ELSEVIER
`
`Polymer 42 (2001) 9583-9592
`
`polymer
`
`www .elsevier.com/locate/polymer
`
`Interactions in poly(ethylene oxide)-hydroxypropyl
`methylcellulose blends
`
`'Department of Materials Science and Metallurgy, University of Cambridge, New Museums Site, Pembroke Street. Cambridge CB2 3QZ, UK
`bP/wrmaceutical R & D, Pfizer Limited, Sittingbourne Research Centre. Sittingbourne, Kent ME9 BAG, UK
`
`Received 27 March 2001; received in revised form 11 June 2001; accepted 19 June 2001
`
`Abstract
`
`Blends of poly(ethylene oxide) and hydroxypropyl methylcellulose in the form of free films are examined for evidence of specific
`polymer:polymer interactions. Such interactions might affect the drug release behaviour of compressed matrices incorporating these poly(cid:173)
`mers. The effect of HPMC on the crystallisation of PEO during casting is investigated using differential scanning calorimetry (DSC) and
`wide-angle X-ray diffraction. Fourier transform infra-red (FT-IR) and Raman spectroscopies are used to examine the possibility of a complex
`between the two polymers. Thermodynamic interaction parameters are calculated for films cast from water and N,N-dimethylacetamide
`(DMAc) using the Flory-Huggins theory of mixing. The interaction parameter calculated is negative, indicating a miscible blend, and a
`hydrogen bonding interaction is detected. This hydrogen bonding is less likely to occur in films cast from water than in films cast from DMAc
`perhaps because residual water can shield the interaction sites.
`Finally, a transition involving a sharp reduction in heat capacity at high temperatures is reported. This transition is characterised using DSC
`and FT-IR and Raman spectroscopies, and is interpreted as a further complexing of the polymers. © 2001 Elsevier Science Ltd. All rights
`reserved.
`
`Keywords: Polyethylene oxide; Hydroxypropyl methylcellulose; Blends
`
`1. Introduction
`
`Hydrophilic polymers are used extensively to formulate
`matrix tablets for controlled drug delivery. The combination
`ofhydroxypropyl methylcellulose (HPMC) and poly( ethylene
`oxide) (PEO), two non-ionic polymers, has been shown to
`give a novel matrix tablet system that allows modification of
`the rate of drug release compared with pure HPMC. For
`example, the HPMC/PEO system can be used to increase
`the release rate at later times [1]. A possible mechanism by
`which drug release is modified is via a directpolymer:polymer
`interaction. Studies by Kondo et al. have established that the
`primary hydroxyl group on cellulose and methylcelluloses
`can form a hydrogen bond to the ether oxygen in PEO [2,3].
`This opens up the possibility of a similar interaction
`between PEO and the hydroxyl groups on hydroxypropyl
`methylcellulose. This study aims to find the nature and
`extent of any interactions between these polymers, and is
`a natural extension of the work of Kondo et al. and Nishio
`et al. [2-4].
`
`* Corresponding author. Tel.: +44-1223-334324;
`334567.
`E-mail address: rec11 @cam.ac.uk (R.E. Cameron).
`
`fax: +44-1223-
`
`Films have been studied because any polymer:polymer
`interaction which occurs in compressed matrix tablets will
`be exaggerated in a more intimately mixed system. Two
`different solvents have been used for film casting: DMAc
`in order to allow a direct comparison with previous studies
`on similar systems; and water because drug release occurs
`via penetration of aqueous fluid ingress into the system.
`Films are sh1died, both in the 'as cast' state, in which
`significant amounts of bound and unbound solvent may be
`present, and, for
`interaction parameter analysis, after
`annealing at elevated temperature. Such annealing might
`be expected
`to
`remove some residual solvent. The
`possible effects of residual solvent on the nature of the
`polymer:polymer interactions are discussed.
`
`2. Experimental
`
`2.1. Materials
`
`HPMC K4M Premium grade was purchased from Dow.
`The nominal molecular weight of this grade is 88,000 and
`the degrees of substitution for CH3 and CH2CHOHCH3
`are 4.12 and 19.24%, respectively. PEO with a nominal
`
`0032-3861/01/$ - see front matter© 2001 Elsevier Science Ltd. All rights reserved.
`PII: S0032-3861(01)00477-3
`
`SUBJDG-0007715
`
`RBP_TEVA05022343
`
`DRL - EXHIBIT 1031
`DRL001
`
`

`
`9584
`
`C.S. Fuller et al. I Polymer 42 (2001) 9583-9592
`
`molecular weight of 200,000 was purchased from Union
`Carbide Corporation. HPLC-grade N.N-dimethylacetamide
`(DMAc) and HPLC-grade water were both supplied by
`Aldrich Chemical Co. Distilled water was supplied by
`the Department of Materials Science and Metallurgy. All
`materials were used without further purification.
`
`2.2. Preparation of samples
`
`Separate solutions of 0.8 wt% HPMC and 1.3 wt% PEO
`were prepared. Aqueous solutions were made by heating
`two thirds of the water to approximately 85°C, stirring on
`a magnetic hot plate stirrer, adding the polymer powder to
`the vortex in a steady stream, then adding the remaining
`water at room temperature. These solutions were then left
`to cool, and stirred for 3 days before mixing in the relevant
`quantities and stirring for a further 3 days. DMAc solutions
`were made in a similar manner without heating the liquid
`prior to adding the polymer. The relative amounts (w/w) of
`the two polymers in the final solutions were 100/0, 67/33,
`50150, 34/66 and 0/100 (HPMC/PEO). Samples will be
`referred to in terms of their PEO content in per cent. After
`mixing, the solutions were poured into Petri-dishes and
`dried at 50cC in air for 3 days followed by 3 days at
`50°C under vacuum. Samples were then stored in vacuum
`desiccators.
`
`2.3. Measurements
`
`Differential scanning calorimetry was carried out on a
`Perkin Elmer DSC-7 in a nitrogen atmosphere. The instrument
`was calibrated with an indium standard. The thermal beha(cid:173)
`viour of the samples was examined by applying controlled
`heating and cooling regimes. Melting temperature was taken
`as the peak of the melting endotherm. The error in each
`measurement was estimated to be ±0.5°C. Where measure(cid:173)
`ments were repeated the error given is the standard devia(cid:173)
`tion divided by the number of measurements minus one.
`Wide-angle X-ray diffraction patterns were obtained
`using a Philips Diffractometer. Samples were placed on a
`silicon substrate to carry out the scans, and each sample was
`measured in duplicate to ensure reproducibility. CuKa
`radiation was produced by a Philips X-Ray generator.
`Raman Spectra were recorded using a 780 nm diode laser
`on a Renishaw Ramascope 1000. Spectra were measured in
`two or three places on each sample and representative data
`are shown. Infra-red spectra were recorded using attenuated
`total reflectance on a Perkin Elmer Infrared Fourier
`Transform Spectrometer.
`
`3. Results and discussion
`
`3.1. Characterisation of as-cast films
`
`In this section, the behaviour of as-cast films is consid(cid:173)
`ered. These films are likely to have small but significant
`
`quantities of residual solvent which may be bound to the
`polymer. Since bound solvent could reduce the extent of
`specific polymer:polymer interactions by occupying inter(cid:173)
`action sites, the nature of the solvent used in casting may
`have an effect on the final properties of the film.
`The films were examined under a polarised, transmitted
`light microscope and similar structural features were
`observed in films cast from DMAc and water. Pure PEO
`has a spherulitic morphology with spherulites measuring
`approximately 0.05 mm in diameter in water-cast, and
`0.5 mm in DMAc-cast films. This difference reflects a
`difference in the balance of spherulite nucleation and
`growth rates in films cast from the two solvents, which
`could be due to different levels of impurities or different
`solvent evaporation rates. All the films cast from blends
`have a much finer scale stmcture with very small (non(cid:173)
`sphemlitic) crystalline domains measuring
`less
`than
`0.01 mm across and no evidence of gross phase separation
`in the amorphous phase. It is possible that there was a degree
`of phase separation in the solutions, which could have lead
`to some residual phase separation in the amorphous phase
`of the as-cast films, although none is detectable by the
`techniques used in this study.
`
`3.1.1. Melt behaviour
`DSC traces of HPMC/PEO blends heated to 90°C at
`1 0°C min -I reveal that the PEO melt temperature decreased
`with increasing HPMC content, as illustrated in Fig. 1. The
`error bars represent the standard deviation divided by the
`number of repeats minus one. The decrease was similar for
`blends obtained from DMAc and water, but films cast from
`DMAc gave lower melting points across the complete
`composition range. This lower melting temperature in
`DMAc-cast films could indicate increased miscibility in
`the amorphous regions of these blends compared with the
`water-cast blends, although since the effect is also seen
`(albeit to a lesser extent) in the pure PEO film, this cannot
`account for the full difference. It would appear that even in
`pure PEO, more stable crystals result from casting from
`water, possibly due to differences in solvent evaporation
`rates or different levels of impurities in each solvent.
`The quantity of PEO melting, as indicated by the melting
`enthalpy, also decreased with increasing HPMC content,
`again with DMAc giving lower values across the complete
`composition range, although the values for pure PEO in this
`case are very similar . .b'ilms cast from both solvents give
`non-zero intercepts on the %PEO axis of the plot of melting
`enthalpy vs. %PEO in Fig. 1, indicating that PEO is unable
`to crystallise below a certain concentration in HPMC. The
`intercept for blends cast from DMAc is at a higher PEO
`concentration than for blends cast from water.
`The melting enthalpy data indicate that HPMC hinders
`the crystallisation of PEO in binary blends cast from either
`solution. This effect is more pronounced for blends cast
`from DMAc than from water.
`
`SUBJDG-0007716
`
`RBP_TEVA05022344
`
`DRL - EXHIBIT 1031
`DRL002
`
`

`
`A
`73
`
`72
`71
`
`70
`
`69 2 68
`~ 67
`66
`
`65
`
`64
`
`63
`62
`20
`
`8
`200
`
`180
`
`160
`
`140
`
`C.S. Fuller et al. I Polymer 42 (2001) 9583-9592
`
`9585
`
`A
`
`25000 . , . - - - - - - - - - - - - - - - - - - - - - - ,
`
`(j) 20000
`
`()j c :::l
`
`~ 15000
`.£
`(/)
`~ 10000
`£
`
`40
`
`60
`%PEO
`
`80
`
`100
`
`5000
`
`l
`
`0
`
`0
`
`I
`
`~L
`
`30
`Angle (28)
`
`'?"'"
`
`20
`
`10
`
`40
`
`50
`
`60
`
`8
`
`18000
`
`16000
`
`:w 14000
`.l!l
`§ 12000
`0
`~ 10000
`~ ·u; 8000
`c
`Ol 6000
`£
`
`4000
`
`2000
`
`0
`
`0
`
`10
`
`20
`
`30
`Angle (28)
`
`40
`
`50
`
`60
`
`01120
`
`"'; :c 1oo
`
`<1
`
`80
`
`60
`
`40
`
`20
`
`0
`
`0
`
`20
`
`40
`
`60
`%PEO
`
`80
`
`100
`
`Fig. 1. (A) Peak melting temperature vs. PEO content and (B) melting
`enthalpy vs. PEO content for as-cast films cast from water (.A.) and
`DIVIAc (e).
`
`Fig. 2. Examples of the X -ray diffractometer scans used to calculate degrees
`of crystallinity: (A) water-cast; (B) Dl\1Ac-cast. Thin line= O%PEO; thick
`line= 50%PEO; medium line = 100%PEO.
`
`3.1.2. Crystallinity
`Measurement of crystallinity was carried out by calculat(cid:173)
`ing the areas under wide-angle X-ray diffractometer scans.
`The results were compared with crystallinities calculated
`from DSC. Examples of the X-ray data used to calculate
`crystallinity are shown in Fig. 2. The formula used to calcu(cid:173)
`late crystallinity was:
`
`crystalline fraction =
`
`total area - area of amorphous halo
`total area
`
`(1)
`
`This calculation assumes that the scatter from each mole(cid:173)
`cule is the same. Melting enthalpies (/1H), measured using
`first heat DSC data, were converted into crystallinities by
`dividing 11H by the melting enthalpy of 100% crystalline
`PEO (197 J g -I) [5].
`that water
`The dependence on solvent suggests
`prevents, to some degree, the interactions between PEO
`and HPMC. In the absence of water, the polymers interact
`more strongly and PEO is less able to crystallise during
`casting. This effect may be due to water interacting with
`the two polymers, preventing them from interacting with
`
`each other, he this on a molecular level, or hy a greater
`degree of phase separation
`in
`the casting solution.
`Alternatively, this effect could be due to differences in
`impurity levels affecting nucleation rates, or a difference
`in solvent evaporation rates; with DMAc evaporating
`more quickly than water, there may be insufficient time
`for PEO to crystallise as fully as it could with a more slowly
`evaporating solvent.
`Fig. 3 shows the crystallinities of the HPMC/PEO blends
`calculated from W AXS and DSC data. The crystallinity
`decreases linearly with decreasing PEO content for films
`cast from both DMAc and water. This relationship is
`expected because the content of crystallisable polymer is
`decreasing linearly. Both plots show non-zero intercepts
`on the %PEO axes indicating that PEO does not crystallise
`above a certain HPMC content;
`the effect is more
`pronounced for films cast from D:MAc,
`the intercept
`occurring at about 40%PEO compared to 20%PEO for
`films cast from water. The DSC and W AXS data are in
`good agreement and demonstrate the differences between
`the films cast from water and from DMAc, that is, the
`PEO in the blend is more able to crystallise when the film
`is cast from water.
`
`SUBJDG-0007717
`
`RBP_TEVA05022345
`
`DRL - EXHIBIT 1031
`DRL003
`
`

`
`C.S. Fuller et al. I Polymer 42 (2001) 9583-9592
`
`"' (.)
`c
`"' -e
`0 en
`.c
`<(
`
`"' (,)
`c
`"' -e
`0 en
`.c
`<(
`
`1075
`1100
`1125
`Wavenumber (em·')
`
`1100
`1075
`1125
`Wavenumber (em·')
`
`A(ii)
`
`B(ii)
`
`9586
`
`A
`
`0.9
`
`0.8
`
`0.7
`-~0.6
`.S
`'ai 0.5
`~ 0.4
`0 0.3
`0.2
`
`0 1
`
`0+-----~~--~~------~------+-----~
`20
`40
`100
`60
`80
`0
`
`%PEO
`
`B
`
`0.9
`
`0.8
`
`0.7
`
`>-
`:~ 0.6
`'ai 0.5
`t) c 0.4
`0 0.3
`0.2
`
`0.1
`
`0
`
`20
`
`40
`
`60
`
`80
`
`100
`
`%PEO
`
`Fig. 3. Crystallinity vs. PEO content measured by (A) wide-angle X-ray
`diffraction and (B) DSC for as-cast films cast from water (.l) and DMAc
`ce).
`
`3.1.3. Vihratinnal spectrnscnpy
`Fourier transform infra-red (FT-IR) and Raman spectro(cid:173)
`scopy of as-cast films were carried out in order to detect any
`peak shifts that could be attributed to weak interactions
`between the two polymers, such as hydrogen bonding or
`complexation.
`The TR peak of interest is the C-0-C asymmetric stretch
`at 1100 em -l [6]. This peak in the PEO spectrum has been
`shown to shift due to hydrogen bonding to methylcellulose
`[2,3]. The spectra obtained for blends are shown in Fig. 4.
`There were no detectable peak shifts for water-cast films,
`but there was a 5 em -l shift to higher wavenumber for
`blends cast from DMAc compared with pure PEO cast
`from DMAc. This strongly supports the idea that a hydrogen
`bond can form between PEO and HPMC. The absence of a
`peak shift in the water-cast films may be because water
`bonds to the interaction sites, thus preventing the interaction
`with HPMC and allowing PEO to crystallise more readily.
`In addition, there could be a greater degree of microphase
`separation in the amorphous regions of blends cast from
`water compared with those cast from DMAc, which
`would also prevent the polymers from interacting.
`The region of the Raman spectra of particular interest is
`
`~ 100%
`c ~ "' -e
`
`"' (.)
`
`0 en
`.c
`<(
`
`66%
`
`50%
`
`~\}
`
`/
`
`33%
`
`t 0%
`
`c
`
`0 en
`<(
`
`\.._---/;
`
`0%
`
`:
`
`()
`
`N~ 100%
`
`66%
`
`50%
`
`"' (.)
`"' -e '~~~
`.c !Vp f',~ 33°/o
`
`1050
`1150
`1250
`Wavenumber (em" 1
`
`)
`
`950
`
`1050
`1150
`1250
`Wavenumber (cm"1
`)
`
`950
`
`Fig. 4. IR spectra of PEO from films cast from (A) water and (B) DMAc,
`showing (i) the range 1075-1125 em - 3 and (ii) the range 950-1250 em - 3
`The percentages refer to the amount of PEO in each blend. The relevant
`proportion of the pure HPMC spectrum has been subtracted from the blend
`spectra to obtain these traces.
`
`100-600 em -I. This region contains peaks attlibuted to
`PEO backbone vibrations (e.g. C-C-0, C-0-C bends
`and C-C, C-0 internal rotations) [7]. Once again, if a
`hydrogen bond is formed to the ether oxygen in PEO then
`these vibrations will be affected. There are no significant
`peaks in the HPMC spectra.
`Pig. 5 shows Raman spectra from as-cast films. The
`measured spectra for the blends are compared with theore(cid:173)
`tical spectra for mechanical mixtures of the two polymers.
`Theoretical spectra were calculated by adding the appropri(cid:173)
`ate fractions (in terms of mass) of the PEO and HPMC
`spectra. In general, the measured spectra show less intense
`peaks across the range of wavenumbers indicating that the
`PEO backbone is being prevented from vibrating. This
`effect is more pronounced for the DMAc-cast films. This
`shows that there is a strong possibility that hydrogen bonds
`have been formed between the hydroxyl groups of HPMC
`and ether oxygens of PEO, and that water prevents this
`interaction to some extent.
`
`SUBJDG-0007718
`
`RBP_TEVA05022346
`
`DRL - EXHIBIT 1031
`DRL004
`
`

`
`C.S. Fuller et al. I Polymer 42 (2001) 9583-9592
`
`9587
`
`A
`
`B
`
`Fig. 5. Raman spectra from films cast from (A) water and (B) DMAc; the
`upper of each pair (thinner line) is calculated from the measured spectra for
`pure HPMC and pure PEO and the lower of each pair is the measured
`spectrum for that blend.
`
`It is clear that the solvent used to cast films is important in
`determining the degree of interaction between the two poly(cid:173)
`mers in the as-cast state. Water has been seen to prevent the
`interaction to some extent, possibly because it can bond to
`the interaction sites on PEO and HPMC. All the evidence
`presented here strongly suggests that the two polymers can
`hydrogen bond to each other in a similar way to PEO and
`2,3-di-0-methylcellulose [2,3], but that the interaction is
`less likely to occur in films cast from water.
`
`3.2. Flory-Huggins analysis of annealed films
`
`An attempt to quantify the interaction between the PEO
`and HPMC cast from DMAc and from water was made
`using the method reported by Kondo et al. [3]. Samples
`were heated in the DSC to 90°C at 1 0°C min - 1 and held
`for 10 min before quenching at 200°C min -l to the iso(cid:173)
`thermal crystallisation temperature, Tic· The samples were
`monitored and held at this temperature for at least 10 min
`after complete crystallisation. The samples were then
`
`cooled to 20°C at 10°C min - 1, held for 2 min then heated
`to 90°C at 1 0°C min -I to measure the subsequent melting
`temperature, Tm.
`The thermal profile applied by the DSC first subjects the
`as-cast film to an annealing treatment at an elevated
`temperature, to allow full interaction of the polymers in
`the amorphous phase before controlled crystallisation.
`Since the annealing temperature of 90°C is considerably
`
`higher than the casting temperature of 50°C, it is likely
`that further residual solvent is driven from the films in this
`stage. However, it is still possible that some solvent remains
`bound to the polymer molecules. One might, therefore still
`expect there to be differences in the behaviour of films
`originally cast from the different solvents, if the quality of
`the residual solvent has an effect.
`It is important to note here that samples were annealed at
`a temperature below the Tg of HPMC. Ideally, the blends
`would be annealed at a temperature above the glass
`transition temperatures of both constituents to allow the
`amorphous phase to interact fully. However,
`thermal
`degradation occurs if the blends are heated above the
`glass transition temperature of HP:MC, invalidating the
`results. Kondo et al. [3] also encounter this problem, and
`adopt a similar solution. By following their method, and
`annealing at 90°C, we enable our results to be directly
`compared with theirs on PEO blends with cellulose and
`methylcellulose. Furthermore, we observed an unpredict(cid:173)
`able transition in the blends studied here at around 130°C,
`which is discussed later. Annealing at 90°C has the
`additional advantage of avoiding the complication of this
`transition occurring in some samples but not others.
`The concept of melting point depression to measure the
`interaction parameter is used because the blends consist of a
`crystalline and an amorphous polymer. However, morpho(cid:173)
`logical effects must also be considered because the degree of
`perfection and size of polymer crystallites, as well as any
`interaction between the polymers, affect the melting point of
`isothermally crystallised polymers. A true Flory-Huggins
`interaction parameter may only be calculated if morphology
`is independent of PEO concentration, that is, melting
`point depression is solely a result of polymer:polymer
`interactions.
`
`3.2.1. Hoffman-Weeks plots
`If morphology is independent of PEO concentration, then
`the stability parameter, ¢, which is a function of crystal
`thickness, will also be independent of PEO concentration.
`ln order to find out the stability, and the equilibrium melting
`temperature of the PEO crystals in the blends, the observed
`melting temperatures, Tm. of isothermally crystallised PEO
`the isothermal crystallisation
`were plotted against Tic'
`temperature for each blend composition. These plots are
`known as Hoffman- Weeks [8] plots and are shown in
`Fig. 6. The lines are lines of best fit calculated by the least
`squares method. Although there is some scatter in the data
`there is a general increase in Tm with Tic· Each data set was
`fitted to the following equation to estimate a value for
`stability parameter, ¢ ( ¢ being equal to the gradient of
`the line):
`
`T'::{ is the equilibrium melting point and ¢, the stability
`parameter which depends on the crystal thickness. The
`
`(2)
`
`SUBJDG-0007719
`
`RBP_TEVA05022347
`
`DRL - EXHIBIT 1031
`DRL005
`
`

`
`9588
`
`A
`
`74
`
`72
`
`70
`
`u 68
`::t
`E
`1-
`
`66
`
`64
`
`62
`
`60
`
`B
`
`74
`
`72
`
`70
`
`u
`::t
`E 68
`
`1-
`
`66
`
`64
`
`62
`
`40
`
`50
`
`60
`
`70
`
`o 33%PEO
`
`"'33%PEO
`
`40
`
`50
`
`60
`
`70
`
`Fig. 6. Hoffman-Weeks plots for (Al DMAc-cast and (Bl water-cast films.
`The broken line is Tm = Tic·
`
`C.S. Fuller et al. I Polymer 42 (2001) 9583-9592
`
`Table 1
`Stability parameter, ¢, of PEO crystals
`
`I
`
`Blend (HPMC/PEO)
`
`DMAc-casl
`
`'vValer-casl
`
`0/100
`34/66
`50150
`67/33
`
`0.20 ± 0.07
`0.17 ± 0.06
`0.16 ± 0.04
`0.14 ± 0.03
`
`0.24 ± 0.07
`0.3 ± 0.1
`0.14 ± 0.03
`0.18 ± 0.08
`
`experiment, and that it is valid to go on to calculate a
`Flory-Huggins interaction parameter.
`The equilibrium melting temperatures, T~, at this heating
`rate were found by calculating the intersection of the
`measured Tm vs. Tic lines with the line Tm =Tic· The results
`of these calculations are plotted in Fig. 7. The measured
`equilibrium melting point decreases with
`increasing
`HPMC concentration. The high uncertainty inherent in the
`data means that the extrapolated lines cross in some
`cases, which should not occur. This uncertainty is carried
`through the calculation to allow an assessment of the
`reliability of the final result. The errors were calculated
`using the standard method for the error in the gradient and
`intercept of a straight line [9].
`
`3.2.2. Meltinf( point depression
`Flory-Huggins interaction parameters [10] for PEO and
`HPMC may he estimated using melting point depression,
`assuming that the depression is due solely to thermo(cid:173)
`dynamic effects (which has been established to be a
`reasonable assumption within the error of the experiment).
`The melting point depression is calculated by subtracting
`the blend equilibrium melting point from the equilibrium
`melting point of 100%PEO. The uncertainties are given
`by standard error combination and are high because of the
`scatter in the melting point data. The values obtained are
`shown in Table 2.
`The conventional formulation for the thermodynamic
`depression of melting point caused by a diluent is as follows
`[ 11]:
`
`1/Tm- liT~
`
`error in qy was calculated using the standard method for the
`error in the gradient of a straight line r9l.
`The values obtained for qy are summarised in Table 1.
`A stability parameter of 1 implies Tm = Tic' and hence
`describes unstable crystals. A stability parameter of zero
`implies T'::{ = Tm, and hence stable crystals at equilibrium.
`The values are all significantly greater than zero, indicating
`that all crystals are fairly unstable.
`There is considerable scatter in the data, giving high
`values of uncertainty in the values of qy. However, within
`the experimental error, the values of qy are not dependent on
`composition for either set of films. Hence, it is assumed
`that morphological effects have been eliminated in this
`
`= - R(Vzui~Hzu)[lnvz/Vz + (1/Vz- 1/VJ)VJ + BviiRTml
`(3)
`
`where 1~~1 is the equilibrium melting point of PEO and 1;n is
`the observed equilibrium melting point of the blended PEO.
`Subscripts 1 and 2 refer to HPMC and PEO, respectively, v
`being the volume fraction of polymer and V being the molar
`volume of the polymer. V2u is the molar volume of the
`repeating units of PEO and ~H2u is the enthalpy of per
`mole of repeating units of PEO. B is the interaction energy
`density and R is the gas constant.
`V1 and V2 are large and hence the entropy term of
`Eq. (3) may be neglected [11]. The equation can hence be
`
`SUBJDG-0007720
`
`RBP_TEVA05022348
`
`DRL - EXHIBIT 1031
`DRL006
`
`

`
`C.S. Fuller et al. I Polymer 42 (2001) 9583-9592
`
`9589
`
`• DMAc
`
`o Water
`
`12
`
`10
`
`6
`
`:E G
`.=
`1
`
`4
`
`2
`
`0
`
`·2
`
`Fig. 7. Plots of the melting point depression.I:!.Tm. vs. the sqnare of volume
`fraction of HPMC. vi. for films casl from waler (0) and DMAc (e).
`
`rearranged to the following form, allowing the evaluation of
`the enthalpic contribution to the melting point depression:
`
`(4)
`
`where 11Tm is the melting point depression of the PEO
`component.
`A Plory-Huggins interaction parameter, x 12, may be
`defined to describe the enthalpy of mixing. It is related to
`the parameter B [10]:
`
`(5)
`
`where V1u is the molar volume of repeating units of HPMC.
`the melting point
`Experimental measurements of
`depression, 11Tm, are plotted against the square of the
`volume fraction of HPMC, vi, in Fig. 7. The volume fraction
`was calculated using 1.3 g em -I as an approximate value of
`the density of HPMC [2] and 1.09 gem -I as the density of
`the PEO melt (at 75°C) [12]. The slopes of these plots are
`equal to R(V2uii1H2u)IB, thus enabling the calculation of B
`and hence x 12 . Other quantities used in this calculation are:
`heat of fusion per unit volume (11H2ufV2u) of PEO =
`2401 cm- 3 [5], V1u = 151.86 cm3 mol- 1
`. V1u was calcu(cid:173)
`lated from the molar mass of HPMC, 197.42 g mol- 1 and
`its density.
`The lines were drawn by the least-squares fitting method
`assuming a linear relationship between 11Tm and vi, and
`including the point at zero. The non-zero intercept may be
`attributed to the entropic contribution to the melting point
`
`Table 2
`Equilibrium melting points obtained from the Hoffman-Weeks plots in
`Pig. 6, and the resulting values of melting point depression, calculated by
`subtracting the blend equilibrium melting point from the equilibrium
`melting point of 100% PEO
`
`Composition
`
`DMAC-cast
`
`Water-cast
`
`100% PEO
`66% PEO
`50% PEO
`33% PEO
`
`T~
`
`73 ± 2
`69 ± 2
`67 ± 1
`66 ± 1
`
`f:!.Tm
`
`T:f
`
`f:!.Tm
`
`0
`-4 ± 3
`-6 ± 2
`-7 ± 2
`
`74 ± 3
`72 ± 5
`70 ± 1
`70 ± 2
`
`0
`-2 ± 5
`-4 ± 3
`-4 ± J
`
`Table 3
`Values of the interaction parameter, X 12, for binary blends with components
`compatible in the melt
`
`Parameter
`
`DMAc-cast
`
`Water-cast
`
`X12 (at 348 K)
`
`-0.6 ± 0.2
`
`-0.4 ± 0.1
`
`depression, which was assumed to be negligible in the
`derivation above. Both sets of data yield positive slopes
`and the values of x12 obtained from them are shown in
`Table 3.
`The errors quoted in Table 3 indicate that although water(cid:173)
`cast films give a lower interaction parameter than DMAc(cid:173)
`cast films, this may be the result of experimental error, and
`the interaction parameters may, in fact, be the same. If the
`interaction parameter for the HPMC/PEO blend cast from
`water is indeed lower than that for the blend cast from
`DMAc, it may be concluded that the heat treatment at
`90°C is not sufficient to remove all residual solvent and
`that although the polymers are miscible when cast from
`either solvent, the presence of bound water renders them
`less so.
`The interaction parameter calculated for PEO/HPMC cast
`from DMAc at 75°C is -0.6 ± 0.2, which may be compared
`with reported values of -0.51 for PE0/2,3-di-0-methyl(cid:173)
`cellulose [3] and -0.67 for PEO/cellulose [4]. A lower
`value for PEO/HPMC would be expected if the interaction
`is partly due to hydrogen bonding to hydroxyl groups at the
`C6 positions on HPMC because some of these OH groups
`have been substituted in HPMC. As no errors are quoted for
`the values obtained by Kondo et al. [3] or Nishio et al. [4] no
`further comparison can be made. We must conclude that the
`interaction parameters are similar for PE0/2,3-di-0-MC
`and PEO/HPMC cast from DMAc.
`
`3.3. Overview of the polymer:polymer interactions
`
`The results for as cast films have shown that HPMC
`hinders crystallisation of PEO during solvent evaporation.
`This effect is greater in DMAc- than water-cast films. Spec(cid:173)
`troscopy indicates the presence of hydrogen bonds between
`the two polymers in the DMAc-cast films which do not form
`in the presence of water, either because water molecules
`block the interaction sites on both polymers or because
`there is a greater degree of phase separation in the aqueous
`solution, which is retained when on drying. The increased
`hindrance of crystallisation in DMAc cast films is likely to
`be because of this increased H bonding between the poly(cid:173)
`mers. It is also possible that different solvent evaporation
`rates or differences in impurity levels when casting from the
`different solvents play an additional secondary role in
`controlling the degree of crystallisation in blends cast
`from different solvents.
`The interaction parameter obtained for HPMC/PEO cast
`from DMAc indicates a similar miscibility in this blend to
`that in the 2,3-di-0-methylcellulose/PEO blend studied by
`
`SUBJDG-0007721
`
`RBP_TEVA05022349
`
`DRL - EXHIBIT 1031
`DRL007
`
`

`
`9590
`
`A
`18
`
`16
`
`14
`
`t-l----------------------~100%
`
`:§12
`5
`;- 10
`0
`"' 8
`"@
`- - - - - - - - - - - - - . . . . ,___ _ __ 166%
`Q) ~
`6
`I
`
`1'--------~------------~ \__
`
`50%
`
`4
`
`2
`
`0
`
`20
`
`70
`
`120
`Temperature CC)
`
`170
`
`8
`
`20
`
`18
`
`16
`Oi 14
`~ 12
`~ 10
`"' "@
`8
`I
`6
`
`(j)
`
`4
`
`0
`
`)~
`
`./
`h
`
`t
`1\
`
`1 00%
`
`66%
`
`-
`
`20
`
`40
`
`60
`
`120
`100
`80
`Temperature ("C)
`
`140
`
`160
`
`180
`
`Fig. 8. DSC scans showing the step transition in films cast from (A) DMAc
`and (B) water. The upper curve of each pair (thinner line) is the second heat
`after isothermal crystallisation at 50°C. Percentages refer to the amount of
`PEO in each blend.
`
`Kondo et al. For the blend cast from water it is not possible
`to determine whether the heat treatment applied has been
`sufficient to eliminate the effect of the solvent, but it is
`possible that the polymers are slightly less miscible in this
`case, suggesting that some bound water remains allowing
`water to interfere with the interactions between these two
`polymers.
`
`3.4. Transition at elevated temperature in as cast films
`
`In this section, the high temperature behaviour of as-cast
`films is considered.
`in the DSC to 180°C at
`On heating as-cast films
`l0°C min - 1 a step-shaped
`transition was sometimes
`observed. This transition was indicated by a sharp reduction
`in heat capacity, the opposite of what would be observed for
`a glass transition. The step never occurred in either pure
`PEO or pure HPMC cast from either solvent, so it is inter(cid:173)
`preted as a further polymer:polymer interaction. The height
`of the step was a maximum for blends containing 50% PEO
`for films cast from both D:MAc and water (Fig. 8). In addi(cid:173)
`tion, the temperature of the step was independent of film
`composition or solvent to within experimental error.
`
`C.S. Fuller et al. I Polymer 42 (2001) 9583-9592
`
`After the first heat, samples were quenched to 50cC at
`200°C min - 1 and allowed at least 30 min to crystallise
`(the crystallisation was observed during the scan and the
`sample was left at 50cC for at least 10 min after the crystal(cid:173)
`lisation exotherm was complete). Once crystallisation was
`complete, the samples were cooled to 20°C then reheated to
`180°C at 1 0°C min - 1 and the enthalpic behaviour measured.
`In films originally cast from DMAc, the melting endotherm
`in the second heat was very small, if it was there at all,
`showing that very little P.EO crystallised after undergoing
`the step transition. The melt in the second heat was also at a
`slightly