pubs.acs.org/langmuir © 2009 American Chemical Society
Influence of Superheated Water on the Hydrogen Bonding and Crystallography of Piperazine-Based (Co)polyamides )
Esther Vinken,†,§ Ann E. Terry,†, Anne B. Spoelstra,† Cor E. Koning,‡,§,^ and Sanjay Rastogi,†,§,#
)
† Laboratory of Polymer Technology, Department of Chemical Engineering, ‡Laboratory of Polymer Chemistry, Department of Chemical Engineering, Eindhoven University of Technology, P.O. Box 513, 5600MB Eindhoven, The Netherlands, §Dutch Polymer Institute (DPI), P.O. Box 902, 5600AX Eindhoven, The Netherlands, ISIS Facility, Rutherford Appleton Laboratory, Chilton, Didcot, Oxfordshire, OX11 0QX, U.K., ^Physical and Colloid Chemistry, Free University of Brussels, Pleinlaan 2, 1050 Brussels, Belgium and # Department of Materials, Loughborough University, Loughborough LE11 3TU, United Kingdom
Received December 9, 2008. Revised Manuscript Received January 20, 2009 Here we demonstrate that superheated water is a solvent for polyamide 2,14 and piperazine-based copolyamides up to a piperazine content of 62 mol %. The incorporation of piperazine allows for a variation of the hydrogen bond density without altering the crystal structure (i.e., the piperazine units cocrystallize with the PA2,14 units (Hoffmann, S.; Vanhaecht, B.; Devroede, J.; Bras, W.; Koning, C. E.; Rastogi, S. Macromolecules 2005, 38, 1797-1803). It is shown that the crystallization of PA2,14 from superheated water greatly influences the crystal structure. Water molecules incorporated in the PA2,14 crystal lattice cause a slip on the hydrogen bonded planes, resulting in a coexistence of a triclinic and a monoclinic crystal structure. On heating above the Brill transition, the water molecules exit from the lattice, restoring the triclinic crystal structure. With increasing piperazine content, and hence decreasing hydrogen bond density, the dissolution temperature decreases. It is only possible to grow single crystals from superheated water up to a piperazine content of 62 mol %. For these single crystals, the incorporation of water molecules in the vicinity of the amide group is seen by the presence of COOstretch vibrations with FTIR spectroscopy. These vibrations disappear on heating above the Brill transition temperature, and the water molecules leave the amide groups. For copolyamides with more than 62 mol % piperazine, no Brill transition is observed, no single crystals can be grown from water, and no water molecules are observed in the vicinity of the amide groups (Vinken, E.; Terry, A. E.; Hoffmann, S.; Vanhaecht, B.; Koning, C. E.; Rastogi, S. Macromolecules 2006, 39, 2546-2552). The high piperazine content (co)polyamides have fewer hydrogen bond donors and are therefore less likely to have interactions with the water molecules. This work demonstrates the relation among the Brill transition, the dissolution of polyamide in superheated water, and its influence on the hydrogen bonds and the amide groups.
Introduction The thermal properties in polyamides are largely due to the directional hydrogen bonds present between recurring amide groups. By varying the density of these hydrogen bonds, it is possible to influence the polyamide’s physical properties;3 fewer hydrogen bonds generally imply a lower melting temperature. One route to changing the hydrogen bond density is to change the length of the aliphatic portions in the linear polyamide chains, which results in a change in the spatial separation between the amide groups and hence an overall change in the hydrogen bond density.4 However, this approach leads to different polyamides with different lengths of repeat units, such as polyamide 4,6 and polyamide 6,6, which inevitably leads to different crystal structures, lamellar thicknesses, and so forth. All of these variables inevitably complicate the study of the hydrogen bond density in polyamides. *Corresponding author. E-mail:
[email protected]. (1) Hoffmann, S.; Vanhaecht, B.; Devroede, J.; Bras, W.; Koning, C. E.; Rastogi, S. Macromolecules 2005, 38, 1797–1803. (2) Vinken, E.; Terry, A. E.; Hoffmann, S.; Vanhaecht, B.; Koning, C. E.; Rastogi, S. Macromolecules 2006, 39, 2546–2552. (3) Ehrenstein, M.; Dellsperger, S.; Kocher, C.; Stutzmann, N.; Weder, C.; Smith, P. Polymer 2000, 41, 3531–3539. (4) Brydson, J. Plastics Materials, 7th ed.; Butterworth-Heinemann: Oxford U.K., 1999.
5294
DOI: 10.1021/la804046r
Therefore, a second route to changing the hydrogen bond density becomes more attractive: replace the diamine residue with a different chemical unit that reduces the possibility of hydrogen bond formation but has similar structural features and size to the original diamine residue. These similarities in structure enable the chemically different repeat units to cocrystallize. A suitable comonomer is piperazine; when varying piperazine contents are incorporated into a polyamide 2,14 (PA2,14) backbone, a set of copolymers as shown schematically in Figure 1 can be obtained. Piperazine residues incorporated into the polyamide chain do not contain any amide hydrogens and are therefore only hydrogen bond acceptors and not hydrogen bond donors.1,2,5 The PA2,14 units can act as hydrogen bond donors and acceptors. By introducing piperazine into the chain, the overall hydrogen bond density decreases. Copolyamides of polyamide 2,14 (PA2,14) and polyamide piperazine,14 (PApip,14)1,5 were synthesized from 1,12-dodecanedicarboxylic acid and variable amounts of 1,2-ethylenediamine and piperazine. A range of copolyamides (PA2, 14-co-pip,14) with a piperazine content ranging from 30 to (5) Vanhaecht, B.; Devroede, J.; Willem, R.; Biesemans, M.; Goonewardena, W.; Rastogi, S.; Hoffmann, S.; Klein, P. G.; Koning, C. E. J. Polym. Sci., Part A: Polym. Chem. 2003, 41, 2082–2094.
Published on Web 3/16/2009
Langmuir 2009, 25(9), 5294–5303
Vinken et al.
Figure 1. Chemical structure of (a) 1,2-ethylenediamine-based and (b) piperazine-based repeat units. 90 mol % were prepared together with homopolymers PA2,14 and PApip,14. The copolymers exhibit a decrease in melting and crystallization temperatures with increasing piperazine content.5 Although the introduction of a rigid cyclic monomer usually leads to an increase in melting temperature with respect to the linear aliphatic homopolymer as a result of increased rigidity of the polymer chain and, consequently, a decrease in the gain in entropy upon melting,5 the reduced possibility of the piperazine units to form hydrogen bonds overrules this effect. Additionally, up to a piperazine content of 62 mol %, the PA2,14 and PApip,14 units cocrystallize into a common crystal lattice, which differs only slightly from the PA2,14 crystal lattice.1 For a piperazine content of 90 mol % and higher, the crystal structure is distorted from that of PApip,14. For intermediate piperazine contents of 70 and 82 mol %, a coexistence of the PA2,14 and PApip,14 crystal structures was observed.1 It was further concluded that the piperazine rings incorporated into the copolyamides are planar to the hydrogen-bonded sheets, the sheets being sheared with respect to one another. The effect of temperature on the (co)polyamides was investigated previously.2 Upon heating, many polyamides show a Brill transition,6,7 which entails a solid-state crystalline transition from the low-temperature triclinic phase to the high-temperature pseudohexagonal phase (i.e., in WAXD, the 100 reflection related to the interchain distance and the 010 reflection related to the intersheet distance merge into a single reflection). The series of (co)polyamides were investigated, and it was found that the Brill transition is independent of the piperazine content,2 despite changes in the melt and dissolution temperatures. It was shown that the Brill transition temperature is not related to the hydrogen bond density but is directly related to and primarily caused by conformational changes occurring in the polyamide main chain, in particular, in the alkane segments. In previous studies8,9 using WAXD, it was shown that polyamide 4,6 (PA4,6) can be dissolved in superheated water at ∼200 C. This is well below the melting point of PA4,6, being ∼295 C.10 Considering this information, it is interesting to learn how the dissolution of a hydrogen-bonded polymer in water is affected by the hydrogen bond density. Additionally, the influence of the hydrogen bond density on the (co)polyamides’ interaction with superheated water is significant. The series of piperazine-based (co)polyamides (6) Brill, R. J. Prakt. Chem. 1942, 161, 49–64. (7) Jones, N. A.; Atkins, E. D. T.; Hill, M. J. Macromolecules 2000, 33, 2642–2650. (8) Rastogi, S.; Terry, A. E.; Vinken, E. Macromolecules 2004, 37, 8825– 8828. (9) Vinken, E.; Terry, A. E.; van Asselen, O.; Spoelstra, A. B.; Graf, R.; Rastogi, S. Langmuir 2008, 24, 6313–6326. (10) Sweeny, W.; Zimmerman, J. In Encyclopedia of Polymer Science and Technology; Mark, H. F.; Gaylord, N. G.; Bikales, N., Eds.; Wiley: New York, 1969; Vol. 10.
Langmuir 2009, 25(9), 5294–5303
Article
available to us presents a unique opportunity to study the effect of reduced hydrogen bond formation on the dissolution process of polyamides in superheated water. The present article aims to give more insight into the effect of incorporating a secondary diamine residue that reduces hydrogen bond formation in the dissolution process in superheated water and the subsequent crystal structure. Initially, high-pressure differential scanning calorimetry (DSC) was performed for a range of (co)polyamide concentrations in water in order to investigate the dissolution process; this was also followed by wide-angle X-ray diffraction. Single crystals grown from an aqueous solution were investigated by transmission electron microscopy (TEM). To determine the influence of water molecules on hydrogen bonding and other conformational changes, Fourier transform infrared (FTIR) spectroscopy studies were performed.
Experimental Section Materials. Homopolymers PA2,14 and PApip,14 as well as copolymer PA2,14-co-pip,14 were synthesized via a polycondensation reaction of 1,12-dodecanedicarbonyl dichloride and varying amounts of 1,2-ethylenediamine and piperazine as described elsewhere.5 The piperazine-based copolyamides used in this study have piperazine molar fractions of 0.30, 0.46, 0.54, 0.62, 0.82, and 0.90. These copolymers are referred to as coPA 0.30 through coPA 0.90, respectively. Differential Scanning Calorimetry (DSC). As-synthesized (co)polyamide and varying amounts of water were placed in large-volume capsules (LVC) and cycled twice between 30 and 220 C at 10 C/min under nitrogen using a Perkin-Elmer Pyris 1 DSC. The amounts of polymer and water were carefully weighed and placed into the capsules, sealed, and reweighed after sealing so as to verify the amount of water in the capsule. After the two heating cycles, the capsules were weighed again to establish if any leakage occurred. In all events, the second heating cycle was used for data analysis. Wide-Angle X-ray Diffraction (WAXD). The as-synthesized (co)polyamide was melted into a glass capillary and cooled to room temperature at 10 C/min on a Linkam TMS94 hot stage. The (co)polyamide was heated in the presence of water at a concentration of 20 wt % polymer in water in an in-house-built pressure device described elsewhere.11 During the heating/cooling cycles, in situ WAXD data was collected on the materials science beamline ID11,12 situated at the European Synchrotron Radian Facility (ESRF), Grenoble, France. Two-dimensional wide-angle X-ray diffraction patterns were collected using a Bruker CCD detector and a 25 keV (λ = 0.4959 A˚) 300 μm X-ray beam with a sample-to-detector distance of ∼40 cm. A diffraction pattern was collected every 12 s with an exposure time of 6 s. The diffraction patterns were corrected for spacial distortion, and a silicon standard was used to calibrate the sample-to-detector distance. WAXD Background Correction. To separate the crystalline peaks from the amorphous component and the background scattering, the halo in the WAXD data originating from the amorphous component of the (co)polyamide, the water encapsulated in the capillary, and the glass capillary itself were subtracted, leaving only the crystalline part of the diffraction pattern that originates from the (co)polyamide. The amorphous halo originating from the water and glass were determined experimentally. A Gaussian distribution was used to model the (co)polyamide’s amorphous component. To account for thermal expansion, the position of the Gaussian distribution (11) Vinken, E. Polyamides: Hydrogen Bonding, the Brill Transition, and Superheated Water. Ph.D. Thesis, Eindhoven University of Technology, Eindhoven, The Netherlands, 2008. (12) Kvick, A. Nucl. Instrum. Methods Phys. Res. B. 2003, 199, 531–535.
DOI: 10.1021/la804046r
5295
Article
Vinken et al.
was allowed to vary. The final diffraction pattern is now no longer superimposed on an amorphous halo. Small-Angle X-ray Diffraction (SAXS). Simultaneous small and wide-angle X-ray diffraction experiments, similar to the WAXD experiments performed on ID11, were performed on the high brilliance beamline ID0213 at the ESRF. The experiments were performed using a 12 keV (λ = 0.9951 A˚) 300 μm X-ray beam. The X-ray patterns were collected every 30 s with an exposure time of 1 s. Lupolen was used to calibrate the SAXS sample-to-detector distance. Transmission Electron Microscopy (TEM). Approximately 1 wt % of (co)polyamide was heated in the presence of water to ∼200 C and cooled to room temperature at 10 C/min. A droplet of the thus-obtained (co)polyamide crystal suspension was placed on a carbon-coated copper TEM grid on which a small amount of gold was deposited as an internal reference for diffraction. The suspension was allowed to dry under ambient conditions. Low-dose diffraction images were collected on a Fei Technai 20 transmission electron microscope operating at 200 kV. Fourier Transform Infrared Spectroscopy (FTIR). Samples were prepared by heating 30 wt % as-synthesized (co)polymer in water to 200 C and cooling to room temperature at 10 C/min. The so obtained suspension was placed on a zinc selenium disk and allowed to dry under ambient conditions. After sample preparation, FTIR spectra, the average of 128 scans, were collected using a Bio-Rad FTS6000 spectrometer with a resolution of 2 cm-1. The samples were heated to the melt on a Linkam TMS94 hot stage at 10 C/min, and spectra were collected every 10 C. During data collection, the temperature was kept constant. The resulting spectra were scaled to the area under the methylene bands between 3000 and 2800 cm-1.
Results and Discussion Dissolution Behavior of Piperazine-Based (Co)polyamides. DSC is used to follow the phase behavior of PA2,14 and its piperazine-based copolymers in water, where the latter acts as a solvent when in the superheated state.8,9,14 Figure 2 shows the results obtained from DSC measurements on the (co) polyamides in the presence of superheated water. Here the end temperatures of the endotherms as a function of polymer content in water are plotted and represent the temperature at which the polymer is completely dissolved in superheated water. This has been further confirmed by WAXD. Figure 2a shows the phase behavior of several piperazinebased copolymers and homopolymers PA2,14 and PApip,14 in water as a function of polymer concentration in water. The lines drawn serve as a guide to the eye in observing the phase transitions. The general observation is that the dissolution temperature increases with increasing polymer content. In a heating run, with increasing temperature, conformational changes in the main chain occur. With the introduction of gauche conformers along the main chain, complemented by the rotation of the methylene units next to the hydrogen bonding motifs, increased strain on the hydrogen bonds is realized.2 Because of the reduced strength of the hydrogen bonding, the superheated water molecules, which are highly mobile in the superheated state,15,16 have the opportunity to (13) Urban, V.; Panine, P.; Ponchut, C.; Narayanan, T. J. Appl. Crystallogr. 2003, 36, 809–811. (14) Wevers, M. G. M.; Pijpers, T. F. J.; Mathot, V. B. F. Therm. Acta 2007, 453, 67–71. (15) Kilburn, D.; Townrow, S.; Meunier, V.; Richardson, R.; Alam, A.; Ubbink, J. Nat. Mater. 2006, 5, 632–635. (16) Dill, K. A.; Truskett, T. M.; Vlachy, V.; Hribar-Lee, B. Annu. Rev. Biophys. Biomol. Struct. 2005, 34, 173–199.
5296
DOI: 10.1021/la804046r
Figure 2. Influence of polymer concentration on the dissolution temperature of the piperazine-based (co)polyamides in water. (a) Phase diagram that evolved from the measured temperatures of the end of the DSC dissolution endotherms for various concentrations of (co)polyamide in water. Here b = PA2,14; O = coPA 0.30; 0 = coPA 0.54; ) = coPA 0.62; * = coPA 0.90; and Δ = PApip,14. (b) Melt and dissolution temperatures as a function of piperazine content. a - melt data taken from Vanhaecht et al.;5 b - dissolution temperature X-ray diffraction data for 30 wt % polyamide in water; and c - plateau dissolution temperature DSC data. All lines serve as a guide to the eye. The heating rate applied during the X-ray diffraction and DSC experiments is 10 C/min. perturb the crystal lattice by sharing the electrons forming the hydrogen bonds between the amide groups. In Figure 2a, the plateau region shows the concentration dependence on the dissolution temperature for the different (co)polyamides. For the (co)polymer concentration of 30 wt % that lies in the plateau region, three different dissolution temperatures can be distinguished; PA2,14 and coPA 0.30 have a dissolution temperature of ∼200 C, coPA 0.54 and coPA 0.62 dissolve at ∼165 C, and coPA 0.90 and PApip,14 have a dissolution temperature of ∼118 C. The three different dissolution temperatures are in accordance with the three distinct crystal structures reported for the copolymers.1 The influence of piperazine content on the dissolution process is illustrated in Figure 2b. The Figure shows the dissolution temperature for 30 wt % (co)polyamide in water, with increasing piperazine content. From a study2 performed on these piperazine copolyamides, it is known that the Brill transition is independent of the piperazine content and remains constant at ∼165 C on heating to the melt. As a reference, the Brill transition is also shown in Figure 2b. The melt temperatures5 of the (co)polymers are also plotted in the Figure. The data are obtained from DSC experiments and in situ WAXD and SAXS/WAXD experiments of the dissolution process performed on the ESRF (Grenoble) beamlines Langmuir 2009, 25(9), 5294–5303
Vinken et al.
ID11 and ID02, respectively. Because of the decrease in hydrogen bond density with the incorporation of the secondary piperazine diamine, the melting temperature decreases with an increase in piperazine content. Though an increase in chain rigidity with the incorporation of piperazine residues may favor a shift in the melting temperature to higher values, experimental observations suggest that the reduced possibility of hydrogen bond formation of the piperazine residues dominates the chain rigidity aspect, resulting in a shift of the melting temperature to lower values. Thus, the reduced melting temperature with increasing piperazine content is the cumulative effect of both opposing phenomena.1,2,5 The difference in melt temperatures is also mimicked by the dissolution temperature. Because of the weakening of the hydrogen bond strength in the (co)polyamides with increasing piperazine content, fewer gauche conformers are needed along the polymer main chain to allow the superheated water to disrupt the crystal lattice. The (co)polyamides rich in piperazine are therefore expected to exhibit a lower dissolution temperature than the (co)polyamides rich in 1,2-ethylenediamine. A closer look at the data in Figure 2b shows that the difference between the melt and dissolution temperature gradually decreases from ∼45 C for PA2,14 to ∼20 C for PApip,14 but does not reach zero (i.e., the melt temperature is not equal to the dissolution temperature for PApip,14). No hydrogen bonds are present in PApip,14; therefore, it is plausible that water molecules will not interact with PApip,14. However, the evidence from the DSC data presented in Figure 2 shows the contrary. Most likely, water molecules interact only with the carbonyl group that is present in the PApip,14 homopolymer as opposed to the complete amide group in the PA2,14 homopolymer. There will be fewer interactions with only the carbonyl group in comparison to the entire amide group; therefore, the dissolution temperature depression is less for PApip,14 in comparison to that for PA2,14. The data presented in Figure 2 conclusively demonstrates that water in the superheated state is a solvent for the piperazine-based (co)polyamides. The data presented here further strengthens the notion that the dissolution of aliphatic polyamides in superheated water is a phenomenon applicable to many polyamides, whether synthetic or biological in nature. Influence of Superheated Water on the Crystallography of PA2,14. Figure 3a,b show the WAXD pattern of PA2, 14 crystallized from the melt (MC) and water (WC). The melt-crystallized sample shows a single sharp reflection at 1.49 (for wavelength 0.05 nm), the 001 reflection, and two broader reflections at 6.83 and 7.17 corresponding to the 100 and 010 reflections, resepctively. The WAXD pattern for the melt-crystallized PA2,14 compares well with other results.1,2,17 However, the WAXD pattern obtained for PA2,14 crystallized from water is remarkably different from the melt-crystallized diffraction pattern. The water-crystallized sample shows several additional sharp and broad reflections (Figure 3a,b), perhaps indicating a different crystalline structure in comparison to that of the meltcrystallized sample. Thus, the unit cell of PA2,14 needs to be considered. (17) Li, Y.; Zhang, G.; Yan, D.; Zhou, E. J. Polym. Sci., B: Polym. Phys. 2002, 40, 1913–1918. (18) Jones, N. A.; Cooper, S. J.; Atkins, E. D. T.; Hill, M. J.; Franco, L. J. Polym. Sci., B: Polym. Phys. 1997, 35, 675–688.
Langmuir 2009, 25(9), 5294–5303
Article
Figure 3. WAXD data of PA2,14. The graphs show the diffraction patterns of the melt-crystallized sample (MC), water-crystallized sample (WC), and after heating the water-crystallized sample to 210 C (WE). All results shown are at 50 C. The calculated WAXD pattern is based on a theoretical calculation of the computed unit cell. The triclinic unit cell17,18 has parameters a = 0.49 ( 0.01 nm, b = 0.51 ( 0.01 nm, c = 2.23 ( 0.01 nm, R = 60 ( 2, β = 77 ( 1, and γ = 59 ( 1. The monoclinic unit cell has parameters a = 0.46 ( 0.01 nm, b = 0.41 ( 0.01 nm, c = 2.31 ( 0.01 nm, and β = 113 ( 1. (c) Triclinic and monoclinic peak indexing in more detail. Here the blue indices are triclinic, the red indices are monoclinic, and the black indices are both for the triclinic and monoclinic structures. The data is plotted for wavelength λ = 0.05 nm. The triclinic unit cell17,18 for PA2,14 has parameters a = 0.49 ( 0.01 nm, b = 0.51 ( 0.01 nm, c = 2.23 ( 0.01 nm, R = 60 ( 2, β = 77 ( 1, and γ = 59 ( 1. The predicted/ calculated WAXD pattern for the above-mentioned unit cell is determined using PowderCell 2.419 and is also shown in Figure 3b; for more information see the Supporting Information file. The positions of the calculated triclinic diffraction pattern compare well with the experimental diffraction pattern of the melt-crystallized sample. However, as mentioned above, the water-crystallized sample shows a different diffraction pattern from that of the melt-crystallized sample. Several reflections in the water-crystallized diffraction pattern are either absent or shifted in the calculated triclinic diffraction pattern. On studying the 001 reflection for the water-crystallized PA2,14 originating from the c axis of the unit cell (Figure 3b), it becomes apparent that next to the predicted/expected peak at 1.91 nm there is also an (19) Kraus, W.; Nolze, G. J. Appl. Crystallogr. 1996, 29, 301–303.
DOI: 10.1021/la804046r
5297
Article
Vinken et al.
additional peak at 2.12 nm. Both peaks must be correlated to the c axis, which is 2.23 nm. This indicates a different crystal structure for the water-crystallized PA2,14. A second, monoclinic unit cell with parameters a = 0.46 ( 0.01 nm, b = 0.41 ( 0.01 nm, c = 2.31 ( 0.01 nm, and β = 113 ( 1 is also shown in Figure 3a,b. The reflections as calculated show a mismatch with the triclinic unit cell but do show a good match with the monoclinic unit cell. It is apparent from the calculated diffraction patterns for both the triclinic and the monoclinic unit cells, in which the reflections arise from the (010) planes and from all planes whose reflections arise at smaller angles, that the calculated intensities are much larger than the observed intensities, whereas the opposite is true for planes perpendicular to the (010) plane (i.e., the (100) planes). We were unable to reduce these discrepancies in the atomic parameters. Two possible explanations exist for this discrepancy, and indeed both could play a role in generating the observed effect.20 One possibility is that the thermal vibrations of the molecules are restricted by the hydrogen bonds present in the sheets suppressing vibrations along the (010) plane. Therefore, the intensity of the reflections arising from the (010) planes will be reduced in comparison to those from other planes. A second explanation is a possible distortion of the crystals in the direction normal to the b axis due to the stacked hydrogen-bonded sheet, which make up the crystals with van der Waals forces between the sheets. The sheets are flexible in directions perpendicular to the sheets but less flexible (and therefore less distorted) within the sheets owing to the hydrogen bonding. The effect on the intensity would be similar to that of thermal vibrations. It is likely that both processes will play a role in influencing the relative intensities. Nonetheless, the WAXD pattern for the water-crystallized PA2,14 suggests the presence of both triclinic and monoclinic crystals. In Figure 4, the possible coexistence of two crystals is demonstrated using TEM and electron diffraction images obtained for PA2,14 single crystals grown from water. Figure 4a shows the single crystals obtained from PA2,14, where two distinct crystal morphologies are observed: lathlike and needlelike crystals. The needlelike crystals are large structures (∼1 mm) that are easily visible under the optical microscope, whereas the lathlike crystals are much smaller in dimension. Figure 4b,c shows the electron diffraction patterns obtained from the water-grown PA2,14 single crystals. Two distinctly different diffraction patterns are observed: a pattern consisting of two (large) arcs oriented approximately orthogonal to each other (Figure 4c) spaced at 0.44 and 0.37 nm and a pattern displaying six-point symmetry (Figure 4b) with spacings of 0.42, 0.44, and 0.42 nm. The difference between these two patterns is striking, again suggesting the presence of two crystal structures. Figure 3b also shows the WAXD pattern obtained after heating the water-crystallized PA2,14 to 210 C and subsequently cooling to 50 C (water-excluded, WE). The heating run of the WAXD patterns of the water-crystallized PA2, 14 is shown in Figure 5a. Here, the reflections arising from the monoclinic and triclinic crystal structures are again apparent. On heating, the reflections remain virtually unchanged until at the Brill transition, for PA2,14 at ∼170 C,2 all of the reflections disappear simultaneously to show only the pseudohexagonal phase. The polymer is heated below its (20) Bunn, C. W.; Garner, E. V. Proc. Roy. Soc. 1947, A189, 39–68.
5298
DOI: 10.1021/la804046r
Figure 4. TEM and electron diffraction images of PA2,14 crystallized from superheated water. (a) Lathlike and needlelike structures. (b) Electron diffraction image of a lathlike structure. (c) Electron diffraction image of a needlelike structure. The outer rings shown in b and c originate from the gold used as a calibration reference. melt temperature and subsequently cooled to 50 C (Figure 5b); the final result is shown in Figure 3 (trace WE). On cooling the water-crystallized PA2,14 from above the Brill transition temperature but below the melt temperature to 50 C, the polymer sample does not return to the crystal structure observed before heating. On cooling, once the Brill transition temperature is crossed, the sample returns to a triclinic structure similar to that of a melt-crystallized sample. The final diffraction pattern now resembles the diffraction pattern of the melt-crystallized sample. The additional reflections noted in the water-crystallized sample do not reappear. Conformational Changes in Water-Crystallized PA2,14. Figure 6 shows the FTIR spectra recorded at 30 C, obtained Langmuir 2009, 25(9), 5294–5303
Vinken et al.
Article
Figure 5. WAXD patterns of water-crystallized PA2,14 (a) heated from 50 to 210 C and (b) cooled from 50 to 210 C. Only every fifth data file is plotted for clarity. Note that the polymer is not heated to the melt during the heating run shown in part a. after crystallizing PA2,14 either from superheated water or from the melt. The data is normalized to the area under the methylene bands between 2800 and 3000 cm-1. Figure 6a shows the spectral range between 3500 and 2700 cm-1, and the band assignments are summarized in Table 1. Figure 6a shows several remarkable differences between the melt and water-crystallized samples. First, although the position of the NH stretch vibration at 3302 cm-1 is identical for the two samples, the shape and area of the two bands are different. The NH stretch vibration of the melt-crystallized sample is much broader than that of the water-crystallized sample, which shows a much sharper, narrower band. This shows that the hydrogen bonds between the two samples are significantly different. Additionally, the amide I and II overtone bands show a large shift in position: 3063 cm-1 in the water-crystallized sample and 3085 cm-1 in the meltcrystallized sample. The fingerprint region of the melt- and water-crystallized spectra between 1800 and 650 cm-1 is shown in Figure 6b,c together with their assignment in Table 1. The most significant differences between the two samples are the amide bands, certain methylene scissoring bands, and the so-called Brill bands. The Figure shows the amide I and II bands moving to lower wavenumbers on crystallization from superheated water. This is an indication of more hydrogen bond formation.21 When more hydrogen bonds are formed by the CO groups in the polyamide, the amide I vibration shifts to lower wavenumbers. This is consistent with the sharp NH stretch vibration at 3302 cm-1 observed for the water-crystallized sample. The (21) Rotter, G.; Ishida, H. J. Polym. Sci.: Part B Polym. Phys. 1992, 30, 489–495.
Langmuir 2009, 25(9), 5294–5303
Figure 6. FTIR spectra of water-crystallized (WC) and melt-crystallized (MC) PA2,14 at 30 C. (a) Spectra from 3450 to 2700 cm-1, (b) from 1800 to 1300 cm-1, and (c) from 1300 to 650 cm-1. presence of more hydrogen bonds in the water-crystallized sample suggests that superheated water is a good solvent for PA2,14 and facilitates crystallization into a perfect crystal packing. Next to the amide I and II vibrations, the amide III vibrations also show significant changes between the two samples. Amide III vibrations are sensitive to changes in the main chain methylene units. In Figure 6a, the symmetric and asymmetric methylene stretch vibrations show the most remarkable difference. The melt-crystallized sample shows the symmetric and asymmetric vibrations at 2850 and 2919 cm-1 respectively. The water-crystallized sample shows two additional vibrations at 2876 and 2945 cm-1. These two bands are also coupled to symmetric (2876 cm-1) and asymmetric (2945 cm-1) CH2 vibrations. Methylene stretch vibrations, as in small cyclic alkanes, are known to shift to higher wavenumbers when strained.22 Almost all of the Brill bands summarized in Table 1 also show a variation between the two samples. The most important differences are the appearance of two bands at 1607 and 1455 cm-1 in the water-crystallized sample. These are asymmetric and symmetric COOvibrations22 which may occur if the water molecules, (22) Lin-Vien, D., Colthup, N. B., Fateley, W. G., Grasselli, J. G. The Handbook of Infrared and Raman Characteristic Frequencies of Organic Molecules; Academic Press: San Diego, 1991.
DOI: 10.1021/la804046r
5299
Article
Vinken et al.
Table 1. Main Amide, Methylene Stretching and Scissoring, And “Brill” Bands Present in Melt Crystallized and Water Crystallized PA2,14a water crystallized
melt crystallized
band assignment
amide, methylene, and Brill bands
NH stretch22,25,26 NH stretch and amide (I+II) overtone22,25 NH stretch and amide II overtone22,25 CH2 asymmetric stretch27 2919 s CH2 asymmetric stretch27 CH2 symmetric stretch27 2850 s CH2 symmetric stretch27 1644 s amide I (CO stretch) 22 COO- asymmetric22 1556 s amide II (in-plane NH deformation, with CO and CN stretches)22,25,26 CH2 scissoring next to NH group, trans conformation27 1466 m CH2 scissoring for all methylenes not next to amide group29 COO- symmetric22 1419 m CH2 scissoring next to CO group, trans conformation27 1386 m CH2 wagging and twist, fold band27 1320 w CH2 wagging or twist27,30 1303 m 1305 m CH2 twist28 1283 m amide III26 1263 sh 1263 m amide III coupled with hydrocarbon skeleton25,26 1243 m skeletal C-C stretch28 1230 m skeletal C-C stretch28 1200 m 1188 w amide III coupled with hydrocarbon skeleton25,26 and crystalline bands28 1048 w 1054 w skeletal C-C stretch17,28,30 907 m 977 w CH2 rocking27-29 943 m 943 m amide IV (C-CO stretch)29,30 735 w 721 m CH2 rocking26,31 a Here, vs = very strong, s = strong, m = medium, w = weak, and sh = shoulder. 3302 vs 3200 w 3063 m 2945 s 2919 s 2876 s 2850 m 1638 vs 1607 sh 1542 s 1480 s 1466 s 1455 sh 1416 vs 1384 m
3302 s 3200 w 3085 m
amide amide amide methylene methylene methylene methylene amide Brill amide methylene methylene Brill methylene Brill Brill Brill amide amide Brill Brill amide Brill Brill amide Brill
incorporated into the crystal lattice, associate with the amide group to form a COO- group.9,23,24 On heating, all of the new or altered bands in the watercrystallized sample disappear or move to a different position at 190 C, (i.e., close to the Brill transition temperature for PA2,142). The asymmetric and symmetric CH2 stretch bands at 2945 and 2876 cm-1 shift and merge with the bands at 2919 and 2850 cm-1 (Figure 7a), and the COO- asymmetric and symmetric vibrations at 1607 and 1455 cm-1 also disappear (Figure 7b). These observations are in agreement with the structural changes in the water-crystallized samples close to the Brill transition temperature, as shown in Figure 5. Melting Behavior of Water-Crystallized PA2,14. Figure 8 shows the DSC traces obtained when heating a dried, watercrystallized PA2,14 sample to the melt together with a (straight) baseline (dashed gray line). In the first heating run, a broad exotherm is observed at 164 C with a heat of fusion of 22 J/g, which is not observed in the second heating run. The melt endotherm for both heating runs occurs at a similar temperature (∼240 C) and a heat of fusion of 82 J/g. The melt endotherm corresponds well with other reported values,1,2,5 whereas the exotherm seen below the baseline during the first heating run is unusual. A similar event was observed for water-crystallized PA4,6.9 (23) Harings, J. A. W.; van Asselen, O.; Graf, R.; Broos, R.; Rastogi, S. Cryst. Growth Des. 2008, 8, 2469–2477. (24) Harings, J. A. W.; van Asselen, O.; Graf, R.; Broos, R.; Rastogi, S. Cryst. Growth Des. 2008, 8, 3323–3334. (25) Cannon, C. G. Spectrochim. Acta 1960, 16, 302–319. (26) Cui, X.; Yan, D. J. Polym. Sci., Part B: Polym. Phys. 2004, 42, 4017– 4022. (27) Vasanthan, N.; Murthy, N. S.; Bray, B. G. Macromolecules 1998, 31, 8433–8435. (28) Cooper, S. J.; Coogan, M.; Everall, N.; Priestnall, I. Polymer 2001, 42, 10119–10132. (29) Jakes, J.; Krimm, S. Spectrochim. Acta 1971, 27A, 19–34. (30) Yoshioka, Y.; Tashiro, K. J. Phys. Chem. 2003, 107, 11835–11842. (31) Yoshioka, Y.; Tashiro, K.; Ramesh, C. Polymer 2003, 44, 6407–6417.
5300
DOI: 10.1021/la804046r
Figure 7. FTIR spectra obtained on heating PA2,14 crystallized from superheated water from 30 C to the melt at 10 C/min. We address this event in the following section. The second heating run of the now melt-crystallized sample does not show the exotherm observed during the first heating run. There is, however, a small shoulder present before the main endothermic melt peak. This shoulder is possibly related to Langmuir 2009, 25(9), 5294–5303
Vinken et al.
Figure 8. First (1H) and second (2H) heating runs of water-crystallized PA2,14. The table gives the peak maxima/minima and heats of fusion observed for the endotherms and exotherms, respectively. The dashed gray line indicates a baseline. a premelting and/or reorganization. This event was also seen previously for PA2,14.5 Influence of Superheated Water on PA2,14. From previous studies on PA4,6,8,9 it is known that when PA4,6 crystallizes from superheated water it is in a monoclinic structure with water molecules incorporated into the crystal lattice in close proximity to the amide groups. The presence of asymmetric and symmetric COO- vibrations confirm the presence of water molecules in the PA2,14 crystal lattice in proximity to the amide groups. It is to be noted that the two methylene groups in the diamine segment of PA2,14 are very rigid and unable to deform, whereas longer methylene sequences in other polyamide families can. The diamine alkane segment is also too short to allow chain folding to occur in this part of the chain.18 Therefore, when water molecules are incorporated in the PA2,14 crystal lattice in proximity to the amide groups, the methylene sequence in the diamine residue is too short and therefore too rigid to allow for deformation of the main chain. Hence, the water molecules force the hydrogenbonded planes to slip. The slip in the hydrogen bonded planes causes an alteration in how the planes align, giving rise to the monoclinic crystal structure. In addition, the scissoring band of the CH2 unit next to the NH group is present at 1480 cm-1 in the water-crystallized sample and absent in the melt-crystallized sample. The scissoring band of the CH2 next to the CO group is virtually unchanged between the two samples at ∼1417 cm-1. This confirms that conformational changes in the methylene chain segments between the melt- and water-crystallized samples indeed occur in the diamine units. The presence of features from the meltcrystallized sample in the spectra and diffraction patterns obtained from the water-crystallized sample confirms the existence of two crystal conformations in the water-crystallized PA2,14. The effect of strain and slip in the hydrogen-bonded planes also implies that (a portion of the) methylene segments in the water-crystallized PA2,14 are distorted. On heating the water-crystallized sample above the Brill transition, the Langmuir 2009, 25(9), 5294–5303
Article
“additionally” present methylene bands disappear. This indicates that the water molecules in the vicinity of the amide groups are released at the Brill transition. The removal of these water molecules from the crystal lattice at a temperature well above the bulk boiling point of water is also observed in DSC, where an exothermic event is observed at the expected Brill transition of PA2,14. This exotherm coincides with the transition of both the monoclinic and triclinic crystal structures seen in the WAXD patterns to the pseudohexagonal phase (i.e., at the Brill transition; see Figure 5). The Brill transition is caused by gauche conformers in the main chain,2 increasing the chain mobility and allowing enough freedom for the water molecules to escape. The incoming of gauche conformers is seen in the decrease of the band at 1480 cm-1 in the water-crystallized sample. In PA2,14 crystallization from superheated water causes the polyamide to crystallize in two crystal structures: a triclinic structure and a monoclinic structure where the triclinic structure is the regular melt-crystallized structure. This is contrary to what we observed for PA4,6,9 where crystallization from superheated water did not result in a different crystal structure from that of the melt-crystallized structure. The reason for this exists in the crystal structure of PA4,6. When PA4,6 is crystallized from acid or annealed at elevated temperatures whereby a perfect crystal structure is formed, the chains arrange into a monoclinic structure with alternating hydrogen bonds with an amide group in the fold.32 This structure is similar to the β bends in proteins. Because crystals of PA4,6 are already in the monoclinic phase, crystallization from superheated water does not cause the chains to slip as is observed in PA2,14 where the chains in the triclinic structure slip to form regions in the monoclinic phase. Considering the X-ray data and FTIR studies, it could be stated that the two crystal structures (monoclinic and triclinic) coexist in the water-crystallized PA2,14 where the monoclinic structure disappears with the restitution of hydrogen bonding on removal of the water molecules. Similar to PA2,14, PA6,6 also shows phase transformations in the presence of water. This arises because of the presence of the triclinic phase in PA6,6.11 Piperazine Copolyamide Water-Grown Single Crystals. The DSC data presented in Figure 2 suggests that superheated water is a solvent for all of the piperazine-based copolymers under investigation. However, it is possible to grow single crystals only up to a piperazine content of 62 mol %. Above 62 mol %, the copolyamide does not form a crystal suspension from the water solution. Instead, the copolyamide forms a crystalline, bulky solid. Therefore, TEM and electron diffraction images are available only for the coPA 0.30 and coPA 0.62 crystals. The TEM images of coPA 0.30 are similar to those of homopolymer PA2,14, showing both needlelike and lathlike crystals; see the Supporting Information file for the data. However, there appears to be only one type of electron diffraction pattern for both types of crystals: a six-point symmetrical pattern. The single crystals of coPA 0.62 show only lathlike morphology and also have a six-point electron diffraction pattern. The electron diffraction patterns conclusively show that single crystals are formed up to a piperazine content of (32) Atkins, E. D. T.; Hill, M.; Hong, S. K.; Keller, A.; Organ, S. Macromolecules 1992, 25, 917–924.
DOI: 10.1021/la804046r
5301
Article
Vinken et al.
62 mol %. For higher piperazine content, it is not possible to grow such single crystals, indicating that the (co)polyamides with high piperazine content do not truly dissolve in superheated water but merely melt in the presence of water. This notion is further strengthened by the small difference between the melt and dissolution temperature observed for PApip,14 (20 C) compared to that for PA2,14 (45 C) in Figure 2b. Influence of Superheated Water on the Crystallography of Piperazine Copolyamides. The WAXD data of coPA 0.30, coPA 0.62, coPA 0.82, and PApip,14 at 50 C crystallized from the melt (MC), from water (WC), and the watercrystallized sample that has been heated to a temperature below the melting point (WE) is shown in the Supporting Information file. The data shows similar observations for all of these (co)polyamides; the WAXD and SAXS data (SAXS data not shown) for the melt-crystallized, water-crystallized, and water-crystallized samples that have been heated are similar. This suggests that in the presence of water the incorporation of piperazine into PA2,14 inhibits the crystallization into two crystal structures (i.e., a triclinic and a monoclinic phase). For the copolyamides shown here, it is likely that once in the dissolved state the hydrogen bonds interact with the water molecules. Because piperazine is only a hydrogen bond acceptor and not a hydrogen bond donor, the interactions between the amide groups in the polymer chain and the water molecules in the superheated state greatly decrease with the introduction of piperazine. On crystallization, fewer water molecules can be incorporated into the crystal lattice, and no slip of the hydrogen-bonded planes occurs. Additionally, the increased rigidity of the main chain by the introduction of piperazine residues also prevents the slippage of the hydrogen bonded planes. It is likely that a much lower piperazine content of e.g. 5 mol % could show the monoclinic structure on crystallization from superheated water. Conformational Changes in Water-Crystallized Piperazine Copolyamides. In a similar procedure to that of watercrystallized PA2,14, the piperazine-based copolyamides were also studied using FTIR spectroscopy. CoPA 0.30 and coPA 0.62 form a crystal suspension on cooling from the dissolved state, which was placed on a ZnSe disk and allowed to dry under ambient conditions. CoPA 0.82 and PApip,14 do not form such a suspension; instead a solid residue is formed. For these two samples a piece of the residue was cut off and analyzed in a similar fashion to coPA 0.30 and coPA 0.62. The FTIR spectra obtained for the piperazine (co)polyamides crystallized from water are also shown in the Supporting Information file. The observed vibrations are in line with those seen for PA2,14, as assigned in Table 1. Additionally there are bands present at 3004, 1365, 1283, 1226, 1252, 1173, 1024, 1011, 986, 836, and 768 cm-1. These bands are all correlated to the piperazine unit incorporated into the polyamide chain2 and all originate from methylene bending, wagging, twisting, and stretching motions. With an increase in piperazine concentration, these peaks increase in intensity. Furthermore, the vibrational bands associated with hydrogen bonding (i.e., the NH stretch vibration at 3300 cm-1 and the amide I vibration at 1640 cm-1) show a strong decrease and broaden significantly with increasing piperazine content. An enlargement of the spectra in the region of 1650 to 1400 cm-1 is present in the Supporting Information file. Of 5302
DOI: 10.1021/la804046r
significance here is the presence of the shoulder at 1455 cm-1 in coPA 0.30 and coPA 0.62, which is absent in coPA 0.82 and PApip,14. This band is related to symmetric COOvibrations. The presence of this band together with the asymmetric COO- vibration at 1620 cm-1 shows the presence of water molecules in the vicinity of the amide groups, similarly to that seen in PA2,14 in the previous sections. These bands are not present in coPA 0.82 and PApip,14. This strongly suggests that the water molecules are present only in the lower piperazine-based copolymers where hydrogen bonding motifs are sufficiently present. On heating (Supporting Information), the COO- symmetric and asymmetric bands at 1620 and 1455 cm-1 disappear near the Brill transition temperature. PApip, 14 shows similar behavior to that of the melt-crystallized sample.2 This suggests that PApip,14 does not dissolve in superheated water but melts in the presence of water. coPA 0.62 shows similar behavior to that of coPA 0.30, and coPA 0.82 shows similar behavior to that of PApip,14. The amorphous bands that are normally identified in the region between 1300 and 1400 cm-1 are not included in the discussion because no considerable changes in the spectral region are observed in the water-crystallized (co)polyamides. Moreover, it is to be realized that water trapped in the amorphous region will be very mobile, thus hardly influencing the hydrogen bonding present in the amorphous region. Furthermore, water molecules present in the amorphous region are anticipated to leave on heating above 100 C.
Conclusions The work presented in this article combines the knowledge obtained from previous studies, where it is shown that the Brill transition is not dependent on the hydrogen bond density2 and that superheated water is a solvent for PA4,6.8,9 In this article, we show that superheated water is also a solvent for polyamide 2,14. The PA2,14 dissolves in superheated water at ∼200 C. On cooling, PA2,14 crystallizes from water in two crystal structures: a triclinic and a monoclinic structure. Water molecules incorporated into the crystal lattice cause slip between the hydrogen-bonded sheets, directly influencing the alignment of the sheets. FTIR spectra show upon crystallization from a water solution that water molecules are incorporated in a portion of the PA2,14 single crystals. The incorporated water molecules cause strain in the diamine part of the methylene chains, resulting in additional CH2 stretch vibrations at 2945 and 2876 cm-1. Also, the presence of symmetric and asymmetric COO- stretch vibrations strongly suggests that water molecules are indeed present in the vicinity of the amide groups. On heating above the Brill transition temperature, the COO- and CH2 stretch vibrations disappear at the Brill transition temperature, resulting in an FTIR spectrum similar to the spectrum of a melt-crystallized sample. On heating water-crystallized PA2,14, the incoming gauche conformers allow for enough translational motion along the main chain for water molecules to escape from the amide group at the Brill transition. Heating the water-crystallized sample above the Brill transition and not melting the sample results in a diffraction pattern similar to the triclinic diffraction pattern. With the introduction of piperazine into the main chain, the dissolution of the copolyamide is only possible up to a piperazine content of 62 mol %. From previous work2 on these piperazine copolyamides, it is known that the Brill Langmuir 2009, 25(9), 5294–5303
Vinken et al.
transition is observed only up to a piperazine content of 62 mol %. This suggests that the dissolution of a polyamide in superheated water is directly related to the Brill transition. For coPA 0.82 and PApip,14, which do not show a Brill transition, no single crystals could be grown. Instead the (co)polyamide shows a melting-point depression of 20 C in the presence of water, and on cooling, a crystalline solid residue is obtained. Similar to PA2,14, coPA 0.30 and coPA 0.62 show the presence of COO- stretch vibrations. These bands are absent in coPA 0.82 and PApip,14. However, these vibrations are considerably weaker in the copolyamides than in PA2,14. Also, in the copolyamides the COO- vibrations disappear at the Brill transition temperature of ∼190 C. Acknowledgment. We thank the ESRF for beam time on the materials science beamline ID11 and the high brilliance beamline ID02 as part of the long-term proposal
Langmuir 2009, 25(9), 5294–5303
Article
number SC1279. We thank Dr. Gavin Vaughn for his help and expertise during the WAXD measurements and Dr. Peter Boesecke for his help during the SAXS/WAXD measurements. This research has been carried out with the support of the Soft Matter Cryo-TEM Research Unit, Department of Chemical Engineering and Chemistry, Eindhoven University of Technology. We thank Jan Devroede for synthesizing the piperazine-based copolyamides used in this study and Professor Ted Atkins for fruitful discussions on the manuscript. Supporting Information Available: Although not presented in this paper, TEM, WAXD, and FTIR data on the water-crystallized piperazine-based copolyamides is also available. A description of the atomic positions of PA2,14 is also provided. This material is available free of charge via the Internet at http://pubs.acs.org.
DOI: 10.1021/la804046r
5303