Rigid rod polymers with flexible side chains
`Synthesis, structure and phase behaviour of
`poly(J-n-alkyl-4-oxybenzoate)s*
`
`R. Stern, M. Ballaufft, G. Lieser and G. Wegner
`Max-Pianck-lnstitut fur Polymerforschung, Postfach 3148, 65 Mainz, Germany
`(Received 26 February 1990; revised 31 July 1990; accepted 31 July 1990)
`
`The synthesis of poly(3-n-alkyl-4-oxybenzoate )s (PAOB-n) with n = 3-18 is reported. The sufficient solubility
`of these comb-like polymers with stiff-chain backbones allows the determination of the Kuhn length
`(100-200 A). The unit cells could be determined for PAOB-n with n=3 and 5. All PAOB-n form
`thermotropic mesophases. For n=3 a nematic phase is found. For n=5 there is a transition from a
`smectic-like, layered mesophase to a nematic mesophase. PAOB-n with n ~ 6 solely form layered mesophases.
`Except for n = 3, all PAOB-n exhibit a transition to the isotropic state. The transition temperature is
`lowered monotomically with increasing n.
`
`(Keywords: rigid rod polymer; synthesis; phase behaviour)
`
`INTRODUCTION
`Rigid rod polymers usually exhibit very low solubility
`and melting points far above the temperatures of thermal
`decomposition. In a number of recent publications it has
`been shown that flexible side chains appended to the rigid
`backbones lower the melting point and increase the
`solubility in a systematic fashion 1. Most of the fully
`aromatic systems studied so far are composed of rather
`symmetric repeating units, i.e. the different conformers
`of these polymers exhibit nearly the same shape. Asym(cid:173)
`metric monomers, on the other hand, should cause a
`further disturbance of crystallization because of the vast
`number of shapes generated by different conformers. In
`this work we present a comprehensive study of the
`poly(3-n-alkyl-4-oxybenzoate )s(PAOB-n)
`
`as an example of a comb-like polymer composed of
`asymmetric repeating units. The number of carbon atoms
`n in the side chains is varied between 3 and 18, thus
`PAOB-12 is poly(3-dodecyl-4-oxybenzoate). The choice
`ofpoly(4-oxybenzoate) as the stiff backbone derives from
`the fact that this polymer has already been the subject
`7
`of a number of exhaustive studies 2-
`• Hence the influence
`of the side chains on the structure can be assessed in
`detail by comparing the results found here with the data
`obtained on the unsubstituted poly(4-oxybenzoate). To(cid:173)
`gether with a previous investigation 8 of defined oligomers
`ofPAOB-3 the present study aims at a full understanding
`of the structure and the phase behaviour of these
`comb-like polyesters.
`
`* Part of the PhD Thesis of R. Stern
`t To whom correspondence should be addressed. Present address:
`Polymer-Institut der Uni'<ersitiit Karlsruhe, Kaiserstr. 12, 7500
`Karlsruhe, Germany
`
`0032-3861/91/112096-10
`© 1991 Butterworth-Heinemann Ltd.
`2096 POLYMER, 1991, Volume 32, Number 11
`
`The synthesis of the monomers, the 3-n-alkyl-4-
`hydroxybenzoic acids (1-n), proceeds along the following
`lines: the monomer bearing the propyl substituent (1-3,
`n = 3) is easily available from the 3-allyl-4-hydroxybenzoic
`acid 9 through catalytic hydrogenation at atmospheric
`pressure. The monomers bearing longer n-alkyl chains
`(1-5 to 1-18) are obtained through a Fries rearrange(cid:173)
`ment10 of the respective 0-acylated 4-hydroxybenzoic
`acids (2-n) and subsequent Clemmensen reduction of the
`ketones 3--n (Scheme 1). The polycondensation of the
`monomers 1-n is performed by heating the monomer
`with an excess of acetic anhydride (cf. ref. 3).
`
`EXPERIMENTAL
`Materials
`All chemicals were purchased from Merck or Fluka.
`They were used without further purification unless
`otherwise stated. All solvents used were of p. a. quality;
`1,1 ,2,2-tetrachloroethane, phenol and a-dichlorobenzene
`employed as solvents for viscosity measurements were
`distilled prior to use.
`
`Synthesis of monomers
`1,3-Propyl-4-hydroxybenzoic acid (1-3). The 3-allyl-4-
`hydroxybenzoic acid was synthesized through Claisen
`rearrangement of ethyl-4-allyloxybenzoate which was
`prepared from ethyl-4-hydroxybenzoate9
`• Hydrogenation
`using palladium on charcoal as catalyst and subsequent
`saponification in 30% aqueous sodium hydroxide solution
`led to 3-propyl-4-hydroxybenzoic acid. The monomer
`
`0
`0
`II
`c9rOH ~ R'
`O-C-R'
`liAlCl3 • Q
`-
`2JNaOH
`COOEt
`
`COOH
`
`c¢J
`
`OH
`
`Zn/HCl &R
`Hg ~
`COOH
`
`Scheme 1
`
`

`
`Table 1 Melting points and elemental analyses of the 3-n-alkyl-4-
`hydroxybenzoic acids (1-n)
`
`Melting
`point
`(OC)
`
`115-118
`126-127
`100-101
`103-104
`98-99
`101-102
`100-102
`100-101
`
`Calc.
`
`Found
`
`C(%)
`
`H(%)
`
`C(%)
`
`H(%)
`
`66.64
`69.20
`70.23
`73.33
`74.45
`75.39
`76.18
`76.86
`
`6.73
`7.76
`8.18
`9.43
`9.89
`10.26
`10.58
`10.86
`
`66.58
`69.39
`70.11
`73.46
`74.38
`75.53
`76.31
`76.91
`
`6.76
`7.75
`8.14
`9.26
`9.77
`10.14
`10.36
`10.77
`
`n
`
`3
`5
`6
`10
`12
`14
`16
`18
`
`1H n.m.r. (acetone): J=0.89 (t, CH 2cth), 1.2-1.7 (m, -(C{j2 ).-), 2.69
`(t, Ar-C{:Jd, 6.93, 7.72 (d, Ar-B), 7.82 (s, Ar-B)
`
`Table 2 Melting points and elemental analyses of the 3-(n-alkyl-2-
`one )-4-hydroxybenzoic acids (3-n)
`
`Melting
`point
`CO C)
`
`191-193
`185-187
`179-180
`174-176
`173-175
`168-170
`
`n
`
`6
`10
`12
`14
`16
`18
`
`Elemental analysis
`
`Calc.
`
`Found
`
`C(%)
`
`H(%)
`
`C(%)
`
`H(%)
`
`66.08
`69.82
`71.20
`72.36
`73.35
`74.20
`
`6.84
`8.29
`8.82
`9.27
`9.66
`9.98
`
`66.13
`70.00
`71.32
`72.22
`73.57
`73.07
`
`6.71
`8.13
`8.70
`9.13
`9.53
`9.90
`
`1H n.m.r. (trifluoroacetic acid): J=0.95 (t, CH 2C{j 3 ), 1.2-2.2 (m,
`-(C{j 2 ).-), 3.24 (t, COq:hCH 2 ), 7.18 and 8.34 (d, Ar-B), 8.81 (s,
`Ar-B)
`
`1-3 was purified from H 20 and dried carefully prior to
`use. The overall yield of the four-step reaction was 45%.
`The melting point and the elemental analysis of 1-3 is
`given in Table 1.
`
`Synthesis of monomers 1-5 to 1-18. The ethyl-4-
`alkanoyloxybenzaotes 2-5 to 2-18 were prepared by
`reacting ethyl-4-hydroxybenzoate with an excess of the
`respective acid chloride according to the standard
`procedures given in the literature 11
`. The Fries rearrange(cid:173)
`ment of 2-n was achieved in the following way (cf. ref.
`10): 0.3 mol of 2-n was dissolved in 500 ml CS 2 and
`1.2 mol A1Cl 3 were added in small portions leading to a
`slightly exothermic reaction and evolution of HCI. After
`the last addition the mixture was refluxed for 3 h. Then
`the solvent was distilled off and the remainder heated to
`140-160°C followed by evolution of HCl and strong
`foaming. After 2 h and cooling to room temperature
`800 ml H 20 were added and subsequently 160 g NaOH
`with caution. The mixture dissolved upon heating to
`110-l20°C and the resulting ketone 3-n could be isolated
`by acidification with 400 ml concentrated HCI. Purifica(cid:173)
`tion was achieved through recrystallization from ethanol,
`toluene or chloroform (yields 40-45% ). Table 2 shows
`the melting points and the elemental analyses of 3-5 to
`3-18.
`The Clemmensen reduction was carried out in the usual
`way following references 10 and 11: 0.1 mol 2-n was
`refluxed for 24 h in 350 ml of a mixture ofH 20/ethanol/HCl
`( 1 : 1 : 2) with zinc amalgam prepared from 200 g zinc
`powder 11
`. During this time 10 ml of concentrated HCl
`were added several times. The resulting 1-n was isolated
`
`Rigid rod polymers: R. Stern et al.
`
`from the cold mixture by extraction with diethyl ether.
`Analysis by 1 H nuclear magnetic resonance (n.m.r.)
`spectroscopy demonstrated that the raw materials still
`contained a side product having a double bond in the a
`position of the alkyl side chain. For further purification
`20 g of the raw material were dissolved in 300 ml ethanol
`and hydrogenated after addition of 0.5 g palladium on
`charcoal (10%) at 50°C for 24-30 h. Then the catalyst
`was filtered off and the monomers 1-n were recrystallized
`three times from CH 30H/H 20 or CH 30H (yields
`50-75%).
`
`Polymerization (cf. ref. 3)
`For the polymerization reaction, 1.5-3.0 g of 1-n were
`added to 1.2 equivalents of acetic anhydride and refluxed
`under an atmosphere of argon at 180°C for 30 min. Then
`the condenser was taken off and the acetic acid removed
`by a slow stream of argon. While raising the temperature
`to 260oC the evolution of acetic acid started again at
`- 245oC. By variation of the reaction time from 45 min
`to 7 h polymers with different molecular weights could
`be obtained (Table 3). Purification of PAOB-3 was
`carried out by refluxing the polymer with acetone;
`polyesters with longer side chains were dissolved in hot
`chloroform and precipitated from methanol. All polymers
`reported in this investigation gave satisfactory elemental
`analyses when taking into account the measured molecular
`weights and the respective end groups.
`
`Determination of' molecular weight through end group
`analysis
`According to Kricheldorf and Schwarz 3
`, 20-30 mg of
`the polyester PAOB-n, -200 mg 40% NaOD in D 20
`and 500 mg CD 30D were weighed carefully in a n.m.r.
`
`Table 3 Characterization of the PAOB-n
`
`Polymer
`
`PAOB-3,1
`PAOB-3,2
`PAOB-3,3
`PAOB-3,4
`PAOB-3,5d
`
`PAOB-5,1
`PAOB-5,2
`PAOB-6,1
`PAOB-6,2
`
`PAOB-10,1
`PAOB-10,2
`PAOB-12,1
`PAOB-12,2
`PAOB-14,1
`PAOB-14,2
`PAOB-16,1
`PAOB-16,2
`PAOB-16,3
`PAOB-16,4
`PAOB-16,5
`PAOB-16,6
`PAOB-18
`
`Reaction time
`(min)
`180°Cj260°C
`
`30/45
`30/60
`30/160
`30/240
`30/240
`30/240
`30(360
`30/240
`30/360
`
`30(240
`30/360
`30/150
`30/300
`30/240
`60/240
`30/60
`30/120
`30/180
`30/240
`30/360
`30/420
`60/180
`
`DP"
`
`16
`22
`39
`123
`74
`23
`40
`16
`37
`
`17
`41
`13
`84
`
`20
`27
`8
`16
`22
`23
`30
`40
`
`35
`
`M:
`
`2650
`3600
`6350
`20000
`12100
`4450
`7650
`3350
`7600
`
`4500
`10750
`3800
`24300
`
`6400
`8600
`2950
`5650
`7550
`7800
`10400
`13900
`13100
`
`['I] (dl g" 1
`
`)
`
`0.310b
`0.490b
`1.620b
`0.860b
`
`0.096'
`0.180'
`0.290'
`0.350'
`0.485'
`
`"Determined by 1H n.m.r. end-group analysis
`b Measured in phenol/a-dichlorobenzene (I :1) at 50°C
`'Measured in tetrachloroethane/a-dichlorobenzene (1:1) at 50oc
`d Polymerized with addition of 5 mol% 4-methoxyl-3-propylbenzoic
`acid
`
`POLYMER. 1991, Volume 32, Number 11 2097
`
`

`
`Rigid rod polymers: R. Stern et al.
`
`tube and heated to 50-60°C. The solution obtained was
`analysed by 300 MHz or 400 MHz 1 H n.m.r. spectroscopy
`to yield the amount of end groups quantitatively (see
`below).
`
`Methods
`1 H n.m.r. spectra were recorded with 80 MHz con(cid:173)
`tinuous wave, 300 MHz Fourier transform (F1) and
`400 MHz FTspectrometers (Bruker A W 80, AC 300 and
`WM 400). Vapour pressure osmometry (v.p.o.) was done
`at 50°C using toluene as a solvent (Wescan 232 A,
`Corona). The concentration of the polymer was 2-10 g 1- 1
`.
`For polarizing microscopy a Zeiss Fotomikroskop III
`equipped with a Leitz hot stage was used. Differential
`scanning calorimetry (d.s.c.) was done with a Perkin(cid:173)
`Elmer DSC-7 calibrated with indium and tin. Thermo(cid:173)
`gravimetric data were obtained by means of a TG 50 of
`Mettler under an atmosphere of oxygen or nitrogen
`employing a heating rate of 10 K min - 1
`• The intrinsic
`viscosity of the polyesters was determined using an
`Ubbelohde capillary viscosimeter. All data are mean
`values of at least four measurements corrected 13 according
`to Hagen bach. High performance liquid chromatography
`(h.p.l.c.) was carried out on Li Chrosorb RP 18 or Li
`Chrosorb diol columns (Knauer) using a mixture of
`diethyl acetate and n-heptane (1: 1) with ultraviolet (u.v.)
`detection at ). = 254 nm. For gel permeation chromatog(cid:173)
`, 103 and
`raphy (g.p.c.) 10 ,urn Styragel columns (10 5
`10 2 , Polymer Laboratories) with tetrahydrofuran
`(1 ml min - 1
`) as the mobile phase (u.v. detection at
`).=254 nm)were used. Wide-angle X-ray analysis (WAXS)
`was performed using Ni-filtered CuKa radiation in
`reflection mode on a Siemens D 500 diffractometer
`equipped with a hot stage. All diffractograms are
`uncorrected. Electron diffraction was done using a
`Philips EM 300 calibrated with thallium chloride. The
`density of polymer films and fibres was determined in a
`density gradient set up from mixtures of H 20 and
`Ca(N0 3 ) 2 at room temperature.
`
`RESULTS AND DISCUSSION
`All polymers investigated were synthesized by the acetoxy
`method 3
`, i.e. by reaction of the unsubstituted hydroxy(cid:173)
`benzoic acid with acetic anhydride. Since the maximum
`temperature applied in the course of the polycondensation
`did not exceed 260°C, no side products as discussed by
`Economy and co-workers 2 have been observed. The
`satisfactory degrees of polymerization achieved and the
`absence of any by-products to be detected in the analysis
`of the saponificated polyesters are furthermore indicative
`of the absence of side reactions. Special care has been
`taken to remove the evolving acetic acid as well as excess
`acetic anhydride by a slow stream of argon. During the
`reaction small amounts of a solid material sublimed out
`of the mixture. Analysis by infra-red (i.r.) and n.m.r.
`spectroscopy unambiguously demonstrated this substance
`to be identical with the acetylated monomer. In the case
`of short side chains (n = 3, 5, 6) the melt turned turbid
`after 40-120 min indicating the formation of a mesophase.
`For monomers having longer side chains the resulting
`melt became increasingly viscous during polycondensation
`but only turned turbid when cooled to temperatures
`below 200oc. As is obvious from Table 3 the degree of
`polymerization (DP) can be adjusted by the length of the
`
`2098 POLYMER, 1991, Volume 32, Number 11
`
`reaction time. The present data indicate that even higher
`molecular weights could be obtained if desired. However,
`raising the DP beyond the values given here resulted in
`difficulties when determining the molecular weight by
`end-group analysis (see below).
`In contrast to the unsubstituted poly(4-oxybenzoate )3
`the PAOB-3 already exhibits sufficient solubility in
`mixtures like a-dichlorobenzene with tetrachloroethane
`or phenol. Polyesters bearing longer side chains may be
`dissolved in toluene or chloroform. Therefore purification
`can be achieved through dissolution in these solvents
`and reprecipitation into methanol. This finding is in
`accordance with observations on a number of other rigid
`rod polymers being similarly substituted by flexible side
`chains 1
`. It indicates that the side chains act as a 'bound
`solvent'. However, investigations of solutions of PAOB-
`16 in solvents like toluene lead to the conclusion that
`these polyesters still have a strong tendency for association
`in solution. This fact may be also inferred from the
`formation of gels by these solutions at room temperature.
`Mixtures of strong polar solvents like a-dichlorobenzene
`and tetrachloroethane prevent the formation of such gels
`if the concentration is not too high. Up to now attempts
`to determine the molecular weight and the radius of
`gyration by light scattering have failed because of either
`strong association or small refractive index increment.
`Thus the DP of the polymers had to be determined by
`1 H n.m.r. end-group analysis according to Kricheldorf
`and Schwarz3
`. For this the magnitude of the triplet of
`the CH 2 group neighbouring the aromatic core of the
`monomer is compared to the signal of the acetic acid
`anion after saponification with NaOD/D 20. When
`evaluating the magnitude of the latter signal, a small
`triplet resulting from partial H-D exchange on the acetic
`acid has to be taken into account. By this method the
`number average DP could be obtained within 10% error
`for smaller molecular weights; higher DPs are determined
`with less accuracy. For PAOB-16 the DP thus measured
`compares favourably with data derived from v.p.o. The
`DPs given by the latter method are in general higher by
`10-15% which certainly can be assigned to association
`in solution. Therefore only DPs derived from end-group
`analysis are used in this work (cf. Table 3).
`It has been observed that substitution of the poly(4-
`oxybenzoate) backbone may lower the thermal stability
`of the resulting polymers considerably 14
`• Thermogravi(cid:173)
`metric measurements on the PAOB-n conducted in
`nitrogen (10 K min - 1 heating rate) showed these polymers
`to be stable up to - 350°C. A residual weight loss at
`lower temperatures can be traced back to the onset of
`further polycondensation. Nevertheless the oxidative
`degradation of the side chains represents a severe
`constraint for the stability at elevated temperatures in
`the presence of oxygen. All thermogravimetric data
`obtained so far indicate the upper limit for long-term
`stability in air to be at temperatures between 160°C and
`220oC. Polyesters having longer side chains are more
`susceptible to degradation. Hence, measurements at
`elevated temperatures have to be conducted in an inert
`atmosphere or in vacuo.
`Another point of great interest is the stiffness of the
`main chain. Since the unsubstituted polyester is not
`soluble in any known solvent there are no experimental
`data on its persistence length. Calculations by Erman
`et al. 15 using the rotational isomeric state model lead to
`the prediction of a rather rigid chain with a persistence
`
`

`
`0
`
`(M~),
`
`[rJ)
`
`750
`
`sao
`
`250
`
`0
`
`50
`
`100
`
`150
`
`200
`
`Figure 1 Bohdanecky plot of the intrinsic viscosities of PAOB-3 ( 0)
`and PAOB-16 ( +)
`
`length of the order of 700 A. However, these authors
`remark that such a high value has to be regarded as an
`upper limit since torsional fluctuation about the ester
`bonds should lower it appreciably 16
`. Molecular dynamics
`calculations performed by Jung and Schiirmann 17 lead
`to a much smaller value of the persistence length ( ~ 60 A)
`for this reason. Experimental data on similar polyamides
`are of limited use for settling this question since the values
`for the persistence length of these systems given by
`various methods are not conclusive 1
`18
`. By virtue of their
`•
`improved solubility in aprotic solvents the PAOB-n are
`much better candidates for an experimental test of the
`above conflicting predictions. In the absence of reliable
`measurements of the radius of gyration by light scattering
`data, the intrinsic viscosity [17] may serve as a first
`estimate of the chain stiffness. Thus [17] has been
`measured for PAOB-3 in phenol/a-dichlorobenzene and
`for PAOB-16 in tetrachloroethane/a-dichlorobenzene at
`50°C (cf. Table 3). The dependence on temperature as
`well as the dependence on concentration is well-behaved
`for these systems whereas solutions of PAOB-16 in
`toluene and in chloroform did not lead to reproducible
`results (see above). The resulting Mark~Houwink relations
`are ([17] in dl g- 1
`):
`
`PAOB-3 [IJ]=6.95x10- 2 M~· 06
`in phenol/a-dichlorobenzene at 50oC
`PAOB-16 [1]]=2.87 x 10- 2 M~· 16
`in tetrachloroethane/a-dichlorobenzene at 50aC
`A quantitative interpretation of the data for [17] may be
`given in terms of the worm-like chain model developed
`by Yamakawa and FujiP 9
`. Based on this
`theory,
`Bohdanecky20 has recently given a simple procedure for
`data evaluation leading to the following relation:
`
`(M2/['7])1i3 =A~+ B~M112
`
`where A~ is a quantity depending on the hydrodynamic
`diameter d of the chain and B~ is expressed by
`B~ = Bo<DO'),f3( <rz >ol M)~ 1/2
`where <r 2
`) 0 is the mean-square end-to-end distance, and
`the subscript oo indicates that the <r2 )/M value obtained
`from B~ is the random coil value.
`The quantity <Do.oo is the viscosity function for infinite
`chain length (<Do,oo =2.86 x 1023
`20
`. The quantity B0
`)
`varies between 1.10 and 1.00 and rna y be set to a mean
`
`Rigid rod polymers: R. Stern et al.
`
`value of 1.05 at the present level of accuracy. Figure 1
`shows a Bohdanecky plot of the present data. The weight
`average molecular weights have been calculated from the
`data given in Table 3 by assuming a most probable
`distribution.
`The Kuhn length is calculated as lK = <r2 )/Nlu where
`N = M I M u with M u and lu being the mass per unit length
`and the length of the monomer unit, respectively. The
`latter quantity is given by 6.2 A in good approximation
`(see below). From these data and the slope of the
`respective curves in Figure 1, lK is ~ 90 A for PAOB-3
`and ~ 190 A for PAOB-16. The intercept A may be used
`to yield first estimates of the hydrodynakic radius r
`according to 20
`:
`
`~0 =(4<D0,00 1.215nNA)(i!/A~)B~
`where d, = djlK is the reduced hydrodynamic radius and
`a; I Ao is related to d, by 20
`ln(d; / A0 ) = 0.173 + 2.158 ln(d,)
`If the partial specific volume v of the polymer in solution
`!s approximated by unity, we obtain for PAOB-3 a value
`m the range 7~8 A, and for PAOB-16 11~13 A. Similar
`values have been found recently for substituted cellulose
`polymers21 and seem to be quite reasonable. One has to
`bear in mind that these values have been derived under
`a number of stringent assumptions 19
`20
`. On the molecular
`·
`level the polymer chain is approximated by a cylinder
`which may be rather questionable when looking at the
`structure of PAOB-n with n > 6. Hence, these data should
`be only regarded as rough estimates and the present
`investigation of chain stiffness must be certainly supple(cid:173)
`mented by other methods.
`Despite these problems it is clear that our findings data
`are more in support of the prediction of Jung and
`Schiirmann 17 . Krigbaum and Tanaka22 recently found
`IK values of similar magnitude for poly(phenyl p(cid:173)
`phenyleneterephthalate) by a variety of methods. Thus
`the present data indicate that the stiffness of the fully
`aromatic polyester chain is significantly smaller than
`anticipated by the calculations neglecting the effect of
`bond-angle fluctuations. The surprising fact that the
`stiffness seems to increase with increasing length of the
`side chains may be explained23 by the notion that steric
`interactions between bulky side chains lead to a significant
`rise in lK. In the absence of further information on the
`shape of the PAOB-n chain as revealed by small angle X-ray
`or neutron scattering no firm conclusions can be drawn.
`
`Phase behaviour of PAOB-3
`In the following it will become obvious that the phase
`behaviour of PAOB-3 is significantly different from the
`phase behaviour of PAOB-n bearing longer side chains.
`In the case of PAOB-3 the main chains control the
`structure and thermal transitions. For longer side chains
`the volume fraction of the main chains is decreasing and
`the most favourable arrangement of the side chains
`determines the packing of the polymer in the solid state
`as well as in the mesophase.
`
`Solid state. Figure 2 shows the W AXS patterns of
`PAOB-3 recrystallized from the melt (Figure 2a) and
`from dioxane (Figure 2b). From this it is evident that
`two different modifications have been formed through
`
`POLYMER, 1991, Volume 32, Number 11 2099
`
`

`
`Rigid rod polymers: R. Stern et al.
`a
`
`.ci
`L.
`
`"' -
`0 ' :!:'
`"' c
`Qj c
`
`VI
`c
`QJ
`c
`
`5
`
`10
`
`15
`
`20
`
`25 2 8 30
`
`b
`
`5
`
`10
`
`15
`
`20
`
`25
`
`28
`
`30
`
`Figure 2 Wide-angle X-ray diffractograms (uncorrected) of PAOB-3
`recrystallized from the melt (a) and from dioxane (b)
`
`absence of meridional reflections with odd 1 indices
`indicates a 21 or a 42 helical structure of the main chain.
`An additional feature is the presence of two series of
`reflections on the different layer lines leading to two
`hyperbolas per layer. This can be seen most clearly from
`the splitting of the 0 0 2 reflection. Variations of the
`c-parameter within a range of 12.4-12.9 A were observed
`in earlier work on poly(4-hydroxybenzoic acid) and were
`shown to be a function of the D P5
`• All reflections found
`in this work may be indexed satisfactorily in terms of
`two tetragonal modifications, Ia and lb, which differ only
`with regard to the c vector of the unit cell. Table 4 gives
`the Bragg distances and their respective indices.
`The presence of two very similar modifications Ia and
`
`Table 4 Calculated and observed X-ray reflections of modifications
`Ia, lb and II of PAOB-3
`
`Reflection
`index
`(hkl)
`
`Calculated
`spacing
`(A)
`
`Observed
`spacing
`(A)
`
`Position/modification
`
`Tetragonal modifications Ia and Ib, a=b=l6.95A, c=l2.36A (Ia),
`c= 12.86 A (Ib)
`110
`020
`
`11.99
`8.48
`
`220
`130
`230
`040
`140
`330
`240
`150
`1 21
`13 I
`231
`031
`231
`041
`241
`1 51
`002
`012
`122
`132
`002
`032
`222
`0 13(a)
`013/11 3(b)
`023(a)
`023/123(b)
`004
`
`5.99
`5.36
`4.70
`4.24
`4.11
`4.00
`3.79
`3.32
`6.46
`4.92
`4.39
`5.17
`4.42
`4.02
`3.64
`3.22
`6.18
`5.81
`4.79
`4.05
`6.43
`5.12
`4.38
`4.0(a)
`4.16j4.04(b)
`3.7l(a)
`3.83j3.73(b)
`3.09(a)
`3.22(b)
`
`ll.99(vs)
`8.47(vs)
`7.16(vw)
`5.96(vw)
`5.34(s)
`4.70(s)
`4.24(vs)
`4.12(m)
`3.99(s)
`3.78(m)
`3.34(w)
`6.50(vw)
`4.86(m)
`4.28(m)
`5.ll(m)
`4.50(m)
`4.03(vw)
`3.68(w)
`3.19(vw)
`6.18(s)
`5.85(w)
`4.69(w)
`3.95(vw)
`6.43(s)
`5.08(vw)
`4.50(vw)
`4.06(m)
`
`Equator (a, b)
`
`1st layer line: a
`
`1st layer line: b
`
`2nd layer line: a
`
`2nd layer line: b
`
`3.77(m)
`
`3rd layer line
`
`3.ll(w)
`
`4th layer line
`
`Figure 3 X-ray fibre diagram of PAOB-3 drawn from the melt
`(modifications Ia and lb ). The Miller indices are given in Table 4
`
`the different conditions of crystallization. The WAXS
`patterns furthermore show both modifications to have a
`high degree of crystallinity. Melt-crystallized samples
`(referred to as modification I fibres) with a high degree
`of orientation may be drawn. The resulting fibre diagram
`is shown in Figure 3.
`The distance between the layer lines demonstrates
`that the identity period along the chain consists of two
`oxybenzoate monomer units. This is in accordance with
`the crystal structure of the dimeric model compound 8 in
`which the planes of subsequent benzene rings are rotated
`through an angle of 90° with respect to each other. The
`
`Reflection
`index
`(hkl)
`
`Calculated
`spacing
`(A)
`
`Observed spacing (A)
`
`X-ray
`
`Electron diffraction
`
`Orthorhombic modification II
`a= 14.20 A, b=9.60 A, c = 12.48 A
`110
`7.95
`7.94(s)
`200
`7.10
`7.07(m)
`0 0 2
`6.24
`6.24(w)
`
`020
`310
`220
`420
`
`4.80
`4.25
`3.98
`2.85
`
`4.72(w")
`4.22(s)
`3.96(s)
`
`7.99(s)
`7.12(s)
`
`4.83(w)
`4.25(s)
`4.00(s)
`2.87(w)
`
`• Shoulder of 3 I 0 reflection
`Abbreviations: vs, very strong; s, strong; m, medium; w, weak; vw, very
`weak
`
`2100 POLYMER, 1991, Volume 32, Number 11
`
`

`
`Rigid rod polymers: R. Stern et al.
`
`by material with lower degrees of polymerization. Since
`crystallites of stiff polymers usually have dimensions
`which are much larger in the direction perpendicular
`rather than parallel to the chains4
`, these crystallites will
`lie flat on the 0 0 1 surface. Thus with this geometry
`(electron beam perpendicular on the crystal surface) only
`h k 0 reflections will be monitored. Figure 5 shows the
`result whereas Table 4 gives the d-spacings of modification
`II which can be indexed by an orthorhombic unit cell
`(a= 14.20 A, b=9.60 A, c= 12.48 A, four chains per unit
`cell). The 6.24 A reflex which is only visible in the W AXS
`diffractogram (Figure 2b) is interpreted as the 0 0 2 reflex
`in analogy to modification I. The length of the repeating
`unit (12.48 A) is the mean value of the data derived from
`the cis and trans chains (see above). The X-ray density
`(1.27 gem - 3 ) is similar to values (1.2-1.3 gem - 3
`) found
`for the solution-crystallized oligomers8
`.
`
`Elevated temperatures. All PAOB-3 samples exhibit a
`broad bimodal melting peak with temperatures depending
`on molecular weight. Since the specimens having a
`smaller DP still contain reactive end groups, heating up
`leads to further polycondensation. Second runs in
`consequence exhibit higher melting temperatures. If the
`materials are annealed at the temperature of the lower
`melting peak a second run only gives a single sharp
`melting transition at T = 294 oc and no indication of a
`glass transition (Figure 6).
`This value nearly coincides with the melting point of
`the oligomers extrapolated to infinite molecular weight 8
`.
`It is thus concluded that the crystallinity in the annealed
`material is very high. This finding is already obvious from
`the fibre diagrams of melt-drawn specimens. In this
`context it is interesting to note that the heat of fusion
`(36 J g- 1
`) is approximately equal to the value found 3 for
`the high temperature solid-solid transition (38 J g- 1
`) of
`the unsubstituted poly(4-oxybenzoate ). WAXS diffracto(cid:173)
`grams and polarizing microscopy however, revealed that
`PAOB-3 is transformed into an ordinary nematic phase.
`No transition to the isotropic fluid occurred within the
`accessible range of temperature. The extrapolation of the
`transition temperatures obtained from oligomers to
`infinite molecular weight indicates8 the Tn; ofthe polymer
`to be located at ~ 430°C, i.e. where thermal decomposition
`prevails.
`
`3: 10.0
`E
`' ~
`0
`-
`0
`Q.l
`.s::.
`
`5.0
`
`0 0 t----,---,---,---,---,---,---.--!
`200
`100
`0
`300
`L.OO
`
`TfOC
`Figure 6 Differential scanning calorimetry analysis or an annealed
`sample or PAOB-3 at 20 K min _ ,
`
`POLYMER, 1991, Volume 32, Number 11 2101
`
`c1s chain
`C ;5 ~12.36A~ O
`II
`-
`II
`c
`c
`0
`0 ~ h c
`0-
`~
`0
`
`\-o-1 ~h \
`
`Figure 4 Representation or the trans and cis conformers or the
`p-oxybenzoate backbone together with the calculated c values (modifi(cid:173)
`cations Ia and lb)
`
`I
`
`\
`
`,
`
`'
`
`.!:
`
`'
`
`,
`
`\
`
`I
`
`Figure 5 Electron diffraction or PAOB-3 recrystallized from dioxane
`(modification II). Table 4 gives the Miller indices
`
`lb may be due to the formation of crystals with different
`conformation of the main chains (Figure 4). The length
`of the identity period along the c direction can be
`calculated from the bond lengths and angles derived from
`the crystal structure of the dimer. This leads to 12.60
`or 12.36 A for the trans and cis chains, respectively.
`The values derived from the analysis of the fibre diagram
`(Ia: 12.36; lb: 12.86) compare favourably with these
`deductions if the slightly higher value of c found for
`Ib is explained in terms of small distortions of bond
`angles. Also the mean value of the theoretical density
`(1.190gcm- 3 ) is in accordance with the experimental
`result ( 1.183 g em- 3
`). The only reflection not being
`indexed by the above unit cell corresponds to a Bragg
`distance of 7.16 A and is the strongest one of a series of
`reflections (21 .3, 10.7, 7.1 ). The most probable explanation
`for this is the assumption of a third orthorhombic
`modification Ic which seems to be formed during solid
`state polycondensation (cf. ref. 4 for similar observations
`on the unsubstituted polyester).
`In the case of solution-crystallized samples (cf. Figure
`2b) the corresponding structure may be elucidated by
`electron diffraction applied to small crystallites formed
`
`

`
`Rigid rod polymers: R. Stern et al.
`
`along the chain consists of two monomer units. However,
`there is no splitting into cis or trans segments, the chains
`seem to contain both forms in statistical sequence. A
`comparison with modification II of PAOB-3 shows the
`magnitude of b to be identical. From Table 5 and Figure
`7b it is obvious that the gradual loss of long-range
`correlation does not allow an unambiguous assignment
`of the unit cell owing to the absence of mixed reflections.
`On the other hand, all reflections of PAOB-6 can be
`satisfactorily indexed in terms of modification III (cf.
`Table 5). No new feature appears in the fibre diagrams
`
`Table 5 Calculated and observed X-ray reflections of orthorhombic
`modification III (PAOB-5 to PAOB-18) a=···, b=9.6 A, c= 12.6 A
`
`Calculated
`spacing
`(A)
`
`Observed
`spacing
`(A)
`
`Reflection
`index
`(hkl)
`
`PAOB-5
`100
`200
`300
`400
`020
`220
`330
`
`Figure 7 X-ray fibre diagram of PAOB-5 (a) and PAOB-10 (b)
`monitored at room temperature. The Miller indices (modification III)
`are given in Table 5
`
`Phase behaviour of PAOB-5 to PAOB-18
`As shown for PAOB-3, the structure of the polyester
`strongly depends on the preparation of the sample. The
`PAOB-n having large side chains exhibit the same feature.
`In the following, three modifications termed Ill, IV and
`V are described. Modification III results from melt, IV
`is prepared through precipitation from solution and V is
`formed when casting films from solutions in chloroform.
`As a common feature all three structures are organized
`in layers, the spacings of which depend on the length of
`the side chains.
`
`Solid state: modification Ill. Oriented samples of this
`modification are obtained by drawing fibres from melt
`or stretching films at temperatures between l20°C and
`170°C. All specimens thus obtained are highly oriented
`and exhibit an uniaxial texture. The resulting fibre
`diagram for PAOB-5 is displayed in Figure 7a.
`The off-meridional reflections point to a well-developed
`three-dimensional order. Longer side chains are followed
`by a loss of these long ra