Synthesis and properties of poly (ether imide) s derived from dihydroxynaphthalenes Geoffrey C. Eastmond" and Jerzy Paprotny Donnan Laboratories, University of Liverpool, PO Box 147, Liverpool L69 3BX, UK Poly(ether imide)s were synthesized from bis(ether anhydride)s derived from 1,5-, 2,3-, 2,6- and 2,7-dihydroxynaphthalenesand various aromatic diamines using a two-stage solution process, normally with chemical imidization. During the synthesis of polymers from the bis(ether anhydride) from 2,3-dihydroxynaphthaleneand 4,4'-oxydianiline (4,4'-ODA), by both chemical and thermal imidization, small proportions of a relatively insoluble, infusible, crystalline solid were produced. It is proposed that this product is a cyclic oligomer. Solubilities of the polymers were assessed and, where sufficiently soluble, molecular weights were determined by gel permeation chromatography.Apart from polymers based on 2,3-naphthalene units, the polymers had limited solubilities. Glass-transition temperatures were determined; all were in excess of 220 "C, some were in excess of 300 "C. Several poly(ether imide)s based on ODA, were found to be thermally stable to 590 "C. Polymers based on the bis(ether anhydride) derived from 2,3-dihydroxynaphthalenegave strong solvent-cast films with high moduli extensions to break were modest except for the polymer from the diamine BAPB which extended to 150% prior to fracture. Naphthalene units have been incorporated into polymers over the years in order to achieve a variety of effects.Early workers used naphthalene units to increase glass-transition tempera- tures of polymers. Subsequently naphthalene units were intro- duced to decrease glass-transition and processing temperatures of polyarylates, especially of liquid crystalline polyesters. Often, naphthalene units are used to replace a proportion of other aromatic residues in order to produce copolymers which are more processable than the parent polymer. Korshak et al.' produced polyesters with high melting points based on 1,2-dihydroxynaphthalene.Jedlinski and Sek' synthe- sized polyesters with bisnaphthylene and binaphthyl residues by reacting their diols with terephthaloyl chloride. The bisnaphthylene units incorporated flexible links (e.g. -CH,-) as well as the rigid naphthylene moieties, while the binaphthyl residues incorporated units with a large dihedral angle which might disrupt chain packing.These various polymers had melting points in the range 340-420°C and a polymer with bisnaphthylene units had a melting point of 340 "C, compared with 270 "C for the corresponding polymer with bis-p-phenyl- ene units. In liquid crystalline polymers, liquid crystallinity can be maintained while incorporating 1,4-, 1,5- or 2,6-substituted naphthalene units in which the bonds incorporating those units into the chain are co-linear or ~arallel.~ Other structures introduce non-linearity into chains and disrupt liquid crystal- linity. The units may be incorporated into the copolymers as diols or as dicarboxylic acids.Introducing these units in place of 1,4-phenylene units provides a rigid structure which disrupts chain packing and reduces thermal transition temperatures in the resulting copolyesters. A copolyester based on 4-hydroxy- benzoic acid and 2,6-dihydroxynaphthoic acid is commer-cialized as Ve~tra.~ The 2,6-naphthalene unit is very effective in reducing melting points of copolymers while 1,4- or 1,5-units give melting points too high for pro~essing.~ Poly(ethy1ene naphthalene-2,6-dicarboxylate)(PEN) has been introduced, as an alternative to poly(ethy1ene terephthal- ate); 2,6-naphthalene units give enhanced physical and mechan- ical proper tie^.^ PEN crystallizes and has low solubility. A polymer equivalent to PEN was one of a series of polyesters based on naphthalene units which were among the first poly- esters to be synthesized.6 Amorphous copolyesters with higher solubilities have been prepared.' In contrast, there are few instances of polyimides incorporat- ing naphthalene residues and little information on or under- standing of the nature and properties of the resulting polymers or copolymers.To date, studies by other workers fall into two main categories: (i) structures containing the imide derived from 1,8-naphthalenedicarboxylicacid units; (ii) polymers in which the amine residue contains a naphthalene moiety. Within category (i), polymers based on 1,4,5,8-naphthalene- tetracarboxylic acid dianhydride have been prepared. This structure, which was converted to its diaminobisimide and used as a diamine monomer, incorporates a large disc-like unit into the polymer backbone, cf.pyromellitic dianhydride; poly(- ether imide)s so formed were found to be soluble and to have high glass-transition temperatures.' Another study used dian- hydrides containing two 1,8-naphthalenedicarboxylicacid dianhydride units with an intermediate link so that two rigid moieties are incorporated per unit." A further study employed 4,4'- binaph thyl- 1,1',8,8'-tetracarbox ylic acid dianh ydride, which also incorporates two rigid moieties per unit, with a large intermediate dihedral angle. In the latter case, polymers were prepared with a variety of aliphatic and aromatic diamines and, except for polymers based on p-phenylene diamine, were mainly found to be soluble in chlorinated as well as aprotic solvents; glass-transition temperatures for aromatic polymers were in excess of 400°C and the polymers were thermally stable." In category (ii), 1,5-diaminonaphthalene was polymerized with 6FDA [2,2-bis( 3,4-dicarboxyphenyl) hexafluoropropane dianhydride] and was found to increase the glass-transition temperature relative to p-phenylene diamine and to give poly- mers with good modulus but reduced extension to break.12 Such polymers of modest molecular weight were found to be soluble in a number of solvents, e.g.tetrahydrofuran and dich10romethane.l~ We recently reported the synthesis of a series of new poly(- ether imide)s based on catechol and its derivatives.This series includes a poly(ether imide) based on the bis(ether anhydride) 2,3-NBA, derived from 2,3-dihydroxynaphthalene.14There is also a report of a diamine derived from the same dihydroxyna- phthalene being used in polyimide synthesis." Following our investigation of nitrodisplacement reactions between 4-nitroph- thalodinitrile and dihydroxynaphthalenes and the identifi- cation of which dihydroxynaphthalenes can be readily converted into bis(ether anhydride),16 we now report on a preliminary study of the synthesis and properties of poly(ether imide)s based on bis(ether anhydride)s derived from 1,5-,2,3-, 2,6- and 2,7-dihydroxynaphthalenes.These bis(ether anhy- dride)s are identified J. Muter. Chem., 1996, 6(9), 1459-1464 1459 1.5-NBA 2,3-NBA 0 0OyJq-p3$ Q 2,6-NBA 0 0 0mqo 0 0 by a code in accordance with that used previo~sly;'~ the numerical component corresponds to the substitution pattern of the naphthalene moiety and polymers are identified as 2,3- NBA/MPD (polymer from 2,3-NBA and MPD) etc.With other dihydroxynaphthalenes nitrodisplacement reactions are so inefficient that the reactions are unrealistic as potential sources of commercial materials, although their study might be fruitful subsequently to provide a more complete knowledge of structure-property relationships. Experimental Bis(ether anhydride)s derived from 1,5-, 2,3-, 2,6- and 2,7- dihydroxynaphthalenes were prepared as described in the preceding paper.16 Diamines were obtained from sources ident- ified previously. l5 Other reagents and solvents were general laboratory reagents.Polymer synthesis Polymer syntheses were carried out by two conventional two- stage syntheses (Scheme 1). In both cases anhydride and 0 0 &O, Py or heatI diamine were reacted in solution to form poly(amic acid). In one case the poly(amic acid) was imidized chemically and in the other thermally. Typical syntheses are described below. Chemical imidization. In a reaction flask fitted with a magnetic stirrer, 4 mmol of diamine was dissolved in 25 cm3 1-methyl-2-pyrrolidinone (NMP), and an exact stoichiometric equivalence of bis(ether anhydride) was added in one portion with stirring at room temperature. After reacting for one day, when the mixture had become very viscous due to the forma- tion of poly(amic acid), the polymer was imidized by addition of an excess of a 50: 50 v/v mixture of acetic anhydride and pyridine.The mixture was allowed to react for at least 6 h and usually overnight. The poly(ether imide) was isolated by drop- wise addition of the solution into a large excess of methanol, at which point the polymer usually precipitated as small yellow balls. The polymer was filtered off and added to a further large volume of methanol and the mixture boiled for several hours to remove residual solvent and imidizing reagents. This process was repeated, after which the polymer was filtered off and dried under vacuum at 110 "C. Yields were approximately quantitative.Thermal imidization. About 8 mmol of diamine was added to a flask fitted with a magnetic stirrer and a nitrogen bleed and was dissolved in 40 cm3 of NMP. An exact equivalence of bis(ether anhydride) was added and the mixture reacted as above to form poly(amic acid). A small sample of poly(amic acid) was removed and imidized chemically. Then 20cm3 xylene was added and the reaction flask was fitted with a Dean-Stark trap. The mixture was brought to the boil at 160 "C and water was removed as its xylene azeotrope in order to effect imidization. If the solution remained homogeneous refluxing was continued for 30 h after which the mixture was cooled and the polymer extracted as described for chemical imidization.When 4,4'-ODA was used as diamine the polymer started to precipitate out in the early stages of imidization and some xylene was removed to improve the solvent quality. This procedure raised the boiling point to about 190°C but some insoluble powder always remained. The insoluble material was separated by filtration. Refluxing was continued to complete thermal imidization, no further precipitate formed and the polymer was isolated from homogeneous solution as described above. Polymer characterization Molecular weights were determined by gel-permeation chroma- tography using DMF with 1 mol dm-3 lithium chloride as the mobile phase, as described previ0us1y.l~ Glass-transition tem- peratures were determined with the aid of a Perkin-Elmer DSC-2.Solubilities of the polymers were determined in a series of solvents by allowing samples of polymer to stand in solvents for a period of two weeks. Samples of some polymers were cast as films from 5-7 wt.% filtered (0.45 pm) homogeneous solutions in chloroform by slow drying at room temperature in flat-bottomed petri dishes and subsequently heating under vacuum at 140 "C. Dumbbell-shaped samples were cut from the films and mechanical properties were determined using an Instron tensile tester. Results and Discussion Poly(ether imide)s were prepared from the four bis(ether Scheme 1 anhydride)s identified (see following page) 1460 J. Muter. Chem., 1996, 6(9), 1459-1464 H2NTYH2 MPD 4,4'-ODA H2N, & ! - D H 2 3,4'-ODA BAA BAP TPE-Q TPE-R BAPB TMB XMBD BAA-GF CF, MPD-CFS and a series of diamines.Polymers investigated were mainly prepared in NMP solution with chemical imidization (Scheme 1). Polymerizationof 2,3-NBA with 4,4'-ODA Some unusual observations were made during the synthesis and handling of 2,3-NBA/4,4'-ODA poly(ether imide)s. Preparation according to the standard procedure produced a homogeneous solution on chemical imidization and the poly- mer was isolated by dropwise precipitation into methanol. During subsequent solvent-casting, from homogeneous solu- tions, of films for mechanical testing, it was found that the polymer gave a cloudy film and, subsequently, was not totally soluble in chloroform but gave a cloudy solution.A white solid was separated by filtration through a 0.45 pm filter; the solid was retained. A film was cast from the clear solution but was hazy when dry; other poly(ether imide)s prepared gave optically clear films except 2,3-NBA/BAPB which also devel- oped some haziness when cast. Samples of 2,3-NBA/4,4'-ODA were also prepared by ther- mal imidization of the poly(amic acid) solution, as described above. During the first hour of heating at 160"C, while refluxing in the presence of xylene, a white precipitate formed. The composition of the solvent mixture was modified by addition of NMP, to improve solvent quality, and the reflux temperature was raised to 190°C; the solid did not dissolve. This powder, which constituted 8 wt% of the total solids, was isolated and retained.Further prolonged heating did not produce any further precipitate. Both white solids had virtually identical compositions, as determined by elemental analysis, and similar to that of polymer prepared from the same constituents. The calculated composition for the poly(ether imide) is C, 74.02; H, 3.52; N, 4.55%. The composition of the solid isolated from the chemi- cally imidized polymer was C, 70.85; H, 3.55; N, 4.60% and that from the thermally imidized sample was C, 70.28; H, 3.26; N, 4.43%; it is normal for carbon contents of aromatic polymers as determined by elemental analysis to be slightly lower than calculated. The solids, which were totally insoluble in chloroform or NMP (both are solvents for the polymer) and even in methanesulfonic acid, were only soluble in concen- trated sulfuric acid.The solids did not exhibit a glass-transition (there was possibly an extremely weak transition at 240 "C) and the solid did not fuse or decompose on heating in air to temperatures up to 550 "C on a hotplate. No useful information was obtained from mass spectrometry of samples in concen- trated sulfuric acid. The solids and polymer had almost ident- ical infrared spectra and contained absorptions at all the characteristic frequencies for polyimides identified by Dine- Hart and Wright." In the fingerprint region all peak frequen- cies, for both the solid and the polymer, were identical to within 2cm-' and all peak intensities were identical except for very small differences at and between 1112.8 and 1076.2 cm-', between 880.7 and 831.5 cm-' and at 672 cm-'.The polymer exhibited small absorptions at about 3500 cm-' which were absent in the white powder and could have arisen from amino groups. It was also demonstrated by X-ray powder diffraction that, while polymer films were amorphous, the white powder was highly crystalline. At present we are unable to make a positive identification of the nature of the solid powders. They do not have the characteristics of relatively insoluble high molecular weight polymer; they showed no signs of swelling, even in hot NMP, in solvents for the polymer, even after standing in those solvents for ten months. It also seems incomprehensible that 8 wt% of a polymer with a peak molecular weight by gel permeation chromatography (approximately the weight-aver- age molecular weight) of 40 kg mol-' could be so insoluble and crystalline when the same polymer, prepared by chemical imidization with a peak molecular weight of 90 kg mol-' was almost totally soluble in chloroform and amorphous.Gel permeation chromatograms of most poly(ether imide) s pre-pared show distinct peaks at low molecular weights corre- sponding to dimer and trimer species and we suggest that the solids isolated might in fact be insoluble cyclic oligomers. This proposal is consistent with small amino absorptions in the infrared spectrum of the polymer. Further, cyclic polycarbonate oligomers are highly crystalline and the polyimide equivalents could be very insoluble.Proof of this proposal requires further study and if we are able to establish the identity if the powder we will report the results separately. J. Mater. Chem., 1996,6(9), 1459-1464 1461 Characterization Table 2 Solubilities of poly(ether imide)s formed from naphthalene bis(ether anhydride)s" Molecular weights. Molecular weights of several polymers soluble in DMF-LiCl( 1M), possibly after dissolution in NMP anhydride and dilution into DMF-LiCl, were determined by gel per- 1,5-NBA 2,3-NBA 2,6-NBA 2,7-NBAmeation chromatography, and the results are presented in amine Table 1 Molecular weights quoted correspond to peaks of the MPD CHCI, p CHCl, s NMP (hot) s NMP (hot)chromatograms, based on polystyrene standards In a few DMA i DMA trg trgcases with 2,7-NBA, poly(amic acid)s were prepared in two NMP s cresol s different solvents [NMP and dimethylacetamide (DMA)] 4,4-ODA cresol trg CHC1, p cresol s DMA trg according to the standard procedure described above It was H,SO,s DMAs NMP s NMP s observed that molecular weights of poly(ether imide)s formed 3,4-ODA CHC1, p CHCl, sin DMA were greater than those of polymers prepared in CHCI, s CHCI, s DMA iNMP, solubilities of the polymers prepared in NMP were NMP s lower, consistent with higher molecular weight BAA CHC1, s -CHCI, s CHCI, s Polymer solubilities were tested in a series of solvents BAP CHCl, s -hot cresol s CHCI, p Normally for such polymers the solvent power is in the order DMA trg NMP s conc H,S04 >cresol >NMP >DMA >DMF >CHC1, TPE-Q -CHCl, s, trg -NMP trg cresol sIn some solvents polymers were only partially soluble (high- TPER -CHC1, s -cresol s molecular-weight fractions were probably insoluble) and poly- BAPB -CHCl, s cresol trg cresol trg -CHCI, smers were only soluble in hot solvents and formed thermally MBXD CHC1, s CHCI, s -CHC1, sreversible gels on cooling, in a few cases polymers swelled in TMB cresol s CHCl, s hot solvents to form gels The results of several tests are MPD CF, CHCl, s CHCl, s NMP s summarised in Table2 For several polymers not all cases of BAA-6F CHCl, s CHCl, s -NMP s CHCl, s CHC1, sinsolubility are recorded Unless otherwise stated, if a polymer is designated to be soluble in one solvent it can be assumed "Key s, soluble, 1, insoluble, p, partially soluble, trg, forms a thermally that it is soluble in the more powerful solvents in the above reversible gel on cooling from hot solution list, otherwise polymers were not soluble in solvents not identified The influences of the different naphthalene units may be compared with different phenylene units Thus, 1,5- and 2,6-naphthalene units may be compared with the hydro- quinone unit but with linkages parallel and offset rather than Table 3 Glass-transition temperaturesrc of poly(ether 1mide)s formed co-linear Similarly, the 2,7-naphthalene unit might be likened from naphthalene bis(ether anhydride)s to a resorcinol unit and 2,3-naphthalene to catechol, 2,3- anhydridedihydroxynaphthalene is a benzannelated catechol The same pattern of solubilities is found for naphthalene as for phenylene amine 1,5-NBA 2,3-NBA 2,6-NBA 2,7-NBA units Thus, 2,3-naphthalene units impart much greater solu- bility than the other units, as do catechol and substituted MPD 260 255 Tg 230 2 54 catechols when compared with hydroquinone or resorcinol 4,4-ODA T, 340 240 235 249 245and their derivatives l9 3,4-ODA 235 226 225 221 BAA 265 -236 247 Tbermal-transition temperatures.Glass-transition tempera- BAP 253 -246 246 tures of the poly(ether imide)s were determined and the results TPE-Q -229 -228 are summarised in Table 3 To a first approximation, the glass- TPE-R -208 -not found -227 230 23 1 transition temperatures of the polymers are independent of the BAPB 28 1 265 -272substitution pattern of the naphthalene units Otherwise, glass- MBXD TMB not found 308 -not found transition temperatures fit with previously established patterns MPD-CF, 250 250 -235 in that glass-transition temperatures increase with the rigidity BAA-6F 276 256 218 256 of the diamine unit and with reduced possibilities of rotation about linkages to the imide units Thus, MPD-based polymers Table 1 Molecular weights (kg mol I) of poly(ether imide)s formed from naphthalene bis(ether anhydride)s anhydride method of amine imidization 1,5-NBA 2,3-NBA 2,6-NBA 2,7-NBA ~~____ a 4 aMPD chemical 70 4.4-ODA chemical a 94 a 46 thermal -41 a3,4-ODA chemical 60 41 39 BAA chemical 54 23 37 BAP chemical 78 -a 146 aTPE-Q chemical 170 4TPE-R chemical 68 aBAPB chemical -127 MBXD chemical 62 97 92 TMB chemical a 56 34 -7MPD-CF, chemical 25 34 BAA-6F chemical 89 52 56 47 "Insoluble in DMF-LiCl (1 mol dm ,) 1462 J Muter Chem, 1996, 6(9), 1459-1464 have slightly higher Tg values than those based on the more flexible 4,4'-ODA unit (10-20 "C) and the latter polymers have slightly higher $ values (5-25 "C) than those based on the unsymmetrical 3,4'-ODA.The lowest G is for a polymer based on TPE-R, which incorporates two flexible ether linkages into a single chain unit and also introduces additional non-linearity. In contrast, higher Tg values are observed for polymers based on MBXD, in which ortho-methyl groups are adjacent to the linkages to phthalimide units, which restricts chain rotation although it does not restrict any hinge in the system.20 The highest Tg values observed are for polymers based on TMB which not only has ortho-methyl groups which restrict group rotations but also incorporates a rigid biphenyl moiety.High G values are also observed for polymers having fluorinated units. In one case (MPD-CF,) there is a pendant trifluorome- thy1 group which increases the bulkiness of groups which might be required to undergo rotation and in the other (BAA- 6F) this group increases the stiffness of a hinge unit in the chain, cf. BAA. To date only one polymer, 2,6-NBA/MPD, has been shown to be crystallizable, with a melting point of 340°C. Thermogravimetric analysis.Polymers prepared from each of the naphthalene bis(ether anhydride)s and 4,4'-ODA, and additionally the polymer 1,5-NBA/BAA, were subjected to thermogravimetric analysis in air; heating rates were 10 "C min-l. All polymers showed good thermal stability, and thermograms of the polymers based on 4,4'-ODA are given in Fig. 1. Polymers 2,3-NBA/4,4'-ODA and 2,7-NBA/4,4'-ODA exhibited the greatest thermal stabilities; their thermograms were practically superimposable. These polymers did not start to decompose below 560°C and they lost 96% of their weight between that temperature and 740 "C; 4% weight loss was gradual on heating from ambient temperature and was prob- ably due to loss of residual solvent.2,6-NBA/4,4'-ODA lost 96% between 547 and 690 "C; there is little distinction between onsets of degradation for 2,3-NBA/4,4'-ODA, 2,6-NBA/4,4'- ODA and 2,7-NBA/4,4-ODA, but 2,6-NBA/4,4'-ODA lost weight more rapidly. The least stable polymer of the series was 1,5-NBA/4,4'-ODA, which showed 98% weight loss between 390 and 667°C. The polymer 1,5-NBA/BAA showed a 95% weight loss between 473 and 713 "C, of which about 10% was lost at about 580°C. It is interesting to compare the thermal stabilities of these polymers with the related poly(ether imide)s, in which the central units of the bis(ether anhydride)s are differently substi- tuted phenylenes; i.e. Nap in I is replaced by phenylene. Takekoshi et uL21 have previously reported the thermal stabilit- ies of poly(ether imide)s derived from 4,4'-ODA and the bis (ether anhydride) s derived from hydroquinone and resorci- I I I L 190 350 510 670 830 TPC Fig. 1 Thermogravimetric data for: (a) 2,3-NBA/4,4-ODA and 2,7-NBA/4,4'-ODA (thermograms are superimposed); (b) 2,6-NBA/4,4-ODA; (c) 1,5-NBA/4,4'-ODA nol, HBA and RBA, respectively, and we have reported on the thermal stability of that based on 4,4'-ODA and catechol bis(ether anhydride) CBA.17,22 Takekoshi et al.reported that, in air, HBA/4,4'-ODA and RBA/4,4'-ODA suffered 5 % weight loss at 553 and 537 "C, respectively; under nitrogen, decompo- sition temperatures were about 20 "C higher. For CBA/4,4- ODA we reported initial weight loss at 450°C and maximum rate of weight loss at 560°C.Thus, 2,3-NBA/4,4'-ODA and 2,7-NBA/4,4-ODA show greater thermal stabilities than CBA/4,4'-0DA, the parent polymer of 2,3-NBA/4,4'-ODA, and have thermal stabilities at least comparable with the corre- sponding poly(ether imide)s based on HBA and RBA. Mechanical properties Because many of the polymers based on naphthalene bis(ether anhydride)s are insoluble in solvents suitable for solvent casting of films, studies of mechanical properties concentrated on polymers based on 2,3-NBA, which is of course a catechol derivative, and these results may be compared with our pre- vious results for poly(ether imide)s based on catechol bis(ether anhydride) (CBA).22 The mechanical properties of several poly (ether imide) s based on 2,3-NBA and several different diamines were deter- mined.The results are summarized in Table4. Samples were cut from solvent-cast films with a dumbbell-shaped cutter. Many poly(ether imide)s are tough and yield before fracture. In this study a number of polymers showed brittle fracture at low elongations (cu. 5% extension) and these results conform with those we reported recently for poly(ether imide)s based on catechol bis(ether anhydride)s.22 However, a few samples yielded at low extensions and exhibited strain-hardening prior to fracture. For some polymers both types of behaviour were observed in different samples. It is difficult in these circum- stances to define true behaviour but we assume that, where observed, yielding is more representative of true behaviour and that failure to yield was a result of premature fracture, possibly because of small defects introduced on sample prep- aration; maximal values of parameters are recorded in Table 4, average values are given in brackets.The molecular weights of the polymers prepared in this study are not optimal and it is probable that better properties are achievable. The initial moduli of all polymers prepared are high and for several are comparable to those of Kapton2, and Ultem (2.96 GPa).Z4 Ultimate strengths are less than that of Kapton (172 MPa) but higher than that of Ultem (105 MPa). In most cases, however, elongations to break are less than those of the commercial materials (60-70%). This range of properties, in terms of modulus and strength, are superior to those of polymers based on the parent catechol bis(ether anhydride) (CBA), reported previously.22 An exception to the low exten- sions to break are the data for 2,3-NBA/BAPB, which showed a maximal extension to break of 135%.This result is quite exceptional compared with the properties of the other polymers synthesized in this study. It is, however, comparable to the data for the corresponding polymer CBA/BAPB, which exhib- ited 170% extension to break.22 We speculated previously on the origins of this behaviour but its actual origin remains obscure. Overall, the results for 2,3-NBA/BAPB confirm our previous conclusions and show that poly(ether imide)s based on catechol and its derivatives are capable of providing materials which are highly processable and have excellent properties.Conclusions Bis(ether anhydride)s based on the 1,5-, 2,3-, 2,6- and 2,7- substituted naphthalene units can be used to incorporate naphthalene residues into poly(ether imide)s by polymerization with any of several aromatic diamines. Synthesis was achieved J. Muter. Chem., 1996, 6(9), 1459-1464 1463 Table 4 Mechanical properties of some poly(ether imide)s based on 2,3-NBAa method of initial modulus/ yield stress/ elongation ultimate stress/ elongation amine imidization GPa MPa (%) MPa (Yo) MPD chemical 2 91 95 1 ODA thermal (2 7) 2 48 (88)104 4,4-ODA chemical (2 24) 2 88 109 6 89 90 3 3,4-ODA chemical (2 65) 2 65 - - 99 8 BAP chemical (2 47) 27 - - 118 2(86 7) BAPB chemical (2 6) 26 117 4 93 120 4 (111) TMB chemical (2 42) 34 120 6 (113 2) 88(9) (107 3) 140 (2 8) (119 7) (8 7) (124) 'Maximal parameters Average values given in brackets by a conventional two-stage solution polymenzation involving intermediate formation of the poly(amic acid) followed by chemical imidization Glass-transition temperatures of the polymers exceed 220 "C and, for some diamines, exceed 300°C or are unobservable below 450 "C The transition temperatures are largely indepen- dent of the substitution pattern of the naphthalene residue in the anhydnde moiety but vary with the structure of the diamine Variations in glass-transition temperature follow pre- viously established patterns in that rigid diamines, especially those with substituents which hinder rotation, raise glass- transition temperatures, while flexible residues lower glass- transition temperatures Of all the polymers investigated, only that based on the anhydnde 2,6-NBA and MPD was found to be crystallizable It was established that the polymers have good thermal stabilities in air Polymers based on 4,4'-ODA and three of the anhydndes are stable to about 550°C, the polymer based on 1,5-NBA has a lower stability, decomposition starts at 390 "C Polymer solubilities vary with the structure of the diamine used Many of the polymers have limited solubilities and several give thermally-reversible gels in solvents such as NMP, DMA and cresol However, polymers based on 2,3-NBA, a catechol derivative, are far more soluble than those based on other naphthalene bis(ether anhydr1de)s This observation parallels the previously found high solubility of poly(ether imide)s based on catechol in comparison with the analogues based on hydroquinone and resorcinol Many polymers based on 2,3-NBA had sufficiently high molecular weights and solubilities to be cast into films Most of these polymers exhibited high moduli and strengths but relatively low extensions to break, failure to yield may have been due to limited molecular weights in some cases The polymer based on 2,3-NBA and BAPB yields and shows a large (135%) extension to break, a similar result was observed previously with the analogous polymer based on the parent catechol bis(ether anhydride) and BAPB The authors wish to thank Dr N C Billingham for thermo- gravimetric analysis data, Valerie Laberthe, an ERASMUS student, for measurements of mechanical properties and Dr M Harding for help with X-ray diffraction The authors also thank the SERC and the DRA for financial support References 1 V V Korshak, S V Vinogradova and M A Iskenderov, Vysokomol Soedzn ,1962,4,345 2 Z Jedlinski and D Sek, J Polym Sci A-1,1969,7,2587 3 W J Jackson, Jr ,Macromolecules, 1983,16,1027 4 G W Calunden, USP 4,161,470, 1979 5 Saturated Polyester Resin Handbook, ed K Yuki, Nikan Industnal Publisher, Japan, 1990, p 874 6 G J Cooke, H P W Huggdl and A R Lowe, unpublished results quoted in ref 7 7 R Hill and E E Walker, J Polym Sci ,1948,3, 609 8 C -S Wang and Y -M Sun, Polym Prepr Am Chem SOC Dzv Polym Chem ,1996,36,197 9 H Ghassemi and A S Hay, Macromolecules, 1994,27, 3 116 10 D Sek, P Pijet and A Wanik, Polymer, 1992,33, 190 11 J P Gao and Z Y Wang, J Polym Sci Part A Polym Chem, 1995,33,1627 12 H H Gibbs and C V Breder, in Copolymer Polyblends and Composites, ed N A J Platzer, Adv Chem Ser 142, ACS, Washington, 1975 13 G R Husk, P E Cassidy and K L Gebert, Macromolecules, 1988, 21,1234 14 G C Eastmond and J Paprotny, Macromolecules, 1995,28,2140 15 C-P Yang and W-T Chen, Macromolecules 1993, 26, 4865, J Polym Scz Part A, 1994,32,5148 16 G C Eastmond and J Paprotny, J Muter Chem, preceding paper17 G C Eastmond and J Paprotny, Polymer, 1994,35,5148 18 R A Dine-Hart and W W Wnght, Makromol Chem, 1971, 143, 189 19 G C Eastmond and J Paprotny, Reactive and Functional Polymers, 1996,30,21 20 G C Eastmond, J Paprotny and I Webster, Polymer, 1993, 34, 2865 21 T Takekoshi, J E Kochanowski, J S Manello and M J Webber, J Polym Sci Polym Symp ,1986,74,93 22 G C Eastmond and J Paprotny, Macromolecules, 1996,29,1382 23 Du Pont Technical Information Bulletin H2, 1966 24 R 0 Johnson and H S Burhlis, J Polym Sci Polym Symp, 1983, 70,129 Paper 6/02622B, Received 15th April 1996 1464 J Muter Chem, 1996, 6(9), 1459-1464