Toughening Polylactide With a Catalyzed Epoxy-Acid Interfacial Reaction.
There are many approaches to toughening polylactide (PLA), but one of the most prevalent and economical methods is reactive blending with an immiscible minority rubbery component [1-8]. A similar approach has been used to toughen other brittle polymers, resulting in super tough nylon [9, 10], high impact polystyrene [11, 12], and other toughened polyesters [13, 14]. Deformation and failure mechanisms have been well-studied for rubber toughened polymer blends. Rubbery droplets can cause multiple crazing, spreading deformation energy to large parts of a glassy matrix, and eventual shear yielding . In addition, Kowalczyk et al. noted PLA blends with poly(l,4-m-isoprene) displayed an extra cavitation step between crazing and shear yielding, resulting in dramatically increased toughness compared to the neat PLA matrix . When interfacial adhesion is strong in rubber toughened blends, microcrack and void formation also serve to dissipate energy of deformation .
To obtain high adhesion, block or graft copolymers must be at the interface. For PLA and linear low-density polyethylene (LLDPE), this usually involves addition of polyethylene-co-glycidyl methacrylate (EGMA), which can react with PLA end groups at the interface to form a block copolymer (Fig. 1). Other reactive polyethylene copolymers have been used, such as ethylene-co-methacrylic ionomers , maleic anhydride grafted polyethylene , and ethylene-co-vinyl alcohol , but EGMA remains the primary method due to rapid kinetics and reasonable cost. PLA has been toughened using a reactive compatibilization system of LLDPE and EGMA, achieving the best toughness with around 20 wt% additives [6, 19-22], Evidence of this reaction was demonstrated by Fourier transform infrared (FTIR) spectroscopy, size exclusion chromatography (SEC), [sup.1]H NMR, droplet size, and rheology, leading to improved extension at break (up to 200%) and Charpy impact strength (up to 70 kJ/[m.sup.2]) [19, 20, 23].
The amount of copolymer formed is strongly related to adhesion and blend toughness, so rapid reactions are desired [9, 24-27], Several studies suggest that faster kinetics result in higher toughness of EGMA/PLA blends. Li et al. demonstrated the effect of reactive group concentration on dispersion and toughness , Oyama showed improvement in impact strength and elongation at break for samples prepared with higher extruder screw speeds, possibly from a higher extent of reaction , Reaction rates are most strongly influenced by the reactivity of reactive groups used, with relative rankings similar to those seen in small molecules , Epoxy-acid reactions are generally fast enough for commercial applications (e.g., Lotader, Joncryl, Biomax Strong), but are almost three orders of magnitude slower than the most commonly used coupling reaction, amine-cyclic anhydride .
Addition of catalyst is a promising method for increasing reaction rates, since catalysts for epoxy-acid reactions have been well-studied [30-34], Feng et al. recently published work on a PLA/PE-co-ethyl acrylate-co-glycidyl methacrylate (EMA-GMA) reactive blend, showing increased reaction rate with catalyst incorporation by mixer torque measurements and FTIR. This led to an impressive 50-fold increase in Charpy notched impact strength with the incorporation of 20 wt% EMA-GMA and 0.2 wt% dimethylstearylamine (DMSA) catalyst over neat PLA, and a ~threefold increase over the analogous uncatalyzed blend , Bai et al. also blended EMA-GMA with PLA in the presence of DMSA catalyst, successfully creating EMA-g-PLA with high conversion , Faster, catalyzed reactions may preclude the necessity of high functional polymer loadings or long processing times.
This work examines the effect of reaction kinetics and subsequent mechanical property changes of PLA/EGMA/LLDPE blends with the addition of N,N-dimethyldodecylamine (DDA, acts as a control similar to DMSA) and cobalt ethylhexanoate (Co[Oct.sub.2]) catalysts. Relative kinetics are quantified using a new method, mixer normal force. We use a smaller amount of functional polymer than Feng et al. or Bai et al., while still achieving a large increase in tensile toughness.
LLDPE ENGAGE[TM] 8200 (Trademark of the Dow Chemical Company) is an ethylene-octene random copolymer with 7.3 mol% octene, a melt flow index (MFI) of 5 g/10 min (190[degrees]C/ 2.16 kg) as measured by ASTM D1238 and density of 0.870 g/[cm.sup.3] as measured by ASTM D792. PLA Ingeo Bioworks 2003D was purchased from Natureworks LLC (MFI = 6 g/10 min (210[degrees]C/ 2.16 kg) and density = 1.24 g/[cm.sup.3]) and dried in a vacuum oven for at least 48 h at room temperature prior to use. EGMA (Lotader AX8840, Arkema, 8 wt% glycidyl methacrylate, MFI = 5 g/10 min (190[degrees]C/2.16 kg), density = 0.94 g/[cm.sup.3]), DDA, and Co[Oct.sub.2] (65 wt% in mineral oil) were purchased from Sigma Aldrich and used as received.
Blends were produced with 80 wt% PLA and 20 wt% combined LLDPE and EGMA. Catalysts were added to select blends at 0.01 M (mol catalyst/L melt blend volume). Three levels of EGMA content (0, 1, and 5 wt% EGMA) and three levels of catalyst (none, 0.01 M DDA, or 0.01 M Co[Oct.sub.2]) were examined. These blends are denoted as "catalyst, wt% EGMA" (e.g., "DDA, 5 EGMA" corresponds to a blend with 80 wt% PLA, 15 wt% LLDPE, 5 wt% EGMA, and 0.01 M DDA).
A 5 mL Xplore twin screw microcompounder produced small scale batch blends (total mass of 4.00 g per batch). 0.01 M catalyst was added by syringe (0.014 g Co[Oct.sub.2] solution in mineral oil or 0.009 g DDA) after being diluted in acetone to roughly 60 |iL solution per batch to reduce viscosity and ease loading. The compounding conditions were 180[degrees]C, 200 rpm, with nitrogen purge, mixing for 5 min before extruding into a liquid nitrogen bath. Normal force, which is proportional to pressure build-up in the twin screw mixer, and thus is related to blend viscosity, was recorded to examine relative interfacial reaction kinetics (in triplicate, to ensure reproducibility). Extrudates were quenched in liquid nitrogen for SEC and morphology measurements.
Large sample batches (300-500 g) of the same compositions were created for tensile tests using a 16 mm twin screw extruder (PRISM, L:D 24:1, Thermo, four heating zones at 180[degrees]C and a feed zone at 140[degrees]C, 40 rpm screw speed). Two mixing zones and three conveying zones were utilized as described in Marie et al. , with flow rates of 6-9 g/min and residence times of about 4 min (residence time distributions ranging from 2 to 7 min for all blends, measured by colored pellets) , Outlet pressures ranged from 0.34 to 2.41 MPa and torques from 5 to 10 N-m, depending on final blend viscosity. After steady state polymer flow was achieved, a syringe pump was used to dispense liquid catalyst (diluted in acetone to reduce viscosity, roughly 10 wt% catalyst solution) into the feed hopper at a controlled rate. Extrudate was chilled in a water bath, dried by air blower, pelletized, and stored in a vacuum oven for 48 h before further processing.
Molecular Weight, Rheological, Morphological, and Tensile Properties Characterization
Rheological properties were measured in shear on an ARES rheometer with 25 mm parallel plates at 180[degrees]C with gaps between 0.5 and 1.0 mm. Samples were compression molded into disks at 190[degrees]C and 2 MPa for 5 min and then loaded between preheated plates. Time sweeps were performed at 3% strain and 3 rad/s to determine melt stability. Strain sweeps (3 rad/s) were used to find the critical strain at the limit of the linear viscoelastic regime, defined as G' ([gamma]) > 0.9 X G' ([gamma] = 0), where y is the strain and G' is the storage modulus. The materials exhibited a critical strain of about 40% with stable readings over several hours. Zero shear rate viscosities were estimated from the Cox-Merz rule and Cross model fits.
All neat materials and blends were analyzed using high temperature SEC (PL-GPC 220, Agilent Systems, 135[degrees]C, 1,2,4-trichlorobenzene eluent, RI detector). This SEC was calibrated with polystyrene standards and reported molecular weights are for equivalent hydrodynamic radius polystyrene.
Blends were microtomed for scanning electron microscopy (SEM) using a Leica EM UC 6 and glass knife at -160[degrees]C. The resultant flat surface was coated with 50 [Angstrom] Pt and imaged using a JEOL 6500 SEM at 5 kV. Droplets were sized at 5000X magnification using ImageJ software (NIH) , The average of two diameters was recorded for each droplet. Aggregate statistics were compiled using Excel and JMP 11 software.
Blends were microtomed at -160[degrees]C using a diamond knife to create 50-70 nm sections for transmission electron microscopy (TEM). Sections were transferred to a copper-Formvar 300 mesh grid (Ted Pella, Inc.) using an eyelash tool. Images were obtained using an FEI Tecnai T12 instrument at 120 kV. Additional images and energy dispersive X-ray spectra (EDS) were collected using an FEI Tecnai G2 F30 instrument at 300 kV operated in scanning (STEM) mode, then analyzed using Oxford INCA and DTSA-II software for the presence of cobalt peaks (0.78 eV L[alpha] and 6.92 eV K[alpha]) , High-angle annular dark field (HAADF) STEM images were collected with an acceptance semi-angle of 59-200 mrad.
ASTM D638 type V tensile specimens (gage length = 25.4 mm, gage width =3.15 mm, thickness = 3.2 mm) were created using a Morgan Press injection molder (barrel temperature = 185[degrees]C, nozzle temperature = 220[degrees]C, mold temperature = 60[degrees]C, ram pressure = 7-34 MPa, pilot valve pressure = 0.7 MPa, clamp force = 13 tons, cycle time = 30 s). Samples were aged for at least 48 h in a room temperature vacuum oven. Compression molded samples did not yield consistent tensile results. An Instron 1101 tensile tester with 5 kN load cell pulled samples to failure at 5 mm/min. Modulus was calculated as the initial slope of the stress-strain curve, yield stress as the maximum stress achieved, and toughness as the integrated area under the stress-strain curve.
RESULTS AND DISCUSSION
Linear viscoelastic characterization of the pure component melts are shown in Fig. 2. The complex viscosities are within an order of magnitude, and G' values are similar. Therefore, droplet breakup and coalescence is expected to occur during melt mixing ,
SEC was used to detect copolymer formation as a function of EGMA content and catalyst. After mixing, blends were dissolved in 1,2,4-trichlorobenzene at elevated temperatures before injection. A linear baseline was created between retention times of 15 and 25 min, and data are normalized by integrated area. All traces showed one broad peak, as shown in Fig. 3. Pure EGMA, PLA, and LLDPE showed integrated changes in RI signal of -2,200, -1,900, and -2,200 mV-s at 1 mg/mL, respectively. Predicted traces are based on traces from the initial blend components, weighted by composition and pure component RI changes. The predicted traces show good agreement with measured molecular weight distributions for the uncatalyzed blends. Compared to predicted traces and uncatalyzed blend traces, Co[Oct.sub.2] blend results show a clear shoulder at high molecular weight. This suggests high molecular weight copolymer is being formed in significant quantities for Co[Oct.sub.2] catalyzed blends.
Polystyrene equivalent molecular weights were calculated and are displayed in Table 1. [M.sub.n] is similar for all blends except those containing Co[Oct.sub.2], which are lower. This is due to Co[Oct.sub.2] accelerating the degradation of PLA in melt mixing (discussed further in toughness results section). [M.sub.w] increases with EGMA content and catalyst, again indicating that reaction occurs more quickly with more functional polymer and in the presence of catalyst. Co[Oct.sub.2] is more effective than DDA for increasing [M.sub.w] and high molecular weight fractions in SEC.
SEM was employed to measure particle-size distributions for all blends, as smaller droplet size is often an indication of copolymer formation and compatibility improvements. Representative SEM images are shown in Fig. 4 and average droplet sizes in Table 2. It should be noted that SEM droplet sizes are accurate only above ~200 nm, so the smallest droplets are not observed. In general, the results show that higher loadings of EGMA decrease droplet size, and catalysts are effective for reducing droplet size with Co[Oct.sub.2] being more effective than DDA. DDA broadens the particle-size distribution, making it almost bimodal, which is why its [d.sub.vv] (volume weighted average diameter) is higher than the uncatalyzed blends (although [d.sub.n], the number weighted average diameter, generally decreases with DDA addition). The Co[Oct.sub.2], 5 EGMA sample seems to have strong adhesion, since very little droplet pullout is observed (Fig. 4f). Also, this sample has irregularly shaped drops, which suggests a lower interfacial tension in mixing.
Select blends were also examined by TEM to verify droplet size rankings, and to explore the possibility of micelle formation (Fig. 5). These low magnification images show a general decrease in average droplet size as EGMA content increases. Also, micelles are observed in the Co[Oct.sub.2], 1 EGMA blend, and especially in the Co[Oct.sub.2], 5 EGMA blend. These micelles are likely spherical with an EGMA core and PLA shell. They were not included in the SEM droplet sizes, as they are only visible via TEM. Micelles are another indication of high copolymer conversion [40-42].
Reaction Kinetics from Mixer Normal Force
Mixer normal force ([f.sub.N]) is used to compare relative reaction rates between blends. We expect viscosity to increase with reaction due to the copolymer generated. This results in an increase in pressure at the bottom of the mixing chamber (due to the drag flow of the screws), which is measured as a normal force between the screws and mixing chamber.
Representative plots of mixer force ([f.sub.N]) versus time (t) are shown in Fig. 6. For the first minute, loading causes a large initial increase in [f.sub.N] for all samples. After the initial force increase, polymer pellet loading is complete. Catalyst loading was performed after pellet loading, and was complete within 100 s of starting the run. Then, there is a steady state region of mixing, followed by extrusion. The small differences in initial forces are due to the higher viscosity of EGMA compared to LLDPE. It is clear that Co[Oct.sub.2] catalyzed blends display a dramatic increase in normal force, indicating Co[Oct.sub.2] is much more effective than DDA at promoting the interfacial acid/epoxy reaction.
There is no exact correlation between [f.sub.N] and viscosity, molecular weight, or size of the dispersed phase. However, relative reaction rates can be estimated between analogous catalyzed and uncatalyzed blends by assuming the reaction follows the same pathway (rate law) and the force is the same for a given conversion and composition. From batch reactor design equations and a generic rate law,
[N.sub.A0] dX/dt = -[r.sub.A]V = -k[C.sup.[alpha].sub.A] [C.sup.[beta].sub.B]V (1)
where [N.sub.A0] is the initial moles of reactant A, X is conversion, t is time, [r.sub.A] is reaction rate, k is reaction rate constant, [C.sub.A] is concentration of reactant A at time t, [aalpha] is the order of reactant A, [C.sub.B] is the concentration of reactant B at time t, [beta] is the order of reactant B, and V is the reaction volume. The reaction takes place almost exclusively at the interface in reactive compatibilization, so V is defined as the interfacial volume (Fig. 7).
Integrating and rearranging yields Eq. 2, where kt depends on V, X, [C.sub.A], [C.sub.B], [alpha], and [beta]. This assumes the same rate law (i.e., a and [beta]) for catalyzed and uncatalyzed blends.
kt = f(V,X,[C.sub.A],[C.sub.B], [alpha], [beta]) (2)
For analogous catalyzed and uncatalyzed blends (e.g., 5 EGMA and Co[Oct.sub.2], 5 EGMA), at a specific [f.sub.N], given the assumptions stated above, the generic function on the right-hand side of Eq. 2 is assumed to be constant. Therefore, the times that analogous blends reach a given force can be used to estimate reaction rate constant ratios. For normalized force ([f.sub.N][f.sub.N0]) versus time plots, slopes (m) can also be compared, as shown in Eq. 3. This method includes data from many time points, and is more robust than comparing individual values.
[k.sub.cat]/k = t ([f.sub.N]/[f.sub.N0])/[t.sub.cat]([f.sub.N]/[f.sub.N0]) = m/[m.sub.cat] (3)
Plots of [f.sub.N]/[f.sub.N0] versus time are shown in Fig. 8, time shifted from Fig. 6 to eliminate force during loading. Table 3 contains aggregate statistics for the slope of [f.sub.N]/[f.sub.N0] versus time. It can be inferred that blends with higher slopes have higher reaction rates, and therefore, higher copolymer contents. For blends that reached a plateau (Co[Oct.sub.2], EGMA 1 and Co[Oct.sub.2], EGMA 5), initial slopes were calculated, avoiding the interface-limited reaction kinetics regime at high conversions.
EGMA content and catalysts have clear effects on reaction rate. Increasing EGMA content from 0 to 20 wt% increases the slope observed from -1.6 X [10.sup.-4] [s.sup.-1] to 4.1 X [10.sup.-4] [s.sup.-1]. DDA further increases the slope, so that DDA, 1 EGMA has a similar slope to 5 EGMA. This efficacy of similar alkylamine catalysts for epoxy-acid interfacial reactions is in agreement with previous reports [19, 30], Co[Oct.sub.2] catalyzed blends show a much more dramatic increase in [f.sub.N], reaching a plateau within 100 s, which suggests complete conversion on this timescale.
Calculations of reaction rate constant ratios are shown in Table 4. DDA catalyst increases reaction rates by less than a factor of two, and Co[Oct.sub.2] increases reaction rate by roughly 90-fold. This dramatic effect agrees with changes seen in SEC and droplet size measurements. Similar values obtained for 1 EGMA and 5 EGMA blends suggest this is a reasonable method for measuring relative reaction rates. Co[Oct.sub.2] is clearly more effective at promoting interfacial reaction rate in this system.
Three weaknesses of this analysis are (1) viscosity cannot be directly related to chemical reaction, (2) interfacial volume is not constant between catalyzed and non-catalyzed blends, and (3) it takes finite time for the catalysts to be dispersed completely in the mixer. The second effect is due to copolymer formation. In blends with faster reactions, copolymer is formed more quickly, reducing interfacial tension and causing finer dispersion, which leads to increased interfacial area and thicker interfaces at any given time. This difference in interfacial reactor volume leads to an overestimation of reactivity ratio. The third effect, time from catalyst addition to having it fully mixed, leads to an underestimation of the reaction rate constant ratio (due to a decrease in effective catalyzed reaction volume at small times). Regardless, mixer normal force is a useful, in situ method for comparing reaction rates, relevant to processing. It clearly shows the effectiveness of Co[Oct.sub.2] over DDA and uncatalyzed blends.
TEM EDS was performed to determine Co[Oct.sub.2] catalyst localization. Past work has shown localization can play an important role in catalyst efficacy for reactive compatibilization [30, 34]. Co[Oct.sub.2] was chosen because Co has clear EDS peaks and should have a solubility parameter similar to stannous octoate, which has been shown to localize at the interface in PE/PLA blends ,
EDS spectra were obtained every 36 nm (100 points, ~20 nm spot diameter). Each spectrum was integrated for Co signal (6.8-7.1 keV) and reported as integrated Co counts as a function of distance along the line scan (Fig. 9). Co[Oct.sub.2] shows measurable concentration in the PLA matrix, and no noticeable increase in concentration at the interface. This result is consistent for multiple Co[Oct.sub.2] blends (0 EGMA, 1 EGMA, and 5 EGMA) and various EDS collection conditions. Normalizing by Cu or C counts does not change the general results. Therefore, another attribute such as activity is contributing to the superior reaction facilitation of Co[Oct.sub.2], rather than localization. Past work has shown that stannous octoate localizes at the interface of PE/PLA blends despite having a solubility parameter close to that of PLA (20.0 [MPa.sup.1/2] estimated by lyoparachor method for stannous octoate versus 20.5 MPa1/2 reported experimentally for PLA) . We hypothesize that changing the metal center from Sn to Co increases the catalyst solubility parameter enough for Co[Oct.sub.2] to be miscible in PLA. DDA is composed of light elements, so EDS is not feasible.
Since Co[Oct.sub.2] is not interfacially localized, it is expected that reaction rate will depend on catalyst concentration. Decreasing the Co[Oct.sub.2] concentration by a factor of 10 with 5 EGMA (0.001 M Co[Oct.sub.2], 5 EGMA) showed only double the reaction rate of uncatalyzed 5 EGMA. This is in contrast to the ~90-fold increase in reaction rate observed for 0.01 M Co[Oct.sub.2].
Copolymer can facilitate stress transfer across the interfaces of blends, which has potential applications in toughening PLA. Tensile tests are one of the most common methods for measuring toughness. Representative stress-strain tensile curves from tensile tests are shown in Fig. 10. The average yield stress, strain at break, tensile toughness, and modulus are displayed in Table 5 (n = 5). All blends showed stress whitening at low strain, in accordance with the cavitation stress dissipation mechanism described by Kowalczyk et al. . Necking was observed for all reactive blends, suggesting shear yielding is also a deformation mechanism.
For rubber toughened blends, improvements in tensile toughness generally come at the cost of decreased elastic modulus and yield strength compared to the neat matrix material. Feng et al. noted a decrease in elastic modulus from ~1800 MPa to ~1300 MPa and yield stress from ~73 MPa to ~40 MPa when adding 20 wt% EMA-GMA to PLA . Oyama found -30% reduction in modulus and ~40% reduction in yield stress when blending 20 wt% EGMA into PLA . Finally, Li et al. added 20 wt% LLDPE + EGMA to PLA, observing decreased modulus (from 3,000 to 1,800 MPa) and yield stress (from ~69 MPa to ~30 MPa) for the blends compared to neat PLA , Modulus and yield stress reduction on all our blends was similar ~25% and ~40%, comparable or slightly less than the blends in the reports previously mentioned.
DDA catalyst blends show slight improvements to tensile toughness over uncatalyzed blends, in accordance with their relative reactivities estimated from mixer normal force measurements. Co[Oct.sub.2], 1 EGMA shows low extension at break, possibly due to the cobalt catalyst accelerating PLA degradation in the extruder and injection molder. This is evidenced by qualitative observations of low melt strength compared to the batch mixed blends and non-Co[Oct.sub.2] catalyzed extruded blends, and by a shift in [M.sub.n] to lower molecular weight in SEC. Co[Oct.sub.2], 5 EGMA shows the best tensile toughness, over an order of magnitude higher than pure PLA and over five times higher than the 0 EGMA blend. The extension at break for this blend is about 200%, comparable to the best PLA blends (with 20% additives) in current literature [1, 2], while using less functional polymer.
Catalysts provide a general method for greatly increasing the reaction kinetics of interfacial reactive compatibilization. Co[Oct.sub.2] is better at promoting an interfacial acid/epoxy reaction compared to previously reported catalysts (e.g., aliphatic tertiary amines), as shown by droplet size, molecular weight increase, and increased blend viscosity as measured by mixer normal force. A new analysis of mixer normal force enables in situ estimation of relative reaction kinetics between blends. Relative interfacial reaction rates correlated well with tensile toughness, a key parameter for PLA blend applications. PLA blends with 15 wt% LLDPE, 5 wt% EGMA, and 0.01 M Co[Oct.sub.2] achieved an extension at break of 200% with only 20 wt% additives. This is comparable to the best-toughened PLA blends in current literature, while using less reactive compatibilizer.
The authors would like to thank the Center for Sustainable Polymers at the University of Minnesota, a National Science Foundation supported Center for Chemical Innovation (CHE1413862). We also acknowledge support from the Industrial Partnership for Research in Interfacial Materials and Engineering (IPRIME) at the University of Minnesota. Part of this work was carried out in the College of Science and Engineering Polymer Characterization Facility, University of Minnesota, which has received capital equipment funding from the NSF through the UMN MRSEC program under Award Number DMR1420013. Parts of this work were carried out in the Characterization Facility, University of Minnesota, a member of the NSF-funded Materials Research Facilities Network (www.mrfn.org) via the MRSEC program. [TM] Trademark of The Dow Chemical Company ("Dow") or an affiliated company of Dow.
[1.] K. Anderson, K. Schreck, and M. Hillmyer, Polym. Rev., 48, 85 (2008).
[2.] H. Liu and J. Zhang, J. Polym. Sci. Part B Polym. Phys., 49, 1051 (2011).
[3.] H. Jin, I. Chin, M. Kim, S. Kim, and J. Yoon, Eur. Polym. J., 36, 165 (2000).
[4.] K.S. Anderson, S.H. Lim, and M.A. Hillmyer, J. Appl. Polym. Sci., 89, 3757 (2003).
[5.] K.S. Anderson and M.A. Hillmyer, Polymer, 45, 8809 (2004).
[6.] H. Liu, F. Chen, B. Liu, G. Estep, and J. Zhang, Macromolecules, 43, 6058 (2010).
[7.] C.H. Ho, C.H. Wang, C.l. Lin, and Y.D. Lee, Polymer, 49, 3902 (2008).
[8.] Z. Su, Q. Li, Y. Liu, G.H. Hu, and C. Wu, Eur. Polym. J., 45, 2428 (2009).
[9.] S. Wu, Polymer, 26, 1855 (1985).
[10.] W.E. Baker, C. Scott, and G.H. Hu, Reactive Polymer Blending, Hanser, Cincinnati, OH (2001).
[11.] A. Ajji and L.A. Utracki, Polym. Eng. Sci., 3, 1574 (1996).
[12.] A.M.L. Magalhles, and R.J.M. Borggreve, Macromolecules, 28, 5841 (1995).
[13.] I. Lee, T.R. Panthani, and F.S. Bates, Macromolecules, 46, 7387 (2013).
[14.] S. Mbarek, M. Jaziri, Y. Chalamet, and C. Carrot, J. Appl. Polym. Sci., 117, 1683 (2010).
[15.] Z. Bartczak and A. Galeski, Polymer Blends Handbook, Vol. 2, Springer, Dordrecht, the Netherlands (2014).
[16.] M. Kowalczyk and E. Piorkowska, J. Appl. Polym. Sci., 124, 4579 (2012).
[17.] Z. Bartczak, A. Galeski, "Mechanical Properties of Polymer Blends," in Polymer Blends Handbook, 2nd ed., L.A. Utracki and C.A. Wilkie, Eds., Springer, Dordrecht, the Netherlands, 1203 (2014).
[18.] C.M. Lee, E.S. Kim, and J.S. Yoon, J. Appl. Polym. Sci., 98, 886 (2005).
[19.] Y. Feng, G. Zhao, J. Yin, and W. Jiang, Polym. Int., 63, 1263 (2014).
[20.] H.T. Oyama, Polymer, 50, 747 (2009).
[21.] D. Li, B. Shentu, and Z. Weng, J. Maeromol. Sci. Part B, 50, 2050 (2011).
[22.] H. Bai, D. Bai, H. Xiu, H. Liu, Q. Zhang, K. Wang, H. Deng, F. Chen, Q. Fu, and F.C. Chiu, RSC Adv., 4, 49374 (2014).
[23.] Y. Xu, J. Loi, P. Delgado, V. Topolkaraev, R.J. McEneany, C.W. Macosko, and M.A. Hillmyer, Ind. Eng. Chem. Res., 54, 6108 (2015).
[24.] E. Boucher, J.P. Folkers, and H. Hervet, L. Le, Macromoleeules, 29, IIA (1996).
[25.] C. Creton, E.J. Kramer, C.Y. Hui, and H.R. Brown, Macromolecules, 25, 3075 (1992).
[26.] Y. Seo and T.H. Ninh, Polymer, 45, 8573 (2004).
[27.] J.S. Lee, K.Y. Park, D.J. Yoo, and K.D. Suh, J. Polym. Sci. Part B Polym. Phys., 38, 1396 (2000).
[28.] C.W. Macosko, H.K Jeon, and T.R. Hoye, Prog. Polym. Sci., 30, 939 (2005).
[29.] C.A. Orr, J.J. Cernohous, P. Guegan, A. Hirao, H.K. Jeon, and C.W. Macosko, Polymer, 42, 8171 (2001).
[30.] P. Guegan, C.W. Macosko, T. Ishizone, A. Hirao, and S. Nakahaha, Macromoleeules, 27, 4993 (1994).
[31.] M.E. Stewart, S.E. George, and R.L. Miller, Polym. Eng. Sci., 33, 675 (1993).
[32.] G. Merfeld, C. Molaison. R. Koeniger, A.E. Acar, S. Mordhorst, J. Suriano, P. Irwin, R.S. Warner, K. Gray, M. Smith, K. Kovaleski, G. Garrett, S. Finley, D. Meredith, M. Spicer, and T. Naguy, Prog. Org. Coatings, 52, 98 (2005).
[33.] W.J. Blank, Z.A. He, and M. Picci, J. Coatings Technol, 74, 33 (2002).
[34.] C.M. Thurber, Y. Xu, J.C. Myers, T.P. Lodge, and C.W. Macosko, ACS Macro Lett., 4, 30 (2015).
[35.] M. Marie and C.W. Macosko, Polym. Eng. Sci., 41, 118 (2001).
[36.] Z. Tadmor and C.G. Gogos, Principles of Polymer Processing, 2nd ed., Wiley-Interscience, Hoboken, NJ (2006).
[37.] M.D. Abramoff, I. Hospitals, P.J. Magalhaes, and M. Abramoff, Biophoton. Int., 11, 36 (2004).
[38.] N.W. Ritchie, Microsc. Microanal., 15, 454 (2009).
[39.] U. Sundararaj and C.W. Macosko, Macromoleeules, 28, 2647 (1995).
[40.] H.Y. Kim, U. Jeong, and J.K. Kim, Macromolecules, 36, 1594 (2003).
[41.] H.K. Jeon and J.K. Kim, Macromolecules, 31, 9273 (1998).
[42.] P. Charoensirisomboon, T. Inoue, and M. Weber, Polymer, 41, 6907 (2000).
Christopher Thurber, (1) Liangliang Gu, (1) Jason C. Myers, (2) Timothy P. Lodge, (1,3) Christopher W. Macosko (1)
(1) Department of Chemical Engineering and Materials Science, University of Minnesota, Minneapolis, Minnesota 55455
(2) Characterization Facility, University of Minnesota, Minneapolis, Minnesota 55455
(3) Department of Chemistry, University of Minnesota, Minneapolis, Minnesota 55455
Correspondence to: C.W. Macosko; e-mail: email@example.com
Christopher Thurber is currently at The Dow Chemical Company, Midland, Michigan, 48667.
Published online in Wiley Online Library (wileyonlinelibrary.com).
Caption: FIG. 1. The dominant interfacial reaction occurring in EGMA/PLA blends. [Color figure can be viewed at wileyonlinelibrary.com]
Caption: FIG. 2. Frequency sweeps of pure blend components, on parallel plates at 180[degrees]C in the LVE regime. (Top) complex viscosity plotted versus frequency with Cross model fits (solid lines) and (bottom) storage modulus versus frequency. [Color figure can be viewed at wileyonlinelibrary.com]
Caption: FIG. 3. High temperature SEC traces for neat polymers and blends (135[degrees]C, 1,2,4-trichlorobenzene, RI detector, normalized by peak area). Predicted traces are calculated from the initial blend component traces, weighted by composition and component RI changes. Curves are shifted vertically for clarity. [Color figure can be viewed at wiIeyonlinelibrary.com]
Caption: FIG. 4. Representative SEM images of blends, with 1 um scale bars, (a) 1 EGMA, (b) 5 EGMA (c) DDA, 1 EGMA, (d) DDA, 5 EGMA, (e) Co[Oct.sub.2], 1 EGMA, (f) Co[Oct.sub.2], 5 EGMA. Droplet size generally decreases with increasing EGMA content and catalyst addition.
Caption: FIG. 5. TEM images of blends at 120 kV. 50-70 nm thick sections created by cryomicrotome then transferred to 300 mesh copper Formvar grids, (a) Co[Oct.sub.2], 0 EGMA, (b) Co[Oct.sub.2], I EGMA, and (c) Co[Oct.sub.2] 5 EGMA blends. Droplet size reduction and micelle concentration increase are observed with increasing EGMA content. Scale bars are 1 [micro]m.
Caption: FIG. 6. Mixer normal force over time for blends of LLDPE, EGMA, and PLA, with various EGMA contents and catalysts. Increasing/decreasing normal force is indicative of blend viscosity increase/decrease, attributed to interfacial reaction and PLA degradation. [Color figure can be viewed at wileyonlinelibrary.com]
Caption: FIG. 7. Schematic of interfacial reaction volume (white). Blue represents the PLA matrix and green represents the LLDPE/EGMA droplets. Interfacial thickness will increase with graft copolymer formation. [Color figure can be viewed at wileyonlinelibrary.com]
Caption: FIG. 8. Representative curves of normalized mixer normal force versus time. Catalyst and EGMA content changes the slope of these curves, indicative of interfacial reaction. [Color figure can be viewed at wileyonlinelibrary.com]
Caption: FIG. 9. TEM EDS line scan of Co[Oct.sub.2], 5 EGMA blend showing that Co[Oct.sub.2] resides in the PLA matrix (300 kV, F30, 70 nm section unstained). (Left) HAADF STEM image with red arrow indicating location and direction of line scan. (Right) Integrated Co counts as a function of distance along line scan. Light and dark background roughly specifies the domains probed in the image. [Color figure can be viewed at wileyonlinelibrary.com]
Caption: FIG. 10. Representative tensile stress-strain curves for PLA/LLDPE/ EGMA/catalyst blends, at 5 mm/min. [Color figure can be viewed at wileyonlinelibrary.com]
TABLE 1. Molecular weights for pure materials and blends. Material [M.sub.n] [M.sub.w] (kg/mol) (kg/mol) PLA 43 130 EGMA 32 190 LLDPE 69 180 0 EGMA 42 130 1 EGMA 42 130 5 EGMA 40 140 DDA, 1 EGMA 41 130 DDA, 5 EGMA 39 150 Co[Oct.sub.2], 1 EGMA 34 150 Co[Oct.sub.2], 5 EGMA 33 210 TABLE 2. Droplet size statistics, 5 min batch mixing. Droplet size ([micro]m, <[d.sub.vv]>/ Blend <[d.sub.vv]>) <[d.sub.n> 0 EGMA 2.7 5.6 1 EGMA 2.2 6.6 5 EGMA 1.8 4.7 DDA, 1 EGMA 2.4 6.5 DDA, 5 EGMA 3.0 10.2 Co[Oct.sub.2], 1 EGMA 1.1 3.5 Co[Oct.sub.2], 5 EGMA 1.0 3.2 Std. Dev. # of droplets Blend ([micro]m) considered 0 EGMA 0.53 687 1 EGMA 0.38 870 5 EGMA 0.33 840 DDA, 1 EGMA 0.36 826 DDA, 5 EGMA 0.54 626 Co[Oct.sub.2], 1 EGMA 0.21 1224 Co[Oct.sub.2], 5 EGMA 0.24 929 TABLE 3. Average starting force (<[f.sub.N0]>) and average slope (<m>) from mixer normal force plots. Blend (<[f.sub.N0]>) [+ or -] S.D. (N) 0 EGMA 1310 [+ or -] 20 1 EGMA 1420 [+ or -] 40 5 EGMA 1550 [+ or -] 30 DDA, 1 EGMA 1380 [+ or -] 10 DDA, 5 EGMA 1530 [+ or -] 20 Co[Oct.sub.2], 1 EGMA 1390 [+ or -] 40 Co[Oct.sub.2], 5 EGMA 1550 [+ or -] 30 Blend <m> [+ or -] S.D. ([10.sup.-4] [s.sup.-1]) 0 EGMA -1.6 [+ or -] 0.9 1 EGMA 0.8 [+ or -] 0.5 5 EGMA 1.9 [+ or -] 0.6 DDA, 1 EGMA 1.5 [+ or -] 0.5 DDA, 5 EGMA 3.2 [+ or -] 0.2 Co[Oct.sub.2], 1 EGMA 80 [+ or -] 11 (a) Co[Oct.sub.2], 5 EGMA 165 [+ or -] 27 (a) Standard deviation reported; n > 3. (a) Initial slope. TABLE 4. Estimated ratio of reaction rates between neat and catalyzed blends. Ratio 1 EGMA 5 EGMA [k.sub.DDA/]k 1.8 1.7 [k.sub.CoOct2/]k 96 88 TABLE 5. Tensile properties of PLA/LLDPE/EGMA/catalyst blends. [[sigma].sub.b] [[epsilon].sub.b] Material (MPa) (%) PLA 78 [+ or -] 1 7 [+ or -] 2 0 EGMA 36 [+ or -] 1 25 [+ or -] 3 1 EGMA 48 [+ or -] 0 47 [+ or -] 3 5 EGMA 47 [+ or -] 1 107 [+ or -] 6 DDA, 1 EGMA 45 [+ or -] 1 116 [+ or -] 9 DDA, 5 EGMA 48 [+ or -] 2 140 [+ or -] 9 Co[Oct.sub.2], 1 EGMA 44 [+ or -] 0 21 [+ or -] 7 Co[Oct.sub.2], 5 EGMA 45 [+ or -] 1 199 [+ or -] 6 Toughness Elastic Material (MJ/[m.sup.3]) modulus (MPa) PLA 3 [+ or -] 0.2 2700 [+ or -] 160 0 EGMA 7 [+ or -] 0.4 1800 [+ or -] 80 1 EGMA 15 [+ or -] 1.0 2100 [+ or -] 90 5 EGMA 16 [+ or -] 0.3 2100 [+ or -] 100 DDA, 1 EGMA 15 [+ or -] 1.0 1900 [+ or -] 70 DDA, 5 EGMA 30 [+ or -] 2.3 2000 [+ or -] 110 Co[Oct.sub.2], 1 EGMA 4 [+ or -] 2.0 1900 [+ or -] 100 Co[Oct.sub.2], 5 EGMA 38 [+ or -] 1.4 1900 [+ or -] 130
|Printer friendly Cite/link Email Feedback|
|Author:||Thurber, Christopher; Gu, Liangliang; Myers, Jason C.; Lodge, Timothy P.; Macosko, Christopher W.|
|Publication:||Polymer Engineering and Science|
|Date:||Jan 1, 2018|
|Previous Article:||Rheological Behavior of Ultrahigh Molecular Weight Polyethylene/Low-Density Polyethylene Blending Gels with High Solid Content.|
|Next Article:||Facile Fabrication of PVA Composite Fibers with High Fraction of Multiwalled Carbon Nanotubes by Gel Spinning.|