Flow-Induced Crystallization of PEEK - ACS Publications - American

Jun 30, 2016 - above the melting temperature of PEEK (Tm), a critical shear rate to ... with the shear rate at which the Cox−Merz rule abruptly begi...
17 downloads 0 Views 2MB Size
Letter pubs.acs.org/macroletters

Flow-Induced Crystallization of PEEK: Isothermal Crystallization Kinetics and Lifetime of Flow-Induced Precursors during Isothermal Annealing Behzad Nazari,† Alicyn M. Rhoades,‡ Richard P. Schaake,§ and Ralph H. Colby*,† †

Department of Materials Science and Engineering, Penn State University, University Park, Pennsylvania 16802, United States School of Engineering, Penn State Behrend, Erie, Pennsylvania 16563, United States § SKF Engineering and Research Centre, 3439 MT Nieuwegein, The Netherlands ‡

S Supporting Information *

ABSTRACT: The role of an interval of shear flow in promoting the flow-induced crystallization (FIC) for poly(ether ether ketone) PEEK was investigated by melt rheology and calorimetry. At 350 °C, just above the melting temperature of PEEK (Tm), a critical shear rate to initiate the formation of flow-induced precursors was found to coincide with the shear rate at which the Cox−Merz rule abruptly begins to fail. In cooling the sheared samples to 320 °C, FIC can be up to 25× faster than quiescent crystallization. Using rheology and differential scanning calorimetry, the stability of FIC-induced nuclei was investigated by annealing for various times at different temperatures above Tm. The persistence of shear-induced structures slightly above Tm, along with complete and rapid erasure of FIC-induced nuclei above the equilibrium melting temperature, suggests that FIC leads to thicker lamellae compared with the quiescently crystallized samples. semicrystalline polymers the structures produced by flow are not yet identified,13 so herein, we refer to these structures as flow-induced precursors. In this Letter, we demonstrate the use of rheology to investigate the role of flow-induced precursors in FIC of PEEK. With the aid of a cone-and-plate rheometer, short- and longterm shearing periods are applied to determine the effect of the applied shear rate and specific work on accelerating crystallization of PEEK. Janeschitz-Kriegl and co-workers have shown the usefulness of the “specific work” parameter in the study of FIC, noting that the external stress component reflects the degree of orientation in a rubber-like fluid and the rate of deformation component represents the measure of successful encounters.14 Therefore, specific work is chosen as a convenient variable for discussion, even while the decoupling of these relationships is beyond the scope of this work. Also, by monitoring the viscoelastic response of the samples undergoing a fast cooling, the thermal stability of structures formed under shear is studied by means of annealing steps at different temperatures above the melting temperature, Tm ≅ 340 °C for PEEK.15 These samples that contain FIC precursors were studied in the DSC to see how stable these nuclei are to

W

ith attributes of excellent chemical resistance, high thermo-oxidative stability, and superb mechanical properties that are retained at elevated temperature, poly(ether ether ketone) (PEEK) merits further investigation for many high-end applications where toughness at high temperature is required.1−4 The crystallization behavior of PEEK has been extensively studied and it has been suggested that the aromatic stiff chain of PEEK makes the dynamics of crystallization different than that of flexible chains such as PE and PP.5 Changes in crystallization conditions are known to result in different crystal morphologies, which influence final product properties.6,7 Flow-induced crystallization (FIC) is ubiquitous to semicrystalline polymers. Brief intervals of either shear or extensional flows can greatly accelerate isothermal crystallization kinetics8 and increase the temperature at which the sample crystallizes when cooled at a constant rate.9 Flow is thought to align and stretch the longest chains in the molecular weight distribution and this lowers the nucleation barrier, leading to faster (or higher temperature) crystallization. For a number of polymers, such as PEEK and poly(ethylene terephthalate) (PET), due to a larger Kuhn length (10.8 and 2.4 nm for PEEK10 and PET,11 respectively), the chain configuration between two entanglements (Me = 1490 and 1170 g/mol for PEEK and PET, respectively,12) is stiffer than for flexible chains such as PE and PP. Do stiff chains show similar FIC effects as the flexible chains that dominate the FIC literature? In all © XXXX American Chemical Society

Received: April 28, 2016 Accepted: June 27, 2016

849

DOI: 10.1021/acsmacrolett.6b00326 ACS Macro Lett. 2016, 5, 849−853

Letter

ACS Macro Letters annealing in the range of 350−400 °C, through the analysis of their recrystallization exotherm and melting endotherm. The PEEK sample used in this work was Victrex PEEK 650G. The rheological measurements were made using a Rheometrics ARES-LS rheometer equipped with a 25 mm cone and plate geometry (5.7° angle and 0.048 mm truncation gap) under a nitrogen purge to minimize degradation. A TA Instruments Q2000 DSC was used to conduct thermal studies on samples weighing 4−6 mg under a 50 mL/min nitrogen flow, for both quiescently crystallized samples and their counterparts that had previously experienced an interval of shear flow in the rheometer. Shear thinning is a common property of all entangled polymer melts and strongly correlated with flow-induced conformational changes.16 Shear flow will cause the longest chains in the molecular weight distribution to align with the flow if the shear rate is larger than the reciprocal of their reptation time τd, resulting in the onset of shear thinning for γ̇τd > 1. Then, at a shear rate that is 6M/Me larger, those longest chains get stretched by the shear flow, as the shear rate then exceeds the reciprocal of the Rouse time of those longest chains. In order to erase shear history of the material, the samples were first annealed at 400 °C for 5 min prior to both the steady and dynamic shear measurements at 350 °C. Steady-shear viscosity of the PEEK sample was measured as a function of shear rate for 0.01−100 s−1 at 350 °C, with no visual evidence of either slip or edge fracture. Linear viscoelastic response of the material in oscillatory shear at strain amplitude 0.005 was also studied at 350 °C. Figure 1 compares the results of these

structural change in the melt induced during the steady shear. One possible explanation would be that 6 s−1 corresponds to the reciprocal of the Rouse time for the longest chains (γ̇τR = 1), where they are expected to begin to be strongly stretched. That would mean the Rouse time of the longest chains is τR = 0.17 s and, combined with their reptation time of τd ≈ 10 s, suggests that those longest chains have M/Me = τd/6τR = 10 entanglements. Yuan et al. (2011) reported two distinct shear thinning regions in the flow curves of different grades of PEEK at 380 °C.18 Similar to our data in Figure 1, for their sample with a high molecular weight they observed a region of mild shear thinning occurring at shear rates ∼0.1−1.0 s−1 with stronger shear thinning at higher rates. They also report another stronger shear thinning region at very low rates (0.01−0.1 s−1) but consistent with White et al.,19 that may indicate mild crosslinking at 380 °C. Mimicking ref 8, we first identified a crystallization temperature of 320 °C, chosen to have an appropriately long crystallization time for a quiescent (not sheared) sample (∼25 min). We use a simple protocol for applying brief intervals of shear at 350 °C (slightly above the melting temperature of PEEK, Tm ≅ 340 °C) before cooling to 320 °C during a 1 rad/s time sweep, defining the time to reach tan δ = 1 at 320 °C as a convenient linear viscoelastic measure of crystallization time.8,20,21 After annealing at 400 °C for 5 min, the sample is equilibrated at 350 °C and a specific interval of shear can be applied for a specific amount of time. After the shear interval, a linear viscoelastic time sweep is started with frequency 1 rad/s and strain amplitude 0.005, and the sample is rapidly cooled to 320 °C, being careful to avoid undershooting the temperature, with temperature equilibrated at 320 °C in roughly 3 min after the start of the time sweep and cooling. Shear can shorten the crystallization time of the PEEK sample to approximately ∼1.0 min at 320 °C. To assess whether the shear interval is adequate in terms of inducing metastable sheared structures, the applied specific work8,9,14,22,23 (energy density in Pa = Jm−3) was calculated t using W = ∫ 0sηγ̇2s dt = tsηγ̇2s , where η is the steady shear viscosity at shear rate γ̇s and the integration is over the entire time of shearing ts. Figure 2 shows the dependence of crystallization time on specific work, at five shear rates of 6, 7, 10, 16, and 25

Figure 1. Comparison of steady-shear and complex viscosities of the PEEK sample at 350 °C.

two experiments, along with cartoons representing the entangled and oriented conformations of the longest polymer chains under low and high shear rates, respectively. The steady shear viscosity in Figure 1 reveals that mild shear thinning for this PEEK sample starts at ∼0.1 s−1, suggesting that the reptation time (τd) of the longest chains is approximately 10 s at 350 °C. The Cox−Merz empirical rule17 expects that the shear rate dependence of the steady shear viscosity and the frequency dependence of the complex viscosity should be identical. That rule is seen to work well in Figure 1 for shear rates up to 6 s−1, but abruptly fails as shear rate is further increased, with steady shear viscosity η ∼ γ̇−1, suggesting a constant shear stress of 20000 Pa for γ̇ > 6 s−1. Such an abrupt failure of the Cox−Merz rule is not usually observed in polymer melts and it suggests the onset of a strong

Figure 2. Effect of applied specific work14 W = ∫ t0sηγ̇2s dt = tsηγ̇2s (using five different shear rates at 350 °C) on crystallization time at 320 °C. As opposed to the quiescent crystallization time of ∼25 min, after applying 25 s−1 shear rate for 16 min, the crystallization time was as short as 51 s. 850

DOI: 10.1021/acsmacrolett.6b00326 ACS Macro Lett. 2016, 5, 849−853

Letter

ACS Macro Letters s−1. Many lower shear rates were also studied but for γ̇ ≤ 5 s−1 the crystallization time was identical to that of the quiescent samples (shown as the dashed line in Figure 2). It is interesting to note that the critical shear rate γ̇ ≥ 6 s−1 for flow-induced crystallization effects in shear coincides with the shear rate at which the Cox−Merz empirical rule abruptly fails (Figure 1). Figure 2 clearly shows that intervals of shear flow can profoundly accelerate crystallization kinetics as long as the shear rate is above the critical shear rate γ̇ ≥ 6 s−1. In previous FIC studies of polyethylenes and isotactic polypropylenes, this critical shear rate has been shown to be the reciprocal of the Rouse time of the longest chains.8,9,23 We observe a general trend that increased specific work does indeed play an important role in the development of FIC, as specific work can create many nuclei for crystallization,14,22 which dramatically alters the morphology of semicrystalline polymers. The spherulites that result from quiescent crystallization disappear completely and are replaced by a far finer crystallite morphology, described as shish-kebabs in polyolefins.23,24 Specific work provides a clear criterion for the formation of FIC precursors.14,22,23 However, shear rate directly plays a second vital role because as shear rate increases, more chains are able to participate in FIC, via the criterion γ̇τR > 1. Our finding of a minimum shear rate to induce FIC is consistent with past work.8,9,23 Based on Figure 2, among the cases with sufficient shear rates (γ̇ ≥ 6 s−1), as either the shear rate or the specific work increases, more nuclei form and the crystallization time lowers. Figure 2 also shows that after applying a certain amount of specific work (∼135 MPa), the crystallization rate saturates, becoming independent of applied specific work for large amounts of work (or long times) at each shear rate, as previously reported for iPP.8,25 However, far lower saturation values have been reported for flexible polymers such as PP (∼7−16 MPa),9 possibly related to the rigidity of PEEK chains in contrast with polyolefins. To investigate the stability of FIC precursors, an annealing step was added to the previous melt-shear-crystallization protocol after an interval of shear at 350 °C. The samples undergoing different shear rates and shearing times were annealed at temperatures in the 350−400 °C range, while varying the annealing time. After each anneal, the sample is briefly equilibrated at 350 °C where the oscillatory time sweep is started, followed immediately by cooling to 320 °C. The shear and thermal histories are summarized in Supporting Information. Figure 3 shows how the crystallization time changes with total annealing time at 350 °C. The sample was sheared at 25 s−1 for 8 min (∼200 MPa of applied specific work) at 350 °C prior to annealing. The annealing periods were applied on seven different samples by heating back to 350 °C and waiting for 2, 4, 8, 16, 32, 64, or 128 min. The data of Figure 3 suggest that the flow-induced precursors are quite stable when subject to annealing at 350 °C. Over 2 h of annealing at 350 °C slightly increased the 3 min crystallization time of the sheared sample to 5 min but that is still by far faster than the quiescent crystallization time (25 min). Annealing at different temperatures for 1 min and for 2 min were also performed and their effects on crystallization time at 320 °C are shown in Figure 4. The annealing periods (1 and 2 min) were done on two different samples after shearing. Figure 4 shows that the FIC precursors can persist up to 375 °C which is fairly close to the equilibrium melting temperature of PEEK4 (T∞ m = 389 ± 4 °C). Short periods of annealing at or above 380 °C between shearing and cooling can completely

Figure 3. Effect of the total annealing time at 350 °C with different durations, on crystallization time (at 320 °C) of a PEEK sample sheared at 25 s−1 for 8 min (W = 200 MPa) at 350 °C. Annealing at 350 °C increases the crystallization time from ∼170 to ∼290 s, but that is still very fast compared to the quiescent crystallization time of 1550 s.

Figure 4. Crystallization time at 320 °C for PEEK sheared at 25 s−1 for 8 min (W = 200 MPa) at 350 °C and then annealed for either 1 or 2 min at different temperatures (350−400 °C).

restore the crystallization time to the value obtained for quiescent PEEK. This result contrasts with the findings for iPP,9 which has very stable FIC-precursors above the equilibrium melting temperature. The difference can be hypothesized by iPP stretched chains adsorbing to the catalyst/support fragments, whereas PEEK stretched chains have no such crud to which they may adsorb. The stability of FIC-precursors to annealing was further studied by DSC on a sample that was sheared in the rheometer at 25 s−1 for 8 min (W = 200 MPa) at 350 °C. The sample was first melted in the DSC at 350 °C and cooled to 100 °C at 5 °C/min, resulting in the blue dashed exotherm in Figure 5, denoted “sheared”. Then the sample was heated at 15 °C/min to the annealing temperature, annealed for 2 min, and cooled to 100 °C at 5 °C/min. This heating, annealing, and cooling cycle was repeated on the same sample at progressively 5 °C higher annealing temperatures, from 350 to 380 °C, at which point the crystallization exotherm approached that of the quiescent sample that had never been sheared (solid blue curve in Figure 5). The crystallization exotherms on cooling are presented in Figure 5 and the melting endotherms are presented in Figure 6. Figure 5 shows that the sheared sample that was never heated above 350 °C after shearing, shows a peak in its crystallization exotherm that is ∼12 °C higher than that of the 851

DOI: 10.1021/acsmacrolett.6b00326 ACS Macro Lett. 2016, 5, 849−853

Letter

ACS Macro Letters

lamellar thickness, as this makes crystals stable to higher temperatures. Successive 2 min anneals at progressively higher temperatures gradually makes the melting endotherm approach that of the quiescently crystallized PEEK, that is, the primary melting peak shifts 2−3 °C lower, back to 340 °C, and the minor melting peak becomes visible again. The cause of the minor melting peak is discussed in the literature and not fully understood.28−35 In summary, brief intervals of shear flow above Tm were shown to greatly accelerate isothermal crystallization kinetics relative to quiescently crystallized samples that were not sheared. A minimum shear rate to initiate FIC was reported and attributed to conditions under which the longest chains in the molecular weight distribution of PEEK might be stretched, coinciding with the abrupt failure of the Cox−Merz rule. The persistence of shear-induced structures at different temperatures was investigated by annealing sheared samples in the rheometer and in the DSC. Annealing above 380 °C for a few minutes is sufficient to fully erase FIC effects, while the flowinduced precursors are far more stable at 350 °C. In the DSC, both significantly higher recrystallization temperature and slightly higher melting temperature suggest that FIC results in thicker lamellae compared to those crystallized without shear. The DSC data also showed that if a PEEK sample with strong FIC, that had been sheared at 25 s−1 for 8 min (W = 200 Pa), is annealed at 385 °C for 2 min, then the crystallization kinetics, recrystallization exotherm and melting endotherm revert to being identical to quiescent, non-sheared PEEK.

Figure 5. Crystallization exotherms of sheared and quiescent PEEK, cooling at 5 °C/min, and for the sheared PEEK sample after annealing for 2 min at six progressively higher temperatures.



ASSOCIATED CONTENT

* Supporting Information S

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acsmacrolett.6b00326. Additional supporting figures (PDF).

Figure 6. Melting endotherms for PEEK that was sheared at 350 °C in the rheometer, followed by 2 min annealing periods at six progressively higher temperatures, compared with that of quiescently crystallized PEEK, all with a heating rate of 15 °C/min.



quiescent sample, qualitatively similar to studies on iPP.9 This enhanced recrystallization temperature only disappears after annealing at temperatures approaching the equilibrium melting temperature of PEEK (>380 °C) where the nucleation precursors formed under shear rapidly disappear. Zhao et al. used wide-angle X-ray diffraction after applying shear rates up to 90 s−1 to PEEK samples at 350 °C that were subsequently crystallized at 330 °C to conclude that the average thickness of crystalline lamellae increases from 5.4 to 6.7 nm.26 While the numbers rely on the Scherrer equation, the increasing trend is robust and consistent with the higher recrystallization temperature reported in Figure 5 and WAXD data are presented in the Supporting Information. We hypothesize that shear flow can increase the lamellar thickness of the crystalline layers in PEEK compared to the typical lamellae formed under quiescent conditions. This hypothesis is consistent with the results reported by earlier works in the 1980s on the role of mechanical drawing of PEEK rods at 330 °C in increasing the melting temperature.27 It is evident in Figure 6 that the melting endotherm of quiescently crystallized PEEK shows a major peak at 340 °C as well as a minor melting peak at roughly 300 °C. For the sheared sample, FIC leads to the disappearance of the minor peak (circled in Figure 6) and to a small increase (2 °C) in the primary melting endotherm peak temperature, which is consistent with our assertion above that shear increases the

AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected]. Notes

The authors declare no competing financial interest.



REFERENCES

(1) Xing, P.; Robertson, G. P.; Guiver, M. D.; Mikhailenko, S. D.; Wang, K.; Kaliaguine, S. J. Membr. Sci. 2004, 229, 95−106. (2) Auimviriyavat, J.; Changkhamchom, S.; Sirivat, A. Ind. Eng. Chem. Res. 2011, 50, 12527−12533. (3) Wang, W.; Schultz, J. M. Macromolecules 1997, 30, 4544−4550. (4) Lee, Y.; Porter, R. S. Macromolecules 1988, 21, 2770−2776. (5) Fougnies, C.; Dosiere, M.; Koch, M. H. J.; Roovers, J. Macromolecules 1999, 32, 8133−8138. (6) van Meerveld, J.; Peters, G. W. M.; Hutter, M. Rheol. Acta 2004, 44, 119−134. (7) Hsiung, C. M.; Cakmak, M.; White, J. Polym. Eng. Sci. 1990, 30, 967−980. (8) Hamad, F. G.; Colby, R. H.; Milner, S. T. Macromolecules 2015, 48, 3725−3738. (9) Hamad, F. G.; Colby, R. H.; Milner, S. T. Macromolecules 2015, 48, 7286−7299. (10) Bishop, M. T.; Karasz, F. E.; Russo, P. S.; Langley, K. H. Macromolecules 1985, 18, 86−93. (11) Imai, M.; Kaji, K.; Kanaya, T.; Sakai, Y. Phys. Rev. B: Condens. Matter Mater. Phys. 1995, 52, 12696−12704. 852

DOI: 10.1021/acsmacrolett.6b00326 ACS Macro Lett. 2016, 5, 849−853

Letter

ACS Macro Letters (12) Fetters, L. J.; Lohse, D. J.; Richter, D.; Witten, T. A.; Zirkel, A. Macromolecules 1994, 27, 4639−4647. (13) Kumaraswamy, G. J. Macromol. Sci., Polym. Rev. 2005, 45, 375− 397. (14) Janeschitz-Kriegl, H Crystallization Modalities in Polymer Melt Processing; Fundamental Aspects of Structure Formation; Springer: Wien, 2010. (15) Blundell, D. J.; Crick, R. A.; Fife, B.; Peacock, J. J. Mater. Sci. 1989, 24, 2057−2064. (16) Xu, X.; Chen, J. L.; An, L. J. Chem. Phys. 2014, 140, 174902. (17) Cox, W. P.; Merz, E. H. J. Polym. Sci. 1958, 28, 619−622. (18) Yuan, M.; Galloway, J. A.; Hoffman, R. J.; Bhatt, S. Polym. Eng. Sci. 2011, 51, 94−102. (19) White, K. L.; Jin, L.; Ferrer, N.; Wong, M.; Bremner, T.; Sue, H. J. Polym. Eng. Sci. 2013, 53, 651−661. (20) Nazari, B.; Nazockdast, H.; Katbab, A. A. Polym. Eng. Sci. 2014, 54, 1839−1847. (21) Pogodina, N. V.; Winter, H. H.; Srinivas, S. J. Polym. Sci., Part B: Polym. Phys. 1999, 37, 3512−3519. (22) Ratajski, E.; Janeschitz-Kriegl, H. Polym. Bull. 2012, 68, 1723− 1730. (23) Mykhaylyk, O. O.; Chambon, P.; Graham, R. S.; Fairclough, J. P. A.; Olmsted, P. D.; Ryan, A. J. Macromolecules 2008, 41, 1901−1904. (24) Somani, R. H.; Yang, L.; Zhu, L.; Hsiao, B. S. Polymer 2005, 46, 8587−8623. (25) Housmans, J. W.; Steenbakkers, R. J. A.; Roozemond, P. C.; Peters, G. W. M.; Meijer, H. E. H. Macromolecules 2009, 42, 5728− 5740. (26) Zhao, G.; Men, Y.; Wu, Z.; Ji, X.; Jiang, W. J. Polym. Sci., Part B: Polym. Phys. 2010, 48, 220−225. (27) Lee, Y.; Lefebvre, J. M.; Porter, R. S. J. Polym. Sci., Part B: Polym. Phys. 1988, 26, 795−805. (28) Kong, Y.; Hay, J. N. Polymer 2003, 44, 623−633. (29) Righetti, M. C.; Laus, M.; DiLorenzo, M. L. Colloid Polym. Sci. 2014, 292, 1365−1374. (30) Verma, R. K.; Hsiao, B. S. Trends Polym. Sci. 1996, 4, 312−319. (31) Lee, Y.; Porter, R. S. Macromolecules 1987, 20, 1336−1341. (32) Jin, L.; Ball, J.; Bremner, T.; Sue, H. Polymer 2014, 55, 5255− 5265. (33) Tan, S.; Su, A.; Luo, J.; Zhou, E. Polymer 1999, 40, 1223−1231. (34) Marand, H.; Alizadeh, A.; Farmer, R.; Desai, R.; Velikov, V. Macromolecules 2000, 33, 3392−3403. (35) Tardif, X.; Pignon, B.; Boyard, N.; Schmelzer, J. W. P.; Sobotka, V.; Delaunay, D.; Schick, C. Polym. Test. 2014, 36, 10−19.

853

DOI: 10.1021/acsmacrolett.6b00326 ACS Macro Lett. 2016, 5, 849−853