Makromol. Chem., Rapid Commun. 5,119 - 124 (1984) 119

Miscibility of poly(viny1idene fluoride) with poly(N-vinyl- 2-pyrrolidone) and poly(N,N-dimethylacrylamide)

Monique Galin

Centre de Recherches sur les Macromoltcules, CNRS 6, rue Boussingault, 67083 Strasbourg-Cedex, France

(Date of receipt: November 30, 1983)

Introduction

During the last decade, poly(vinylidene fluoride)-poly(methy1 or ethyl methacry- late) homogeneous blends have received much attention as outstanding model systems for the analysis of polymer compatibility, and as polymeric materials of definite technological interest * -3). Moreover poly(vinylidene fluoride) (PVF,) was also found miscible with poly(methy1 or ethyl acrylate), poly(viny1 acetate) and poly(methy1 vinyl ketone): Paul et a1.4) suggested that the major driving force for compatibility results from dipolar interactions between PVF, and the carbonyl functions of the amorphous polymers. Since tertiary amide structures are well known for their strong dipolar character (dipole moment p = 3,7 -4,l Da) for representative simple species), it may appear of fundamental and practical interest to check PW, compatibility with poly(N-vinyl-2-pyrrolidone) (PNVP) and with poly(N,N-dimethylacrylamide) (PDMA) as typical dipolar polymers. Before conducting a differential calorimetric analysis of PVF,-PNVP and PVF,-PDMA blends, we performed a gas chromato- graphic study of the interactions between liquid PVF, and N-methyl-2-pyrrolidone or N,N-dimethylacetamide, as low molecular weight model compounds for PNVP and PDMA, in order to estimate the importance of specific interactions in such systems.

Experimental part

Solvents and polymers

N,N-Dimethylformamide (DMF), N,N-dmethylacrylamide (DMA), and N-methyl-2-pyr- rolidone (NMP) were twice vacuum distilled over CaH, and stored over molecular sieves (4 A). Poly(vinylidene fluoride) (PVF,) was an unfractionated Kinar-461 sample (weight-average molecular weight M, = 2,75 1 6 ) from Pennwalt Corporation: it was purified by precipitation from DMF solution into water. Atactic poly(N-vinyl-2-pyrrolidone) (PNVP) (M, = 0,758.16) and poly(N,N-dimethylacrylamide) (PDMA) (m, = 1,62.16) were obtained by radical-initiated polymerization of the corresponding monomers.

a) In SI units 1 D = 3,3356. C . m.

0173-2803/84/$01 .MI

120 M. Galin

Gas liquid chromatography (GLC) measurements

Apparatus and general experimental procedures were the same as already described j). PVF, was deposited from a 3 wt.-% DMF solution on glass beads of specific surface area of 0,016 m2 . g- ' , previously treated with hexamethyldisilazane. After vacuum drying at 100"C, the coated support (0,3 wt.-Vo loading) was packed into a stainless steel tube column (inner diameter = 0,25 inch (=6,35 cm), length I = 1,20 m). Preliminary experiments carried out within the range 165 -200°C showed that the retention times of NMP or DMA were essentially independent of the gas vector flow rate (5 -20 ml . min-I ) for the very low amounts of sample we used (g0 , I pl, limited by the catharometer detector sensitivity).

Differential scanning calorimetry (DSC) measurements

Polymer blends were obtained as homogeneous films by vacuum evaporation (< 10- mmHg) of 3 wt.-To DMF solution first at room temperature and then at 100 "C for three days, followed by slow cooling to room temperature. The DSC measurements were carried out on a Perkin- Elmer DSC-2C calorimeter after previous calibration with indium. Two consecutive DSC cycles were systematically performed on samples of ca. 5 mg with a heating rate of lO"C/min: the first cycle directly on the as cast films; the second one on samples melted and maintained at 200OC for 10 min, then cooled at 40 "C/min to room temperature and maintained at this temperature for 30 min. The glass transition temperature, Tg , and the melting temperatures, T, , were taken at the base line inflexion point and at the endotherm peak, respectively.

Physical data related to sohents and polymers

The relevant physical data for solvents and polymers were taken from literature: vapour pressures and molar volumes of NMP and DMA6s7), calculation of the second virial coefficients of the solvent vapourss), specific volumes and melting enthalpy of PVF, 9); specific volumes of PNVP and PDMA were approximated by their values measured in NMP and DMA solution, respectively (digital densitometer, Anton Paar KG, model DMA-02).

Results and discussion

Thermodynamics of interactions between NVP or DMA and liquid PVF,

The GLC retention times of NVP and DMA on liquid PVF, between 160 and 200 "C were converted to the specific retention volumes, corrected t o 0 "C, e, in the classical way lo). The weight fraction activity coefficient Q", the Flory-Huggins interaction parameter x and the partial molar heat of mixing A-G were derived from

values according t o the equations lo):

(5) - 1 x = log, Q" - log,

where subscripts 1 and 2 refer t o the model compound and the polymer, respectively.

Miscibility of poly(viny1idene fluoride) with poly(N-vinyl-2-pyrrolidone) and. . . 121

M is the molecular weight, P; is the saturation vapour pressure at temperature T, , V and v are the liquid state molar and specific volumes, respectively, and B,, is the gas state second virial coefficient at temperature T.

values for both systems: specific interactions between the model compounds and PVF, may arise from the high dipole moments of the two tertiary amides. These features, which may be correlated with the high solvating power of NMP and DMA for PVF, in dilute solution at 125 OC1,), suggest a potential bulk miscibility of PVF, with either PNVP or PDMA.

Our results collected in Tab. 1 clearly point out strongly negative x and

Tab. 1 . D i p o l e moments p, Flory-Huggins interaction parameter x , and partial molar heat of mixing K of poly(vinylidene fluoride) (PVF, )/model compound systems and PVF,-PNVP or PVF,-PDMA blends

Model compounds and polymers

Model compounds: N-Methyl-2-pyrrolidone (NMP) 4909 - 0,69 b, - 8,37 N,N-Dimethylacetamide (DMA) 3,72 - 0,72b) - 5,44

Poly(N-vinyl-2-pyrrolidone) (PNVP) - -0,7c) - Poly(N,N-dimethylacrylamide) (PDMA) - -0,7'=) -

Polymers:

a) In SI units: 1 D = 3,3356. b, GLC analysis of PVF,-NMP or PVF2-DMA systems. c, Analysis of the melting point depression of the as cast films of polymer blends.

C * m.

Compatibility studies of PNVP or PDMA-PVF, blends

DSC measurements on the polymer blends were analyzed from two complementary view-points: glass transition temperature Tg , melting temperature T, , and degree of cristallinity T calculated from the DSC melting endotherm, using the literature value of 6,7 kJ . mol-1 for the molar heat of fusion of 100% cristalline PVF, 9) . The results are collected in Tab. 2 and illustrated in Figs. 1 and 2.

For blends containing no more than 60 wt.-qo of PVF, , the DSC curves show no melting endotherm and only a single Tg intermediate between those of the corre- sponding two homopolymers: Thus, these amorphous blends may be considered to be homogeneous.

Blends containing more than 60 wt.-Yo of PVF, are semi-crystalline: the DSC curves show a single well defined melting endotherm which depends both on the thermal history of the sample and on its composition. For as cast films (isothermally crystallized at 100°C for 3 days), the PVF, degree of cristallinity T (only a-form crystal) in the blends including more than 75 wt.40 of PVF, is verysimilar to that of the pure homopolymer, T = 0,49 k 0,02: the amorphous polymers PNVP or PDMA actually behave as pure diluents in the binary systems.

122 M. Galin

Tab. 2. Glass transition temperatures T g , melting temperatures T, , and degrees of crystallinity r of homopolymers poly(N-vinyl-2-pyrrolidone) (PNVP), poly(N,N-dimethylacrylamide) (PDMA), poly(viny1idene fluoride) (PVF,) and their blends PVF,-PNVP and PVF,-PDMA having different compositions

~~ ~~~~

Polymer and Composition ofa) , T,/"C T, /"C r T,,,/"C r polymer blend blend in J L

Y u of cast films of recrystallized W1 @I films

PNVP 1 1 PDMA 1 1 PVF, 0 0

PVF,-PNVP 0,lO 0,130 0,25 0,304 0,40 0,472

PVF,-PDMA 0,lO 0,140 0.15 0,205

170 120

-40b) 160,5 0,475 160,5 0,475

158,O 0,504 156,s 0,425 153,5 0,480 150,5 0,235

80 144,5 0,161 ncC) 0

158,O 0,490 156.5 0,405 156,O 0,470 155,5 0,375

0,40 0,493 50 141,s 0,080 nee) 0

w , and @, are the weight fraction and volume fraction, resp., of amorphous polymer in blend. Literature value9). nc = non-crystalline.

Weight fraction of PVF,

Fig. 1

1

0.1 0.2

~ i ~ . 2 @,? (PNVP or PDMA)

Fig. 1. Degree of crystallinity z of poly(viny1idene fluoride) (PVF,) vs. weight fraction of PVF, in blends of PVF, with poly(N-vinyl-2-pyrrolidone (A) and poly(N,N-dimethylacryl- amide) (0) as cast films as well as with poly(N-vinyl-2-pyrrolidone) (A) and poly(N,N-di- methylacrylamide) ( 0 ) as recrystallized films

Fig. 2. in as cast film blends vs. square of volume fraction @; of the amorphous polymer: poly(N-vinyl-2-pyrrolidone (PNVP) ( A ) or poly(N,N-dimethylacrylamide) (PDMA) ( 0)

Melting point depression AT, of poly(viny1idene fluoride) (PVF 2 )

Miscibility of poly(vinylidene fluoride) with poly(N-vinyl-2-pyrrolidone) and. . . 3 23

For higher weight fractions of these polymers, the crystallinity degree of PVF, is strongly decreased, especially for the blends with PDMA. On the other hand, the observed T values of the samples recrystallized at room temperature (see the specific thermal treatment in the Exptl. part) is a strongly decreasing function of the PNVP or PDMA fraction in the blends within the whole range of composition. Such a behaviour is quite similar to that previously observed for PMMA or PEMA-PVF2 blends 12* 13).

The melting points T, of the crystalline as cast films are systematically lower than that of PVF,, and the melting point depression AT,,, is quite parallel for PNVP and PDMA blends. The experimental T,,, values were tentatively analyzed according to Scott's equation'):

AT,= - v2 x12

Lw, Vl T,,,,,-B@f with B = R T -

where @1 is the volume fraction of the amorphous polymeric diluent. T,,,,, is the equilibrium melting temperature of pure PVF,, Vl and V, are the molar volumes per repeating unit of the amorphous diluents and PVF, polymers, AH2 is the molar heat of fusion of 100% crystalline PVF,. The experimental data collected in Tab. 2 are illustrated in Fig. 2.

The variations of AT, versus @: are linear over the whole composition range and nearly identical for the two systems, allowing the derivation of the polymer-polymer interaction parameters x12: their strongly negative values x12 = -0,7 at 170°C are actually very similar to that measured for the model systems PVF2-NVP or PVF2- DMA, (Tab. 1). Moreover, they are still more negative than those calculated in an identical way for PVF,-PMMA of PVF2-PEMA blends, x = -0,30 at 160°C1z*13). This feature suggests stronger interactions in case of tertiary amide polymers, in good agreement with their greater dipolar character. On the other hand, the intercept of the linear variations of AT,,, versus @: is = 1 "C, and thus cannot be neglected. Such a deviation from Scott's equation cannot be taken into account by entropic effects, because of the too high molecular weights of the polymers4): it more probably results from morphological and kinetic effects on the crystallization of the blends 14). The calculated x12 are rough, underestimated values, but this inaccuracy does not weaken significantly the general semi-quantitative conclusions.

Conclusion

The miscibility of PVF, with PNVP or PDMA may be considered as well ascertained on the basis of the occurrence of a single glass transition temperature lying intermediate between those of the parent homopolymers for the amorphous blends (PVF, : < 60 wt.40) and of the analysis of the melting point depression for the semi-crystalline blends. With respect to recent literature compilations -3), these two miscible polymer pairs appear quite original, and as an incidental point, both systems involve water-soluble polymers. On the other hand, it is worth to reemphasize the definite interest of the modelization of a polymer l/polymer 2 system by an

124 M. Galin

analogous solvent l/polymer 2 system, where the solvent molecules are structurally similar to the monomeric unit of the corresponding macromolecular chain. The partial molar heat of mixing KH; of the volatile model compound with the polymer is a direct measure of the real specific interactions, and -a < 0 values suggest potential miscibility for the corresponding polymer blends: this approach may be of interest for finding new compatible systems. Finally, we will show in a forthcoming publication that hydrogen bonding together with dipolar interactions probably contributes to the specific miscibility of the polymers under investigation.

’) 0. Olabisi, L. M. Robeson, M. T. Shaw, Polymer-Polymer Miscibility, Academic Press,

2, S. Krause, in: “Polymer Blends”, edited by D. R. Paul and S. Newman, Academic Press,

3, D. R. Paul, J. W. Barlow, J. Macromol. Sci., Rev. Macromol. Chem. 18, 109 (1980) 4, D. R. Paul, S. W. Barlow, R. E. Bernstein, D. C. Wahrmund, Polym. Eng. Sci. 18, 1225

3 M. Galin, Macromolecules 10, 1239 (1977) 6, 0. Olabisi, J. Appl. Polym. Sci. 22, 1021 (1978) ’) G. J. Welch, R. L. Miller, J. Polym. Sci., Polym. Phys. Ed. 14, 1683 (1976) 8, J. P. O’Connell, J. M. Prausnitz, Ind. Eng. Chem., Prod. Res. Dev. 6, 245 (1967) 9, K. Nakagawa, Y. Ishida, Kolloid Z. Z. Polym. 251, 103 (1973); K. Nakagawa, Y. Ishida, J.

lo) D. G. Gray, Prog. Polym. Sci. 5, 1 (1977) 11) S. Ali, A. K. Raina, Makromol. Chem. 179, 2925 (1978) 1 2 ) T. Nishi, T. T. Wang, Macromolecules 8, 909, 1975 1 3 ) T. K. Kwei, G. D. Patterson, T. T. Wang, Macromolecules 9, 780 (1976) 14) B. S. Morra, R. S. Stein, J. Polym. Sci., Polym. Phys. Ed. 20, 2243 (1982)

New York 1979

New York 1978, vol. 1, p. 15

(1978)

Polym. Sci., Polym. Phys. Ed. 11, 2153 (1973)