Controlling the Morphology of Immiscible ... - ACS Publications

May 5, 2016 - They can form conductive plastics with improved mechanical ... sheets, which can be used as filters or membrane supports. ... coarsening...
2 downloads 0 Views 5MB Size
Article pubs.acs.org/Macromolecules

Controlling the Morphology of Immiscible Cocontinuous Polymer Blends via Silica Nanoparticles Jammed at the Interface Sijia Huang,† Lian Bai,† Milana Trifkovic,‡ Xiang Cheng,*,† and Christopher W. Macosko*,† †

Department of Chemical Engineering and Materials Science, University of Minnesota, Minneapolis, Minnesota 55455, United States Department of Chemical and Petroleum Engineering, University of Calgary, Calgary, Alberta, Canada T2N 1N4



S Supporting Information *

ABSTRACT: Cocontinuous polymer blends have wide applications. They can form conductive plastics with improved mechanical properties. When one phase is extracted, they yield porous polymer sheets, which can be used as filters or membrane supports. However, the cocontinuous morphology is intrinsically unstable due to coarsening during static annealing. In this study, silica nanoparticles, ∼100 nm diameter, with different wetting properties were melt compounded in polyethylene/poly(ethylene oxide) blends. Calculated wetting coefficients of these particles match well with their phase contact angles and their locations in the blends. We demonstrated that a monolayer of particles jamming at interfaces can effectively suppress coarsening and stabilize the cocontinuous morphology. We also correlated the wettability of individual particles at interface to their coarsening suppression ability and found that the most hydrophobic silica nanoparticle is the most effective to arrest coarsening. Moreover, during annealing, we used the rheological dynamic time sweep, a facial but sensitive method, to relate the morphology change with particle dispersion on the interface. We further corroborated these measurements by scanning electron microscopy and confocal microscopy imaging.

1. INTRODUCTION

A different strategy is localization of nanofillers at the interface of the cocontinuous blends. Over a century ago, Pickering23 and Ramsden24 showed that emulsion droplets can be stabilized by jamming of particles that are trapped at the interface. Enlightened by this idea, naonfillers jammed at the interface have also been proposed to stabilize cocontinuous polymer blends. Compared with percolated particle networks within one of the two phases of blends, the interfacial particles can be much more effective in suppressing the coarsening process. The interfacially jammed nanofillers have been realized in blends with clay,25,26 carbon nanotubes, 3,27 carbon blacks,28,29 fumed silica nanoparticles,30−33 and functionalized graphene oxide.34,35 However, the interfacial nanofillers in previous studies appear to form multilayers25,26,34,35 or aggregates,28−32 which means more particles are required to stabilize the blends and also makes observation of individual particles difficult. Because of this formation of multilayers or aggregates, important questions regarding the stabilization effect of interfacial nanofillers on cocontinuous polymer blends have yet to be answered. For example, it is still not clear how to relate the coarsening suppression ability of interfacial particles to the wettability of an individual particle at an interface. It also

Melt blending of immiscible polymers provides an efficient way to obtain high performance materials, since polymer blends can greatly improve properties of their component homopolymers.1−3 Depending on chemical and physical properties of polymers, blend compositions, and processing conditions, immiscible blends can form various morphologies such as droplet, fiber, lamella, and cocontinuous structures.4−6 Among these, cocontinuous structure is particularly interesting due to its unique applications (e.g., the manufacture of conductive plastics7). With extraction of one polymer phase, the remaining porous structure can also be used, for example, as a membrane for filtration or as a tissue scaffold.8−12 However, cocontinuous morphology is inherently unstable; it coarsens to larger pore sizes during annealing and eventually transforms into a dropletmatrix structure. Thus, stabilization of a cocontinuous morphology is crucial for its applications. One method to stabilize cocontinuous polymer blends is adding nanofillers in the polymer blends. If the nanofillers prefer to stay in one phase of the cocontinuous blend, they can form a percolated particle network within one phase, greatly increasing its viscosity which slows down coarsening.13 This strategy has been widely used to stabilize cocontinuous blends with silica nanoparticles,13−15 nanoclay,16,17 carbon black,18,19 carbon nanotubes,20,21 and graphene.22 © XXXX American Chemical Society

Received: January 29, 2016 Revised: April 26, 2016

A

DOI: 10.1021/acs.macromol.6b00212 Macromolecules XXXX, XXX, XXX−XXX

Article

Macromolecules Table 1. Surface Property of Silica Nanoparticlesa

remains controversial how the change of morphology during coarsening affects the dispersion of nanofillers on the interface and the rheology of the blends. In this study, three kinds of colloidal silica nanoparticles with different surface treatments were used to suppress the coarsening of polyethylene/poly(ethylene oxide) cocontinuous polymer blends. As the size of the particle is ∼100 nm, the wettability of different individual particles and the formation of particle monolayers at interface during annealing can be clearly identified by scanning electron micrographs. The morphology change was further quantified by the characteristic pore sizes of blends measured by laser scanning confocal microscopy. By comparing the change in particle distribution at the interface with characteristic pore size change during annealing, we provide unambiguous evidence illustrating the quantitative relation between the arresting of coarsening and the jamming of particles at the interface. Moreover, rheological changes during annealing were correlated with the change of morphology and particle migration in the blends.

θwater is particle contact angle with air and water phase; θLDPE/PEO is measured particle contact angle with LDPE and PEO phase from Figure 3.

a

that the water contact angle of the particles equals to the literature contact angle for the flat dry silane.39 Silanization agents used for surface modification and the corresponding surface properties are summarized in Table 1. 2.2. Differential Scanning Calorimetry. Crystallization temperatures (Tc) during fast quench and melting temperatures (Tm) were determined by differential scanning calorimetry (DSC). Samples (5−9 mg) were placed in hermetically sealed aluminum pans and analyzed with a Q1000 differential scanning calorimeter (TA Instruments). Specimen were first heated at a rate of 10 °C/min to 150 °C and held for 3 min to erase any thermal history in the materials. Samples were then quenched to −100 °C at 60 °C/min to mimic the liquid nitrogen quench after extrusion and heated once more to 150 °C at 10 °C/min. Crystallization and melting temperatures were determined during the cooling and the following heating runs, respectively. DSC results for pure components and neat blends are shown in Figure S3. 2.3. Melt Compounding and Coarsening. All the blends were mixed at 150 °C using a conical twin-screw micro-compounder (Xplore MC 5) at 200 rpm. Silica nanoparticles were premixed with LDPE for 3 min in the compounder. PEO was then added and mixed for additional 5 min with nitrogen purge. The product was extruded and quickly quenched in liquid nitrogen to preserve the morphology of the blends. For all the blends, the concentration of LDPE was kept at 45 wt %. Representative force curves during the blending are shown in Figure S4. A small piece (∼0.5−1.0 g) of extruded sample was placed in between two sheets of fluoropolymer-coated fabric (Premium 6 Mil, American Durafilm) in a steel mold. The sample was then annealed at 150 °C for 3, 5, 10, and 30 min in a Wabash hydraulic press without applied pressure. After the annealing, the sample was cooled down to room temperature in a second water-cooled press. 2.4. Rheology. A 25 mm parallel plate, rotational rheometer (ARES, TA Instruments) was employed to study viscoelastic properties. Pure components and blends were measured with sinusoidal oscillations from 0.01 up to 100 rad/s. All the frequency sweep tests and time sweep tests were performed at blend processing temperature of 150 °C and strain of 5%. 2.5. Morphology Analysis. Samples were cryo-microtomed (Reichert UltraCut S ultramicrotome) at −140 °C with a glass knife and then sputter-coated with 50 Å of platinum. The morphology of the blends was observed using scanning electron microscopy (SEM, JEOL6500) with 5 kV accelerating voltage. In order to study interfaces of the LDPE/PEO system, after cryo-microtoming, samples were immersed in DI water for 3 days to extract PEO phase. Laser scanning confocal microscopy (LSCM, Olympus Fluo View 1000) was performed to obtain 3D images. Samples were cut into ∼200 μm thick slices using a razor blade and then immersed in DI water for 3 days to wash out the PEO phase and generate porous LDPE. Pores were then filled with hydroxyethyl methacrylate (HEMA, Sigma-Aldrich) and 0.1 wt % azobis(isobutyronitrile) (AIBN, Sigma-

2. EXPERIMENTAL SECTION 2.1. Materials. We used low-density polyethylene (LDPE) and poly(ethylene oxide) (PEO) as the two polymer components. LDPE (LDPE 955I, ρ = 0.925 g/cm3, Tm ∼ 112 °C) and a specially prepared silica-free PEO (ρ = 1.21 g/cm3, Tm ∼ 62 °C) were provided by the Dow Chemical Company. Figure 1 shows the frequency-dependent

Figure 1. Storage modulus and complex viscosity of pure materials at 150 °C. storage modulus (G′) and complex viscosity (η*) of LDPE and PEO with the fit of the Cross model of the viscosity. Molecular weights and viscosity profiles are given as Supporting Information (Table S1 and Figure S1) along with a comparison of this PEO to a commercial product (Polyox N10). To test the effect of hydrophobicity, we used particles with three different surface treatments. Untreated, i.e., hydrophilic, silica nanoparticles (A-SNP) were synthesized in our lab using the standard Stober method.36,37 A-SNP particles were labeled with a fluorescent dye, fluorescein isothiocyanate (FITC). A detailed synthesis method has been described in our previous paper.38 Partially hydrophobic (BSNP) and hydrophobic (C-SNP) silica particles were provided by the Cabot Corporation, which are hydrophilic silica particles treated with different silanization agents as shown in Table 1. All were colloidal particles with density of 2.2 g/cm3. Number-average particle diameters were determined by measuring at least 50 particles in SEM images. The particles size distributions were fit to Gaussian distribution (Figure S2). Standard deviations of the size distributions give the particle-size dispersity. To quantify the relative hydrophobicity of particles, we assumed that the particles are fully covered with silane so B

DOI: 10.1021/acs.macromol.6b00212 Macromolecules XXXX, XXX, XXX−XXX

Article

Macromolecules Aldrich) as thermal initiator.25,40 The refractive index of polyHEMA (refractive index, n = 1.5141) and LDPE (n = 1.5141) match well, which makes polyHEMA suitable for confocal imaging. The porous LDPE samples filled with fluorescent labeled HEMA (Rhodamine B, Sigma-Aldrich) and AIBN were heated at 80 °C until the samples became transparent, indicating that the HEMA has polymerized to polyHEMA. Samples were observed under LSCM with a 563 nm laser under a 20× dry objective lens or 60× oil-immersion objective lens, depending on the characteristic pore size of the blends. Under LSCM, LDPE was black, whereas polyHEMA with Rhodamine B that replaced the original PEO phase was bright red. The 3D image stacks from LSCM were analyzed using Avizo software. The program fitted a triangular mesh to the 3D interface and calculated the interfacial area, Aconfocal, by adding up the area of the triangles. The average characteristic pore size, d, was then determined by

d = Vs/Aconfocal

Table 2. Interfacial Tension and Silica Localization of the Blendsa

a

The asterisk indicates the wetting parameter for B-SNP was calculated based on literature of water contact angle and surface energy as shown in the Supporting Information.

(1)

where Vs is the total volume of the probed sample. Further details of this calculation method were given in previous studies.4,25,40

3. RESULTS AND DISCUSSION 3.1. Wetting and Contact Angle. The interfacial location of silica nanoparticles can be understood from the wetting coefficient, ω42 ω = cos θ = (γSi − PEO − γSi − LDPE)/γPEO − LDPE

(2)

where θ is the particle contact angle at the interface and γSi−LDPE, γSi−PEO, and γLDPE−PEO are the interfacial energies between silica nanoparticles and LDPE, between silica nanoparticles and PEO, and between LDPE and PEO, respectively. Because of the difficulties in measuring the interfacial energy between each phase, they were calculated based on the component surface energies and harmonic mean of the dispersive and polar distribution.43,44 For example γPEO − LDPE = γPEO + γLDPE − 2(γPEOdγLDPE d)1/2 − 2(γPEO pγLDPE p)1/2

(3)

where γd and γp are the dispersive and the polar part of the surface energy of the components at processing temperature, respectively, which were obtained from literature44,45 and shown in Table S2. Nanoparticles prefer to locate at the interface between LDPE and PEO when −1 < ω < 1. If ω < −1, the particles locate in the PEO phase, whereas silica particles prefer to stay in the LDPE phase when ω > 1. The interfacial tensions and wetting coefficients are shown in Table 2. Because of the high interfacial tension of PEO and the hydrophilic silica, the hydrophilic particles are predicted to aggregate in PEO phase. But the two hydrophobic silica nanoparticles should locate at the interface because the calculated wetting coefficients are within the range between −1 and 1. Confocal and SEM images of the blend with 4 wt % A-SNP particles without annealing in Figure 2 confirmed the prediction from wetting coefficients. The bright green dots in the confocal image (Figure 2a) are the FITC-labeled A-SNP, suggesting that these particles are aggregated and all located in one phase. This is confirmed by the SEM images (Figure 2c,d) where particles are only observed in the PEO phase and many formed into aggregates as shown by the red circles in Figure 2c,d. Moreover, few interfacial particles were observed in Figure 2b. As our focus is on how interfacial particles can effectively suppress the coarsening and stabilize the cocontinuous

Figure 2. Confocal image (a) and SEM image (b−d) of LDPE/PEO/4 wt % A-SNP without annealing. Images c and d show particle aggregates circled in red in the PEO phase.

morphology, A-SNP was not further studied in this work. Hence, only the two hydrophobic particles, B-SNP and C-SNP, were selected for further experiments. By extracting the PEO phase from the blend without annealing with DI water, silica nanoparticles lying in the LDPE/ PEO interface could be observed via SEM. We found that the B-SNP particles were preferentially wetted by the PEO phase in Figure 3a, while the C-SNP particles were more equally wetted by both phases in Figure 3b. Quantitatively, we measured the embedded diameter, l, from 2D isolated particles on flat interfacial regions, which allowed us to estimate the contact angle, θ (Figure 4): l /2 (4) R Averaging over 20 particles, we calculated the contact angles of B-SNP and C-SNP, as shown in the last column of Table 1. As all the A-SNP particles appeared to remain in the PEO phase, the contact angle of A-SNP particles was assumed to be 0°. Particles with contact angle closer to 90° are predicted to more θ = sin−1

C

DOI: 10.1021/acs.macromol.6b00212 Macromolecules XXXX, XXX, XXX−XXX

Article

Macromolecules

The blends with hydrophobic silica show very different coarsening from the neat blends. For the LDPE/PEO/4 wt % B-SNP blend, the characteristic pore size increases in the first 5 min of annealing, similar to the neat blend. However, after 5 min annealing, the characteristic pore size reaches a plateau of ∼35 μm. By comparison, with 4 wt % C-SNP particles, the characteristic pore size increases slightly and stops at ∼6 μm after only 3 min annealing. Therefore, the hydrophobic C-SNP particles are more effective to suppress the coarsening of cocontinuous polymer blends than the partially hydrophobic BSNP particles, even though both particles prefer to locate at interface from the wetting coefficient calculation. Figure S6 compares the coarsening of silica-stabilized blends with that of nanoclay-stabilized LDPE/PEO blends.25 With only 2 wt % Cloisite 20A, the coarsening effect is completely suppressed and the final pore size is smaller than 4 wt % C-SNP particles. The large aspect ratio of nanoclay allows for a more effectively coverage of the interface. Similarly, we expect silica nanoparticles of smaller sizes to be more effective for stabilizing the cocontinuous structures. The SEM images of blends with 4 wt % C-SNP particles before and after annealing are shown in Figure 6 to show the interfacial localization of particles. The gap between LDPE and PEO phases before extraction is due to the different thermal contraction of LDPE and PEO. During liquid nitrogen quenching after melt mixing, the LDPE solidifies at ∼89 °C

Figure 3. SEM micrographs of silica nanoparticles on LDPE/PEO interface without annealing, after PEO phases extracted: (a) LDPE/ PEO/B-SNP, contact angle ∼60°; (b) LDPE/PEO/C-SNP, contact angle ∼78°.

Figure 4. Schematic for contact angle measurements.

efficiently stabilize cocontinuous blends since they can bind better to both phases. 3.2. Coarsening. Figure 5 shows the characteristic pore size as a function of annealing time at 150 °C for B-SNP and C-

Figure 5. Characteristic pore size, d, as a function of annealing time at 150 °C for LDPE/PEO neat blend and blends with different kinds of hydrophobic silica nanoparticles. The corresponding 3D structures from confocal microscopy after 10 min annealing are also shown.

SNP particles. The neat LDPE/PEO blend shows a continuous increase of the pore size. The coarsening rate is qualitatively similar to the coarsening rate of LDPE/PEO of previous studies25,40 and matches reasonably well with the coarsening rates of other cocontinuous polymer blends (Figure S7).46 Note that although the commercial 100K PEO (Polyox N10) used in previous studies contains ∼2 wt % fumed silica, the coarsening rate of the neat blends with particle-free PEO and commercial PEO are very similar (Figure S5). Hydrophilic fume silica remains in the PEO phase and therefore has a negligible effect on the cocontinuous structures. Moreover, the similar low shear viscosities of PEO (Figure S1) indicate that the 2 wt % extra fumed silica in commercial PEO do not modify the dynamics of the melt blending or coarsening processes.

Figure 6. SEM micrographs of LDPE/PEO with 4 wt % C-SNP: (a, b) without annealing; (c, d) after 3 min annealing at 150 °C; (e, f) after 30 min annealing at 150 °C; (b, d, f) after PEO extraction. D

DOI: 10.1021/acs.macromol.6b00212 Macromolecules XXXX, XXX, XXX−XXX

Article

Macromolecules while PEO crystallizes at ∼36 °C (Figure S3). As such, PEO shrinks more than LDPE in the temperature range between 89 and 36 °C. As shown in Figure 6a,b, before annealing, C-SNP particles form into small patches on the interface due to the attraction of capillary forces between neighboring particles.47 Similar patches have also been observed on the interface of nonpolar drops48 and nonpolar bijels.38 With the reduction of interfacial area during coarsening, the patches of C-SNP particles on the interface quickly merge and prevent any further area reduction. As a result, monolayers of particles formed at interface and the coarsening stopped after 3 min annealing (Figure 6c,d). With further annealing, particles rearranged along the interface as shown in Figure 6e,f. SEM images for blends with B-SNP particles are shown in Figure S8. 3.3. Concentration Dependence and Final Coverage. We also investigated the effect of particle concentration on morphology change during annealing, which allowed us to determine the minimum amount C-SNP particle loading that can stabilize the LDPE/PEO blends. We varied the concentration of C-SNP particles from 1 to 4 wt %. Figure 7

particle. The number of particles in the blends is given by N = (Vsϕ)/(4/3πR3), where Vs is the volume of samples. Thus, if all N particles are in the interface A silica = N ·2 3 R2 = 3 3 Vsϕ/(2πR )

The final relation in eq 5 provides an upper limit of the surface coverage. Because of the finite resolution of confocal microscopy (∼500 nm), the interfacial area measured from the confocal images is smaller than the actual interfacial area, Aconfocal < Atrue. SEM images in Figure 6e,f and Figure S8c,d show significant submicron roughness, which is clearly missed in Aconfocal. Although such submicron roughness does not influence the estimate of characteristic pore size on the order of a few microns (eq 1), it affects the evaluation of n, since the size of silica particles is on the order of 100 nm. Table 3 gives characteristic pore size at 30 min, after coarsening has been arrested, for each volume fraction. The Table 3. Final Coverage Calculation Based on Aconfocal for LDPE/PEO with Different C-SNP Particle Loadings after 30 min Annealing at 150 °C

shows characteristic pore size as a function of annealing time. For all blends, characteristic pore size reaches a plateau at long annealing times. However, blends with higher particle loading take less time to reach the plateau and show smaller pore sizes. Coarsening stops when particles fully cover the interface. As a result, the interfacial area in a stabilized cocontinuous blend should be larger at higher particle loadings, leading to smaller characteristic pore sizes. To quantitatively understand the concentration dependence stabilization of LDPE/PEO blends, we estimated the final coverage of the silica particles at the interface, n, which is defined as the number of particle layers at the LDPE/PEO interface. n is calculated as the silica surface area divided by the blend interfacial area A silica A silica 3 3 dϕ < = = nmax A true Aconfocal 2πR

C-SNP (wt %)

ϕ (vol %)

di (μm)

df (μm)

nmax

1.0 2.0 3.0 4.0

0.5 1.0 1.5 2.0

5.6 5.3 5.0 4.2

33.5 16.3 12.6 6.1

2.5 2.4 2.9 1.9

calculated upper limit of the number of layers based on eq 5 is nmax ∼ 2 independent of initial volume fraction, which is consistent with the hypothesis that coarsening stops when the particles jam. Our direct SEM images (Figures 6c,d) show monolayers of particles, i.e., n ∼ 1. The difference is due to the fact that Aconfocal misses the roughness and is thus smaller than Atrue. 3.4. Rheological Changes. While the dynamics of structural variation during annealing can be studied with SEM and confocal microscopy, the role of interfacial particles on the rheology of polymer blends is still unknown. Shear modulus was measured as a function of time to explore the change of rheological properties during annealing. The suppression of coarsening by interfacial particles was also confirmed by dynamic frequency sweeps done after 1 h annealing at 150 °C. After 1 h, the cocontinuous morphology of the samples has already been stabilized by the interfacial particles. The frequency change of the storage modulus of the blends is shown in Figure 8. At low frequency region, a terminal

Figure 7. Characteristic pore size, d, as a function of annealing time at 150 °C for LDPE/PEO/C-SNP, with different loadings of C-SNP particles.

n=

(6)

(5)

where Asilica is the total surface area covered by silica particles, Atrue is the actual interfacial area of blends, d is the characteristic pore size, ϕ is the volume fraction of silica nanoparticles, and R is the radius of particles. Here, we use eq 1 for Aconfocal. For Asilica, we assume particles pack in a hexagonal lattice along the interface. The area of a 2D Voronoi cell of the hexagonal lattice, 2√3R2, gives the interfacial area that is covered by a single

Figure 8. Storage modulus as a function of frequency for LDPE/PEO neat blend and blends filled with silica nanoparticles after 1 h annealing at 150 °C. The plots are measured at 150 °C and strain 5%. E

DOI: 10.1021/acs.macromol.6b00212 Macromolecules XXXX, XXX, XXX−XXX

Article

Macromolecules

Figure 9. (a) Top: characteristic pore size, d, as a function of annealing time at 150 °C for LDPE/PEO/C-SNP and B-SNP. The same plot is shown in Figure 7. Bottom: interfacial storage modulus, G′inf, as a function of annealing time at 150 °C for LDPE/PEO/C-SNP and B-SNP and neat blends. The rheology plots are measured at 1 rad/s and 5% strain. (b) SEM images of 4 wt % B-SNP particle blend with 10 min annealing without PEO extracted. (c) SEM images of 1 wt % C-SNP particles blend with 10 min annealing with PEO extracted. The red circles indicated the bare regions, which are not covered by SNP particles.

from 7 to 15 min annealing, which is different from the 4 wt % C-SNP particles blend. We attribute this sharp increase of G′inf to patch densif ication during coarsening, which leads to the jamming of particles on the interface. As shown in Figure 9 b,c, patches of silica nanoparticles have formed on the interface. However, bare regions without interfacial particles still exist as indicated by the red circles. The further shrinkage of interfacial area during coarsening, as shown by the increase of characteristic pore size between 7 and 15 min in Figure 9a, drives neighboring patches together to remove bare regions and form larger patches until particles fully jam on the interface. Once the interface has been fully covered by particles and the morphology has been stabilized after ∼15 min for blends with 4 wt % B-SNP and 1 wt % C-SNP particles, the sharp increase of G′inf changes to a gradually slower increase. This gradually increase of G′inf is due to particle rearrangement, which is also found in the blend with 4 wt % C-SNP particles and our related study of bijels.38 Particle rearrangement along the interface is at submicron scale and cannot be observed directly by confocal microscopy. At this stage, G′inf shows a slow increase after particle jamming on the interface, even though few changes of the characteristic pore size can be observed. In summary, the change in G′inf during annealing is an indirect yet more sensitive and facile rheological method compared with SEM and can provide us more information about the particle dispersion on the interfaces, e.g., patch densification, particle jamming, and particle rearrangement.

slope of 0.96 is observed for the neat blends, indicating the formation of drop-matrix morphology after 1 h annealing.49−51 However, the blends filled with B-SNP and C-SNP particles show power-law dependences with smaller terminal slopes (0.47 and 0.40 for blends with B-SNP and C-SNP particles, respectively) in the low frequency region, which is characteristic for cocontinuous structure.49−51 As described in our previous work,38 the storage modulus of blends can be decomposed into a contribution from the components and the interface. G′blend = G′component + G′inf

(7)

Figure 9 shows the change of interfacial storage modulus (G′inf) with annealing time at 150 °C. The time evolution of G′inf of the neat blends is similar to previous studies of cocontinuous blends: G′inf decreases during annealing due to the decrease of interfacial area.49−51 As the sample loading time is ∼7 min, the neat blend has already coarsened to a large characteristic pore size after 7 min annealing. The slow decrease of G′inf indicted the breakup of network structure at late times of annealing.49 The blend with 4 wt % C-SNP particles shows the highest G′inf, which is consistent with a higher particle loading leading to larger interfacial area and smaller characteristic pore size. However, the sample loading time for each measurement is ∼7 min. As shown in Figure 9a, pore size has already reached a plateau and stabilized after 7 min annealing for blend with 4 wt % C-SNP particles. Therefore, the rheology time sweep test missed the coarsening stage of all blends, which explains the absence of the initial decrease of G′inf which was seen in previous studies.38,49−51 For the blends with 4 wt % B-SNP and 1 wt % C-SNP particles, the initial decrease of G′inf also cannot be observed. However, the G′inf in these two blends shows a sharp increase

4. CONCLUSIONS In this study, we showed that hydrophobic silica nanoparticles jammed at interfaces in the form of monolayers can effectively suppress coarsening and stabilize cocontinuous LDPE/PEO blends. A simple model based on wetting coefficients and phase F

DOI: 10.1021/acs.macromol.6b00212 Macromolecules XXXX, XXX, XXX−XXX

Article

Macromolecules

University of Minnesota), NSERC (Natural Sciences and Engineering Research Council of Canada), and PRF (Petroleum Research Fund) of the American Chemical Society (54168-DNI9). The authors thank Garrett Swindlehurst and John Fruehwirth for their help in particle synthesis, Liangliang Gu for his help in image processing and polymer processing, Dr. Guillermo Marques and John Oja (University Imaging Center) for their help in image acquisition, and Dr. Yaoying Wu and Anatolii Purchel for their help collecting GPC data. The authors acknowledge the Minnesota Supercomputing Institute (MSI) at the University of Minnesota for providing resources that contributed to the research results reported within this paper. The authors also thank Dr. Angelos Kyrlidis of Cabot Corp. for the silica and Dr. Matthew Hansen from the Dow Chemical Company for providing silica free poly(ethylene oxide) for this study.

contact angles of individual particles successfully predicts the locations of particles. With appropriate surface treatment, 4 wt % hydrophobic silica nanoparticles (C-SNP particles) can stop coarsening within 3 min. Because of the shrinkage of interfacial area during coarsening, particles merge and jam at the interface to prevent further interface reduction. SEM images directly confirm that the treated silica particles reside in the interface in patches which come together into a hexagonally packed monolayer as the area shrinks during annealing. By comparing hydrophobic particles with different hydrophobicity (B-SNP and C-SNP particles), we found that blends filled with C-SNP particles show a smaller characteristic pore size and faster suppression of coarsening, demonstrating that particles with θLDPE/PEO close to 90° are more effective in stabilizing the cocontinuous blends. By varying particle concentration, we found that 1 wt % CSNP particle loading is enough to suppress coarsening within 10 min annealing. With higher particles loading, blends stabilize in a shorter time and the characteristic pore size is smaller. Our estimates based on interfacial area from confocal microscope images showed that the blends coarsen to approximately the same final coverage, independent of particle loading. The calculation indicates two layers of particles on the interface, although we believe that due to the finite resolution, confocal microscopy misses submicron roughness and underestimates the true interfacial area. SEM images confirm this roughness and that the interfaces in these blends are covered with hexagonally close packed monolayers of silica particles. Frequency sweeps of storage modulus of the blends data shows a drop-matrix structure after 1 h annealing for the neat blend whereas the blends with hydrophobic silica nanoparticles display power-law dependences, characteristic of cocontinuous structures. The change of the interfacial storage modulus during annealing obtained from rheology time sweeps allows us to relate the morphology change with particle dispersions on the interface, consistent with the SEM images. The rheological measurements further confirm that hydrophobic particles effectively stabilize the LDPE/PEO cocontinuous structures and significantly modify the mechanical properties of polymer blends.





ASSOCIATED CONTENT

S Supporting Information *

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acs.macromol.6b00212. Molecular weight and viscosity of the polymers, the DSC data for the polymers and the neat blend, literature surface energies for polymers and particles, melt blending force curves, and coarsening comparisons with the previous studies (PDF)



REFERENCES

(1) Schürmann, B. L.; Niebergall, U.; Severin, N.; Burger, C.; Stocker, W.; Rabe, J. P. Polyethylene (PEHD)/Polypropylene (iPP) Blends: Mechanical Properties, Structure and Morphology. Polymer 1998, 39, 5283−5291. (2) Daux, V.; Prochazka, F.; Carrot, C. Cocontinuous Morphology in Vinylidene Fluoride Based Polymers/Poly (Ethylene Oxide) Blends. J. Appl. Polym. Sci. 2013, 128, 265−274. (3) Xiang, F.; Shi, Y.; Li, X.; Huang, T.; Chen, C.; Peng, Y.; Wang, Y. Cocontinuous Morphology of Immiscible High Density Polyethylene/ Polyamide 6 Blend Induced by Multiwalled Carbon Nanotubes Network. Eur. Polym. J. 2012, 48, 350−361. (4) Lopez-Barron, C. R.; Macosko, C. W. Characterizing Interface Shape Evolution in Immiscible Polymer Blends via 3D Image Analysis. Langmuir 2009, 25, 9392−9404. (5) Scott, C. E.; Macosko, C. W. Morphology Development During the Initial Stages of Polymer-Polymer Blending. Polymer 1995, 36, 461−470. (6) Tucker, C. L., III; Moldenaers, P. Microstructural Evolution in Polymer Blends. Annu. Rev. Fluid Mech. 2002, 34, 177−210. (7) Pötschke, P.; Paul, D. R. Formation of Co-continuous Structures in Melt-Mixed Immiscible Polymer Blends. J. Macromol. Sci., Polym. Rev. 2003, 43, 87−141. (8) Yang, Q.; Adrus, N.; Tomicki, F.; Ulbricht, M. Composites of Functional Polymeric Hydrogels and Porous Membranes. J. Mater. Chem. 2011, 21, 2783−2811. (9) Ulbricht, M. Advanced Functional Polymer Membranes. Polymer 2006, 47, 2217−2262. (10) Slaughter, B. V.; Khurshid, S. S.; Fisher, O. Z.; Khademhosseini, A.; Peppas, N. A. Hydrogels in Regenerative Medicine. Adv. Mater. 2009, 21, 3307−3329. (11) Van Vlierberghe, S.; Dubruel, P.; Schacht, E. Biopolymer-Based Hydrogels as Scaffolds for Tissue Engineering Applications: A Review. Biomacromolecules 2011, 12, 1387−1408. (12) Esquirol, A. L.; Sarazin, P.; Virgilio, N. Tunable Porous Hydrogels from Cocontinuous Polymer Blends. Macromolecules 2014, 47, 3068−3075. (13) Li, L.; Miesch, C.; Sudeep, P. K.; Balazs, A. C.; Emrick, T.; Russell, T. P.; Hayward, R. C. Kinetically Trapped Co-continuous Polymer Morphologies through Intraphase Gelation of Nanoparticles. Nano Lett. 2011, 11, 1997−2003. (14) Liu, X. Q.; Li, R. H.; Bao, R. Y.; Jiang, W. R.; Yang, W.; Xie, B. H.; Yang, M. B. Suppression of Phase Coarsening in Immiscible, Cocontinuous Polymer Blends under High Temperature Quiescent Annealing. Soft Matter 2014, 10, 3587−3596. (15) Liu, X. Q.; Wang, Q. Y.; Bao, R. Y.; Yang, W.; Xie, B. H.; Yang, M. B. Suppressing Phase Retraction and Coalescence of CoContinuous Polymer Blends: Effect of Nanoparticles and Particle Network. RSC Adv. 2014, 4, 49429−49441.

AUTHOR INFORMATION

Corresponding Authors

*E-mail [email protected] (X.C.). *E-mail [email protected] (C.W.M.). Notes

The authors declare no competing financial interest.



ACKNOWLEDGMENTS This project is funded by IPRIME (the Industrial Partnership for Research in Interfacial and Materials Engineering at the G

DOI: 10.1021/acs.macromol.6b00212 Macromolecules XXXX, XXX, XXX−XXX

Article

Macromolecules (16) Filippone, G.; Dintcheva, N. T.; La Mantia, F. P.; Acierno, D. Using Organoclay to Promote Morphology Refinement and CoContinuity in High-Density Polyethylene/Polyamide 6 Blends−Effect of Filler Content and Polymer Matrix Composition. Polymer 2010, 51, 3956−3965. (17) Filippone, G.; Dintcheva, N. T.; Acierno, D.; La Mantia, F. P. The Role of Organoclay in Promoting Co-continuous Morphology in High-Density Poly (Ethylene)/Poly (Amide) 6 Blends. Polymer 2008, 49, 1312−1322. (18) Gubbels, F.; Jérôme, R.; Teyssie, P.; Vanlathem, E.; Deltour, R.; Brédas, J. L. Selective Localization of Carbon Black in Immiscible Polymer Blends: A Useful Tool to Design Electrical Conductive Composites. Macromolecules 1994, 27, 1972−1974. (19) Wu, G.; Li, B.; Jiang, J. Carbon Black Self-Networking Induced Co-continuity of Immiscible Polymer Blends. Polymer 2010, 51, 2077−2083. (20) Li, Y.; Shimizu, H. Conductive PVDF/PA6/Cnts Nanocomposites Fabricated by Dual Formation of Cocontinuous and Nanodispersion Structures. Macromolecules 2008, 41, 5339−5344. (21) Pötschke, P.; Bhattacharyya, A. R.; Janke, A. Morphology and Electrical Resistivity of Melt Mixed Blends of Polyethylene and Carbon Nanotube Filled Polycarbonate. Polymer 2003, 44, 8061− 8069. (22) Mao, C.; Zhu, Y.; Jiang, W. Design of Electrical Conductive Composites: Tuning the Morphology to Improve the Electrical Properties of Graphene Filled Immiscible Polymer Blends. ACS Appl. Mater. Interfaces 2012, 4, 5281−5286. (23) Pickering, S. U. Cxcvi. J. Chem. Soc., Trans. 1907, 91, 2001− 2021. (24) Ramsden, W. Separation of Solids in the Surface-Layers of Solutions and ’Suspensions’ (Observations on Surface-Membranes, Bubbles, Emulsions, and Mechanical Coagulation).–Preliminary Account. Proc. R. Soc. London 1903, 72, 156−164. (25) Trifkovic, M.; Hedegaard, A. T.; Sheikhzadeh, M.; Huang, S.; Macosko, C. W. Stabilization of PE/PEO Cocontinuous Blends by Interfacial Nanoclays. Macromolecules 2015, 48, 4631−4644. (26) Dhibar, A. K.; Kim, J. K.; Khatua, B. B. Cocontinuous Phase Morphology of Asymmetric Compositions of Polypropylene/HighDensity Polyethylene Blend by the Addition of Clay. J. Appl. Polym. Sci. 2011, 119, 3080−3092. (27) Cohen, E.; Zonder, L.; Ophir, A.; Kenig, S.; McCarthy, S.; Barry, C.; Mead, J. Hierarchical Structures Composed of Confined Carbon Nanotubes in Cocontinuous Ternary Polymer Blends. Macromolecules 2013, 46, 1851−1859. (28) Gubbels, F.; Blacher, S.; Vanlathem, E.; Jérôme, R.; Deltour, R.; Brouers, F.; Teyssie, P. Design of Electrical Composites: Determining the Role of the Morphology on the Electrical Properties of Carbon Black Filled Polymer Blends. Macromolecules 1995, 28, 1559−1566. (29) Tan, Y.; Song, Y.; Cao, Q.; Zheng, Q. Characterization of Carbon Black-Filled Immiscible Polypropylene/Polystyrene Blends. Polym. Int. 2011, 60, 823−832. (30) Feng, J. M.; Liu, X. Q.; Bao, R. Y.; Yang, W.; Xie, B. H.; Yang, M. B. Suppressing Phase Coarsening in Immiscible Polymer Blends using Nano-Silica Particles Located at the Interface. RSC Adv. 2015, 5, 74295−74303. (31) Chung, H. J.; Ohno, K.; Fukuda, T.; Composto, R. J. SelfRegulated Structures in Nanocomposites by Directed Nanoparticle Assembly. Nano Lett. 2005, 5, 1878−1882. (32) Gam, S.; Corlu, A.; Chung, H. J.; Ohno, K.; Hore, M. J.; Composto, R. J. A Jamming Morphology Map of Polymer Blend Nanocomposite Films. Soft Matter 2011, 7, 7262−7268. (33) Chung, H. J.; Kim, J.; Ohno, K.; Composto, R. J. Controlling the Location of Nanoparticles in Polymer Blends by Tuning the Length and End Group of Polymer Brushes. ACS Macro Lett. 2011, 1, 252− 256. (34) Tan, Y.; Fang, L.; Xiao, J.; Song, Y.; Zheng, Q. Grafting of Copolymers onto Graphene by Miniemulsion Polymerization for Conductive Polymer Composites: Improved Electrical Conductivity

and Compatibility Induced by Interfacial Distribution of Graphene. Polym. Chem. 2013, 4, 2939−2944. (35) Imperiali, L.; Clasen, C.; Fransaer, J.; Macosko, C. W.; Vermant, A. Simple Route Towards Graphene Oxide Frameworks. Mater. Horiz. 2014, 1, 139−145. (36) Stö ber, W.; Fink, A.; Bohn, E. Controlled Growth of Monodisperse Silica Spheres in the Micron Size Range. J. Colloid Interface Sci. 1968, 26, 62−69. (37) Bogush, G. H.; Tracy, M. A.; Zukoski, C. F. Preparation of Monodisperse Silica Particles: Control of Size and Mass Fraction. J. Non-Cryst. Solids 1988, 104, 95−106. (38) Bai, L.; Fruehwirth, J. W.; Cheng, X.; Macosko, C. W. Dynamics and Rheology of Nonpolar Bijels. Soft Matter 2015, 11, 5282−5293. (39) Lim, S. C.; Kim, S. H.; Lee, J. H.; Kim, M. K.; Zyung, T. SurfaceTreatment Effects on Organic Thin-Film Transistors. Synth. Met. 2005, 148, 75−79. (40) Trifkovic, M.; Hedegaard, A.; Huston, K.; Sheikhzadeh, M.; Macosko, C. W. Porous Films via PE/PEO Cocontinuous Blends. Macromolecules 2012, 45, 6036−6044. (41) Brandrup, J.; Immergut, E. H.; Grulke, E. A. Polymer Handbook, 4th ed; John Wiley& Sons Inc.: New York, 1998; Chapter VI, p 571. (42) Vandebril, S.; Vermant, J.; Moldenaers, P. Efficiently Suppressing Coalescence in Polymer Blends using Nanoparticles: Role of Interfacial Rheology. Soft Matter 2010, 6, 3353−3362. (43) Fenouillot, F.; Cassagnau, P.; Majesté, J. C. Uneven Distribution of Nanoparticles in Immiscible Fluids: Morphology Development in Polymer Blends. Polymer 2009, 50, 1333−1350. (44) Van Krevelen, D. W.; Te Nijenhuis, K. Properties of Polymers: Their Correlation with Chemical Structure; Their Numerical Estimation and Prediction from Additive Group Contributions, 4th ed.; 2009; Chapter 8, p 235. (45) Kasemura, T.; Yamaguchi, S.; Hata. Studies on the Adsorption Behavior of Polymers by Interfacial Tension. II. Molecular Weight Dependence of Interfacial Tension of Poly(Ethylene Glycol)S. Kobunshi Ronbunshu 1987, 44, 657−662. (46) Hedegaard, A. T.; Gu, L.; Macosko, C. W. Effect of Extensional Viscosity on Cocontinuity of Immiscible Polymer Blends. J. Rheol. 2015, 59, 1397−1417. (47) Stamou, D.; Duschl, C.; Johannsmann, D. Long-range attraction between colloidal spheres at the air-water interface: The consequence of an irregular meniscus. Phys. Rev. E: Stat. Phys., Plasmas, Fluids, Relat. Interdiscip. Top. 2000, 62, 5263. (48) Cheng, H. L.; Velankar, S. S. Controlled Jamming of ParticleLaden Interfaces using A Spinning Drop Tensiometer. Langmuir 2009, 25, 4412−4420. (49) López-Barrón, C. R.; Macosko, C. W. Rheological and Morphological Study of Cocontinuous Polymer Blends During Coarsening. J. Rheol. 2012, 56, 1315−1334. (50) Yu, W.; Zhou, W.; Zhou, C. Linear Viscoelasticity of Polymer Blends with Co-continuous Morphology. Polymer 2010, 51, 2091− 2098. (51) Vinckier, I.; Laun, H. M. Manifestation of Phase Separation Processes in Oscillatory Shear: Droplet-Matrix Systems versus Cocontinuous Morphologies. Rheol. Acta 1999, 38, 274−286.

H

DOI: 10.1021/acs.macromol.6b00212 Macromolecules XXXX, XXX, XXX−XXX