Sage Journals: Discover world-class research

Abstract

The effects of varying the alkali metal cation in the high-temperature nucleophilic synthesis of a semi-crystalline, aromatic poly(ether ketone) have been systematically investigated, and striking variations in the sequence distributions and thermal characteristics of the resulting polymers were found. Polycondensation of 4,4′-dihydroxybenzophenone with 1,3-bis(4-fluorobenzoyl)benzene in diphenylsulphone as solvent, in the presence of an alkali metal carbonate M ₂CO₃ (M = Li, Na, K, or Rb) as base, affords a range of different polymers that vary in the distribution pattern of two-ring and three-ring monomer units along the chain. Lithium carbonate gives an essentially alternating and highly crystalline polymer, but the degree of sequence randomization increases progressively as the alkali metal series is descended, with rubidium carbonate giving a fully random and non-thermally crystallizable polymer. Randomization during polycondensation is shown to result from reversible cleavage of the ether linkages in the polymer by fluoride ions, and an isolated sample of alternating sequence polymer is thus converted to a fully randomized material on heating with rubidium fluoride.

Keywords

Polymer synthesis nucleophilic aromatic substitution sequence randomization poly(ether ketone)s crystallinity tailor-made polymers

Introduction

Composite materials for aerospace applications have traditionally been based on thermosetting matrix polymers such as the epoxies and bismaleimides,^1,2 but in more recent years the potential advantages of thermoplastic matrices (increased speed of fabrication and greater toughness) have begun to be realized,³ notably with the introduction of long-fibre composites based on semi-crystalline engineering polymers such as poly(1,4-phenylene sulphide)⁴ and the aromatic poly(ether ketone)s (PEKs), poly(ether ether ketone) (PEEK)⁵ and poly(ether ketone ketone) (PEKK; Figure 1).⁶

Figure 1.

Some thermoplastic aromatic matrix polymers used in carbon fibre composites.

The high crystalline melting points (T _m) of PEKs (typically 340–380°C) result in the retention of significant mechanical strength and stiffness even at temperatures well above their glass transition temperatures (T _gs).⁷ However, such T _ms also require correspondingly high composite fabrication temperatures – up to 420°C.^3,8 In the present article, we report a study of a lower melting but still crystallizable PEK matrix polymer (N1) derived from the nucleophilic polycondensation of 4,4′-dihydroxybenzophenone with 1,3-bis(4-fluorobenzoyl)benzene (Figure 2 ). The synthesis of this polymer (T _g = 152°C; T _m = 285°C) has been briefly noted in a conference paper,⁹ and its combination of a T _g somewhat higher than that of PEEK (T _g = 143°C; T _m = 343°C) and a very much lower T _m – potentially enabling more facile processing – suggested to us that it would be worth investigating further as a possible composite matrix.

Figure 2.

Synthesis of polymer N1, also symbolized as PEKEKmK. (M in M ₂CO₃ = alkali metal). Atom labels refer to ¹H NMR assignments; see Experimental section.¹H NMR: proton nuclear magnetic resonance; PEKEKmK: poly(ether ketone ether ketone ketone)s containing meta-phenyl links.

The polymer that might naively be expected from the above polycondensation would comprise an alternating sequence of two-ring and three-ring monomer residues. A rigorously alternating structure of this type has been obtained from the electrophilic polycondensation of 4,4′-diphenoxybenzophenone with isophthaloyl chloride (Figure 3).^10,11 The resulting semi-crystalline polymer (E1) shows thermal characteristics (T _g = 147°C; T _m = 310°C) similar to that of the ‘nucleophilic’ polymer N1 (T _g = 152°C; T _m = 285°C), though the T _m of E1 is noticeably higher. This difference in T _m may well be significant, as we now report that the nucleophilic synthesis, involving the use of an alkali metal carbonate as base, affords polymers with a range of different T _m and degrees of crystallinity depending on the nature of the alkali metal cation. This variability is shown to relate to the degree of sequence randomization during polycondensation, an effect resulting from reversible cleavage of the ether linkages during the growth of the polymer chain.^12,13

Figure 3.

Synthesis of polymer E1.¹⁰ Note that the structures shown for N1 (Figure 1) and E1 are merely different representations of the same polymer chain sequence.

Experimental

Materials, instrumentation and analysis

Monomers, solvents, alkali metal carbonates and other reagents were obtained from Sigma Aldrich (UK) and were used without further purification. Inherent viscosities (η _inh) were measured at 25°C with 0.1% w/v polymer solutions in 96% sulphuric acid (H₂SO₄) using a Schott Instruments CT 52 viscometer (Mainz, Germany). No insoluble gel fractions were present in any of the polymers described. Phase transitions (glass transitions, cold crystallizations and melting points) were identified from the second heating cycles of differential scanning calorimetry (DSC) traces using a TA DSC Q2000 instrument (New Castle, Delaware, USA; 4–12 mg samples, 10°C min⁻¹ under nitrogen atmosphere). A slight excess of alkali metal carbonate was used in each polycondensation to ensure quantitative conversion of the bisphenol to bisphenoxide. Yields of polymers were essentially quantitative and were diminished only by mechanical losses during the milling stage. Proton (¹H) and carbon (¹³C) nuclear magnetic resonance (NMR) spectra were obtained on Bruker Nanobay 400 MHz or 700 MHz NMR spectrometers (Billerica, Massachusetts, USA) using polymer solutions in (deuterated chloroform; CDCl₃/hexafluoro-2-propanol; (CF₃)₂CHOH (6:1 v/v) or CDCl₃/trifluoro acetic acid (CF₃COOH) (6:1 v/v). Mass spectra (electrospray ionization (ESI)) were obtained from 0.1% (w/v) sample solutions in methanol using a ThermoScientific LTQ OrbiTrap XL instrument (Waltham, Massachusetts, United States) equipped with an ACCELA LC autosampler.

Synthesis and characterization

Polymer N1a

A mixture of 1,3-bis(4-fluorobenzoyl)benzene (4.60 g, 14.28 mmol), 4,4′-dihydroxybenzophenone (3.00 g, 14.00 mmol), sodium carbonate (Na₂CO₃; 1.63 g, 15.4 mmol) and diphenylsulphone (35 g) was heated with stirring to 300°C under argon atmosphere. After 3 h, the polymer solution was poured onto a sheet of aluminium and allowed to cool. The resulting solid was ground to a powder in an ultracentrifugal mill and then stirred in acetone (200 mL) at room temperature for 30 min. The powder was filtered off, washed with acetone and dried. The powder was next extracted with 4× 200 mL of refluxing acetone, and then overnight in a Soxhlet extractor with refluxing acetone. The powder was extracted with 5× 200 mL of boiling water and then finally with 4× 200 mL of refluxing acetone. The resulting, purified material was dried at 110°C under vacuum overnight, affording polymer N1a (5.65 g, 81.3% yield). T _g = 149°C; T _m = 300°C; η _inh (H₂SO₄) = 0.62 dL g⁻¹; IR ν_max cm⁻¹: 2997 (C–H), 1655 (C=O), 1588 (C–C), 1238 (C–O–C), 1161 (C–O–C); ¹H NMR (400 MHz, CDCl₃/(CF₃)₂CHOH 6:1 v/v) δ 8.18 (s, 1H_c), 8.07 (d, J = 7.6 Hz, 2H_b), 7.91 (m, 8H_d,d′), 7.75 (t, J = 7.6 Hz, 1H_a), 7.26 (d, J = 8.0 Hz, 8H_e,e′) ppm; ¹³C NMR (100 MHz, CDCl₃/(CF₃)₂CHOH 6:1 v/v)) δ 199.08, 198.77, 161.35, 160.47, 137.45, 134.22, 133.08, 132.84, 132.60, 131.57, 130.80, 129.00, 119.10, 118.81 ppm.

Polymer N1b

This polymer was obtained using the procedure described for polymer N1a, but with potassium carbonate (2.13 g, 15.4 mmol) replacing Na₂CO₃ to give polymer N1b (6.00 g, 86.3%). T _g = 151°C; η _inh (H₂SO₄) = 0.78 dL g⁻¹; IR ν _max cm⁻¹ 2985 (C–H), 1656 (C=O), 1591 (C–C), 1240 (C–O–C), 1163 (C–O–C); ¹H NMR (400 MHz, CDCl₃/(CF₃)₂CHOH 6:1 v/v) δ 8.19 (s, 1H), 8.08 (d, J = 7.6 Hz, 2H), 7.92 (m, 8H), 7.76 (t, J = 7.6 Hz, 1H), 7.26 (d, J = 8.0 Hz, 8H) ppm; ¹³C NMR (100 MHz, CDCl₃/(CF₃)₂CHOH 6:1 v/v) δ 199.37, 199.00, 161.42, 161.13, 160.82, 160.54 134.25, 133.05, 132.84, 132.82, 132.60, 131.75, 131.57, 130.83, 129.00, 119.75, 118.79 ppm.

Polymer N1c

The same procedure as described for polymer N1a was used, but replacing Na₂CO₃ with rubidium carbonate (3.56 g, 15.4 mmol) and using a 5 mol% excess of 1,3-bis(4-fluorobenzoyl)benzene (4.74 g, 14.70 mmol) to control molecular weight (MW), afforded polymer N1c (5.35 g, 77.0%). T _g = 153°C; η _inh (H₂SO₄) = 0.73 dL g⁻¹; IR ν _max cm⁻¹ = 3000 (C–H), 1654 (C=O), 1588 (C–C), 1497 (C–C), 1236 (C–O–C), 1160 (C–O–C); ¹H NMR (400 MHz, CDCl₃/(CF₃)₂CHOH 6:1 v/v) δ 8.12 (s, 1H), 8.01 (d, J = 7.6 Hz, 2H), 7.86 (m, 8H), 7.70 (t, J = 7.6 Hz, 1H), 7.21 (d, J = 8 Hz, 8H) ppm; ¹³C NMR (100 MHz, CDCl₃/(CF₃)₂CHOH 6:1 v/v) δ 199.04, 198.72, 161.36, 161.07, 160.76, 160.47, 137.47, 134.22, 133.07, 132.84, 132.63, 132.46, 131.78, 131.59, 130.80, 129.00, 119.10, 118.89 ppm.

Polymer N1d

Polymer N1a (2.20 g), rubidium fluoride (1.49 g, 14.28 mmol) and diphenylsulphone (35 g) were heated with stirring at 300°C under argon for 2 h. Using the same workup procedure as described for polymer N1a afforded polymer N1d as a tan powder (1.05 g, 47.0%), T _g = 153°C; η _inh (H₂SO₄) = 0.44 dL g⁻¹; IR ν _max cm⁻¹ 2982 (C–H), 1655 (C=O), 1589 (C–C), 1497 (C–C), 1238 (C–O–C); ¹H NMR (100 MHz, CDCl₃/(CF₃)₂CHOH 6:1 v/v) δ 8.13 (s, 1H), 8.01 (d, J = 7.6 Hz, 2H), 7.87 (m, 8H), 7.70 (t, J = 7.6 Hz, 1H), 7.20 (d, J = 8.0 Hz, 8H) ppm; ¹³C NMR (100 MHz, CDCl₃/(CF₃)₂CHOH 6:1 v/v) δ 197.56, 197.39, 161.00, 160.73, 160.43, 160.14, 137.57, 134.00, 133.07, 132.81, 131.89, 131.71, 130.85, 129.78, 128.94, 119.10, 118.90 ppm.

Polymer N1e

The same procedure as described for polymer N1a was used, but replacing Na₂CO₃ with lithium carbonate (1.13 g, 15.4 mmol) and using an additional 2× 200 mL of boiling water at the extraction stage to ensure removal of lithium fluoride (LiF). This gave polymer N1e (5.53 g, 78.2%). T _g = 143°C; T _m = 304°C; η _inh (H₂SO₄) = 0.40 dL g⁻¹; IR ν _max cm⁻¹ = 1653 (C=O), 1588 (C–C), 1496 (C–C), 1235 (C–O–C), 1161 (C–O–C); ¹H NMR (400 MHz, CDCl₃/(CF₃)₂CHOH 6:1 v/v) δ 8.13 (s, 1H), 8.01 (d, J = 7.6 Hz, 2H), 7.87 (t, 8H), 7.70 (t, J = 7.6 Hz, 1H), 7.20 (dd, J = 8, 4 Hz, 8H) ppm; ¹³C NMR (100 MHz, CDCl3) δ 196.75, 196.63, 160.46, 159.63, 137.13, 133.45, 132.57, 132.31, 131.32, 130.40, 129.32, 128.45, 118.61, 118.36 ppm.

Bisphenol 2

A solution of 4-hydroxy-3-methylbenzoic acid (3.00 g, 19.7 mmol) and o-cresol (2.27 g, 21 mmol) in a mixture of trifluoromethanesulphonic anhydride (3.36 mL, 20 mmol) and trifluoromethanesulphonic acid (30 mL) was stirred under nitrogen for 18 h. The solution was then added dropwise with stirring into 800 mL of cold water. The precipitate was filtered off, washed repeatedly with boiling water, then with a 0.01M sodium hydroxide solution, and finally with cold water until a neutral filtrate was obtained. The solid was dried in a vacuum oven at 90°C for 5 h to afford bisphenol 2 as a colourless powder (3.20 g, 67%). Melting point (m.p.; DSC) 248°C (literature 247–248°C)¹⁴; FTIR ν _max cm⁻¹ 3424 (O–H), 3229 (O–H), 1639 (C=O), 1579 (C–C), 1373 (C–O); ¹H NMR (400 MHz, CDCl₃/CF₃COOH 6:1 v/v) δ 7.65 (d, J = 2.0 Hz, 2H), 7.58 (dd, J = 2.0, 8.4 Hz, 2H), 6.91 (d, J = 8.4 Hz, 2H), 2.31 (s, 6H); ¹³C NMR (100 MHz, CDCl₃/CF₃COOH): δ 200.36, 158.81, 134.68, 131.75, 129.26, 124.74, 115.06, 15.51. ESI-MS m/z = 243.1016 [M+H], 243.1021 calculated.

Polymer 3

A mixture of bisphenol 2 (2.42 g, 10 mmol), 1,3-bis(4-fluorobenzoyl)benzene (3.28 g, 10.2 mmol), Na₂CO₃ (1.16 g, 11.0 mmol) and diphenylsulphone (35 g) was heated with stirring to 300°C under argon atmosphere. The workup procedure for polymer N1a was followed to afford polymer 3 (3.40 g, 64%). T _g = 168°C, η _inh (H₂SO₄) = 0.44 dL g⁻¹; ν _max cm⁻¹ 1736 (C–H), 1657 (C=O), 1584 (C–C), 1496 (C–C), 1233 (C–O–C); ¹H NMR (400 MHz, CDCl₃/CF₃COOH 6:1 v/v) δ 8.18 (s, 1H), 8.06 (d, J = 7.6 Hz, 2H), 7.91 (d, J = 7.2 Hz, 4H), 7.82 (s, 2H), 7.71 (m, 3H), 7.24 (d, J = 5.6 Hz 1H), 7.11 (m, 5H) 2.34 (s, 6H) ppm; ¹³C NMR (100 MHz, CDCl₃/(CF₃)₂CHOH 6:1 v/v) δ 215.24, 198.85, 197.87, 161.56, 157.54, 137.26 133.66, 133.02, 132.88, 130.71, 130.54, 129.90, 128.58, 119.19, 115.80, 117.16, 29.83 ppm.

Results and discussion

Polycondensation of 4,4′-dihydroxybenzophenone with 1,3-bis(4-fluorobenzoyl)-benzene, in diphenylsulphone as solvent, at 300°C in the presence of an alkali metal carbonate M ₂CO₃ as base (Figure 2; M = Na, K, or Rb), afforded high-molecular-weight PEK N1a, N1b and N1c respectively, with inherent viscosities in the range 0.6–0.8 dL g⁻¹. A slight molar excess of the difluoroketone was used to control the final MW. Following exhaustive extraction of diphenylsulphone and inorganic salts, the polymers were dried and analysed by DSC. After heating to 350°C, the samples were cooled at 10°C min⁻¹, but none showed evidence of crystallization on cooling from the melt. However, on reheating at the same rate (Figure 4), polymer N1a underwent a glass transition (onset at 149°C), followed by a cold crystallization exotherm peaking at 245°C, and finally a crystal melting endotherm at 300°C. The other two polymers (N1b and N1c) showed only glass transitions at 151 and 153°C, respectively.

Figure 4.

DSC scans (2nd heating, 10°C min⁻¹) for polymers N1a, N1b and N1c. DSC: differential scanning calorimetry.

It seemed possible that the observed variation in crystallizability between the three polymers could result from differences in their sequence distributions since transetherification with sequence randomization is known to occur during the nucleophilic synthesis of aromatic polyethers in which both monomer residues in the chain are activated towards nucleophilic attack adjacent to the ether linkage.^12,13 This possibility was confirmed by ¹³C NMR analysis (Figure 5), which showed useful diagnostic resonances in the range δ = 160–162 ppm, corresponding to the aromatic carbons attached directly to ether oxygens. Polymer N1a shows only two peaks in this region, corresponding to the two different carbons of this type that would be expected in the simple alternating structure (EKEKmK) _n (c.f. polymer E1), whereas polymers N1b and N1c show two additional ‘inner’ peaks in the ¹³C–O–C region, with the relative intensity of these increasing substantially from N1b to N1c (Figure 5).

Figure 5.

¹³C NMR resonances in the C–O–C region for polymers N1a, N1b and N1c. ¹³C NMR: carbon nuclear magnetic resonance.

In order to aid the assignment of ¹³C–O–C peaks to the specific two-ring and three-ring residues, a dimethyl-substituted polymer (3), analogous to N1a, was synthesized. Direct condensation of o-cresol with 4-hydroxy-3-methylbenzoic acid in trifluoromethanesulphonic (triflic) acid gave very pure 3,3′-dimethyl-4,4′-dihydroxybenzophenone (2).¹⁴ This condensed with 1,3-bis(4-fluorobenzoyl)benzene in the presence of Na₂CO₃ (Figure 6) to give polymer 3.

Figure 6.

Synthesis of bisphenol 2 and its derived PEK 3. PEK: poly(ether ketone).

In the ¹³C NMR spectrum of polymer 3, two ¹³C–O–C resonances are still evident, but the lower field ¹³C–O–C peak is shifted only very slightly (ca. 0.25 ppm) relative to its position in the spectrum of 1a, whereas the other peak moves substantially upfield by some 2.9 ppm (Figure 7). This result strongly suggests that the lower field resonance represents the ¹³C–O–C carbon associated with the three-ring residue – which is chemically unchanged in the dimethyl polymer 3 – and that the strongly shifted, higher field resonance can be assigned to the two-ring residue, in which the methyl substituents are ortho to its ¹³C–O–C carbons.

Figure 7.

¹³C NMR resonances in the C–O–C region for polymers N1a and 3. ¹³C NMR: carbon nuclear magnetic resonance.

The ¹³C NMR spectra of polymers N1b and N1c however show two additional inner peaks in the ¹³C–O–C region (Figure 5). The increased multiplicity of peaks suggests sequence randomization could be occurring, and indeed the new peaks proved assignable to sequences comprising two adjacent two-ring residues and two adjacent three-ring residues. These assignments were achieved by doping samples of polymer N1b with the homopolymers PEK⁷ and PEKmK⁹, which resulted in the enhancement of the intensities of the higher field and lower field inner peaks, respectively (Figure 8).

Figure 8.

¹³C NMR resonances in the C–O–C region for (a) polymer N1b, and the same polymer doped with either (b) PEKmK or (c) PEK. ¹³C NMR: carbon nuclear magnetic resonance; PEKmK: poly(ether ketone) containing meta-phenylene links.

The ¹³C NMR spectrum of polymer N1c, produced using rubidium carbonate, showed even more extensive sequence randomization than in N1b, now with four peaks of equal intensity in the ¹³C–O region (Figure 5). Analysis of the probability distribution for the three possible dimer sequences around the ether linkages in polymer N1 shows that a completely random polymer would contain KEK, KEKmK and KmKEKmK sequences in the relative proportions 1:2:1. This distribution would indeed give rise to four ¹³C NMR resonances of equal intensity (1:1:1:1) in this region of the spectrum since the unsymmetrical KEKmK sequence contains two inequivalent C–O–C carbons. A similar calculation indicates that the intensity ratios observed in the spectrum of polymer N1b (ca. 3:1:1:3) correspond to approximately 50% randomization, relative to a fully alternating sequence. It is worth noting that the original report of polymer N1 indicated that it was synthesized using a mixture of sodium and potassium carbonates and that its melting point was 285°C,⁹ significantly lower than that that observed for polymer N1a. It thus seems very probable that a significant degree of sequence randomization of N1 had also occurred in that work. In fact, even N1a is unlikely to be 100% alternating as its melting point (300°C) is still slightly lower than that of the “electrophilic” polymer E1 (305°C). Moreover, the ¹³C NMR spectrum of N1a, shown in Figure 5, reveals a very weak but still detectable inner resonance corresponding to the symmetric, non-alternating sequence KmKEKmK. The relationships between polymers N1a–N1d, identified in the present work, are summarized in Figure 9.

Figure 9.

Representations of the chain sequences in polymers N1a (almost entirely alternating), N1b (semi-randomized) and N1c (fully randomized), arising from the use of different alkali metal carbonates M ₂CO₃ in the nucleophilic polycondensation shown in Figure 1. Treatment of N1a with RbF results in its conversion to polymer M1d, having the same, fully random, sequence as N1c. RbF: rubidium fluoride.

A number of possible mechanisms have been proposed for transetherification during the synthesis of aromatic PEKs, but all depend on reversible, nucleophilic cleavage of the ether linkages (Figure 10). Candidate nucleophiles in the system include the carbonate and fluoride anions, and indeed potassium carbonate has previously been shown to induce a small degree of sequence randomization in an aromatic PEK, albeit requiring very high reaction temperatures (340°C) and long reaction times (6 h).¹⁵ The fluoride ion can be a very strong nucleophile in dipolar aprotic solvents,^16,17 but its effectiveness in the present context would depend both on the solubility of the fluoride salt involved and on the extent of pairing with its counterion in solution. The larger the counterion, the weaker the ion pairing and the more soluble the salt, so RbF should be much more effective than sodium fluoride, with potassium fluoride somewhere in between (r _ionic = 1.16, 1.52, 1.66 Å for 6-coordinate Na⁺, K⁺, and Rb⁺, respectively).¹⁸ This is fully consistent with our experimental results for sequence randomization in the synthesis of N1.

Figure 10.

Partial mechanism for fluoride-catalysed sequence randomization in the synthesis of polymer N1. The initial chain cleavage by fluoride can occur at either the two-ring or three-ring monomer residue, and the fluoro end-group resulting from this reaction can subsequently regenerate a fluoride anion by reaction with a phenoxide monomer or end-group.

In the present work, sequence randomization catalysed by fluoride ion was demonstrated conclusively by treatment of the alternating polymer N1a with RbF in diphenylsulphone, at the same concentrations, temperature and time as in polymer synthesis. The result was completely clear-cut, with the diagnostic ¹³C–O–C resonances for the product N1d changing from two equal intensity resonances in N1a (alternating structure) to four equal intensity resonances (random sequence structure), exactly as found in polymer N1c.

As a final test of the proposed mechanism for sequence randomization, the polycondensation shown in Figure 2 was carried out using lithium carbonate as base. The extremely low solubility of lithium fluoride in organic solvents¹⁹ should strongly inhibit fluoride-catalysed transetherification and indeed, as shown in Figure 11, resonances arising from sequence randomization were scarcely discernable in the ¹³C NMR spectrum of the resulting polymer (N1e).

Figure 11.

¹³C NMR resonances in the C–O–C region for polymers N1c (fully randomized) and N1e, illustrating the virtual absence of sequence-randomization in the latter polymer. ¹³C NMR: carbon nuclear magnetic resonance.

As shown in Figure 12, polymer N1e also crystallized from the melt (T _c = 215°C) – unlike the other polymers described in this work – and showed a slightly higher T _m value than N1b (304 vs. 300°C) and a much higher degree of crystallinity (ΔH _m = 49 vs. 16 J g⁻¹), presumably the consequences of a more perfectly alternating chain sequence.

Figure 12.

DSC scans (2nd heating and subsequent cooling scan, 10°C min⁻¹) for polymer N1e. DSC: differential scanning calorimetry.

An intriguing observation is that, although DSC analysis customarily discounts the first heating scan, a consistent feature of the first (but not subsequent) DSC heating scans for polymers N1a, N1b, and N1c is the presence of a melting endotherm at ca. 174°C – in addition to a conventional polymer melting peak in the range 230–300°C – that increases in intensity with the degree of sequence randomization. The lower melting peak is however essentially absent from the DSC trace of the fully alternating polymer N1e. This correlation seems to imply the existence of a low-melting crystalline phase in the ‘as-isolated’ polymers that is associated specifically with the packing of random sequences. This could be possible, in principle, because the three ‘parent’ polymers (KEKmK, KEK and KmKEKmK) have almost identical unit cells in cross-section perpendicular to the chain direction (orthorhombic, a = 7.67 ± 0.05, b = 6.04 ± 0.07 Å, two chains per cell).^10,20,21 Moreover, the X-ray powder patterns from ‘as-isolated’ samples of polymers N1c and N1d indicate substantial degrees of crystallinity, despite the high levels of sequence randomization. Crystallization of random sequence copolymers is not of course unknown when the comonomer residues are isomorphic, but it is not yet clear how isomorphism arises in the present system. Computational modelling studies are under way in our laboratory to investigate this problem further.

Conclusions

Sequence randomization, via transetherification, during the nucleophilic synthesis of an aromatic PEK involving fluoride displacement from a bis(4-fluoroaryl)ketone can be controlled by varying the alkali metal cation (Li⁺, Na⁺, K⁺ or Rb⁺) present during polycondensation. The degree of transetherification increases with the ionic radius of the alkali metal involved, and a proposed mechanism in which fluoride ions reversibly cleave the growing polymer chain is substantiated by a direct demonstration of sequence randomization in the presence of RbF. The crystallizability of the polymer from the melt declines markedly as the degree of sequence randomization increases, although crystallization of the more highly randomized polymers from solution in diphenylsulphone affords an unusually low-melting crystalline phase whose nature remains to be established.

Footnotes

Authors’ Note

Underlying data for this article may be requested from the corresponding author.

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This work was supported by Cytec Aerospace Materials and the UK Engineering and Physical Sciences Research Council, who jointly funded a PhD studentship for K.J.C.L. through the University of Reading.

References

Morgan

. Structure-property relations of epoxies used as composite matrices. Adv Polym Sci 1985; 72: 1–43.

Landman

. Advances in the chemistry and applications of bismaleimides. In: Pritchard

(ed.) Developments in Reinforced Plastics. Vol. 5. London: Elsevier, 1986, pp. 39–81.

Cogswell

. Thermoplastic Aromatic polymer composites. Oxford: Butterworth-Heinemann, 1992.

Diez-Pascual

Naffakh

. Inorganic nanoparticle-modified poly(phenylene sulphide)/carbon fiber laminates: thermomechanical behaviour. Materials 2013; 6: 3171–3193.

Wood

Pandey

Walling

. Manufacture of high performance thermoplastic composite structures. In: SAMPE Conf. Proceedings, Long Beach, USA, Vol 53, 2008, 325/1-325/11.

Mazur

Oliveira

Rezende

. Environmental effects on viscoelastic behavior of carbon fiber/PEKK thermoplastic composites. J Reinf Plast Compos 2014; 33: 749–757.

Staniland

. Poly(ether ketones). In: Allen

Bevington

(eds) Comprehensive Polymer Science. Vol. 5. Oxford: Pergamon, 1989, pp. 483–497.

Schmitt-Thomas

Yang

Z-G

Malke

. Failure behavior and performance analysis of hybrid-fiber reinforced PAEK composites at high temperature. Compos Sci Technol 1998; 58: 1509–1518.

Staniland

. Synthesis and properties of novel polyetherketones and polyethersulfones. Bull Soc Chim Belges 1989; 98: 667–676.

10.

Qiu

Sheng

. Crystal structure and variation in unit cell parameters with crystallization temperatures of poly (ether ketone ether ketone ketone)s containing meta-phenyl links (PEKEKmK). Macromol Chem Phys 2000; 201: 2756–2759.

11.

Qiu

Yang

. Isothermal and nonisothermal melt crystallization kinetics of a novel poly(aryl ether ketone ether ketone ketone) containing a meta-phenyl linkage. J Appl Polym Sci 2006; 102: 4775–4779.

12.

Colquhoun

Dudman

Blundell

. An aromatic polyether in which sequence-randomization leads to induction of crystallinity: X-ray structure of the crystalline phase [–OArCOArArCOAr–] _n (Ar = 1,4-phenylene). Macromolecules 1993; 26: 107–111.

13.

Rao

Sabeena

Rao

. Synthesis and characterization of poly(ether sulfone) and poly(ether sulfone ketone) copolymers. J Appl Polym Sci 1999; 73: 2113–2121.

14.

Hubacher

MH.

3,3’-Dimethyl-4,4’-dihydroxybenzophenone. J Am Chem Soc 1939; 61: 2664–2665.

15.

Fukawa

Tanabe

Hachiya

. Trans-etherification of aromatic polyetherketone and trans-etherification between aromatic polyetherketone and aromatic polyethersulfone. Polym J 1992; 24: 173–186.

16.

Finger

Kruse

. Aromatic fluorine compounds. VII. Replacement of aromatic –Cl and –NO₂ groups by –F. J Am Chem Soc 1956; 78: 6034–6037.

17.

Markezich

Zamek

Donahue

. Reactions of 4-nitrophthalic anhydride with potassium fluoride and potassium nitrite. J Org Chem 1977; 42: 3435–3436.

18.

Cotton

Wilkinson

. Advanced Inorganic Chemistry. 5th ed. Chichester: John Wiley, 1988.

19.

Wynn

Roth

Pollard

. The solubility of alkali-metal fluorides in non-aqueous solvents with and without crown ethers, as determined by flame emission spectrometry. Talanta 1984; 31: 1036–1040.

20.

Dawson