539e553.[33] M. Lee, C. Tzoganakis, C.B. Park, Polym. Eng. Sci. 38 (7) (1998) 1112e1120.[34] S. Merlet, C. Marestin, F. Schiets, O. Romeyer, R. Mercier, Macromolecules 40

(6) (2007) 2070e2078.[35] Y.-T. Shieh, J.-H. Su, G. Manivannan, P.H.C. Lee, S.P. Sawan, W. Dale Spall,

J. Appl. Polym. Sci. 59 (4) (1996) 695e705.[36] Y.-T. Shieh, J.-H. Su, G. Manivannan, P.H.C. Lee, S.P. Sawan, W. Dale Spall,

J. Appl. Polym. Sci. 59 (4) (1996) 707e717.[37] M.J. Bortner, D.G. Baird, Polymer 45 (10) (2004) 3399e3412.[38] M. Lee, C. Tzoganakis, C.B. Park, Adv. Polym. Technol. 19 (4) (2000) 300e311.[39] J.R. Royer, Y.J. Gay, J.M. Desimone, S.A. Khan, J. Polym. Sci. Part B: Polym. Phys.

38 (23) (2000) 3168e3180.[40] M.J. Wingert, S. Shukla, K.W. Koelling, D.L. Tomasko, L.J. Lee, Industrial Eng.

Chem. Res. 48 (11) (2009) 5460e5471.[41] J. Wang, D.F. James, C.B. Park, J. Rheology 54 (1) (2010) 95e116.[42] A.K. Chaudhary, K. Jayaraman, Polym. Eng. Sci. 51 (9) (2011) 1749e1756.[43] U.A. Handge, V. Altst€adt, J. Rheology 56 (4) (2012) 743e766.[44] N. Najafi, M.-C. Heuzey, P. Carreau, D. Therriault, C. Park, Rheol. Acta 53

(10e11) (2014) 779e790.[45] Y.-M. Corre, A. Maazouz, J. Duchet, J. Reignier, J. Supercrit. Fluids 58 (1) (2011)

177e188.

[46] Z. Xu, Z. Zhang, Y. Guan, D. Wei, A. Zheng, J. Cell. Plastics 49 (4) (2013)317e334.

[47] A. Maani, H.E. Naguib, M.-C. Heuzey, P.J. Carreau, J. Cell. Plastics 49 (3) (2013)223e244.

[48] J. Stange, H. Münstedt, J. Rheology 50 (6) (2006) 907e923.[49] T. Gil Haenelt, P. Georgopanos, C. Abetz, S. Rangou, D. Alisch, A. Meyer,

U.A. Handge, V. Abetz, Korea-Australia Rheology J. 26 (3) (2014) 263e275.[50] C.D. Han, D.M. Baek, J.K. Kim, S.G. Chu, Polymer 33 (2) (1992) 294e305.[51] S. Rangou, K. Buhr, V. Filiz, J.I. Clodt, B. Lademann, J. Hahn, A. Jung, V. Abetz,

J. Membr. Sci. 451 (2014) 266e275.[52] J. Hahn, J.I. Clodt, V. Filiz, V. Abetz, RSC Adv. 4 (20) (2014) 10252e10260.[53] F. Wode, L. Tzounis, M. Kirsten, M. Constantinou, P. Georgopanos, S. Rangou,

N.E. Zafeiropoulos, A. Avgeropoulos, M. Stamm, Polymer 53 (20) (2012)4438e4447.

[54] Y. Sato, T. Takikawa, S. Takishima, H. Masuoka, J. Supercrit. Fluids 19 (2)(2001) 187e198.

[55] E. Aionicesei, M. �Skerget, �Z. Knez, J. Supercrit. Fluids 47 (2) (2008) 296e301.[56] J. Crank, G.S. Park, Diffusion in Polymers, Academic Press, New York, 1968.[57] J. Honerkamp, J. Weese, Rheol. Acta 32 (1) (1993) 57e64.[58] Freiburg, Materials Research Center, Service Group Scientific Data Processing,

NLREG (Non-linear-regularization), Software Version, May 2008.[59] J. Honerkamp, J. Weese, Rheol. Acta 32 (1) (1993) 65e73.[60] M. Matsen, F. Bates, Macromolecules 29 (4) (1996) 1091e1098.[61] E. Helfand, Z.R. Wasserman, Macromolecules 9 (6) (1976) 879e888.[62] V.M. Hoffmann, Die Makromol. Chem. 144 (1) (1971) 309e321.[63] J.S. Chiou, J.W. Barlow, D.R. Paul, J. Appl. Polym. Sci. 30 (6) (1985) 2633e2642.[64] A.F. Ismail, W. Lorna, Sep. Purif. Technol. 27 (3) (2002) 173e194.[65] J.J. Shieh, T.S. Chung, J. Polym. Sci. Part B: Polym. Phys. 37 (20) (1999)

2851e2861.[66] W. Zha, C.D. Han, D.H. Lee, S.H. Han, J.K. Kim, J.H. Kang, C. Park, Macromole-

cules 40 (6) (2007) 2109e2119.