How flow affects crystallization in a heterogeneous polyethylene oxide melt

Abstract

Extension-induced crystallization in a heterogeneous polyethylene oxide (PEO) melt is investigated by small angle X-ray scattering (SAXS) and extensional rheology. The crystalline complex of PEO and sodium bromide is the heterogeneous component, which simulates the roles of crystal or strongly interacted additives during flow. It is aimed to demonstrate how the altered chain relaxation by the heterogeneous particle affects the crystallization. The main findings are listed as follows: (i) strain hardening occurs in heterogeneous melts but not in pure PEO. This indicates that the crystalline complex strengthens the entanglement network of free chains, leading to slower chain relaxation. (ii) In morphology, the orientation of lamellae becomes easier with the existence of the crystalline complex. A jump in the long period is also observed with large strain, accompanied by changes in the evolutional trend at the early stage. (iii) The crystallization kinetics converge with increasing strain; even the stress response significantly differs in different samples. Based on these findings, a quasi-network consisting of the crystalline complex and the entangled free chain is anticipated. Nucleation induced by the stretch of the quasi-network is supposed to lead to the change in crystallization observed.

---

Introduction

The rheological responses of polymer melts reflecting the microscopic chain deformation is a vital concern in the research into flow-induced crystallization (FIC). It is widely accepted that chain deformation during flow determines the subsequent crystallization of the polymer, in either the crystallization kinetics or the morphology.^1–5 The potential for tuning the crystallization makes correlating chain deformation with the crystallization greatly urgent. Nevertheless, it is still a great challenge for current rheological theories to predict chain deformation and relaxation in the non-linear region. The coupling between relaxation and nucleation further hinders the theoretical description of FIC. On the other hand, experimentally it is convenient to tune the rheological properties of a polymer melt through the chain structure design, e.g. long and short chain blend,^6–10 branched chains¹¹ and crosslinked networks.^12,13 The method allows us to qualitatively determine the chain deformation from chain dynamic parameters in the linear viscoelastic region, and correlate it with the crystallization behavior.^9,14,15 This approach has provided plenty of information and in turn promotes more quantitative descriptions of FIC. It is desirable to find out the coupling between chain deformation and nucleation, through purposely changing the rheological properties of a polymer melt.

Besides altering the chain structure, the industrially important particle filled system (PFS) also provides a feasible way to tune the rheological properties. In a PFS, particles commonly have an attractive interaction with a polymer to get reinforcement of the mechanical properties. Chain diffusion in the PFS is thus changed. The heterogeneous melt is supposed to be a combination of a free melt matrix and a confined melt near to the particle’s surface. Rheological investigation shows a second plateau in the storage modulus in the low frequency region,^16–21 indicating a slowed relaxation of the polymer chain. With other techniques it is found that a confined shell about several nanometers exists around the particle. Chain segments in this shell have a lower mobility.^22,23 Based on these works, the long-term diffusion of polymer chains in PFSs is well recognized, while the dynamic response during flow is still lacking characterisation.

The importance of change in the rheological properties of a PFS has not drawn a lot of attention in FIC research, in contrary to the intensive studies on chain dynamics. FIC is important in PFSs because flow-induced crystal orientation and morphology change still affect or even determine the properties of the products.²⁴ Several works on filled polyolefin have been reported. Anisotropic particles, e.g. carbon nanotubes or graphene, are added as fillers. With the same flow conditions, a higher orientation of crystals and easier formation of shish is found.^25–30 This change is explained as a result of an easier chain stretch and higher stability of the nuclei, induced by the strong surface interactions. The explanation focuses on regions near to the particle surface, and so ignores the change in chain relaxation of the matrix. Two factors lead to this ignorance. On one hand, shish formation is the concern of most FIC research, which is highly inhomogeneous in space.^31,32 The localized structure formation naturally makes an explanation concerning the immediate vicinity of particle preferred where flow field differs in space. On the other hand, the critical concentration for the formation of a percolated network of anisotropic particles is low, leading to difficulties in separating the effect of change in the matrix from the formation of a percolated network of particles. For isotropic particles, a change in matrix relaxation happens and a polymer-particle network forms.¹⁸ It is expected that crystallization will be substantially different when the interaction is strong. Our previous work may give an indication of this, where crystals formed during extension transiently crosslink the melt matrix.³³ Strain hardening in the engineering stress–strain curve is found, indicating the occurrence of chain stretch. The orientation is further enhanced, and the evolution of the long period also changes. The result shows that flow can activate the correlation of polymer chains on a large scale, but not localized near to formed crystals. A similar effect is expected in PFS.

Formation of a polyethylene oxide (PEO)–salt crystalline complex (PEO–SCC) supplies a novel way to simulate a PFS. PEO has the ability to solvate cations with its ether oxygen atoms. This ability increases with the anion size of the salt. Three phases, namely the pure crystalline PEO, a salt-rich crystalline complex, and an amorphous PEO phase, coexist at room temperature. Depending on the concentration and type of the salt used, the melting point of the PEO–SCC can be either higher or lower than a PEO crystal.^34–36 The difference in the melting point allows selective melting of the pure PEO crystal. The remaining PEO–SCC will act as a cross-linking point and change the response of free chains to the flow. Compared to the common PFS, the resulting heterogeneous melt has a great advantage in that the conformational change of the chain segment induced by surface interactions is minimized, since the PEO–SCC maintains a folded-chain character. Meanwhile, the strong entanglement with the melt matrix is retained. The difference makes it possible to separate the effect originating in the surface interactions, and the change in the matrix.

In this work, extension-induced crystallization of a heterogeneous PEO melt is investigated with a combination of rheological measurements and SAXS. Sodium bromide (NaBr) is added to PEO through solution blending and a PEO–SCC with a high melting temperature is found to form. NaBr is used as a compromise between a high solubility and a low X-ray adsorption of sodium salt. With the pre-existing PEO–SCC, the PEO melt shows a higher modulus and stronger strain hardening with increasing salt concentration. The long period changes correspondingly, similar to our previous work.^33,37 The crystallization kinetics is accelerated with the existence of the PEO–SCC, while the difference between different samples reduces with increasing strain. A quasi-network composed of the crystalline complex and entangled free chains is supposed, which leads to the change in crystallization behavior.

Experimental

PEO used in this work was purchased from Lian Sheng Chemical Co. of Shanghai with a commercial brand P20. Before use the powder was dried under vacuum at 55 °C for 24 h. Sodium bromide (NaBr) was purchased from Sinopharm Chemical Reagent Co. and dried under vacuum at 100 °C for 24 h before use. The PEO/NaBr blend was prepared by mixing appropriate molar ratios of PEO and NaBr in methanol. The mixture was stirred at 63 °C for 12 h. Then the solvent was vaporized under vacuum at room temperature for 3 days and then at 50 °C for 2 days. Two concentrations, namely 4.9% and 2.0% in weight fraction, were used in this work. For convenience, the two blends and the pure PEO will be referred to as S5 (4.9 wt%), S2 (2.0 wt%) and S0 (pure PEO) hereafter. All the samples were molded to a plate with a thickness of 1 mm by a vulcanizing press. The pure PEO was molded at 85 °C. For S2 and S5, the temperature used was 73 °C, to prevent precipitation of salt under pressure. The samples to be studied were cut into a rectangular shape with dimensions of 28 × 20 × 1 mm³.

The melting point of each sample was determined by differential scanning calorimetry (DSCQ2000, TA Instrument) at a heating rate of 10 °C min⁻¹.

A home-made extensional rheometer³⁷ was used to impose extension, and a Hencky strain³⁸ was obtained. Briefly, the stretch is imposed by two-geared drums rotating in opposite directions. The length of the sample being stretched is equal to the center distance of two drums, thus the strain rate does not change during a stretch with a constant motor speed. Note that in this work the time for acceleration and deceleration of the motor was 95 ms, leading to a non-constant strain rate in a short time extension (less than 260 ms). For convenience the apparent strain rate of 25 s⁻¹ will still be used in descriptions of flow conditions.

Each sample was first heated to 95 °C and held for 10 min to erase the thermal history. Then it was cooled to 64 °C at a rate of 2 °C min⁻¹. A nitrogen gas flow helped to homogenize the temperature and prevent the sample from degrading. The temperature fluctuations were within ±0.5 °C. The extension was imposed on the supercooled melt immediately after reaching 64 °C. The set strain rate was 25 s⁻¹ and the strain was varied from 1.2 to 4.0. Immediately after the cessation of extension, crystallization was monitored by in situ SAXS measurements at the beamline BL16B1 of the Shanghai Synchrotron Radiation Facility. The X-ray wavelength was 0.124 nm and a Mar165 CCD detector (2048 × 2048 pixels with pixel size 80 μm) was employed to collect time-resolved two dimensional (2D) SAXS patterns. The sample-to-detector distance was calibrated to be 5570 mm. Fit2D software from the European Synchrotron Radiation Facility was used to analyze the SAXS patterns in term of the scattering vector q = 4π sin θ/λ, with 2θ as the scattering angle and λ as the X-ray wavelength. A Lorentz correction was applied to all the one-dimensional intensity profiles.

Results and discussion

Fig. 1 shows the thermograms of all the samples by differential scanning calorimetry (DSC), and quantitative results are summarized in Table 1. The curves are shifted vertically to give a better view. During each scan, the sample was heated from 0 °C to 150 °C, cooled to 0 °C, and reheated to 150 °C at a rate of 10 °C min⁻¹. In the first heating run, only one melting peak is observed in S0 at around 68 °C. With addition of salt, a new broad endothermic peak appears above 120 °C in S2 and S5, indicating formation of a PEO–SCC. The enthalpy increases with the salt concentration while the ratio is a little different from that of the weight fraction of salt. This may be a result of the structural differences in PEO–SCCs with different salt contents. In the second heating run, the thermograms show no peak besides the melting of pure PEO crystal, irrespective of the salt concentration. This indicates an irreversible melting of the PEO–SCCs. The irreversibility limits the temperature we can use when erasing the thermal history of the samples. The melting enthalpy of the PEO crystal of all the samples in the second heating run shows a similar value, indicating that the salt does not significantly affect the crystallinity of PEO for the content used.

Fig. 1 Differential scanning calorimetry thermograms of samples: (a) the first heating run. (b) The second heating run.

Table 1 Melting enthalpy and melting point of the samples^a

Sample	TPEO–SCC (°C)	H (J g^−1)	TPEO1st (°C)	H1st (J g^−1)	TPEO2nd (°C)	H2nd (J g^−1)
a TPEO–SCC, TPEO1st and TPEO2nd are the melting points of the PEO–SCC and the PEO crystal in the first and second heating run, respectively. The corresponding enthalpies are labeled as H, H1st and H2nd, respectively. Note that the results are from samples after they were molded by the vulcanizing press.
S5	129.7	17.8	68.8	140.2	59.3	150.2
S2	121.4	6.2	68.4	151.4	61.3	153.7
S0	—	—	68.7	154.3	61.9	156.2

---

Fig. 2 presents the engineering stress–Hencky strain curves of different samples. For S0, the stress increases nearly linearly first, and passes a maximum around 0.19 MPa. After a short decrease, the stress increases again at around strain 2.0. The apparent stress increase is induced by the clamps, as the force of the arm increases when the sample is passing through it. Finally a decrease is observed due to the reduced cross-sectional area of the sample and also the deceleration of the motor. For S5 and S2, a similar stress response is observed, while the absolute value of the first stress maximum increases with an amplitude of up to 50%.

Fig. 2 The engineering stress–Hencky strain curves of different samples. (a) S5. (b) S2. (c) S0. The dashed line indicates the calculated critical stress for strain hardening.

To distinguish the occurrence of strain hardening, the ratio of the two stress maxima in the stress–strain curves is compared. The ratio should be equal in all the samples if no strain hardening happens. Here we choose the pure PEO as the reference. No structural change during stretch will happen at such a high temperature. The ratio with a strain of 3.6 is used, since sample S0 will break during stretch with a strain of 4.0. The critical stress for strain hardening is calculated through eqn (1), as shown by the dashed line in Fig. 2a and b.

It is clear that strain hardening occurs in S5 when strain is larger than 3.3. For S2, the stress fluctuates around the calculated level and only weak strain hardening can happen. The occurrence of strain hardening confirms our expectation that the PEO–SCC will act as a cross-linking point. A quasi-network composed of entangled matrix and the PEO–SCC forms. Chain stretch in the quasi-network is supposed to be the origin of the strain hardening.

The two-dimensional scattering patterns of different samples are presented in Fig. 3, where results with a strain of 3.6 are selected as the representative. Though the salt concentrations and stress responses differ, the evolution of the scattering patterns is similar to each other. Before extension (t = 0 s) there is no periodicity signal of the PEO–SCC, which greatly simplifies the extraction of scattering by PEO. Shortly after stretch, a meridian streak near to the beamstop appears. It gradually transforms to clear maxima, accompanied by a shift of peak position to the large q side in the late stage. Clear second order scattering is also observed in the late stage, indicating a slow perfection of the lamellae periodicity. It is worth pointing out that no equator streak is observed with all the strains, indicating no shish forms. The absence of shish is different from works in shear with the existence of either particles or nuclei.^28,32,39 The limited strain in extension is supposed to be the reason for this.

Fig. 3 The two-dimensional scattering patterns of different samples with a strain of 3.6. The extensional direction is horizontal, as shown by the arrow.

Fig. 4 shows the orientation parameter of different samples. The orientation of lamellae in the final crystallized samples is characterized by Herman's orientation parameter f, which is defined as

Here φ is the angle between the reference direction (extensional direction) and the normal direction of the lamellae. Thus f has a value of 1 when all lamellae are perpendicular to the flow direction, a value of 0 when the lamellae have no preferred orientation and a value of −0.5 when all lamellae are parallel to flow direction.

Fig. 4 The orientation parameters of different samples.

With a strain of 1.2 the orientation parameter of S0 is much lower than that of S2 and S5. Given the relatively small strain, it is concluded that the existence of the PEO–SCC significantly enhances the orientation of the lamellae. The enhancement can be attributed to an easier chain deformation and slower relaxation in the quasi-network. With an increase in strain, the orientation of S0 increases rapidly while that of S2 and S5 changes little. Finally, the orientation parameters of all the samples show a similar value.

Fig. 5a shows the integrated intensity of different samples with a strain of 3.6. It is found that the higher the salt concentration is, the faster the crystallization is. The half-crystallization time t_1/2, defined as the time at which the normalized intensity is 0.5, is presented in Fig. 5b to give a quantitative description. The t_1/2 of S0 with a strain of 1.2 is not given here, since the intensity maximum does not appear during the observation lasting 4880 s. A monotonic decrease of the t_1/2 is observed in all the samples. It is very clear that the addition of salt can accelerate crystallization, while with large strains the difference becomes rather small. This seems similar with the addition of a nucleation agent, where a transition from nucleation agent domination to flow domination is found.^40,41 It is easy to understand this, since flow can strongly enhance the nucleation of a polymer melt, leading to a much higher nuclei density when a weak nucleation agent is used. The underlying assumption is that the nucleation agent does not significantly affect the rheological properties, and that the chain deformation is determined only by the flow conditions. Nevertheless, in this work the stress response clearly shows a different chain deformation with different salt concentrations. The apparent contradiction may also come from the change in chain mobility. The crystallization kinetics is determined by two factors, namely the nucleation rate and the lateral growth rate. If only the change in nucleation rate is considered, a plateau should appear in the kinetics, since a different chain deformation gives the same kinetics only when saturation happens. However, no plateau is observed in any of the samples. This indicates that the growth rate should be taken into account. In the quasi-network, the diffusion of the polymer is reduced. It is reasonable that in S5 the nucleation rate is higher while the growth rate is lower. Thus, similar kinetics as that in S2 is observed with large strain.

Fig. 5 (a) The integrated intensity of different samples with a strain of 3.6. (b) The half-crystallization time of all the samples.

The evolution of the long periods of all the samples are presented in Fig. 6, where the results of S5, S2 and S0 correspond to Fig. 6a–c, respectively. For S5, the long period evolution can be divided into three cases: (I) the long period decreases rapidly at the beginning, followed by a plateau and then a second decrease. This case includes strain 4.0, 3.6 and 3.3. (II) A slight increase of the long period happens before the final decline. Strain 3.3 and 2.7 can be assigned to this case. (III) The long period shows an increase first with larger amplitude and then decreases monotonically. This case includes all the strains below 2.7. The initial value of the three cases shows a step-like increase with a step around 8–10 nm. It is also found that case I and III corresponds to strong strain hardening and no strain hardening, respectively. For case II, it is difficult to determine whether strain hardening happens from the stress, which fluctuates around the calculated critical value. While compared to previous work,³³ the time evolution of the long period indicates occurrence of weak strain hardening.

Fig. 6 The time evolution of the long period of different samples. (a) S5. (b) S2. (c) S0.

The evolution of the long period in S2 is similar to that of case II and III in S5. With a strain of 4.0 and 3.6, the long period increases slightly from about 100 s and begins to decrease at around 400 s. The initial value is also larger than that of other strains. This is similar to case II in S5. With a decrease in strain, the long period shows a more remarkable increase initially and then decreases monotonically, corresponding to case III. The initial increase of the long period is partially unobservable when the strain is less than 2.7, which is the result of slow crystallization kinetics.

The time evolution of the long period in S0 shows a similar behavior, irrespective of the strain used. A marked increase followed by continuous decrease behavior is observed in all the conditions. The uniform evolution is in line with the absence of strain hardening in stress. Again, the evolution can be assigned to case III of S5. The final long period is not observed as the temperature is rather high, which also occurs in the crystallization of S2 and S5.

It is clear that all the evolution modes of the long period appear in S5, which are interestingly the same as with our previous work.^33,37 Similar structural evolution is expected and the main findings in previous works are given as below: crystallization during stretch leads to the occurrence of strain hardening. These crystals act as cross-linking points and further stretch will increase the distance between them, resulting in a jump in the long period. Relaxation of the chain stretch after flow cessation leads to the initial decrease of the long period. Nucleation between two adjacent nuclei is assumed to be suppressed, so that insertion hardly happens at the early stage. With a small strain, the chain orientation dominates. The linear nucleation density is supposed to be the inverse of the long period. With an increase in time the chain orientation relaxes and the linear nucleation density decreases. Consequently, the time-averaged long period shows an increasing trend. For weak strain hardening, an intermediate jump of the long period happens. The evolution also changes from decrease to slight increase, showing a transition state. The final decrease of the long period is induced by the insertion of lamellae into the existing stack. Clearly, the relaxation after flow cessation leads to the long period change at the early stage of crystallization. The trend is determined by whether chain stretch or orientation dominates.

Compared to previous work, the difference is that the temperature used in this work is higher, so the PEO–SCC will be the only effective cross-linking point during extension. Clearly, chain stretch in the strain hardening region can only raise the distance between the PEO–SCC but not that of nuclei forming after the stretch. This means the jump of amorphous thickness or the long period is simultaneous with the formation of nuclei in the free matrix. In other words, the nucleus seems to exclude formation of new ones near to itself. The exclusion originates in a constraint on chain segments by the existing nucleus. This is why an amorphous part always exists in the lamellae stack. The role of chain stretch in the long period jump is that it significantly amplifies the exclusion effect.

One interesting feature of the evolution of the long period in S2 is the short decrease at the beginning, which gradually disappears with decreasing strain. The maximum decrease is about 3 nm when the strain is 4.0, much smaller than that in S5. The result seems in line with our previous assumption that, in conditions of weak strain hardening, a faint decrease of the long period will happen first, corresponding to the relaxation of the weak chain stretch. With a smaller strain the faint decrease is also observed in the heterogeneous melt, which shows no measurable strain hardening. This suggests that the initial decrease is possibly induced by the chain stretch in the vicinity of the PEO–SCC, where the flow field is amplified.

The whole structural evolution can be qualitatively described as follows. In the heterogeneous melt, the free chain interacts with the PEO–SCC through the entanglement network. The melt will respond to flow as a quasi-network, where the PEO–SCC acts as cross-linking point. With increasing strain the free matrix will undergo a transition from orientation to stretch, which leads to the occurrence of strain hardening. The network character also enhances the orientation, as the chain is easier to deform, and the relaxation is slowed down. After flow cessation, relaxation and crystallization happen, leading to a decrease or increase of the long period at the early stage. The trend is determined by whether chain orientation or chain stretch happens. A correspondence between strain hardening and chain stretch is assumed here. After chain stretch, the initial value of the long period increases abruptly. The significant increase is attributed to an amplification of exclusion between adjacent nuclei. Relaxation of the chain stretch leads to a decrease of the long period. In the case of orientation, chain relaxation induces a decrease of the linear nucleation density, and thus the time-averaged long period increases first. When only weak strain hardening happens, an intermediate state, where the long period shows a moderate jump and a weak increase trend, will appear. After the relaxation dominated change, the lateral growth gives a plateau in the evolution of the long period. At the late stage, insertion happens and the long period monotonically decreases. The crystallization kinetics is also affected by the pre-existing PEO–SCC. The reduction of chain diffusion seems to lead to a higher nucleation rate but a lower growth rate in the two blends. This change results in similar kinetics in S5 and S2 with a large strain, even if the chain stretch differs.

Conclusion

In this work, the crystallization of a heterogeneous PEO melt was investigated by SAXS and extensional rheology. The heterogeneous component was the crystalline complex of PEO and NaBr, which has a higher melting point than pure PEO. The crystalline complex was introduced to change the rheological properties of the PEO melt, aiming to simulate FIC of the melt with low crystallinity or additives. A quasi-network with the crystalline complex as a cross-linking point is assumed. An easier orientation of lamellae supports this assumption. Several meaningful findings can be obtained from the crystallization of the heterogeneous melt: (i) different to shear flow, the existence of the heterogeneous particle does not induce shish formation in the extensional flow. The limited strain may be the main reason for this. (ii) The convergence of crystallization kinetics with large strain is attributed to the reduction of chain diffusion but not flow domination, as the stress response differs in different samples. A higher nucleation rate but lower lateral growth rate is supposed in the heterogeneous melt. (iii) The long period jump, which depends on the amplitude of the stress increase in strain hardening region, is a result of chain stretch. The relaxation of the stretch will lead to a decrease of the long period, showing coupling between nucleation and chain relaxation. (iv) Suppression of nucleation as proposed in previous work is generalized as exclusion between nuclei, which is a result of the restricted chain conformation. It is proved that chain stretch does not induce the exclusion but only amplifies it.

Acknowledgements

The authors thank the team in the Synchrotron Radiation Facility of Shanghai for support during the SAXS measurement. This work is supported by the National Natural Science Foundation of China (51033004, 51120135002, 51227801), 973 program of MOST (2010CB934504). The research is also in part supported by the China Postdoctoral Science Foundation (Grant no. 2012M521233), “the Fundamental Research Funds for the Central Universities” and the Project 2013BB05 supported by NPL,CAEP.

References

1. R. H. Somani, B. S. Hsiao, A. Nogales, S. Srinivas, A. H. Tsou, I. Sics, F. J. Balta-Calleja and T. A. Ezquerra, Macromolecules, 2000, 33, 9385–9394.
2. R. R. Lagasse and B. Maxwell, Polym. Eng. Sci., 1976, 16, 189–199.
3. M. R. Mackley, F. C. Frank and A. Keller, J. Mater. Sci., 1975, 10, 1501–1509.
4. A. J. Pennings, J. M. Vanderma and A. M. Kiel, Kolloid-Z. u. Z. Polymere, 1970, 237, 336–358.
5. S. Coppola, N. Grizzuti and P. L. Maffettone, Macromolecules, 2001, 34, 5030–5036.
6. G. Kumaraswamy, J. A. Kornfield, F. J. Yeh and B. S. Hsiao, Macromolecules, 2002, 35, 1762–1769.
7. S. Kimata, T. Sakurai, Y. Nozue, T. Kasahara, N. Yamaguchi, T. Karino, M. Shibayama and J. A. Kornfield, Science, 2007, 316, 1014–1017.
8. M. Seki, D. W. Thurman, J. P. Oberhauser and J. A. Kornfield, Macromolecules, 2002, 35, 2583–2594.
9. O. O. Mykhaylyk, P. Chambon, R. S. Graham, J. P. A. Fairclough, P. D. Olmsted and A. J. Ryan, Macromolecules, 2008, 41, 1901–1904.
10. M. Okura, O. O. Mykhaylyk and A. J. Ryan, Phys. Rev. Lett., 2013, 110, 087801.
11. E. L. Heeley, C. M. Fernyhough, R. S. Graham, P. D. Olmsted, N. J. Inkson, J. Embery, D. J. Groves, T. C. B. McLeish, A. C. Morgovan, F. Meneau, W. Bras and A. J. Ryan, Macromolecules, 2006, 39, 5058–5071.
12. A. Keller and M. R. Mackley, Pure Appl. Chem., 1974, 39, 195–224.
13. D. Liu, N. Tian, K. Cui, W. Zhou, X. Li and L. Li, Macromolecules, 2013, 46, 3435–3443.
14. A. Elmoumni, H. H. Winter, A. J. Waddon and H. Fruitwala, Macromolecules, 2003, 36, 6453–6461.
15. S. Acierno, S. Coppola and N. Grizzuti, J. Rheol., 2008, 52, 551–566.
16. G. Capuano, G. Filippone, G. Romeo and D. Acierno, Langmuir, 2012, 28, 5458–5463.
17. A. S. Sarvestani, Eur. Polym. J., 2008, 44, 263–269.
18. Q. Zhang and L. A. Archer, Langmuir, 2002, 18, 10435–10442.
19. A. Kelarakis and E. P. Giannelis, Polymer, 2011, 52, 2221–2227.
20. A. S. Sarvestani and C. R. Picu, Polymer, 2004, 45, 7779–7790.
21. T. Chatterjee and R. Krishnamoorti, Soft Matter, 2013, 9, 9515–9529.
22. T. Glomann, G. J. Schneider, J. Allgaier, A. Radulescu, W. Lohstroh, B. Farago and D. Richter, Phys. Rev. Lett., 2013, 110, 178001.
23. T. Glomann, A. Hamm, J. Allgaier, E. G. Hubner, A. Radulescu, B. Farago and G. J. Schneider, Soft Matter, 2013, 9, 10559–10571.
24. W. R. Hwang, G. W. M. Peters, M. A. Hulsen and H. E. H. Meijer, Macromolecules, 2006, 39, 8389–8398.
25. M. C. Garcia-Gutierrez, J. J. Hernandez, A. Nogales, P. Pantine, D. R. Rueda and T. A. Ezquerra, Macromolecules, 2008, 41, 844–851.
26. Y.-H. Chen, G.-J. Zhong, J. Lei, Z.-M. Li and B. S. Hsiao, Macromolecules, 2011, 44, 8080–8092.
27. W.-S. Tung, V. Bird, R. J. Composto, N. Clarke and K. I. Winey, Macromolecules, 2013, 46, 5345–5354.
28. N. Patil, L. Balzano, G. Portale and S. Rastogi, Macromolecules, 2010, 43, 6749–6759.
29. P. W. Zhu, A. W. Phillips and G. Edward, J. Chem. Phys., 2012, 136, 054903.
30. J. Z. Xu, C. Chen, Y. Wang, H. Tang, Z. M. Li and B. S. Hsiao, Macromolecules, 2011, 44, 2808–2818.
31. Y. Zhao, G. Matsuba, T. Moriwaki, Y. Ikemoto and H. Ito, Polymer, 2012, 53, 4855–4860.
32. B. Shen, Y. R. Liang, J. A. Kornfield and C. C. Han, Macromolecules, 2013, 46, 1528–1542.
33. N. Tian, D. Liu, X. Li, Z. Wang, S. Zhu, K. Cui, W. Zhou and L. Li, Soft Matter, 2013, 9, 10759–10767.
34. P. R. Sorensen and T. Jacobsen, Polym. Bull., 1983, 9, 47–51.
35. S. Lascaud, M. Perrier, A. Vallee, S. Besner, J. Prudhomme and M. Armand, Macromolecules, 1994, 27, 7469–7477.
36. C. Vachon, M. Vasco, M. Perrier and J. Prudhomme, Macromolecules, 1993, 26, 4023–4031.
37. N. Tian, W. Q. Zhou, K. P. Cui, Y. P. Liu, Y. Y. Fang, X. Wang, L. B. Liu and L. B. Li, Macromolecules, 2011, 44, 7704–7712.
38. M. L. Sentmanat, Rheol. Acta, 2004, 43, 657–669.
39. S. Yamazaki, K. Watanabe, K. Okada, K. Yamada, K. Tagashira, A. Toda and M. Hikosaka, Polymer, 2005, 46, 1675–1684.
40. M. D'Haese, F. Langouche and P. Van Puyvelde, Macromolecules, 2013, 46, 3425–3434.
41. M. D'Haese, P. Van Puyvelde and F. Langouche, Macromolecules, 2010, 43, 2933–2941.

---